I build a Very Specific Stepper Motor the Wrong Way
Вставка
- Опубліковано 27 чер 2024
- Just some early testing on an idea I'm messing with. Already have a few ideas on how to clean this up, Excited to see what kind of performance I can get with coils that aren't made the completely wrong way.
And as always, My social media:
Instagram: / wesleykagan
Patreon: / wesleykagan
Website: www.wesleykagan.com
For promotional inquiries: wesley.kagan@gmail.com - Авто та транспорт
This is what 3d printing is for, so many people want it to be final product, and maybe one day? But it’s best work is in prototyping, imagine building all this from billet.
Exactly... I'm working on some aluminum manufacturing abilities soon, but It's so nice to be able to prototype out of PLA
@@WesleyKagan slingshot channel uses plywood for prototyping, really strong stuff if used right. Might be worth a look
I don't think it has reached consumer level useability and affordability but titanium 3D printing seems to have reached the point where it by is an useable for industry. There are several cycling brands using it for frame parts and other small, complex parts where the ability to print complex shapes with voids in them that would be hard, if not impossible, to machine or cast is used to create light, strong titanium parts.
3D printing is past prototyping. Many end use parts being used in multi million dollar projects in many applications I've seen.
Well well well... Look what just showed up at exactly the time that I was getting bored at work.
Glad I can be a quality distraction!
All of your projects are absolutely fascinating... id love to see what you could come up with with a unlimited budget and all the tools in the world 🤣🤣
Same here, although all these projects might be the reason I don't have an unlimited budget.. hah
@@WesleyKagan I feel you man.
Limited budget increases creativity., unlimited increases productivity. Source: experience
Sooooo... what if you use explosions to power the valves? Basically a rudimentary small engine on top of the main engine where the valve stems of the main engine are the pistons of the small engine? "Yo dawg, I heard ya liked cars so we put an engine on your engine so you can rev while you rev!" Ok ok, I'm gonna go take my meds now.
Wait please, I want more of whatever this was
Or maybe drive the valves of the IC engine with steam controlled by cams driven by small steppers or servos. Use an electric heating element to make steam initially, then switch over to exhaust heat once warmed up. Ditch the transmission in favor of an alternator and traction motors...
Result: Digitally controlled steam-actuated gasoline-electric car. The model name has to be "Rube", of course ;-)
On air jet textile weaving machines the machines use overexcitation to make the solenoids work quickly. We have gotten up to 1250 PPM. The MAC solenoid rep said the have solenoids on cigarette machines that can do 12000 PPM. Thanks for the time you spend sharing your projects.
Interesting- Might look into that a bit!
@@WesleyKagan do you have any interest in SMCs high speed solenoid? The rep sent me a pdf I could forward.
5:50 declaring "it works" is bit optimistic... Also with high inductance coils, you can use voltages above the nominal rated value to force saturation/desaturation quicker, but the power control would have to be much more precise. Also you seem to have 1 x mosfet per coil - try extend that to a H-bridge so you can optimise trailing push forces with your leading pull forces. Linear rotary encoder could also help here - as you could just create a lookup table by incremental index to determine mosfet states specifically on your actuator position. Minimuse all your air gaps (to a point, becareful of excessive cogging), and I also suggest you order some soft iron rod from eBay (high permeable iron) - all this should maximise your flux vs. the power input.
I would think moving to a slide valve or rotary valve would be easier than using tappit valves when going to electric drive
I like the "understand the theory but try and see" approach. I also like the cat.
Congratulations! Astonishing work!
In the Miata project you have full control of valve timing and duration, but, as you stated previously, you still cant control lift, is 1 or 0.
What if you keep everything as you did in the Miata engine, but add some movable stops? The pneumatic valve actuator will stall when the valve hits it.
These stops can be as simple as a threaded tube between head and valve spring retainer. This tube would be spun via stepper motor to increase/decrese gap betwen it and spring retainer, modifying valve lift. It doesnt need to move superfast as the valves do. These stops have to move according to throttle position and engine RPM only.
Thank you for sharing your work!
You have mastered the whiteboard. 🤙
This (and many of your experiments) is marvelous in the truest sense.
I always enjoy watching your videos!
Add Iron yokes to your coils, that will improve field strength instead of loosing the magnetic field of the back side of the coil. There should always be a closed loop out of iron / magnets for the field lines to travel. It may be even possible to connect the backside of each left coil with the back side of the corresponding right side coil
Signing up as a student of the School of Wesley once it opens.
Great vid as always.
My man flexin with the casual Lagavulin 16.
also solid plan, decent execution. starting to look like its time to get Tormach or Haas to sponsor a build.
My idea 7 years ago was to use stepper motors to directly move a spherical valve with middle cut out 90degrees offset to allow full flow in and out rather than interrupted by the head of the stem, with the coding on the intake doubling up to be the throttle body.
It's super exciting. Thank you
My mans has his computer backgrounds set as 2 928's and Tyler's purgatory album. Top tier taste. 11/10
Dude you make smart stuff easy for guys like me who are like 95% capable of understanding the concept and you push it to that last little bit toward weakened grasp it and I thank you for it. Plus you do some pretty wicked projects and I am absolutely impressed by it as a mechanic I can totally see your genius.
Thank you for the kind words, I'm really enjoying putting some of these smaller tests together!
I agree with Justin Baker. Perfect a better valve design. A rotating ball design would be easier to control because you don't need to reverse direction. Just need position and speed. Might need an oval or tapered oval opening through the ball for best flow and overlap, but all intakes could be driven on a single shaft, and then all exhausts on a single shaft.
The problems to overcome would be sealing and thermal expansion. Might need to use more of a ceramic ball and seats to reject heat, and have seats on both the chamber and port side of the valve. Maybe even with an controlled gas porting of the ball valve seats?
Wow that's really cool! Looking forward to what you have planned!
Thank you!
Great video 👍 Like how you explain the pros and cons of components used or not.
Miata man returns
Interesting project, but I think this particular rabbit hole could be a lot deeper than most people think.
The problem (As you know) will be eventually setting this idea up to work on a car engine that may be able to run at over 6000rpm. At 6000rpm the valves need to cycle 50 times a second, but as they stay still for 3 of the 4 strokes, it needs to be able to accurately move at the equivalent of 200 times a second, and I just can't think of any readily available electrical operated rotary or linear actuator that could do that.
It could be that the only way this could work would be with an entirely custom cylinder head with a radically different valve design. Passive vacuum operated inlets and some sort of servo motor controlled rotary disc valves for the outlets could possibly work as the timing of the cut out on the outlet valve disc is all that needs controlling rather than the direction of travel. Adding some extra steps to half of the servo's rotation and dropping a few on the other half of the single direction of travel has to be easier than reversing travel of the motor and fighting the systems inertia by using the original more traditional valve arrangement, but again, it would be ludicrously hard to get to work properly, and prohibitively expensive to make.
I do think he is slowly showing why Freevalve went with the electrohydraulic setup it did.
@@1320crusierThe Freevalve is also pneumatic
My thoughts, i don't this solution will do it
I have no suggestions just really interested in what you are doing and trying to accomplish. Yay Wesley👏
That wall pc is epic.. Who needs a case anyway
@3:27 Put a resistor in series with your freewheel (discharge) diode that matches the impedance of your coil. This will allow the field to collapse faster.
Tyler Childers!! Such a good artist.
Great video! I love all of your projects
this style of stepper motor was used in Hard Disks however they now use a voice-coil set up. its not too different you could easily use a audio amp to get the power you need.
edit - spelling
I do have an idea I'd like to share with you, but I'm not sure how to describe it. Basically if you take your prior idea of the rotary ball screw, but make it so 1/2 a revolution, or more even, is the down stroke, and the rest of the revolution is the up stroke, this would solve the issue with having to reverse direction. You could even have it so there is a flat section of the ball path which keeps the valve closed or open to compensate for some inaccuracy or spinning momentum issues. Now granted this is essentially trying to infer camshaft motion in a different dimension, but of course being attached to an encoder motor it would still allow you to stop anywhere along the path up or down stroke for infinite variability.
Basically the ball bearing path would start and end at the same point. If you stretched it out straight it would make a V. One full revolution would open the valve to its lowest point and close it, just like a cam. The major advantage being not having to reverse the motor. Going from forward to reverse very quickly being the most difficult part timing wise, I would think. The speed of the engine would basically dictate how fast the motor spins the valve open and closed and also how long it dwells at any point. Not one of my original ideas for free valve, but based off your last video. I'm sure you'll probably get what I'm saying or may have already thought of it and I'm sure I'm also missing considerations. Just a thought!
but then you cant change how much the valves open... but that can be fixed with a lifting mechanism.. but the thing with freevalve is that ju can generate something like a square wave.
I think the main con to that it goes back to a fixed valve lift height, which means adding the throttle back plate in and all the associated pumping losses of forcing the gases through a restriction.
Even if you can get through the "V" very quickly, It's still a lot of work to move it all the way down and all the way back up in a short enough period of time to idle or do low rpm/low power cruising.
@@bullzebub Of course you can. Exactly the same as his last video. The amount you spin determines how open it is in however many steps the motor has between fully open and fully closed. That's the point. ^_^
@@nickopedia5669 No matter what there is a fixed amount of maximum lift and the benefits there of are determined by valves size anyways. No such thing as infinite lift. The point is just to have it be infinitely variable and not driven by camshafts.
@@802Garage
You added a word in there, "maximum".
While that is also a concern for peak power, the bigger concern isn't maximum but minimum. Being able to opening the valve to a low enough height AND with a low enough duration that it can idle, as well as go full throttle too of course.
Which is why the fixed lift is a problem - you have to go all the way down before you can come back up. To compensate for opening so far you have to do it even faster than normal, yet because it has to move so far it makes it even harder to make that motion in a short period of time.
That's going to be neat!
I absolutely love it when he shows us what he's thinking. No matter how it turns out.
I appreciate you watching!
You’re my spirit animal.
Pft, not only does UA-cam not tell me when you made another valve related video, it doesn't tell me you followed it up before I comment on it bringing up all the concerns you addressed here. Of course if this were LTT they would not be able to revise and consider all these details. Looking forward to seeing what you come up with
magnet coils are a science.
i tried using a coil calculator to make a linear actuator and in on paper it should have worked but it couldn't overcome the light spring. i was getting less flux strength than calculated for some reason.
I used the same wire dia, turns, and voltage i used to calculate the strength of it.
10 points for your choice of Scotch 👍
To replace the nails you can take apart an induction motor of a large enough diameter and cut out a piece of the stator.
Then for the pivot arm you could use regular flat neodymium magnets.
So in a sense it would be like a piece of a brushless motor.
I don't think this is the best way to do it but i'm still excited to see how this experiment goes.
Getting rid of camshafts has been in my to-do list since forever, but it's buried under many other things i want to do also.
Very very cool.
Hey, whatever happened to that V12 Vette?
Keep this stuff coming.
If a smooth transition and precision is more important than holding power you could offset the left from the right could by 1/2 the width of a magnet. If would mean two left and one right would be position A. Then to move one step. Only one left that was already on stays on and now one additional right turns on. May have an issue with sticky magnets having false positioning but it seems much smoother.
Beware small resistance with inductive loads. They have resonant frequencies which can cause higher then expected voltages in the coil.
Connector resistance around 0.1 ohms when driven with 12v can cause more then 600v on the coil. So beware the breadboard connector resistance.
O_o whoa
Yeah, I should note that a breadboard is the WRONG application for this test.
@@WesleyKagan if I had a dollar for every time I made a wrong decision I'd be rich.
Anyhow just keep in mind that your power supply has a capacitance front end, then you have connector resistance to an inductor. That configuration is a "tank circuit"
Tyler Childers background? Nice! That’s a good album
Thank you for covering
solenoids!
No problem! I needed to address it
Thank you for answering my previous question of "why not just use a solenoid?" :)
So I'm thinking a high speed motor would be best. Fast rotating with a flywheel to help smooth it out for taking load. Now the linear part should have steep threads kind of like a lead screw but larger pitch. Then control opening and closing with clutches. Now the clutches are going to be controlled by rotating electrical fields. The motor aways turns in one direction, but we can get two outputs rotating in opposite directions and energize the one that is turning the direction we want the clutch to lock so we have slip. Slip allows us to hold the valve closed, while atmosphere is trying to force it open, without hitting a stop and halting the motor.
So motor rpms stay high and valve can travel to stop and reverse direction and close very quickly also. Plus i believe a rotating gizmo is going to be WAY smaller than that thing.
Amazing work as always, hope to see this long-shot project to success! Anyone knows what's the wonderful clock on bike in the outro?
A tough problem to solve. If it were me, I think I'd be looking at using strong spring tension and an over-center mechanism to 'drive' the valves open/shut. Then the drive speed/strength is reduced to bringing the spring close to over-center, with a final nudge at the correct time.
That one dead fan on the wall PC upsets me for no reason.
put a brush on the long side and you've got a nice back scratcher !
all jokes asside, I love your videos !
Making a basic CNC Coil Winder may be worth it if you plan on making a bunch of coils! There are some good Open Source designs out there. Granted a drill rig like that is great too.
I didn't even think of that, but you're right! Definitely another tool I need to build around the shop. Looking into some other cnc stuff as well!
Looks very similar to a mechanical hard drive read head... There's some insane speeds those heads move at. Granted, the mass is tiny...
im LOVING this adventure into camless engine control. i was fascinated with freevalve.. and then they went kinda hush-hush about it and it never really materialized on a large scale.. :/
very promising and very disappointing that it didn't deliver on its promises. at least not in mass produced engines. Koenigsegg TFG shows that it's technically possible but mass production still far away
I was too impressed by the video wizardry w/ the white board and had to rewatch that whole part several times 🤣
The white board has accounted for my greatest increase in production quality to date.
To get a stronger electromagnet without higher inductance, wind one layer with leads, then another, then another. Then connect them in parallel. The total inductance will be only slightly more than a single layer coil and with less resistance. This will still create a strong magnetic field, and you can drive it with much more current due to more wire cross area. If you need more resistance overall, just add an in-line resistor.
You need a flux capacitor on your computer component wall. Would be e a nice little easter egg.
It says Wesly is gonna make stuff the wrong way. This is gonna be good!!
Usually me doing things wrong is the more entertaining option for sure.
Idea of something like BMW drivetrain. Cam that restricts the amount that opens and something like a voice coil to slam it open and a spring to close the valve.
You should use hydrologic! You can push and pull the valves!
Have you looked at any direct injection style coils? I don't know if they would have the capability of lift you need but the Rotax Etec system for 2 strokes are apparently pretty quick.
If you want to charge up a coil faster, just use a higher voltage with current chopping. :)
I was actually doing some reading on that! Might work it into the strategy.
Ford Powerstroke injectors use 120 to 240 drive voltage.
@@WesleyKagan I'd suggest trying the hardware you've built with a bog-standard off-the-shelf 5-phase stepper motor driver. I.e. Wire up the coil pairs in parallel and rely on the driver's constant current control and higher voltage rather than using separate up and down coils. You could also stagger the coils on one half by half a step and use either a 2-phase or 5-phase stepper motor driver depending on how you wire them.
# 1👍🏻!! 💪🏻
a viable way to make this work could be to use a crank-shaft like armature to do the oscillating whilst a stepper servo manages rpm to vary timing. this would allow the stepper servo to maintain a fairly constant if not slowly changing magnetic field and delay from inductance could go down. but at this point it would probably be more efficient to have independently driven cams that are connected to individual steppers. mechanical oscillation is difficult to maintain via electronic means especially at high frequencies. u start to deal with impedance and back emf which is another obstacle in this method of valve timing. maybe try redesigning the valve mechanism so that it doesnt depend on controlled oscillation?
also keep in mind that using magnetic fields like this will produce a exponentially decaying SHO which will add uncertainty in the positioning of the lever and timing the following motions especially with the abrupt movements that a valve requires.
Brushless controllers like odrive were recommended to me, and that's what my thought is with them. i'm curious to see how much adjustment I can actually get with speeding up and slowing down on a rotation.
You make some good stuff, I like (Borat thumbs up)
Why not have 2 rows of coils , 1 for up stroke, 1 for down?
nice
Soon nice 😎😎😎
What's the longest throw piezoelectric actuator they make? The issue so far is converting electricity to motion through magnetic fields, so if you can bypass the magnetic fields things should move faster. Of course, the problem there is those are kinda meant more for things like focusing the JWST than ramming valves up and down, but there might be something available.
Maybe there are some cues to be taken from coil gun or rail gun designs.
I need a video on that broken graphics card fan
YES.
I didn't want to notice it. I was really trying to pay attention. But now that's going to be the screensaver in my brain for a while.
i think its a water cooler but same
Your setup looks similar to a hard drive coil that actuates the read/write head
Wesley could be wolfman jack’s grandchild
Neat
"Breadboards are not rated to 3 amps but they are handling 3 amps" ... and I instinctively cringe for an ElectroBoom moment.
I love your free-form approach and willingness to show the mistakes, even the ones that are patently foolish in hindsight.
You can collapse a magnetic field in a solenoid just as fast as you can build it. You simply use the same voltage in the opposite direction.
Could you do multiple solenoids interacting with a single shaft?
What cad and FEA package are you using?
So, you've built an electric desmodromic... I like it. Didn't BMW fail at this a couple decades ago? Or was it spring retained single action poppet valves vial longitudinal solenoid? I can't remember. Desmo via a lever action seems far more obvious, but it's up in the higher RPMs where this stuff tends to fall apart :)
Most auto and equipment manufaturer tried camless design at some point. Some prototypes did work (at valeo I am certain) but cost/benefit in the context of cheap oil meant it never made it to production.
They built valvetronic instead. Get full control of valve lift with valvetronic and get partial control over cam phasing with VANOS.
@@zoppp621 It's not full control of the lift, and certainly not independently of duration, because of how the lost motion mechanism works. It also affects the phase, so some of the VANOS range is needed to correct the Valvetronic peculiarities and recover some of the lost duration under some circumstances, and remove it where it isn't wanted. The two systems likely respond at different rates so there needs to be some dulling of response (i.e. ignition timing to reduce output and prevent surges) under dynamic situations.
This results in severe drawbacks in terms of the induction and exhaust efficiency, that naturally impacts combustion and overall efficiency. But it obviously offers far more flexibility than fixed lift and duration, just nothing like the utterly free timing camless designs are usually chasing.
looks like a hard disk head :-) are you going to make the coils flat?
That's actually a pretty apt description except it's trying to move the disk instead of reading it!
When Wozniak invented the floppy drive he also simultaneously invented this stepper motor. So it's not a coincidence that they look similar.
I don't understand why :x
But, the amount of current change with BLDC ?
What kinds of shortcomings does a basic galvo have? Same as a solenoid? At least in light effect fixtures they seem to be super quick, smooth and precise (given a good control electronics), though the mass and needed power on those applications (like HDD's) is surely minuscule compared to something like valve actuation.
Servos are the next expirement when they come in- Theoretically they will be more effective, but it's still an incredible amount of force moving very fast.
Is the plan to make this radial, or just linear? If you did make it radial, you could make multiple arms to act as a stepper motor in the opposite side as well. It would be double actuated, and could be made offset to get more precise control. Otherwise, if you did make it linear, you could double the arms and make it like a fork actuating one lever point. Interesting concept for sure. (still reminds me of a HDD head arm, like someone else stated).
Just realized what you're thought process for this was after watching your last video with the steppers. I have an alternate idea. It would take some doing, but you could develop a voice coil (like for a subwoofer), that would be able to open and close the valve at the needed speed, which would be 50 hz. If you build them strong enough, you'd have a completely variable opening option, as you would just adjust the power (volume) to whatever setting you would need. Not sure how you would seal them though. You could use a standard valve, but would have to be halfway open to start. So that way, the full travel of the coil is realized. You would just need to figure out a way to send the right waveform to the amplifier that would power the coils. They could also operate off of supercapacitors charged off the alternator.
I think that separating to two sets of coils will not solve the energy storage problem as they are still mechanically coupled. Because conservation of energy is a thing; the "e" coils of the left side will still be collapsing as the "a" coil in the right side energizes, and the net effect will be largely the same as not splitting the coils in the first place. Since you literally need more energy in less time to change magnetic field faster, I think the only way is to increase your control power. This would probably look like overdriving the voltage and using 'current control' instead of 'voltage control' to ensure you don't exceed the power ratings of the stepper coils and melt them.
Further thought: This is all just electro-mechanical advantage. If you're already driving the steppers at max power, why not just machine the cams to have a steeper angle so you don't need 180 degrees of rotation?
@@AndrewLeCain
Because then you need more torque to accelerate the valve, which necessitates a larger stepper motor, which can't accelerate as fast. And so the cycle continues.
I think a solution here may be to do what industrial servos do, which is operate at 10x the voltage that would normally be used to operate the motor at normal load conditions.
i.e., a peerless brushed servo motor I was looking at is rated at 24VDC for continuous full load operation, but is insulated for and intended for use with a 180VDC current-limiting driver to achieve the acceleration values on the dataplate.
So maybe some 6V power wheels motors with a 100V drive with overcurrent limiting (not just "protection") will get them to accelerate significantly faster than otherwise.
One example of this type of drive is the BTS7960 which is very cheap and commonly available on amazon etc, but its current limiter is set to 43 amps which is way too high for many purposes I want to use it for. I sure wish it was adjustable.
@Nick Wolford that's what I was suggesting originally; use higher voltage but limit power by controlling the current to not melt the coils.
Step 1: Idea happens
Step 3: ?????
Step 47: PROFIT
So essentially an alternating electromagnetic rocker
“I’ve wanted to have one around and figured it would be good to have one on hand”
“This is a fun little science experiment that I just wanted to try”
Wesley is me.
The problem is is I have a garage and house filled with things I like to have on hand haha
@@WesleyKagan man I wish I could share with you a photo of just the tiny corner near my desk. Over 1000 discrete items for a huge range of projects I have planned. Technical books, 3D printers, tools… all the things for ALL THE PROJECTS!
I think that’s what condensers are for or capacitors
👍😊👍
All hail the algorithm!
All Hail!
What is the board to the left of your white board?
uh, shouldn't the coil windings go the one way only ? Not back-and-forth ?
question why can't you just stick a motor on each camshaft?
I think he wants to independently control the cylinders. I suppose that allows things like cylinder shut down at cruising speeds for fuel savings.
Man please fix your radiator fan on that Pc 😅 My OCD is killing me from the previous video... So much that I can’t concentrate on what you’re saying 😅 had to watch the video 3 times 😂
It's one of my most requested things, It will be fixed by next video I promise hah.
@@WesleyKagan Thank you man 😇 Keep it up, you’re doing great things 😇
What if you modified a stock engine, where it's basically the polar opposite of a freevalve technology, but to make it even less reasonable, the camshafts are replaced with gears.
For the record, I suggested "winding your own" in the comments on the last vid xD
have you removed the springs? i dony think you need em...
In this case I won't need them, No.
yup, fan is still broken...
Could capacitors be used to rapidly charge the solenoids?
increased voltage can boost solenoid charge speed. Just adding capacitors - no, it doesn't work this way.
@@SergeiSugaroverdoseShuykov how could you achieve the faster charging and discharging of the solenoids? I’d really like to learn more about electronics.
@@ronniejarvis2679 solenoid is just specific case of brushless motor so reading about inductance coil is good way to understand it
Have you thought about Magnetic Gears where the teeth don't actually touch? I found a YT channel on it once.
Why not linear actuators
One of the fans on your GPU isn't spinning in the video.
So basically you're going to make a segment of a stepper motor? And use it as a lever control
make it linear instead like a solenoid