Wow. Thank you for describing a real world way of measuring suspension geometry. My daily is 122s (Amazon) and I am planning upgrades. I have watched many suspension videos and no one has really explained how to actually measure these values. It is really hard when the irrational bits of the grey matter work against to crush motivation. All the best and thanks for sharing your build, struggles and achievements. -Todd
Good video, when I did mine, I built the controls arms out of PCV plumbing pipes in the mock up stage and tried out dozens pick-up points and control arm lengths. Made the virtual FVSA (Front View Swing Arm) with string to capture the roll center. Was manual and took awhile but had everything worked out before cutting and welding. Keep up the good work.
That's like the Stainforth string computer - works a treat from what I understand. I'm going to try some computer simulations in the near future. The major learning in this exercise was to drill into my head what happens when pivots change. I needed to experience this and get some hands-on before computer analysis output makes any sense. I learn mechanical things through my hands more than my head it seems.
Long time fan of your content. I've used your videos to help me tune my own skills over the years. You sure do make it look a lot easier then it really is! Thanks for the great work.
This is a great tutorial. I have been building race car suspensions since the 70s and am self taught. I have been frustrated trying to explain to people why youtube videos don't show the real process for scratch build suspensions. Henceforth I will just suggest anyone look at this first.
This was great. Thank you for going into the weeds a bit for us. I'm following the graphs well enough. The one thing I'm missing is what you've been changing to get those new curves. Adjusting position of control arms? Angle? It would be interesting to know/see in the next suspension video. Or in a comment, whatever works. Otherwise, hey, anyone who's done enough work has run into "ope" moments like this. C'est la vie. We're here to see the process, and no one ever learned by doing something perfectly the first time.
Sorry - I'm changing the location of the links as well as caster/camber. In the graphs - the UCA/LCA both went down 2" (a little more for the front UCA pivot).
just remember this task is usually done by a person that has at least 8 years of schooling and then years and years of hands on experience and they STILL get it wrong... keep on it, it will eventually all fall into place, or apart... either way it makes good content!
I played with the Corvette suspension a few years back. When I used the standard upright positions on bump and droop the wheel angle was terrible. This left me really confused then I re ran the tests with the body of the car rolling slightly and the suspension worked perfectly. What the designers realized that I had considered is that when you are driving straight it doesn't matter what angle the wheel patch is during bump etc. But in a corner is were it counts and that means you need to take into account body roll. Love your videos keep up the great work.
I've got that in mind - it just doesn't pack into a video very well...I didn't show the side view geometries or any of the bump steer correction work (though watching my laser line move might be fun if you have a cat). So body roll is part of the overall vehicle design.
Love your work and and remember, everybody make mistakes, good people admit them, great people share them to help prevent others making the same mistake. Thanks for sharing.
This pretty much blew my mind about roll centers, great content as always! You took up several things I've never heard mentioned in all the suspension tips from the "pros". Definitely invaluable information before I start hacking up my own frontend! :)
incredible work. Thanks for sharing your struggle with the world, its gonna make things way easier for the people who follow. No one talks about suspension in such detail. Never change TSG.
Excellent explanation. I too had wondered about 2 diamesion instead of 3 . I did note at position approx 17.30 on video . When jacking up to test the front top mount bolt moves . Hopefully that didn't make a difference . Good work keep it up
Sharp eye! I had some issues but in the end I measured and remeasured...then optimized and remeasured again. It was solid with little variability. I did do a standard error of the measurement analysis (not reported) - I'm a scientist - so I can do boring things virtually forever.
@@ThrottleStopGarage Glad to see you able to keep up your enthusiasm for the project after you determined the mistakes/problems with the suspension. Many a mortal would have been crushed!
I measured the roll center in a similar way when setting up the front suspension on my Volvo 142 race car to make sure I wasn't messing anything up by lowering it. Removed the springs and raised and lowered the car with the front tires on some rolling dollies, measuring the distance changes between the dollies in the process (and camber changes). That process made me realize that the slope of the CPC's trajectory is the real metric we're after, in addition to being easier to understand. A 10:1 slope, means 10% of the lateral loads are added to the tire vertically before the car even begins to lean and load the springs.
Typically roll centers are tuned for weight transfer dynamics and caster/camber curves are tuned for tire patch optimization independently. Also, roll centers migrate, so I'm not sure using your method actually calculates the roll center at curb position. You may be right, I definitely want to look into it, but when I think about adding significant length to a single arm, you will massively change the results. Like if your upper arm is shorter than the lower, you'll camber in. But if it's longer, you'll camber out. You can be at the same RCH with both conditions, but have significantly different kinematics. I would say this is a good method for measuring the camber curve for sure.
It's a method that can be used and it works. All roll centres are instantaneous...and can be found using forced based or geometric methods. There are problems with definition, I'll cover this in my next video.
@ThrottleStopGarage I guess the issue I have with the method is, I can change the kinematics to get different outputs that you measure, but without changing the roll center height. What I'm not sure of, is if the outputs, though different, will still produce the same RCH. Regardless of that, you seem to insinuate tuning your caster camber curves by adjusting RCH, but I typically treat them as completely decoupled metrics and tune them to do different things as they are responsible for different aspects of dynamics. I guess my comment just stems from your early reference of being critical of other RCH videos, but not actually teaching what they do or how to tune them, but what I got from this video seems to be very different than what's practiced in the vehicle dynamics community. Again, not saying your measurement method is wrong as I've never used that method to measure RCH and I plan on running the math as it's intriguing, but you seem to be the type of person who would take pleasure in technical discussions and I also enjoy learning new methods and techniques.
@@badinvestmentbuilds3823 Interesting - I'll try to explain the method first and then we can chat about the rest. The criticism of the other videos is that most share a simple 2D "join the dots" geometric method (nothing wrong there) with the assumption that measuring the intersections of these points is trivial. In practice, most suspensions are complex shapes and it's difficult to measure these points. If we're making some quick adjustments to arm chassis mount points, then there is an "easier" way. I glance off the topic of what a roll centre might represent as it opens up a bit of a pandora's box and I am not qualified to argue these points. There are many research articles and I think a deep dive into this is beyond what I would be comfortable trying to explain. In the vehicle dynamics community, they know what they're doing and they have studied the subject and worked with it - they have data-driven validation. No question. I guess I just didn't need to see another "join the dots" video. Lots of those, most are a little shy of what happens when you do these things. If I can only replicate what others have done, I'd just leave it. I also didn't find the locating of these points to be effecient to measure on a jig. The accumulation of error in their location (uncertainty) was pretty large (while it may have only been a few mm total). Camber curves are a reflection of the relative lenth of the links and their angular relationship. That's it as far as I know. As my links were fixed, then what I can do is change their angular relationship. The effect of this change is to alter the location of the location of the instantaneous centre - as, at that instant in space, the suspension is moving as a swing arm fixed at that point. The validity of the method I present (which is not my creation - I was told to implement this method and the geometry makes sense) stems from the geometric reality that measuring the slope of the arc at the ground plane yields the line where the IC is located without having to measure the specifics of the arms that produced the IC location. The line that connects the contact patch centre to the IC is what is used as a geometric location for the RCH and for a symetrical suspension, the RCH is going to fall at the centre of the vehicle. For any given RCH location for any specific combination of links and angular relationships between these links, there is a single IC. In other words, if you hold the angular relationship constant - there is only one possible RCH. The arms could be any length up to an including their intersection at the IC. If you fix RCH and vehicle track width - you can have many different arrangements that will generate the RCH - but they all must lie on the line described by the contact patch intersection and the RCH. This method seeks to measure the angle of that line. If you change the angular relationship of the links to alter the camber curve...you will change the RCH. They are completely dependent. I hope I have explained this a little better.
@ThrottleStopGarage I think the part my brain is messing with me is that I can change the kinematics without changing the RCH, which will change the output the curves, but the new curves will likely point to the same location. I'll put a little math model together comparing them without measurement error. I'm the field, we would use inclinometers on the arms, articulate it on an alignment rack, and pop those numbers into a calculator, but that requires a bit more up front work. I'll reach out again if I can find the time to make the model and hopefully post up a video complimenting yours. You're correct that there's a lot of 2D dot connecting that doesn't do much good to the average tuner. That has been a bit of inspiring for me to want to make a "Vehicle Dynamics for a 2 Car Garage" series lol. High level specific development with the average resources a DIYer has.
Just restarted redesigning my front end so well timed sir! Almost done with seat mounts so almost time for wiring which I’m not quite but almost dreading.
I like the different approach, if your adapting a suspension to a car this is a great method to get readings. One thing I would have touched on is the relationship of the roll center to center of gravity.
Love your work! Very informative, keep it up! At one point you mention you don't want that binding in the suspension but I think what you're seeing is the anti-squat working, unless some other factor is affecting it. When your arms are parallel to each other they use the spring and damper to react the forces only. When they aren't parallel, you introduce a fore/aft movement of one ball joint relative to the other which causes the bind. For example, say your bottom arm is parallel to the ground and your upper arm is angled up at the front, like on your rig. When you look at the side view the lower ball joint will move vertically. The upper ball joint will move vertically but also rearwards. If the upper and lower joints weren't connected, it wouldn't be a problem but the upright has a fixed length. The distance has to remain fixed but the arms want to separate which causes binding. More compression, more bind. Essentially you're creating a geometric resistance, ie a stiffness proportional to the compression, similar to a variable rate spring that increases in stiffness with more bump.
What was happening was because the suspension arms were in the "wrong" place the ball joint was locking and that was the problem. I also had issues with the caster and anti-dive and the spherical joint on the tie rod at one point. Taking the caster to a more sensible value (based on the camber rate target and other factors) and dialling out the anti-dive to a value for this car as well as changing the bearing to another model provides zero bind over more than the range of travel for the suspension. This was over a month of solid work to get done.
Phenomenal video! Thank you so much for making it. I agree with your assessment of the information that's available on the internet on roll centers. Everybody is theorizing and nobody is actually testing anything . And as we all know mechanical suspension modifications are generally very expensive. I'm wondering through your analysis of the lateral movement of the track width. Wouldn't that increased lateral movement increase the track width and therefore improve handling?
Would love to steel a moment of your time to chat about suggestions that may help the install of my C4 components into my '54 Pontiac. Your videos are great, they've helped significantly
Very interesting video. I learned a lot and am doing a similar build with the C4 components I’m hoping to narrow into a Triumph GT6. Keep up the great videos.
static RC critique: totally agree. sounds like Carroll Smith's 'Tune to Win', great book with same story; or Carroll's bump steer+measure jig he describes in 'Prepare to Win'
23:58 You meant ideas for videos? In line with the suspension topic: dampers? Depending on the characteristics of your car and the driving characteristics, how do you choose your damper? Valving, oil, ... ?
Ideas for the suspension...but thanks for the video ideas. We'll do these as we roll along. For sure full suspension characteristics - once I drive it. Then it's on to tuning suspensions for specific use.
Around 18:00 during compression, your camber is getting more positive, not negative. Is that expected? From your graphs, and maybe I wasn't paying attention, it's not clear if ride height positive is up or down.
@@ThrottleStopGarageit's not. Unless yours is different, I've owned that camber gauge for 12 years. It's going positive during compression. Once you cross zero while going up with the jack, you can see the little triangle on the inside is on top, and on the outside it's on bottom. That indicates positive.
@@grant1984 at 18:00 to 18:14 the suspension is being cycled through it's extension (droop) not compression and you're right it is gaining camber (positive) relative to ride height. That is expected. It crosses back through static ride height at 18:14 and from there it is in compression relative to ride height and is correctly reading negative camber. This is expected.
@@ThrottleStopGarage at 18:01 when you start to droop the suspension the gauge does indicate the outside ( to the left of the gauge reading) is higher than the inside (to the right of the reading). This is the opposite of my expectations.
@6:40 isnt this technically considered Caster? The angle of the upper arm dictates caster while the lower arms angle will dictate anti-dive? Maybe thats just for RC car suspension tuning.
No - caster is the displacement from vertical of the steering axis. UCA angle is use to create anti-dive. View the suspension in side view - there is also an IC in that dimension. So in this case the UCA is behind the centre defined by the LCA giving caster and the tilt of the UCA gives us anti-dive.
Hi, thanks. Just a few questions. Have you tried changing the hight location for the power steering rack and different rod lengths? Also, under full compression, Is it better to have the lower control arms at a level position? I know this requires moving the suspension and pivot pick up points but my theory is there will be less camber and toe deflection. Would this be the case?
Interesting question. I did alter the height of the rack and that does change bump steer characteristics - it has to sit on the plane described by the UCA/LCA pivots. Then there is where it sits in and out - this changes the Ackermann. I'm just working on this again in the new series...it's taken me a long time, but I'm back on the suspension right now. At the static ride height, for this suspension, the LCA upper surface is meant to be level. So as they compress, the angle increases. The issue is there is a fixed curve described by the length of the arms, the upright and the pick-up points. There is only so much you can change if you're working around the parts.
Ok thanks for that. Do you try different control arm lengths on a cardboard cutout like the old school hot rod guys used to do it or do you prefer making and trying different arms?
@aaronpitts2762 I model the suspension in a software package. As long as I can hit the goal for camber gain and other geometric characteristics, then the arms are good.
Interesting. You are determining roll center through physical measurements of suspension movement, Not the usual line drawing method. I have never seen this before, Is there any way to confirm its accuracy ?
I've validated the measurements using two different 3D suspension packages (one was used for the final optimization) and it's bang on (there is less error in this method than in working out the pivots) and it's a lot easier to do. I'm just about to jump back into the frame build and make this come alive - I'll explain it in that video.
Tires are flexible and therefor so is the location of the contact patch. Is it a matter of putting it where it SHOULD be? Because the methods of drawing the lines to get the roll center height doesn't take flexure into consideration either. Comments?
In the first approximation you would ignore tire data because it's very difficult to come by the information you'd need for a street tire. We also ignore the bushing compliance. With a fully developed force model, these can be evaluated. So this is a first approximation - it gets you in the game. Tuning is the next step. In truth, this is about as far anyone can get in their garage.
at the end you mention you had a certain camber movement in mind per inch of wheel bump/droop. Was that number based on load transfer in curves and subsequent movement of the control arms?
Doesn't the lower control arm being pivoted forward of the centerline create anti-dive, similar to the rear lower arm being tilted rearward? The upper arm on an angle shouldn't have the same affect, but being rotated backwards from the main pivot position of the lower arm as you have it would create caster gain throughout the travel. Unless I'm missing something I can't see the relation to the upper arm affecting the dive characteristics.
Not sure I understand what you mean - the LCA relationship to the hub centreline creates caster not anti-dive. Anti-dive is created by having the rear upper control arm bushing lower than the front. This creates a side view swing arm (like a roll centre if you like - but in side view).
@@ThrottleStopGarage Castor is the angle between the 2 pivot points on the hub centerline, the upper ball joint being rearward of the lower ball joint will give positive caster. With the upper arm having that much of an angle difference, on compression the upper ball joint will move further rearward from the lower creating more caster. In this specific example it is also likely to effect bump steer by increasing toe on that wheel under compression, and toe out would likely happen on the inside wheel on extension. Having driven track cars the bump steer is undesirable in that situation. - But this might actually do the opposite, would be interesting to test. I probably don't fully understand the relation of the UCA in this equation but If you picture the extreme example of the LCA mounted at a 45degree with the front pivot above the rear, the car would compress the arm heavily under breaking from it pushing backwards on the arm. If the arm was rotated 45 degrees forward the braking action wouldn't move the arm in compression, it would in fact extend the arm. (at the extreme 45 degree angle) It would be fascinating to see measurements of the toe as well as castor change throughout the front suspension travel as well with your setup. It is very well thought out for a visual reference.
@@CBOCalgary Yes, of course - I was trying to work out what you were getting at. There is castor gain with compression and loss with extension. This is by design - that's what the Corvette C4 suspension is designed to do (as is the C5, C6. C7, C8...and most others I have seen). Bump steer is virtually zero. I've measured it. I could add this to the next instalment. You could also tilt the LCA up at the rear - but this suspension was not designed to have a tilted LCA. It's complicated - but generally anti-dive is used to control body pitch during braking by transferring load through the suspension links and arms instead of through the springs/damper. If you go above the 100% line you can induce jacking under braking. Sure - not sure why you'd do that. I'm just finishing optimizing the suspension and will see what works for explaining some of these behaviours. I'm more concerned with packaging right now to be honest.
Love the video. I did notice you never went into an in depth discussion as to what caused the binding the first time you put it on the bench. I thought I might give a little explanation. If you know this already then just ignore me. But in order for a double wishbone to articulate freely without binding it is required that the 4 frame mount pivot points (2 on upper and 2 on lower control arms) must be exactly in the same plane. If any 1 is slightly misaligned from the other 3 then there will always be some binding just as you demonstrated.
Sorry Alan - I thought I'd responded to you already. If I understand what you're saying (it's been a long week - so apologies for the foggy brain). Most SLA suspensions do not have the two planes coplanar. What happens in that case is as the suspension say goes into bump, the upright must rotate in order to account for this - so in my case you gain caster in bump (as well as camber - but we're looking at the side view). The binding in my case was caused by the suspension locating positions being in a position that the upper ball joint locked (the pin goes through a slot in the cup) and it was hitting that point due to too much caster. I dialed that out a little at a time until it was happy (and we were on the targets) - I also had issues at one point with the tie rod end spherical joint running out of travel (that is the picture I used for the thumbnail). So I changed a few more angles and got a different spherical bearing and it is now fine. The entire suspension is free of bind through an arc that is greater than the shock travel. This is optimal. Suspension member must never be stressed due to mechanical binding.
@@ThrottleStopGarage I get what your saying. Totally get how yours bound up now. I wasn’t referring to the camber gain. The point I was referring too was the upper and lower control arms frame mount pivot points (not the spindle side) must always lie within the same plane in order to articulate. Basically if one of those point gets moved to a point they are no longer on the same plane then your suspension will bind. The farther off plane it is the less it will articulate before binding. It’s like adding a flat to a 4 link to add preload, more than 2 flats and you essentially shift one point too far out of plane and it starts to bind.
@@ThrottleStopGarage I'm still not seeing it for some reason. It's 0.060" horizontal movement for 1" vertical movement so I see where it would be 1.7" inboard for 28" vertical? But is it getting mirrored or rotated 90deg somewhere?
@@jimdaug It's how much the tire contact patch is moving laterally. So we're measuring the slope of the force line. The tire movement creates the force. You can mirror the line if you like - but it intersects at the vehicle centre line unless your suspension is asymmetrical. So the distance is from the contact patch to the centre line.
Yes - the basic geometries are summed up by this point...it's understood that the contact patch is an area and is mobile etc. The geometry still applies to this point.
Hi sir how can I get in touch with you I need some advice and help... I swapped my front axle assy with bigger and heavier ones like the whole front axle assembly(bigger brakes calipers n hub stub axle) weight increased by almost double or triple even...i also used new n heavier ball joints in exact same position after I did this the car is very much unpredictable over bumps even minor unevenness of smooth roads I need to know where is the problem as I m using exact same joint positions
Oh my - that doesn't sound good. My guess is you've introduced some bump steer. Check to see if you have any change in toe when you raise a wheel. If you do - then the unevenness is steering your car.
Thank you so much for the reply...i might have introduced it I am not sure.... Here mechanics aren't that knowledgeable... But the alignment guy says the angles are ok i hv to jack it up and check for this(toe) .... How do I get around this problem.... Will stiffer springs help... I hv a rigid damper already and this helped a lot.... But didn't fix it completely
You are brave for taking this on as an amateur! I think the engineering perspective is a whole lot more complicated than what you are presenting here. Roll center height doesn't mean much by itself. The first engineer may not have been completely wrong in his assessment that your original geometry is perfectly adequate for a classic car. I thought you were building a "stock" looking classic car, with a little something extra, right? Not a full on race car? The optimization comes down to knowing precisely what you are trying to achieve. If you are planning on having stiff racing suspension, you need a low roll center to maintain compliance/ allow some down jacking while cornering. But if you want a softer, more comfortable ride with controlled roll characteristics, you need a higher roll center than the race car. (You said in the video that lower roll center meant less down jacking, but I think you got that backwards. The further apart the roll center and the center of gravity is, the more jacking you get. When they are the same, you have zero jacking...In my mind the "roll center" is more like "Moment of zero roll". As a casual onlooker, I'm just observing that you are diving into measuring the geometry, but you didn't explain what end result is that you are trying to achieve. I've been watching for awhile, but I don't think I've seen you explain the overall vision for the car and how you expect it to perform. Race car in old car body? Old car that handles like a modern car?
Fredy - you missed the "clueless" in front of amateur! Thanks for the detailed feedback - I learn best hands-on and I really soak up the feedback...I am here to LEARN - so I really appreciate the sharing. I had a long sit-down (well a drive through Saskatchewan) with a great friend and he asked the same question - no BS - it's not a race car, it's not meant to be a classic car that you drive to show off either. The "Vision" is to do two things - it must be able to make me smile at a track day (and this car was a restored version with all the normal suspension mods before all of this) and it must be a quick road car. So in your words - "Old car that handles like a modern car." that's the vision. Look at Cyan Racing's P1800 - or a Singer Porsche 911 - that's the goal. Whatever that is called now.
You bet Chris - I'm just optimizing right now and will cover how that was all done and formalizing the design. Working on it this weekend (the car...the video will follow).
Hi John, It would alter the outboard (wheel side) pick up points. Now if the tire is just a little larger the impact would be "small" - though it's hard to say without looking at the suspension. The inside points (frame) would move up a corresponding amount - so I'm going to guess that the largest effect would be the translocation of all points up by the difference in the radius of the tire.
Trick question - if you're going with a suspension transplant like I'm doing, then things like the shapes of the curves are set by the lengths of the links. You can tune the location of the curve. The rest really requires some experience that most don't have. Knowing that RCH is a summary of the links and their location is important. In general, you can look to what the donor car had for RC's and work from there. Then there is the RC height different front and rear - the roll axis and it's inclination. Most books tell you how to compute it - they don't explain how to optimize it. Camber curves are a little easier as there will be body roll and you design in the roll angle (it's not as complicated as I'm making it sound) depending on what you're designing the car to do (race car - street car - family car). We worked out an "optimum" camber rate of 0.6º/ inch and allowed the body to roll the same amount for 1.2º/inch which is sporty but not race car.
@@ThrottleStopGarage Not necessarily doing a transplant, but changing the geometry. To give you a parallell, though might seem as a silly one, i've been doing allot of rc racing. Might seem as toy cars, but you can basically set them up and change everything with little effort (springs, dampening, pack, ackerman, rollcenter, camber, camber gain, weight distribution etc.). Though, these racing cars are usually set up quite good from factory and then work your way from that to suit track, conditions and driving style. But my 1:1 car does not have a good baseline. So is there a way to guesstimate/calculate this? Guidelines maybe? How did you land on your result? My goal is also trackday/sporty. Issues i know i need to fix i.e. is its understeer tendencies, both from bumpsteer and positive camber gain.... (mcpherson) But would be nice to know if there is some good practices here. One with rc cars is do one change at a time and keep a log, but that's after the fact, hehe
As a novice, I'm still confused. You say the 2D method isn't helpful to you, but I'm having trouble understanding either the 2D or your method. What exactly are you showing is better from one versus the other? And what's the problem that you solved? I see that the graphs have different lines, but shifting those did what exactly for you suspension?
The idea is that the roll centre give you information on how the car will handle by summarizing the effect of the links. You CAN think of it as a point about which the car rolls - if that's helpful at the start. Now, if you're tuning, you'd have to work out the link location to try to figure out what is going to work. That ends up being a bit of a bother if you're working with a 3D suspension. As others have pointed out - you can always jam the numbers in a computer program...but they most likely won't tell you anything unless you understand what they mean. The method I've used (it's not my innovation) is to just measure the contact patch centre point's motion. You don't, for example, want the tire to move laterally 1" over a bump - the car would be impossible to drive. So I moved the suspension links to dial these properties out of the suspension and achieve better geometric properties.
@@capnthepeafarmer Unfortunately - the answer is it depends. It depends on the style of suspension you are working with, use of vehicle etc. For example, most strut suspensions have a RC that is at ground level. It's fine - but they will need an anti roll bar because the weight is transferred through the springs in roll. It is best to think of them as being the result of suspensions not the start of suspensions if that makes sense. So the RC I had from my report for the rear would indicate (at over 6") that the suspension was going to behave like a swing arm (so it will fold up under hard cornering or braking events) - I didn't get that picture when the report came in...working with a real suspension engineer...it was his first observation. That estimate turned out to be incorrect (thankfully). But I hope this makes the point - what those numbers are telling us is not what I've seen explained. It's not a dark-art - it's just poorly explained.
Capnthepeafarmer; To really simplify it. A good suspension allows your tires to have good ground contact in all positions. Your car have about four handsize contact surfaces to the ground, and all forces you apply to the car are depending on the friction to those surfaces. It's the most important part of the car. When you turn the wheels, you still want the contact surface to renmain good. If you apply some speed to the scenario, and add some roll so the outer wheel is pressed down , while the inner wheel springs out, you still want an optimum grip. Different cars look for different behavior. If you look at drift cars, they have a very optimized suspension setup for maximum grip at full turn, but not so good at normal driving.
The RC summarizes the links at any position of the suspension. You'd like the RC to move a little but not a lot as the car goes over bumps or is in roll. So what I had before would have been an adventure to drive, where what I'm trying to get is going to work very well.
One major thing I didn’t hear mentioned was the movement of the roll center. This is a huge part of vehicle stability. As the suspension moves (heave, roll, single wheel bump) the roll centers move. And if you don’t account for that they can move several inches in any direction. So yes the static ride height roll center location is important but the other side of the roll center tuning coin is the movement of that roll center.
The reason must books use the 2d method: it works. I've looked at suspension by calculating jacking force (looking at lateral tire force and instant centers, the roll center doesnt even need to exist for this analysis), and even tested them on a racecar. What I found: for the autocross racecar i was working on, a higher roll center than what i started with (all RCs were above ground) had the fastest lap times. There was a point of diminishing returns, where a higher roll center didn't improve lap times. We stopped here, as we expected lap times to start increasing (this was expected based on calculations, theory, etc). NOTE: I said lap times decreased. BUT, lateral acceleration actually went DOWN. What? It's because we reached the peak acceleration FASTER; the change of acceleration per unit time, aka jerk. Would you rather reach 1.7G in .5 seconds or 1.9G in .8 seconds? Of course, this depends on the track youre racing on. How was peak acceleration reached faster? From jacking force. To lift something up, there's an equal but opposite reaction. Lift the cars chassis up, there's an equal force down in the tire. More normal load on the tire = more potential for lateral grip. Why was peak lateral acceleration down? Because of jacking force, the CG was a bit higher, and more of the normal load was going through the suspension links (very stiff) instead of the springs/ damper (relativity less stiff). Anyway, a high roll center isn't necessarily bad. A lot or race cars have adjustable suspension pick up points (upper a arm to chassis, upper/lower a arms to chassis, or if you're complicated upper/lower a arm adjustments on chassis and upright) to help tune from track to track. If the car is "lazy", increases the IC height to increase jacking force. If it's too "twitchy" or "nervous", reduce the IC height. So why isn't there a way to simplify calculating jacking force, and adjust the cars "lazy" or "twitchy" handling? But there is: the 2d roll center! Most suspension books (Milliken, Carol Smith) are RACECAR oriented. Most racecars aren't built, they're bought with pre designed (possibly adjustable) suspension. It's easier to calculate roll stiffness in 2d for the whole car(roll stiffness taking into account springs, anti roll bar, roll center) than to calculate jacking force for the whole car. Car too lazy? Increase the RC height, and reduce ARB stiff to keep the same roll stiffness distribution, all calculated from simple 2d formulas. You may have a spreadsheet that does this quickly, or pre calculated before getting to the track. Its not exact, but useful and pretty close. Complicated simulations not needed
Thanks - great insight. My autocross car had three different suspensions in it in my first year! Trying to learn this material is not easy. Given that the RC is dictated by the links and their geometry - working the tire "harder" can produce faster times - no question. It's always a question of balance.
Having spent a fair bit of time trying to understand suspension I can understand your frustration, seemingly there are so many contradictory opinions. In abandoning the kinematics I think you've thrown the baby out with the bath water. You mention target camber gains but where did these come from? Are they calculated from your vehicle or a rule of thumb? You can work out an ideal camber curve by calculating the positive camber due to roll, this is what you're trying to cancel out to maintain an upright tyre and greatest contact patch. Most modern cars have a roll axis that slopes upwards towards the rear of the car, rear RC higher than the front, this allows the rear to catch up in roll to the front in taking a turn as the front initiates the turn because it steers, I've definitely felt this with changes I've made in rear roll centre. A roll centre closer to the CG is going to increase roll stiffness and decrease the job of anti roll bars later on, ARBs are in my opinion a case of robbing Peter to pay Paul and should be the last stage resort of suspension tuning, they don't fix the underlying problem. I think I'd calculate your ideal camber curves based on your track and suspension travel and try to match these as closely as possible, if the camber change is too rapid move the instant centre outwards and vice versa, look for a roll centre that doesn't move around too much, and target roll centres with the front lower than the back. With hubs, arms, and wheels already set I doubt you will be left with much choice. I'd not stress about the front rear slope of the top arm, not sure but take the average, roll centre can ignore this slope just as pitch centre calculations, anti squat/anti dive calculations ignore the angle of the arms viewed from the front. What historically have people done to tune your car? If it ran a beam axle and they've tuned them by moving watts linkages or panhard rod mountings these will tell you where your roll centre might be best, do they typically add anti-roll bars to the front or rear? If more rear bar move your roll centre up at this end if possible to increase roll stiffness. Can you add provision for multiple mounting locations particularly on the LCA to allow you to tune the roll centre and camber curves. Avoid toe out at the rear in bump, it means snap oversteer, some toe in will help stability. Long post...
First - thanks for taking the time to offer such helpful feedback - really appreciated. It was hard to decide the scope of the video - and I struggled with how to do the presentation to answer questions that I had when I started (failed...then restarted) this phase of the build. So I don't have any arguments - I did calculate the camber rates from body roll etc. The rear RC is below the front (for the reasons you indicated). I computed the anti-dive of the front and anti-squat of the rear and bump steer for both suspensions (rear will be toed in for sure). The suspension that was in the car was 1950's in design - the links were positioned to have as little camber gain as possible -lots of body roll - corner on the door handles (I did get used to that - but it's frightening for passengers). People have moved the UCA's down, but most have opted for lowering the car and adding a big ARB to what results in a suspension that is neither compliant or fun to drive. Steering was recirculating ball and not powered. So I decided to try to advance the design a few decades and get better handling. People also add ARB's to the live rear axle (panhard rod is stock). Anyway, I had all of this on the car and I tried different spring and shock combinations and it was fine - but not good by modern standards.
@@ThrottleStopGarage rear RC should be higher not lower is my understanding and this worked in my case, classic mini, this changes how quickly the back reacts to the turn, called roll frequency, before the change I felt through a longer corner the rear wouldn't decide whether it was going to grip or not, the raised rear RC fixed this and I've got some negative camber gain as well, previously none and rear RC at groundlevel. With a rwd be conservative with camber gain in the rear as otherwise as you squat on acceleration you'll lose contact patch as the camber goes negative and therefore losing traction, I'm used to thinking about fwd. Lowering a car on double wishbones and macpherson struts lowers the RC faster than the CG so the roll moment is actually increased, so gets more body roll, there are roll centre adapters that space the bottom ball joint down to correct this, this may explain the big rollbar on the front but they also might have been trying to solve oversteer problems. I've found explanations by the author and UA-camr Julien Edgar useful, he still has articles available at autospeed.com.au but it's a bit of a nightmare to navigate, otherwise he has a new book out and a UA-cam channel, I didn't find much in his previous "diy suspension development" though, but maybe because I'd already read a lot of it on autospeed.
@@arrindaley3714 This car is rear wheel drive/front engine with an independent rear suspension. The rear RC is often lower in this situation to give more roll to the rear and more grip in cornering. This is a pretty standard configuration.
Lowering the front roll center gives more steering under acceleration, but the car is less responsive. Ideal for smooth high grip tracks, with long sweeping corners. Raising the front roll center provides less steering when accelerating out of the corner, but the car feels more responsive and is less prone to traction roll. Best for high grip twisty tracks. Lower rear roll center improves grip under acceleration, but reduced grip when breaking. Helpful to avoid traction roll as you enter the corner and tracks with low grip to increase traction. Higher rear roll center gives you less grip under acceleration, but the car is more responsive. Works for high grip twisty tracks to reduce traction roll.
@@d.natrop4495 Sure for whatever car you're modifying...again - my point to making this video is that you'd need to know the specific kinematics of your car before making these sorts of generalizations.
It was a combination of a few things (isn't it always) one of the problems was in the bearing I was using...the rest was in the geometry. I fixed both - simple enough, but I'm glad I found it before I discovered this issue on the road.
You know, that's not a bad idea... small oven.. do the small stuff... reasonable plan that preserves garage space. I suspect powders never go bad for use years later... I wonder how the Harbor Freight kit (here in the US) compares with Eastwoods...
I think this comment is on the wrong video...I'm sure the Eastwood and HF powder coating systems are about the same. Everything packed into the oven and it's all in a box in the rafters.
I don't entirely understand how you calculate your roll center height using just the lateral positons of the contact patch center point in different suspension heights. You still need to know your instant center right? What's the formula you use? Great project, though! Keep it up! I like how you deal with all these problems.
Could I also add. Looking at the video the jack is trying to lift the bottom arm off the table after a certain point. Hence binding. Perhaps the jack under the spindle would be alot better
Fair point. I know when I was doing this that the forward jack position was harder to get a consistent purchase on the arm. The jack can't slip position (you'll notice that I've clamped the jack to the table to combat this). The binding was solved. Still waiting on parts (unbelievable but true) and I'd like to get the new design build in the next few months.
Nice one make a jig understand the way it works and will react. Work from data. Poor suspension can destroy a complete build. Have been worries about you being quiet and thanks goodness you seem to have been healthy.
Thanks - Now I'm going to see what I can get from a computer simulation. At least I will have an idea what's going on. I've spent a ton of time on this - but most is pointless for a video!
@@ThrottleStopGarage will need to watch the video again as I know nothing about IFS suspension I am a 4x4 solid axle type of guy. But newer vehicles are now all IFS so find this interesting. Like your ideas and thoughts.
There are other reasons that control lengths for formula cars. Yes, the relative length of the arms and separation of ball joints and mounting points control camber gain.
@@ThrottleStopGarage might take you a few trys to get it the way you want it but any of them would be better that how I'd do it. Although I don't seem to have the patience or maybe the attention span long enough to take on a project to this degree.
@@ThrottleStopGarage hopefully the pain will be worth it in the end. I might jmhave said something like this to you before but I tell my wife: I don't so much enjoy the working stuff, I enjoy the result. Also I couldn't have half what I have if I had to pay someone else to do all this one off stuff or just maintain my stock stuff.
Your key term is "instant!" At best a computer program/test mule can only predict what should happen under a specific set of conditions . . . then the real world happens. (For example, under a certain cornering load, the car body/frame/suspension points move in a certain way. Then it rains! So you don't get the same input loads at the lower speed and the car doesn't twist/roll so much and the "instant" that was calculated never happens. Etc., etc., etc..)
Yes - this is the real "nut that needs cracking" is "instant" - it applies for that geometry in that position...that's what you get. I think this point does get a little lost - instant things have a place but in reality it's things like RC migration and control that matter dynamically. I'm hoping to do a little computer simulation of this but without forcing myself to go through the exercise of working that suspension, I would not have gotten much out of the computer output (as is evidenced by me not understanding the report I got in the first place). I'm starting to get it now...slowly.
My current neighbour comes by when he needs to feel better about his project and the money he's spent. He also chirps me if there hasn't been a video in a while. He's fantastic.
Ok hold on. This is what you put on screen when you are talking about the rear RCH (0.55-0.43)/2=0.06 RCH =1.7 Where are you coming up with the RCH number ? You don't explain this at all, just say that the 2 dimensional renderings are wrong. I get that, I have a Mercedes 450slc drift car with a very similar SLA front end that's been desecrated by being lowered, the RCH is underground as far as I can tell and I'm extending the lower ball joint to compensate. I have a full subframe getting set up for control arm extension jigging and so I can run it through it's travel with a rack and pinion conversion to make sure bump steer is minimal. I just don't see where your measurements tell you where the RCH is.
@@ThrottleStopGarage where are you getting the slope (some call this the n-line btw) from? The slope is contact patch to IC, so how are you getting the IC from just lateral movement?
@@ThrottleStopGarage if you haven't seen a K&C rig before, they also get RC's from vertically displacing the wheel. But, i believe, they look at how the camber is changing. The wheel camber changes around the IC, and using a little geometry, they get the RC height/ lateral location. They're able to plot how the RC changes in bump and roll. Youre method seems super simple, but I'm not wrapping my head around it. Perhaps a short video/picture explaining?
On tarmac I like to have more Caster than KPI to save more camber during tight turns. A car with a low RC is easyer to ballance due to more use of swaybars because the lever from the CG is longer and make more roll. A solid axle with a high RC will allways try to flip the axle and leave the inner wheel in the air. On tarmac driving with a slight or more drift Ackermann is useless due to making toe out during drift and the car unstable in steering. Ackermann is for street use and posers only.
I couldn't get lower KPI's to package with brakes and things - the issue was that I wanted OTS parts for the brake hats, rotors, calipers and calipers mounts. I had to push the wheel out to 45 mm offset to get it to work. I would have liked a little less KPI for the reasons you mentioned.
Watch everything this guy has on suspension design, XF Motorsports ua-cam.com/video/VbReLNi2JP4/v-deo.html I am thinking your upper control arm pivot locations are in the wrong place, they look a little high to me judging from the video.
@@ThrottleStopGarage Shrugs...dunno just doesn't look right. look at a lot of sports car suspension pick up points and race cars. Like road course nascars. Easy to see. Watch how your camber changes through the stroke. It should go smoothly more and more negative.
@@ThrottleStopGarage just watched again, you shortened the arms to fi the track of the car? did you keep the same pick up points and the ratio of the two arms length compared to each other? looks like the upper control arm is too long compared to the lower arm. and the connection point is too high. How long are each of the control arms? from pivot to ball joint?
@@electronsmove The arms are stock - track was shortened by moving pick-up points. The ratio of the arms did not change - that's the shape of the curve. It's fine. I've optimized the package and am satisfied that it's fine.
@@electronsmove What I'm saying is that the optizing process did involve moving arms in 1 mm increments with as little as 5 mm needed to generate changes in how it all worked.
I get it. It’d be cool to do a video on king pin inclination. Very few people understand why it’s so important or how it affects your car when you steering and how it tries to make your steering return to center. Most people think it’s just the caster alone that causes this.
@@ezacher4634 Both are important as they are the distances to the centre of the unsprung mass. In the case of this car - the wheel base is the same as the Corvette/Supra that donated parts. In both instances, I did have to narrow the track to make them fit the car's outer dimensions.
No idea why you've chosen the least precise but most torturous approach possible. You can model a fully parametric 5-link suspension in Fusion in less than an hour. Couple minutes even, if you leave accurate bushing dimensions and pretty details for later. Have you even determined rough values for caster and anti-dive yet? If you're already running into binding issues you should probably sort out the basic geometry first - and replicate it in your test jig - then fine tune stuff like camber curves or bump steer. (which would also only take a fraction of the time in CAD) If you're aiming for a specific roll center height then you must have done all necessary calculations already. So why not use those results and get an actual base line for further tweaks. It's good that you mostly ignore that silly kinematic model, but it still looks like you're putting the cart before the pony.
Thanks for the feedback - super appreciated. Some details got left on the cutting room floor I'm afraid. The reason for doing this the way I'm doing it is to make a computer simulation make sense. I'm not a Fusion expert (really novice) - and my issue is that I learn best by doing. I'm not opposed to torture. I'm getting a good idea (far from expert) now about what's going on. I know where the caster and anti-dive should be and I've worked through all the gory details. I really could not have understood this without doing it this way - taking it into Fusion would have been a real struggle for me. Bump steer has been done for front and rear and while I have footage...I didn't see the point in showing it.
Ok, sounds good^^. I'd still highly recommend that you take the time and model the frame and suspension in Fusion. As you certainly know, you can't change a single parameter of a suspension without affecting everything else to some degree - I mean, some parameters even affect both axles. And that's where a proper cad model will save you countless hours. It is likely that at some point you'll have to make small adjustments to pickup points or the rack position for clearance/packaging or maybe just ease of fabrication. With everything modeled in CAD you can immediately see every change in the geometry caused by the modification you want to make and whether you can/need to compensate for it elsewhere. You also get the updated dimensions of any mounting bracket etc you might need for that change - without having to crawl around the car with a marker and pieces of cardboard. Just export it and have it water cut, or print out a template. The hours you may have to spend watching Fusion tutorials about parametric designs and master sketches etc will absolutely be worth it and pay dividends quickly. With an ambitious/complex and time/money consuming project like this, I would consider a good CAD design the most valuable tool.
@@protator Understood. I've got several designs under consideration right now in Fusion. I think I have the LCA mount worked out and I have about 4 different versions of the UCA mount. I just had a flash the other day that may solve the packaging problem. I ran into a bit of trouble with the AC Compressor when the suspension went down. I've got the frame components in Fusion and I'm trying to improve my CAD skills so I can easily model the suspension. I can pull suspension dimensions from the jig easily - I've been toying with the idea of buying a suspension analysis package that I can enter the points into and move them around and get an output that I understand.
@Throttle Stop Garage Can't really say much about specialized software ... a lot of people in the offroad/ hill climb community seem to use such programs with good results, but most posts I've seen in the past were about solid axle 3 or 4 link setups ... so no idea what's out there that works well for double A-arm/ind. 5 link. My somewhat generic advice would be to stick to the book ... estimate wheel loads, pick spring rates and a brake bias somewhere between 50/50 and the car's weight distribution, then choose anti-dive/squat values based on moderately sporty cars with similar weight and power. I've seen those numbers on all corners of the web at least for models like C3-C6 Corvettes, some Camaros and other american muscle. Then I'd block out the rear suspension and prioritize low scrub over a perfect camber curve unless the car's supposed to be a track toy. Toe link can go pretty much anywhere, you can always correct the toe curve by modifying the uprights ... on a street car I'd probably go with fixed toe-in and simply bolt it to the lower a-arm, tho. That detail can always be changed easily if necessary. So, yeah. Basically ignore the front end and focus on the rear suspension to get quick progress and keep the motivation up. And when it comes to the basic design, KISS is always a good choice. Try to avoid high torque loads on the frame rails, add additional box sections instead of excessively long mounting tabs ... again, the usual generic advice you'd expect from a rando on the net ^^.
@@ThrottleStopGarage just remembered a little something ... if you already know the width and offset of the rear wheels, check the scrub radius before you commit to the design. Seen a lot of bent or ripped off toe links on widebody conversions with too much offset. The forces on that little rod can get pretty high.
Wow. Thank you for describing a real world way of measuring suspension geometry. My daily is 122s (Amazon) and I am planning upgrades. I have watched many suspension videos and no one has really explained how to actually measure these values. It is really hard when the irrational bits of the grey matter work against to crush motivation. All the best and thanks for sharing your build, struggles and achievements. -Todd
Hi Todd - So happy this was helpful for you. Good luck in your project.
Good video, when I did mine, I built the controls arms out of PCV plumbing pipes in the mock up stage and tried out dozens pick-up points and control arm lengths. Made the virtual FVSA (Front View Swing Arm) with string to capture the roll center. Was manual and took awhile but had everything worked out before cutting and welding. Keep up the good work.
That's like the Stainforth string computer - works a treat from what I understand. I'm going to try some computer simulations in the near future. The major learning in this exercise was to drill into my head what happens when pivots change. I needed to experience this and get some hands-on before computer analysis output makes any sense. I learn mechanical things through my hands more than my head it seems.
Long time fan of your content. I've used your videos to help me tune my own skills over the years. You sure do make it look a lot easier then it really is! Thanks for the great work.
Thanks Josh - we're all works in progress! It's good to know that these videos have helped in some small way.
This is a great tutorial. I have been building race car suspensions since the 70s and am self taught. I have been frustrated trying to explain to people why youtube videos don't show the real process for scratch build suspensions. Henceforth I will just suggest anyone look at this first.
Thank you for the kind words. I hope the video proves useful.
This was great. Thank you for going into the weeds a bit for us. I'm following the graphs well enough. The one thing I'm missing is what you've been changing to get those new curves. Adjusting position of control arms? Angle? It would be interesting to know/see in the next suspension video. Or in a comment, whatever works. Otherwise, hey, anyone who's done enough work has run into "ope" moments like this. C'est la vie. We're here to see the process, and no one ever learned by doing something perfectly the first time.
Sorry - I'm changing the location of the links as well as caster/camber. In the graphs - the UCA/LCA both went down 2" (a little more for the front UCA pivot).
@@ThrottleStopGarage no need to apologize! Thanks for the insight. Hope the setup works well for you! Thanks for sharing.
Yeah, this was what I was missing as well, thank you for asking this question.
just remember this task is usually done by a person that has at least 8 years of schooling and then years and years of hands on experience and they STILL get it wrong... keep on it, it will eventually all fall into place, or apart... either way it makes good content!
Thanks. It's important to have that perspective. I'm still working away at it and will get it right in time.
Theres children figuring this out on RC cars… dont make it harder than it needs to be.
I played with the Corvette suspension a few years back. When I used the standard upright positions on bump and droop the wheel angle was terrible. This left me really confused then I re ran the tests with the body of the car rolling slightly and the suspension worked perfectly. What the designers realized that I had considered is that when you are driving straight it doesn't matter what angle the wheel patch is during bump etc. But in a corner is were it counts and that means you need to take into account body roll.
Love your videos keep up the great work.
I've got that in mind - it just doesn't pack into a video very well...I didn't show the side view geometries or any of the bump steer correction work (though watching my laser line move might be fun if you have a cat). So body roll is part of the overall vehicle design.
Love your work and and remember, everybody make mistakes, good people admit them, great people share them to help prevent others making the same mistake. Thanks for sharing.
Thanks. It's been a difficult experience and I'm learning. I just hope others will find the information useful.
I've run into the same frustration with 2D theory vs 3D reality. I'm glad you found a solution and shared your path. I'm still searching for mine.
You'll find it!
Good on you for taking this on and inspiring me to dream about spending months in the garage working through the madness also.
I hope it won't take months - I had a lot of learning to do and had no reference. This video should speed things up a little.
This pretty much blew my mind about roll centers, great content as always!
You took up several things I've never heard mentioned in all the suspension tips from the "pros".
Definitely invaluable information before I start hacking up my own frontend! :)
Glad it was helpful!
Was looking forward to this. Didn’t disappoint. Always nice to learn something new!
Ordered a mug
Thanks! This was a hard one to put together. Over a month of footage in 20 minutes.
incredible work. Thanks for sharing your struggle with the world, its gonna make things way easier for the people who follow. No one talks about suspension in such detail. Never change TSG.
I appreciate that!
Excellent explanation. I too had wondered about 2 diamesion instead of 3 . I did note at position approx 17.30 on video . When jacking up to test the front top mount bolt moves . Hopefully that didn't make a difference . Good work keep it up
Sharp eye! I had some issues but in the end I measured and remeasured...then optimized and remeasured again. It was solid with little variability. I did do a standard error of the measurement analysis (not reported) - I'm a scientist - so I can do boring things virtually forever.
Always look forward to learning something new from your videos. Thank you very much from the States
Glad you like them!
I expect this video will be very moving and dynamic!!!
ROFL !
@@ThrottleStopGarage Glad to see you able to keep up your enthusiasm for the project after you determined the mistakes/problems with the suspension. Many a mortal would have been crushed!
@@kentmckean6795 It's your support and the others that view this content that keep me pushing. So thanks.
I measured the roll center in a similar way when setting up the front suspension on my Volvo 142 race car to make sure I wasn't messing anything up by lowering it. Removed the springs and raised and lowered the car with the front tires on some rolling dollies, measuring the distance changes between the dollies in the process (and camber changes). That process made me realize that the slope of the CPC's trajectory is the real metric we're after, in addition to being easier to understand. A 10:1 slope, means 10% of the lateral loads are added to the tire vertically before the car even begins to lean and load the springs.
Absolutely.
Typically roll centers are tuned for weight transfer dynamics and caster/camber curves are tuned for tire patch optimization independently. Also, roll centers migrate, so I'm not sure using your method actually calculates the roll center at curb position. You may be right, I definitely want to look into it, but when I think about adding significant length to a single arm, you will massively change the results. Like if your upper arm is shorter than the lower, you'll camber in. But if it's longer, you'll camber out. You can be at the same RCH with both conditions, but have significantly different kinematics. I would say this is a good method for measuring the camber curve for sure.
It's a method that can be used and it works. All roll centres are instantaneous...and can be found using forced based or geometric methods. There are problems with definition, I'll cover this in my next video.
@ThrottleStopGarage I guess the issue I have with the method is, I can change the kinematics to get different outputs that you measure, but without changing the roll center height. What I'm not sure of, is if the outputs, though different, will still produce the same RCH.
Regardless of that, you seem to insinuate tuning your caster camber curves by adjusting RCH, but I typically treat them as completely decoupled metrics and tune them to do different things as they are responsible for different aspects of dynamics. I guess my comment just stems from your early reference of being critical of other RCH videos, but not actually teaching what they do or how to tune them, but what I got from this video seems to be very different than what's practiced in the vehicle dynamics community. Again, not saying your measurement method is wrong as I've never used that method to measure RCH and I plan on running the math as it's intriguing, but you seem to be the type of person who would take pleasure in technical discussions and I also enjoy learning new methods and techniques.
@@badinvestmentbuilds3823 Interesting - I'll try to explain the method first and then we can chat about the rest. The criticism of the other videos is that most share a simple 2D "join the dots" geometric method (nothing wrong there) with the assumption that measuring the intersections of these points is trivial. In practice, most suspensions are complex shapes and it's difficult to measure these points. If we're making some quick adjustments to arm chassis mount points, then there is an "easier" way. I glance off the topic of what a roll centre might represent as it opens up a bit of a pandora's box and I am not qualified to argue these points. There are many research articles and I think a deep dive into this is beyond what I would be comfortable trying to explain. In the vehicle dynamics community, they know what they're doing and they have studied the subject and worked with it - they have data-driven validation. No question. I guess I just didn't need to see another "join the dots" video. Lots of those, most are a little shy of what happens when you do these things. If I can only replicate what others have done, I'd just leave it. I also didn't find the locating of these points to be effecient to measure on a jig. The accumulation of error in their location (uncertainty) was pretty large (while it may have only been a few mm total).
Camber curves are a reflection of the relative lenth of the links and their angular relationship. That's it as far as I know. As my links were fixed, then what I can do is change their angular relationship. The effect of this change is to alter the location of the location of the instantaneous centre - as, at that instant in space, the suspension is moving as a swing arm fixed at that point. The validity of the method I present (which is not my creation - I was told to implement this method and the geometry makes sense) stems from the geometric reality that measuring the slope of the arc at the ground plane yields the line where the IC is located without having to measure the specifics of the arms that produced the IC location. The line that connects the contact patch centre to the IC is what is used as a geometric location for the RCH and for a symetrical suspension, the RCH is going to fall at the centre of the vehicle.
For any given RCH location for any specific combination of links and angular relationships between these links, there is a single IC. In other words, if you hold the angular relationship constant - there is only one possible RCH. The arms could be any length up to an including their intersection at the IC. If you fix RCH and vehicle track width - you can have many different arrangements that will generate the RCH - but they all must lie on the line described by the contact patch intersection and the RCH. This method seeks to measure the angle of that line.
If you change the angular relationship of the links to alter the camber curve...you will change the RCH. They are completely dependent. I hope I have explained this a little better.
@ThrottleStopGarage I think the part my brain is messing with me is that I can change the kinematics without changing the RCH, which will change the output the curves, but the new curves will likely point to the same location. I'll put a little math model together comparing them without measurement error. I'm the field, we would use inclinometers on the arms, articulate it on an alignment rack, and pop those numbers into a calculator, but that requires a bit more up front work. I'll reach out again if I can find the time to make the model and hopefully post up a video complimenting yours.
You're correct that there's a lot of 2D dot connecting that doesn't do much good to the average tuner. That has been a bit of inspiring for me to want to make a "Vehicle Dynamics for a 2 Car Garage" series lol. High level specific development with the average resources a DIYer has.
@badinvestmentbuilds3823 when you get the video done, let me know. Always interested in learning.
Just restarted redesigning my front end so well timed sir!
Almost done with seat mounts so almost time for wiring which I’m not quite but almost dreading.
Great - hope it goes well.
Thanks!
Thank you.
All hail the algorithm!
Thanks for showing us all the intricate details ;)
Details are where the learning takes place.
I like the different approach, if your adapting a suspension to a car this is a great method to get readings. One thing I would have touched on is the relationship of the roll center to center of gravity.
I'm sure I'll cover this in a little more detail when I get the new suspension built. The roll moment is critical to handling.
Great to see you back in the thick of it.
Thanks Hugh - I didn't stop...It's just so boring to film it all!
Love your work! Very informative, keep it up!
At one point you mention you don't want that binding in the suspension but I think what you're seeing is the anti-squat working, unless some other factor is affecting it. When your arms are parallel to each other they use the spring and damper to react the forces only. When they aren't parallel, you introduce a fore/aft movement of one ball joint relative to the other which causes the bind. For example, say your bottom arm is parallel to the ground and your upper arm is angled up at the front, like on your rig. When you look at the side view the lower ball joint will move vertically. The upper ball joint will move vertically but also rearwards. If the upper and lower joints weren't connected, it wouldn't be a problem but the upright has a fixed length. The distance has to remain fixed but the arms want to separate which causes binding. More compression, more bind. Essentially you're creating a geometric resistance, ie a stiffness proportional to the compression, similar to a variable rate spring that increases in stiffness with more bump.
Apologies, I meant your anti-dive, given your rig is the front suspension.
What was happening was because the suspension arms were in the "wrong" place the ball joint was locking and that was the problem. I also had issues with the caster and anti-dive and the spherical joint on the tie rod at one point. Taking the caster to a more sensible value (based on the camber rate target and other factors) and dialling out the anti-dive to a value for this car as well as changing the bearing to another model provides zero bind over more than the range of travel for the suspension. This was over a month of solid work to get done.
Wiseman...very genius gentleman...from malaysia...
Thanks - very kind. I've been to Malaysia and enjoyed my time and the durian!
@@ThrottleStopGarage 🤣🤣🤣🤣
Phenomenal video! Thank you so much for making it. I agree with your assessment of the information that's available on the internet on roll centers. Everybody is theorizing and nobody is actually testing anything . And as we all know mechanical suspension modifications are generally very expensive. I'm wondering through your analysis of the lateral movement of the track width. Wouldn't that increased lateral movement increase the track width and therefore improve handling?
Glad it was helpful! The track width does increase but only while in motion. The amount is very small.
Would love to steel a moment of your time to chat about suggestions that may help the install of my C4 components into my '54 Pontiac.
Your videos are great, they've helped significantly
Anytime. I'm just waiting on parts so I can finish up the install of the new design.
This is beautiful science! Great job! Subbed!
Thanks for the sub!
Very interesting video. I learned a lot and am doing a similar build with the C4 components I’m hoping to narrow into a Triumph GT6. Keep up the great videos.
Glad it was helpful!
static RC critique: totally agree. sounds like Carroll Smith's 'Tune to Win', great book with same story; or Carroll's bump steer+measure jig he describes in 'Prepare to Win'
I've read all of Carroll's books and keep his cheat sheet in my race car for track tuning.
23:58 You meant ideas for videos?
In line with the suspension topic: dampers? Depending on the characteristics of your car and the driving characteristics, how do you choose your damper? Valving, oil, ... ?
Ideas for the suspension...but thanks for the video ideas. We'll do these as we roll along. For sure full suspension characteristics - once I drive it. Then it's on to tuning suspensions for specific use.
Really look forward to your videos. Thank you so much !
Glad you like them!
8:45 I’m cracking up watching the magnet bounce off of everything. 😂
I've now got a fancy arrow pointer tip for the magnet wand!
Around 18:00 during compression, your camber is getting more positive, not negative. Is that expected? From your graphs, and maybe I wasn't paying attention, it's not clear if ride height positive is up or down.
It's hard to see the negative symbol...but it's going negative in bump past ride height and gaining camber in droop as normal.
@@ThrottleStopGarageit's not. Unless yours is different, I've owned that camber gauge for 12 years. It's going positive during compression. Once you cross zero while going up with the jack, you can see the little triangle on the inside is on top, and on the outside it's on bottom. That indicates positive.
@@grant1984 at 18:00 to 18:14 the suspension is being cycled through it's extension (droop) not compression and you're right it is gaining camber (positive) relative to ride height. That is expected. It crosses back through static ride height at 18:14 and from there it is in compression relative to ride height and is correctly reading negative camber. This is expected.
@@ThrottleStopGarage at 18:01 when you start to droop the suspension the gauge does indicate the outside ( to the left of the gauge reading) is higher than the inside (to the right of the reading). This is the opposite of my expectations.
@@ThrottleStopGarage at 18:20 it's reading 1.2 degrees positive based on my experience, and the owners manual for the gauge. Do you agree?
What is the intended use for this car? Thanks for a very useful video.
It will be driven on the street with occasional track days.
Bros over here giving ample information for free and people out here trying to charge ya for it 😂love the content
Just keeping it real.
@@ThrottleStopGarage much appreciated even the boring info is intriguing always trying to learn more
I’m curious if the stock C4 mounting points were utilized- would narrowing it have any affect on the roll center?
Absolutely - if all other geometries are kept equat - decreasing the track would lower the roll centre height.
@@ThrottleStopGarage thank you. I guess it’s time to learn the math and figure it out. I Enjoy your channel, keep it up. Thanks
This is such great detail! I can't imagine how it would have handled like that? Would it be on the limit you would notice!
I was trying to work that out...like what would my confidence be going into a fast corner at a track day?
@6:40 isnt this technically considered Caster? The angle of the upper arm dictates caster while the lower arms angle will dictate anti-dive? Maybe thats just for RC car suspension tuning.
No - caster is the displacement from vertical of the steering axis. UCA angle is use to create anti-dive. View the suspension in side view - there is also an IC in that dimension. So in this case the UCA is behind the centre defined by the LCA giving caster and the tilt of the UCA gives us anti-dive.
Hi, thanks. Just a few questions. Have you tried changing the hight location for the power steering rack and different rod lengths? Also, under full compression, Is it better to have the lower control arms at a level position? I know this requires moving the suspension and pivot pick up points but my theory is there will be less camber and toe deflection. Would this be the case?
Interesting question. I did alter the height of the rack and that does change bump steer characteristics - it has to sit on the plane described by the UCA/LCA pivots. Then there is where it sits in and out - this changes the Ackermann. I'm just working on this again in the new series...it's taken me a long time, but I'm back on the suspension right now. At the static ride height, for this suspension, the LCA upper surface is meant to be level. So as they compress, the angle increases. The issue is there is a fixed curve described by the length of the arms, the upright and the pick-up points. There is only so much you can change if you're working around the parts.
Ok thanks for that. Do you try different control arm lengths on a cardboard cutout like the old school hot rod guys used to do it or do you prefer making and trying different arms?
@aaronpitts2762 I model the suspension in a software package. As long as I can hit the goal for camber gain and other geometric characteristics, then the arms are good.
Interesting. You are determining roll center through physical measurements of suspension movement, Not the usual line drawing method. I have never seen this before, Is there any way to confirm its accuracy ?
I've validated the measurements using two different 3D suspension packages (one was used for the final optimization) and it's bang on (there is less error in this method than in working out the pivots) and it's a lot easier to do. I'm just about to jump back into the frame build and make this come alive - I'll explain it in that video.
Did you also work on the steering geos while on the bench ie bump steer and even ackermen
Yes - it's all been done. I have video of the bump steer, but it just didn't fit in this video. Everything was fine.
Keep going!! Good stuff!!
Thanks!
Tires are flexible and therefor so is the location of the contact patch. Is it a matter of putting it where it SHOULD be? Because the methods of drawing the lines to get the roll center height doesn't take flexure into consideration either. Comments?
In the first approximation you would ignore tire data because it's very difficult to come by the information you'd need for a street tire. We also ignore the bushing compliance. With a fully developed force model, these can be evaluated. So this is a first approximation - it gets you in the game. Tuning is the next step. In truth, this is about as far anyone can get in their garage.
at the end you mention you had a certain camber movement in mind per inch of wheel bump/droop. Was that number based on load transfer in curves and subsequent movement of the control arms?
Yes an approximation. At this level, getting it close is based on making reasonable assumptions about the rest of the vehicle dynamics.
Doesn't the lower control arm being pivoted forward of the centerline create anti-dive, similar to the rear lower arm being tilted rearward?
The upper arm on an angle shouldn't have the same affect, but being rotated backwards from the main pivot position of the lower arm as you have it would create caster gain throughout the travel.
Unless I'm missing something I can't see the relation to the upper arm affecting the dive characteristics.
Not sure I understand what you mean - the LCA relationship to the hub centreline creates caster not anti-dive. Anti-dive is created by having the rear upper control arm bushing lower than the front. This creates a side view swing arm (like a roll centre if you like - but in side view).
@@ThrottleStopGarage Castor is the angle between the 2 pivot points on the hub centerline, the upper ball joint being rearward of the lower ball joint will give positive caster. With the upper arm having that much of an angle difference, on compression the upper ball joint will move further rearward from the lower creating more caster. In this specific example it is also likely to effect bump steer by increasing toe on that wheel under compression, and toe out would likely happen on the inside wheel on extension. Having driven track cars the bump steer is undesirable in that situation. - But this might actually do the opposite, would be interesting to test.
I probably don't fully understand the relation of the UCA in this equation but If you picture the extreme example of the LCA mounted at a 45degree with the front pivot above the rear, the car would compress the arm heavily under breaking from it pushing backwards on the arm. If the arm was rotated 45 degrees forward the braking action wouldn't move the arm in compression, it would in fact extend the arm. (at the extreme 45 degree angle)
It would be fascinating to see measurements of the toe as well as castor change throughout the front suspension travel as well with your setup. It is very well thought out for a visual reference.
@@CBOCalgary Yes, of course - I was trying to work out what you were getting at. There is castor gain with compression and loss with extension. This is by design - that's what the Corvette C4 suspension is designed to do (as is the C5, C6. C7, C8...and most others I have seen). Bump steer is virtually zero. I've measured it. I could add this to the next instalment. You could also tilt the LCA up at the rear - but this suspension was not designed to have a tilted LCA. It's complicated - but generally anti-dive is used to control body pitch during braking by transferring load through the suspension links and arms instead of through the springs/damper. If you go above the 100% line you can induce jacking under braking. Sure - not sure why you'd do that. I'm just finishing optimizing the suspension and will see what works for explaining some of these behaviours. I'm more concerned with packaging right now to be honest.
Love the video. I did notice you never went into an in depth discussion as to what caused the binding the first time you put it on the bench. I thought I might give a little explanation. If you know this already then just ignore me. But in order for a double wishbone to articulate freely without binding it is required that the 4 frame mount pivot points (2 on upper and 2 on lower control arms) must be exactly in the same plane. If any 1 is slightly misaligned from the other 3 then there will always be some binding just as you demonstrated.
Sorry Alan - I thought I'd responded to you already. If I understand what you're saying (it's been a long week - so apologies for the foggy brain). Most SLA suspensions do not have the two planes coplanar. What happens in that case is as the suspension say goes into bump, the upright must rotate in order to account for this - so in my case you gain caster in bump (as well as camber - but we're looking at the side view). The binding in my case was caused by the suspension locating positions being in a position that the upper ball joint locked (the pin goes through a slot in the cup) and it was hitting that point due to too much caster. I dialed that out a little at a time until it was happy (and we were on the targets) - I also had issues at one point with the tie rod end spherical joint running out of travel (that is the picture I used for the thumbnail). So I changed a few more angles and got a different spherical bearing and it is now fine. The entire suspension is free of bind through an arc that is greater than the shock travel. This is optimal. Suspension member must never be stressed due to mechanical binding.
@@ThrottleStopGarage I get what your saying. Totally get how yours bound up now. I wasn’t referring to the camber gain. The point I was referring too was the upper and lower control arms frame mount pivot points (not the spindle side) must always lie within the same plane in order to articulate. Basically if one of those point gets moved to a point they are no longer on the same plane then your suspension will bind. The farther off plane it is the less it will articulate before binding. It’s like adding a flat to a 4 link to add preload, more than 2 flats and you essentially shift one point too far out of plane and it starts to bind.
I don't understand how you go from 0.06" contact patch movement to 1.7" roll center height at 12:11
That's the rate of change per inch - so then multiply the 0.06" by 1/2 the track (so for my car 28") - that gives you 1.68" - so 1.7" if you round.
@@ThrottleStopGarage I'm still not seeing it for some reason. It's 0.060" horizontal movement for 1" vertical movement so I see where it would be 1.7" inboard for 28" vertical? But is it getting mirrored or rotated 90deg somewhere?
@@jimdaug It's how much the tire contact patch is moving laterally. So we're measuring the slope of the force line. The tire movement creates the force. You can mirror the line if you like - but it intersects at the vehicle centre line unless your suspension is asymmetrical. So the distance is from the contact patch to the centre line.
Excellent video!
Thank you very much!
When you say "the center of the contact patch", that's the center of a theoretical ideal contact patch without tire flex, right?
Yes - the basic geometries are summed up by this point...it's understood that the contact patch is an area and is mobile etc. The geometry still applies to this point.
@@ThrottleStopGarage Thanks
Hang in there.
I still can’t get over the fact that corvette has such nice suspension. Many European exotics run with McPherson strut.
I know - the Corvette really is a great platform.
Hi sir how can I get in touch with you I need some advice and help...
I swapped my front axle assy with bigger and heavier ones like the whole front axle assembly(bigger brakes calipers n hub stub axle) weight increased by almost double or triple even...i also used new n heavier ball joints in exact same position after I did this the car is very much unpredictable over bumps even minor unevenness of smooth roads
I need to know where is the problem as I m using exact same joint positions
Oh my - that doesn't sound good. My guess is you've introduced some bump steer. Check to see if you have any change in toe when you raise a wheel. If you do - then the unevenness is steering your car.
Thank you so much for the reply...i might have introduced it I am not sure.... Here mechanics aren't that knowledgeable... But the alignment guy says the angles are ok i hv to jack it up and check for this(toe) .... How do I get around this problem.... Will stiffer springs help... I hv a rigid damper already and this helped a lot.... But didn't fix it completely
Pls let me know how I can get in touch with you over a call?
You are brave for taking this on as an amateur! I think the engineering perspective is a whole lot more complicated than what you are presenting here. Roll center height doesn't mean much by itself. The first engineer may not have been completely wrong in his assessment that your original geometry is perfectly adequate for a classic car. I thought you were building a "stock" looking classic car, with a little something extra, right? Not a full on race car?
The optimization comes down to knowing precisely what you are trying to achieve. If you are planning on having stiff racing suspension, you need a low roll center to maintain compliance/ allow some down jacking while cornering. But if you want a softer, more comfortable ride with controlled roll characteristics, you need a higher roll center than the race car. (You said in the video that lower roll center meant less down jacking, but I think you got that backwards. The further apart the roll center and the center of gravity is, the more jacking you get. When they are the same, you have zero jacking...In my mind the "roll center" is more like "Moment of zero roll".
As a casual onlooker, I'm just observing that you are diving into measuring the geometry, but you didn't explain what end result is that you are trying to achieve. I've been watching for awhile, but I don't think I've seen you explain the overall vision for the car and how you expect it to perform. Race car in old car body? Old car that handles like a modern car?
Fredy - you missed the "clueless" in front of amateur! Thanks for the detailed feedback - I learn best hands-on and I really soak up the feedback...I am here to LEARN - so I really appreciate the sharing. I had a long sit-down (well a drive through Saskatchewan) with a great friend and he asked the same question - no BS - it's not a race car, it's not meant to be a classic car that you drive to show off either. The "Vision" is to do two things - it must be able to make me smile at a track day (and this car was a restored version with all the normal suspension mods before all of this) and it must be a quick road car. So in your words - "Old car that handles like a modern car." that's the vision. Look at Cyan Racing's P1800 - or a Singer Porsche 911 - that's the goal. Whatever that is called now.
@@ThrottleStopGarage "restomod"
Great video 👍 Are you planning on going over what changes you made to get the RC where you wanted it?
You bet Chris - I'm just optimizing right now and will cover how that was all done and formalizing the design. Working on it this weekend (the car...the video will follow).
how much and in what way does your final setup change if you went to larger/smaller tire radius?
Hi John, It would alter the outboard (wheel side) pick up points. Now if the tire is just a little larger the impact would be "small" - though it's hard to say without looking at the suspension. The inside points (frame) would move up a corresponding amount - so I'm going to guess that the largest effect would be the translocation of all points up by the difference in the radius of the tire.
How do you calculate what rollcenter height, camber curves etc are ideal for you?
Trick question - if you're going with a suspension transplant like I'm doing, then things like the shapes of the curves are set by the lengths of the links. You can tune the location of the curve. The rest really requires some experience that most don't have. Knowing that RCH is a summary of the links and their location is important. In general, you can look to what the donor car had for RC's and work from there. Then there is the RC height different front and rear - the roll axis and it's inclination. Most books tell you how to compute it - they don't explain how to optimize it. Camber curves are a little easier as there will be body roll and you design in the roll angle (it's not as complicated as I'm making it sound) depending on what you're designing the car to do (race car - street car - family car). We worked out an "optimum" camber rate of 0.6º/ inch and allowed the body to roll the same amount for 1.2º/inch which is sporty but not race car.
@@ThrottleStopGarage
Not necessarily doing a transplant, but changing the geometry. To give you a parallell, though might seem as a silly one, i've been doing allot of rc racing. Might seem as toy cars, but you can basically set them up and change everything with little effort (springs, dampening, pack, ackerman, rollcenter, camber, camber gain, weight distribution etc.). Though, these racing cars are usually set up quite good from factory and then work your way from that to suit track, conditions and driving style. But my 1:1 car does not have a good baseline. So is there a way to guesstimate/calculate this? Guidelines maybe? How did you land on your result? My goal is also trackday/sporty. Issues i know i need to fix i.e. is its understeer tendencies, both from bumpsteer and positive camber gain.... (mcpherson) But would be nice to know if there is some good practices here. One with rc cars is do one change at a time and keep a log, but that's after the fact, hehe
Thank You for sharing your knowledge
Cheers
Thank you ….. this is incredibly helpful
Glad it was helpful!
Your mustache is trying to escape! I can't stop looking at it. Someone get hair and makeup in here! Good videos! :D
No man🤣🤣🤣 why???? Now I can't stop looking
LOL - I trimmed it a little for some segments. Now you have to find them. Winter is coming fast in Canada.
As a novice, I'm still confused. You say the 2D method isn't helpful to you, but I'm having trouble understanding either the 2D or your method. What exactly are you showing is better from one versus the other? And what's the problem that you solved? I see that the graphs have different lines, but shifting those did what exactly for you suspension?
The idea is that the roll centre give you information on how the car will handle by summarizing the effect of the links. You CAN think of it as a point about which the car rolls - if that's helpful at the start. Now, if you're tuning, you'd have to work out the link location to try to figure out what is going to work. That ends up being a bit of a bother if you're working with a 3D suspension. As others have pointed out - you can always jam the numbers in a computer program...but they most likely won't tell you anything unless you understand what they mean. The method I've used (it's not my innovation) is to just measure the contact patch centre point's motion. You don't, for example, want the tire to move laterally 1" over a bump - the car would be impossible to drive. So I moved the suspension links to dial these properties out of the suspension and achieve better geometric properties.
@@ThrottleStopGarage ok, so maybe I need to step back a little further, what is a GOOD roll center? Why is that good versus what you started with?
@@capnthepeafarmer Unfortunately - the answer is it depends. It depends on the style of suspension you are working with, use of vehicle etc. For example, most strut suspensions have a RC that is at ground level. It's fine - but they will need an anti roll bar because the weight is transferred through the springs in roll. It is best to think of them as being the result of suspensions not the start of suspensions if that makes sense. So the RC I had from my report for the rear would indicate (at over 6") that the suspension was going to behave like a swing arm (so it will fold up under hard cornering or braking events) - I didn't get that picture when the report came in...working with a real suspension engineer...it was his first observation. That estimate turned out to be incorrect (thankfully). But I hope this makes the point - what those numbers are telling us is not what I've seen explained. It's not a dark-art - it's just poorly explained.
Capnthepeafarmer; To really simplify it. A good suspension allows your tires to have good ground contact in all positions. Your car have about four handsize contact surfaces to the ground, and all forces you apply to the car are depending on the friction to those surfaces. It's the most important part of the car. When you turn the wheels, you still want the contact surface to renmain good. If you apply some speed to the scenario, and add some roll so the outer wheel is pressed down , while the inner wheel springs out, you still want an optimum grip. Different cars look for different behavior. If you look at drift cars, they have a very optimized suspension setup for maximum grip at full turn, but not so good at normal driving.
After watching the video I have one big question. How roll center height matter or effect on supension performance or car handling?
The RC summarizes the links at any position of the suspension. You'd like the RC to move a little but not a lot as the car goes over bumps or is in roll. So what I had before would have been an adventure to drive, where what I'm trying to get is going to work very well.
That table is certainly earning its keep
Priceless really.
20:10 *This* you need a video about this: how the roll center influences...everything?
I'm not sure I have the knowledge to make this video.
One major thing I didn’t hear mentioned was the movement of the roll center. This is a huge part of vehicle stability. As the suspension moves (heave, roll, single wheel bump) the roll centers move. And if you don’t account for that they can move several inches in any direction. So yes the static ride height roll center location is important but the other side of the roll center tuning coin is the movement of that roll center.
True - those dynamical values were checked in the computer model. The migration as well as the rate of migration are both really important.
The reason must books use the 2d method: it works.
I've looked at suspension by calculating jacking force (looking at lateral tire force and instant centers, the roll center doesnt even need to exist for this analysis), and even tested them on a racecar. What I found: for the autocross racecar i was working on, a higher roll center than what i started with (all RCs were above ground) had the fastest lap times. There was a point of diminishing returns, where a higher roll center didn't improve lap times. We stopped here, as we expected lap times to start increasing (this was expected based on calculations, theory, etc).
NOTE: I said lap times decreased. BUT, lateral acceleration actually went DOWN. What? It's because we reached the peak acceleration FASTER; the change of acceleration per unit time, aka jerk. Would you rather reach 1.7G in .5 seconds or 1.9G in .8 seconds? Of course, this depends on the track youre racing on.
How was peak acceleration reached faster? From jacking force. To lift something up, there's an equal but opposite reaction. Lift the cars chassis up, there's an equal force down in the tire. More normal load on the tire = more potential for lateral grip.
Why was peak lateral acceleration down? Because of jacking force, the CG was a bit higher, and more of the normal load was going through the suspension links (very stiff) instead of the springs/ damper (relativity less stiff).
Anyway, a high roll center isn't necessarily bad. A lot or race cars have adjustable suspension pick up points (upper a arm to chassis, upper/lower a arms to chassis, or if you're complicated upper/lower a arm adjustments on chassis and upright) to help tune from track to track.
If the car is "lazy", increases the IC height to increase jacking force. If it's too "twitchy" or "nervous", reduce the IC height.
So why isn't there a way to simplify calculating jacking force, and adjust the cars "lazy" or "twitchy" handling? But there is: the 2d roll center!
Most suspension books (Milliken, Carol Smith) are RACECAR oriented. Most racecars aren't built, they're bought with pre designed (possibly adjustable) suspension. It's easier to calculate roll stiffness in 2d for the whole car(roll stiffness taking into account springs, anti roll bar, roll center) than to calculate jacking force for the whole car. Car too lazy? Increase the RC height, and reduce ARB stiff to keep the same roll stiffness distribution, all calculated from simple 2d formulas. You may have a spreadsheet that does this quickly, or pre calculated before getting to the track. Its not exact, but useful and pretty close. Complicated simulations not needed
Thanks - great insight. My autocross car had three different suspensions in it in my first year! Trying to learn this material is not easy. Given that the RC is dictated by the links and their geometry - working the tire "harder" can produce faster times - no question. It's always a question of balance.
Having spent a fair bit of time trying to understand suspension I can understand your frustration, seemingly there are so many contradictory opinions. In abandoning the kinematics I think you've thrown the baby out with the bath water.
You mention target camber gains but where did these come from? Are they calculated from your vehicle or a rule of thumb? You can work out an ideal camber curve by calculating the positive camber due to roll, this is what you're trying to cancel out to maintain an upright tyre and greatest contact patch.
Most modern cars have a roll axis that slopes upwards towards the rear of the car, rear RC higher than the front, this allows the rear to catch up in roll to the front in taking a turn as the front initiates the turn because it steers, I've definitely felt this with changes I've made in rear roll centre.
A roll centre closer to the CG is going to increase roll stiffness and decrease the job of anti roll bars later on, ARBs are in my opinion a case of robbing Peter to pay Paul and should be the last stage resort of suspension tuning, they don't fix the underlying problem.
I think I'd calculate your ideal camber curves based on your track and suspension travel and try to match these as closely as possible, if the camber change is too rapid move the instant centre outwards and vice versa, look for a roll centre that doesn't move around too much, and target roll centres with the front lower than the back. With hubs, arms, and wheels already set I doubt you will be left with much choice.
I'd not stress about the front rear slope of the top arm, not sure but take the average, roll centre can ignore this slope just as pitch centre calculations, anti squat/anti dive calculations ignore the angle of the arms viewed from the front.
What historically have people done to tune your car? If it ran a beam axle and they've tuned them by moving watts linkages or panhard rod mountings these will tell you where your roll centre might be best, do they typically add anti-roll bars to the front or rear? If more rear bar move your roll centre up at this end if possible to increase roll stiffness.
Can you add provision for multiple mounting locations particularly on the LCA to allow you to tune the roll centre and camber curves.
Avoid toe out at the rear in bump, it means snap oversteer, some toe in will help stability.
Long post...
First - thanks for taking the time to offer such helpful feedback - really appreciated. It was hard to decide the scope of the video - and I struggled with how to do the presentation to answer questions that I had when I started (failed...then restarted) this phase of the build. So I don't have any arguments - I did calculate the camber rates from body roll etc. The rear RC is below the front (for the reasons you indicated). I computed the anti-dive of the front and anti-squat of the rear and bump steer for both suspensions (rear will be toed in for sure). The suspension that was in the car was 1950's in design - the links were positioned to have as little camber gain as possible -lots of body roll - corner on the door handles (I did get used to that - but it's frightening for passengers). People have moved the UCA's down, but most have opted for lowering the car and adding a big ARB to what results in a suspension that is neither compliant or fun to drive. Steering was recirculating ball and not powered. So I decided to try to advance the design a few decades and get better handling. People also add ARB's to the live rear axle (panhard rod is stock). Anyway, I had all of this on the car and I tried different spring and shock combinations and it was fine - but not good by modern standards.
@@ThrottleStopGarage rear RC should be higher not lower is my understanding and this worked in my case, classic mini, this changes how quickly the back reacts to the turn, called roll frequency, before the change I felt through a longer corner the rear wouldn't decide whether it was going to grip or not, the raised rear RC fixed this and I've got some negative camber gain as well, previously none and rear RC at groundlevel.
With a rwd be conservative with camber gain in the rear as otherwise as you squat on acceleration you'll lose contact patch as the camber goes negative and therefore losing traction, I'm used to thinking about fwd.
Lowering a car on double wishbones and macpherson struts lowers the RC faster than the CG so the roll moment is actually increased, so gets more body roll, there are roll centre adapters that space the bottom ball joint down to correct this, this may explain the big rollbar on the front but they also might have been trying to solve oversteer problems.
I've found explanations by the author and UA-camr Julien Edgar useful, he still has articles available at autospeed.com.au but it's a bit of a nightmare to navigate, otherwise he has a new book out and a UA-cam channel, I didn't find much in his previous "diy suspension development" though, but maybe because I'd already read a lot of it on autospeed.
@@arrindaley3714 This car is rear wheel drive/front engine with an independent rear suspension. The rear RC is often lower in this situation to give more roll to the rear and more grip in cornering. This is a pretty standard configuration.
Lowering the front roll center gives more steering under acceleration, but the car is less responsive. Ideal for smooth high grip tracks, with long sweeping corners.
Raising the front roll center provides less steering when accelerating out of the corner, but the car feels more responsive and is less prone to traction roll. Best for high grip twisty tracks.
Lower rear roll center improves grip under acceleration, but reduced grip when breaking. Helpful to avoid traction roll as you enter the corner and tracks with low grip to increase traction.
Higher rear roll center gives you less grip under acceleration, but the car is more responsive. Works for high grip twisty tracks to reduce traction roll.
@@d.natrop4495 Sure for whatever car you're modifying...again - my point to making this video is that you'd need to know the specific kinematics of your car before making these sorts of generalizations.
So what was causing the suspension to bind?
It was a combination of a few things (isn't it always) one of the problems was in the bearing I was using...the rest was in the geometry. I fixed both - simple enough, but I'm glad I found it before I discovered this issue on the road.
You are having way to much fun
Not really. It's more like hitting your head against something until you either understand it or take the car to the wreckers.
@@ThrottleStopGarage I understand, been there, done that. You'll get through it and the car and you will be better for it
What software are you using?
Racing Aspirations for 2D along with others. CAD is Fusion.
You know, that's not a bad idea... small oven.. do the small stuff... reasonable plan that preserves garage space. I suspect powders never go bad for use years later...
I wonder how the Harbor Freight kit (here in the US) compares with Eastwoods...
I think this comment is on the wrong video...I'm sure the Eastwood and HF powder coating systems are about the same. Everything packed into the oven and it's all in a box in the rafters.
@@ThrottleStopGarage Thanks! You're right! I think the video auto-advanced and I didn't notice. Thanks for the reply!
Awesome stuff. Thank you man.
Glad you liked it.
I'd *love* to see you build a bike. Like a 3 cyl twin turbo cafe racer
Few more cars need building first...oh and my day job. It would be fun though.
You're a magician if you can get bodyroll with zero steering input
Yes you would be...and I never said you can get body roll without steering input.
I don't entirely understand how you calculate your roll center height using just the lateral positons of the contact patch center point in different suspension heights. You still need to know your instant center right? What's the formula you use? Great project, though! Keep it up! I like how you deal with all these problems.
Got it now. I'm amazed at how simple this method is
@@Everude3000 how?
Could I also add. Looking at the video the jack is trying to lift the bottom arm off the table after a certain point. Hence binding. Perhaps the jack under the spindle would be alot better
Fair point. I know when I was doing this that the forward jack position was harder to get a consistent purchase on the arm. The jack can't slip position (you'll notice that I've clamped the jack to the table to combat this). The binding was solved. Still waiting on parts (unbelievable but true) and I'd like to get the new design build in the next few months.
Have you reached out to Quinn from BlondiHacks about some of your machining and other automotive projects? She’s a fellow Canadian and very into cars.
Nope - not to cast shade, but most YT people I've contacted never get back to me.
Love it!
Thanks!!
Thanks
You're welcome
Nice one make a jig understand the way it works and will react. Work from data.
Poor suspension can destroy a complete build.
Have been worries about you being quiet and thanks goodness you seem to have been healthy.
Thanks - Now I'm going to see what I can get from a computer simulation. At least I will have an idea what's going on. I've spent a ton of time on this - but most is pointless for a video!
@@ThrottleStopGarage will need to watch the video again as I know nothing about IFS suspension I am a 4x4 solid axle type of guy.
But newer vehicles are now all IFS so find this interesting. Like your ideas and thoughts.
Isn't the camber curve a function of the length of the control arms? That's why formula cars have as long arms as possible.
There are other reasons that control lengths for formula cars. Yes, the relative length of the arms and separation of ball joints and mounting points control camber gain.
You'll get it perfect
I hope.
@@ThrottleStopGarage might take you a few trys to get it the way you want it but any of them would be better that how I'd do it. Although I don't seem to have the patience or maybe the attention span long enough to take on a project to this degree.
@@AndyFromm Indeed - while it's been a struggle, I do think the learning is worth the effort. It's not worth the pain, but that's another story.
@@ThrottleStopGarage hopefully the pain will be worth it in the end. I might jmhave said something like this to you before but I tell my wife: I don't so much enjoy the working stuff, I enjoy the result. Also I couldn't have half what I have if I had to pay someone else to do all this one off stuff or just maintain my stock stuff.
Your key term is "instant!" At best a computer program/test mule can only predict what should happen under a specific set of conditions . . . then the real world happens. (For example, under a certain cornering load, the car body/frame/suspension points move in a certain way. Then it rains! So you don't get the same input loads at the lower speed and the car doesn't twist/roll so much and the "instant" that was calculated never happens. Etc., etc., etc..)
Yes - this is the real "nut that needs cracking" is "instant" - it applies for that geometry in that position...that's what you get. I think this point does get a little lost - instant things have a place but in reality it's things like RC migration and control that matter dynamically. I'm hoping to do a little computer simulation of this but without forcing myself to go through the exercise of working that suspension, I would not have gotten much out of the computer output (as is evidenced by me not understanding the report I got in the first place). I'm starting to get it now...slowly.
I wish we were next door neighbors lol. We would get into pretty solid nerd sessions
My current neighbour comes by when he needs to feel better about his project and the money he's spent. He also chirps me if there hasn't been a video in a while. He's fantastic.
Ok hold on. This is what you put on screen when you are talking about the rear RCH
(0.55-0.43)/2=0.06
RCH =1.7
Where are you coming up with the RCH number ? You don't explain this at all, just say that the 2 dimensional renderings are wrong. I get that, I have a Mercedes 450slc drift car with a very similar SLA front end that's been desecrated by being lowered, the RCH is underground as far as I can tell and I'm extending the lower ball joint to compensate. I have a full subframe getting set up for control arm extension jigging and so I can run it through it's travel with a rack and pinion conversion to make sure bump steer is minimal. I just don't see where your measurements tell you where the RCH is.
I was also wondering what the actual equation is.
You then take the slope (0.06) and multiply it by half the track to get vehicle centre. I did cover this.
(slope) * 1/2 track
@@ThrottleStopGarage where are you getting the slope (some call this the n-line btw) from? The slope is contact patch to IC, so how are you getting the IC from just lateral movement?
@@ThrottleStopGarage if you haven't seen a K&C rig before, they also get RC's from vertically displacing the wheel. But, i believe, they look at how the camber is changing. The wheel camber changes around the IC, and using a little geometry, they get the RC height/ lateral location. They're able to plot how the RC changes in bump and roll.
Youre method seems super simple, but I'm not wrapping my head around it. Perhaps a short video/picture explaining?
Lol, it may be painful, time consuming, and tedious, but you know what it will also be when you're done? "Correct."
I sure hope so.
Hi mate, I have been subscribed for ages. I noticed today that I had been unsubscribed, maybe you tube playing games?
For sure - the do periodic "clean out" sweeps. As long was we all hail the algorithm we'll be fine.
NURD, I get it I cut my race car in half and rebuilt too lol/
One day it will drive again. One day.
On tarmac I like to have more Caster than KPI to save more camber during tight turns. A car with a low RC is easyer to ballance due to more use of swaybars because the lever from the CG is longer and make more roll. A solid axle with a high RC will allways try to flip the axle and leave the inner wheel in the air. On tarmac driving with a slight or more drift Ackermann is useless due to making toe out during drift and the car unstable in steering. Ackermann is for street use and posers only.
I couldn't get lower KPI's to package with brakes and things - the issue was that I wanted OTS parts for the brake hats, rotors, calipers and calipers mounts. I had to push the wheel out to 45 mm offset to get it to work. I would have liked a little less KPI for the reasons you mentioned.
@@ThrottleStopGarage Just maximize caster and join the gym :-)
Haha, I'm still confused. 🤔
All hail Al Gore’s rhythm!
LOL
Don't say "nobody wants to watch this", I would watch a 1500 hour playlist on roll center if you were teaching it
Thanks - much appreciated. I'm working my way through optimizing now. Not sure if there is much to say about that either, but I'll do something.
Watch everything this guy has on suspension design,
XF Motorsports
ua-cam.com/video/VbReLNi2JP4/v-deo.html
I am thinking your upper control arm pivot locations are in the wrong place, they look a little high to me judging from the video.
I've watched his stuff. How could you possibly tell if my UCA was in the wrong place by eye!
@@ThrottleStopGarage Shrugs...dunno just doesn't look right.
look at a lot of sports car suspension pick up points and race cars. Like road course nascars. Easy to see.
Watch how your camber changes through the stroke. It should go smoothly more and more negative.
@@ThrottleStopGarage just watched again, you shortened the arms to fi the track of the car? did you keep the same pick up points and the ratio of the two arms length compared to each other? looks like the upper control arm is too long compared to the lower arm.
and the connection point is too high.
How long are each of the control arms? from pivot to ball joint?
@@electronsmove The arms are stock - track was shortened by moving pick-up points. The ratio of the arms did not change - that's the shape of the curve. It's fine. I've optimized the package and am satisfied that it's fine.
@@electronsmove What I'm saying is that the optizing process did involve moving arms in 1 mm increments with as little as 5 mm needed to generate changes in how it all worked.
Your drawing accidentally has a negative king pin inclination.
It's just a doodle.
I get it. It’d be cool to do a video on king pin inclination. Very few people understand why it’s so important or how it affects your car when you steering and how it tries to make your steering return to center. Most people think it’s just the caster alone that causes this.
You mean you can't just take a Corvette suspension and throw it on another chassis and say "RACE CAR!"? Hmmm.who knew?
If you get it in the right location and the "professionals" didn't mess it up...then yup.
@@ThrottleStopGarage How much does chassis width and wheelbase matter?
@@ezacher4634 Both are important as they are the distances to the centre of the unsprung mass. In the case of this car - the wheel base is the same as the Corvette/Supra that donated parts. In both instances, I did have to narrow the track to make them fit the car's outer dimensions.
A standard Volvo chassi with standard suspension, but tuned by some swedish guy will beat you mismatch of a Volvo anytime
Nope - Swedes don't have different laws of physics I'm afraid.
No idea why you've chosen the least precise but most torturous approach possible.
You can model a fully parametric 5-link suspension in Fusion in less than an hour. Couple minutes even, if you leave accurate bushing dimensions and pretty details for later.
Have you even determined rough values for caster and anti-dive yet? If you're already running into binding issues you should probably sort out the basic geometry first - and replicate it in your test jig - then fine tune stuff like camber curves or bump steer. (which would also only take a fraction of the time in CAD)
If you're aiming for a specific roll center height then you must have done all necessary calculations already. So why not use those results and get an actual base line for further tweaks.
It's good that you mostly ignore that silly kinematic model, but it still looks like you're putting the cart before the pony.
Thanks for the feedback - super appreciated. Some details got left on the cutting room floor I'm afraid. The reason for doing this the way I'm doing it is to make a computer simulation make sense. I'm not a Fusion expert (really novice) - and my issue is that I learn best by doing. I'm not opposed to torture. I'm getting a good idea (far from expert) now about what's going on. I know where the caster and anti-dive should be and I've worked through all the gory details. I really could not have understood this without doing it this way - taking it into Fusion would have been a real struggle for me. Bump steer has been done for front and rear and while I have footage...I didn't see the point in showing it.
Ok, sounds good^^.
I'd still highly recommend that you take the time and model the frame and suspension in Fusion. As you certainly know, you can't change a single parameter of a suspension without affecting everything else to some degree - I mean, some parameters even affect both axles. And that's where a proper cad model will save you countless hours. It is likely that at some point you'll have to make small adjustments to pickup points or the rack position for clearance/packaging or maybe just ease of fabrication. With everything modeled in CAD you can immediately see every change in the geometry caused by the modification you want to make and whether you can/need to compensate for it elsewhere. You also get the updated dimensions of any mounting bracket etc you might need for that change - without having to crawl around the car with a marker and pieces of cardboard. Just export it and have it water cut, or print out a template.
The hours you may have to spend watching Fusion tutorials about parametric designs and master sketches etc will absolutely be worth it and pay dividends quickly.
With an ambitious/complex and time/money consuming project like this, I would consider a good CAD design the most valuable tool.
@@protator Understood. I've got several designs under consideration right now in Fusion. I think I have the LCA mount worked out and I have about 4 different versions of the UCA mount. I just had a flash the other day that may solve the packaging problem. I ran into a bit of trouble with the AC Compressor when the suspension went down. I've got the frame components in Fusion and I'm trying to improve my CAD skills so I can easily model the suspension. I can pull suspension dimensions from the jig easily - I've been toying with the idea of buying a suspension analysis package that I can enter the points into and move them around and get an output that I understand.
@Throttle Stop Garage Can't really say much about specialized software ... a lot of people in the offroad/ hill climb community seem to use such programs with good results, but most posts I've seen in the past were about solid axle 3 or 4 link setups ... so no idea what's out there that works well for double A-arm/ind. 5 link.
My somewhat generic advice would be to stick to the book ... estimate wheel loads, pick spring rates and a brake bias somewhere between 50/50 and the car's weight distribution, then choose anti-dive/squat values based on moderately sporty cars with similar weight and power. I've seen those numbers on all corners of the web at least for models like C3-C6 Corvettes, some Camaros and other american muscle. Then I'd block out the rear suspension and prioritize low scrub over a perfect camber curve unless the car's supposed to be a track toy. Toe link can go pretty much anywhere, you can always correct the toe curve by modifying the uprights ... on a street car I'd probably go with fixed toe-in and simply bolt it to the lower a-arm, tho. That detail can always be changed easily if necessary. So, yeah. Basically ignore the front end and focus on the rear suspension to get quick progress and keep the motivation up. And when it comes to the basic design, KISS is always a good choice. Try to avoid high torque loads on the frame rails, add additional box sections instead of excessively long mounting tabs ... again, the usual generic advice you'd expect from a rando on the net ^^.
@@ThrottleStopGarage just remembered a little something ... if you already know the width and offset of the rear wheels, check the scrub radius before you commit to the design. Seen a lot of bent or ripped off toe links on widebody conversions with too much offset. The forces on that little rod can get pretty high.
Z