I Built a Static Pressure Machine | Fan Showdown Season 6 Pilot
Вставка
- Опубліковано 30 лис 2023
- I was looking through the Fan Showdown submissions which got me thinking, maybe its time again to shake things up a bit. I have always wanted to compare fans based on which one can produce the most static pressure. This was the idea behind season 5 test bench, each fan pushed air through a radiator then using a anemometer we measured that airflow to try and infer what fan was producing the most static pressure. However this wasn't a perfect solution, to get a real static pressure reading we need to use a monometer. Now, before I go out and try and find a monometer that will work with PC cooling fan I wanted to try and create a DIY solution. So today we look at what I came up with and I ask you to voice your opinion on how well you think this will or will not work.
My Twitch
/ majorhardware
Fan Showdown Instructions
• EVERYTHING you need to...
Fan Dims
www.thingiverse.com/thing:455...
Submit your design (at least .STL or .STP / .STEP file)
fanshowdown@gmail.com
Cameras used.
BlackMagic Pocket cinema 6k pro
amzn.to/3m2RLVw
Reddit
/ fanshowdown - Наука та технологія
My primary thought is to somehow make the measuring arm longer so you can have finer measurements. You're going to have a lot of measurements that are barely off from each other, and your current scale may not be able to show the difference very well.
This is a really cool way to do this! I'm in favor of this shift, overall.
Exactly my thoughts as well
also maybe consider a logarythmic scale rather than a linear, as the effective surface area in the forw of the Flap decreases the wider it opens.
To have more precision without making the arm too long, you could consider using a Vernier Scale.
He's got some electronics chops, probably could scrap the dial and mount a capacitive encoder on the end of the flap shaft piped to an arduino and display an actual value. Find an encoder with 1024 gets a precision of .35 degrees per count.
You, make the lever thinner and longer for higher resolution and build a tunnel around the flap so that the air does not escape to the sides but only at the bottom. (mostly)
I think you would get better resolution of results if the flap was completely enclosed within the tube so that the air had to push the flap completely out of the way to move through instead of just far enough to jump out the side
I like this idea as well; completely enclose the flap in the tunnel.
and perhaps test each fan more than 1ce to eliminate "random flukes"?
if the flap were all the way open the end of the tube should be flush with the 'bottom' of the door, and have a square profile so it would have a decent seal at any angle
the shape of the exhaust would always be rectangular and would have to increase sensitivity as there would be less wasted pressure.
up
Please enclose the flap. I would expect higher results with it enclosed, not that it should change the ending results of the fans that have been submitted. But the range of the scores will be higher than 1-6, which is probably going to be the case with the current set up. (Not many fan designs have beat the stock a12x25.)
Test Engineer here. I think you're going to see a very tight grouping in your results because you have three things going on:
1) As the door opens further, it takes a highly non-linear amount of air flow to push it further. This is because of the reduced flow restriction combined with larger opening.
2) As the door moves to horizontal, the airflow's mechanical advantage is reduced (look at where the force vectors are pointing at 0, 45, and 70deg)
3) As the door becomes closer to parallel with the direction of air flow, the air's ability to impart force is reduced (the sail moves from crosswind to headwind)
If you want to measure pressure, you could just use a good ol' analog gauge. They're actually pretty easy to make in addition to being inexpensive to buy.
And as addition if you want to compare the values over the seizoen you might keep track of temperature and humidity
Tell that to bmw who used a similar design for the air flow sensor in the e36
I was going to make a similar comment. Also, the scale (the physical scale) has a small radius, and therefore a small range of motion. I think that will also contribute to the tight clustering of the results.
Edit: Also, shouldn't the scale on top be logarithmic? Because it's going to take exponentially more force to open the flap further.
can you provide a link to the device you're talking about "good ol' analog gauge" doesn't help me google search. TIA
@@wrenchturners It's just a manometer. Analog ones are like $10 to buy.
A proper meter for this would normally have one control orifice, basically having the flap inside of a square tube so it only opens on one side. The current design has the orifice increase exponentially as it opens.
good point. hope it gets more votes :)
what about one opening at the end with the "flap" perpendicular, but going up and down against the force of gravity with no actual opening? its basically an analog pressure gauge then
My thoughts too.
Yeah this was my worry of this design
So... Some kind of linkage system?
you can use actual water to measure static pressure. Add a U shaped glass pipe at the end of the wind tunel, add water and calibrate:) It will not have any part which can add friction to the equation. Some additional thought: diameter of the pipe can change sensitivity of the sensor.
I thinks that's just how noctua did their testing thus the mmH2O units
@@tenns mm H2O is just a standard unit of measure of static pressure (also mm Hg). Even digital meters have the ability to output with these units, just because they're so entrenched in the field.
@@RobBulmahn Huh, never used mmH2O myself, in physics labs we always use mbar or Pa with scientific notation.
I have one on my radon exhaust system. So they are probably pretty easy to find and cheap at least compared to the manometer
@@matthewpiotrowski2677Same here, works a treat and seems like it'd be pretty easy to setup. Could always use mineral oil or something with a lower specific gravity to get more height/make it easier to see.
Your setup currently has nonlinear scale. It can probably be linearized if you add a semi-circular shroud for the door, with a single continuous slot along the bend, so that the amount of air escaping before it can exert pressure on the flap is directly proportional to the flap height. You'd need to manufacture another flap with a counterweight such that its center of gravity lies on the hinge, and drive it closed using a spring. In turn, to avoid the same nonlinear forces problem as with gravity, connect the spring to the flap through a wire and a wheel.
True that the pressure changes the more the flap is open but all fans are being tested the same and the only variable changing is the fan. So measures are still going to show which is the better fan (using this measuring tool)
Cool explanation
yeah but the idea of this setup is to be quantitative, while nonlinear increase will make this more of a semiquantiative test (bad fan, good fan, great fan), while the subtle differences will be lost on the high end@@Bikerbarrie
My main concern with this measurement device is that it is introducing a lot of unknowns. Conceptually, I certainly agree that the results seem like they would be non-linear, and I also agree with another criticism that the resolution needs be increased. I would feel more comfortable if he could reach out to a lab that has proper equipment and calibrate the measurement device, so that at least would would understand the difference between 'a flap-openning of 6' vs 'a flap-openning of 6.5'.
I would just add a side walls to the flap so the aur escapes just down.
I motion to have the top three of all seasons tested as well, for a more complete set of data.
As for another tool to measure the static pressure, use U-pipe Manometers, three of them to be more precise. One measures the differential pressure of zone 1-2, one of 1-ambient the last of 2-ambient. Basically you have enough data to compare since you have the 1-2 one measure the delta directly and with the other two based on the ambient pressure you can check that in case of suction effects occuring.
I came here to suggest something very similar. This is simple, reliable, and repeatable.
I like the concept. I’ve always advocated that each fan should be directly compared to the others by max rpm, specific rpm values, and specific decibel values.
The main issue I would think is as the angle increases the less the flap is in the way meaning it won't be getting the full force
You could accurately calculate the pressure with an exponential function and a couple known pressures to calibrate. He could increase or decrease the doors weight so long as he adjusts his formulas. Increased weight would more accurately test higher pressure and the opposite for lighter.
That might not be an issue, we just have to think of the measurement as more logarithmic than linear. Like decibels.
Ah, so it's not the device more the measurement
This was my thought too. And, to further piggy-back off of "Windows_2000," the scale is not going to be linear because of this. I wonder if it might be possible to turn this contraption vertical and instead of having the "lid" swing away, have it raise up. The simplest way to do this is have the lid sit around 4 corner posts that guide the lid as it rises, but I think this would cause way too much friction. My suggestion instead would to be not actually have the lid attached to the rest of the wind tunnel, but instead suspended by a sort of crane that is counter-balanced so that the lid is essentially weightless. You could then use the counter-balance to measure the distance as it will descend at the same rate/distance that the lid rises (assuming 1:1 gear ratio). This still has the issue of allowing air to escape in all directions as the lid rises away from the tunnel. To get around that, you'd need to create a much longer tunnel that the lid can fit inside (essentially creating a piston). The problem here is that it introduces more friction and turns this into a much more complicated machine than you started out with. Not sure if these ideas are helpful, I'm mostly just spit-balling here, but maybe they'll spark ideas for someone smarter than I and it will lead to building a better mousetrap.
yeah, I agree. Maybe extend the tunnel to limit the side leakage maybe even ramp/ curve the bottom some to limit that too?
You may be able to set up a hydraulic static pressure tester by having a water reservoir that the fan blows down into, with an elbow out the side that connects to a vertical, clear acrylic tube with graduations on it. The higher the water rises up the tube from your baseline, the more static pressure you have.
Or suck, as it were. With a reduced acting area at the end of the tube I think you could get more displacement and therefore resolution. Finding a fluid with low surface tension could also help reduce measurement error from a wobbly wobbly meniscus.
@@knibknibknibyou dont want a reduced area. It will start to have capilarity effect on the water and change the actual value.
I'm not an expert in any of this but this one actually makes the most sense to measure pressure. The water would keep air from flowing so it would only respond to an increase of static pressure. It would be really easy to engineer. Could also add a little foam float to the measuring side that's just big enough to stick to the walls when the fluid drops.
It would be more complicated, but you could mount a mirror (or laser) to the door near its axis and record deflection of a beam. This would allow you to make the deflection angle much larger than a mechanical lever. Or you could use an optical range finder and measure the movements of the door. A potentiometer may be able to do something similar, but I'm less certain of their sensitivity.
The problem i can see with this is that just measuring the angle is going to be more representative of flow rate since the door will open more with more flow. say if you put a tiny fan that had really high sp it would still only open it a tiny amount making it seem like it actually has very little sp. What might be a good idea that could still use this design is to use the flap to push against a set of fine scales and have the flap set to only open to a set position to keep the opening the same size for different fans. And to add to that you could also mount your anemometer inside the box so you could set the opening so each fan was producing the same flow rate and then measure how much force that takes to hold back the flap to get a somewhat normalised flow vs pressure test.
i agree because the movement of the flap can be described as a funktion of the Airflow and static pressure in this setup. the sp must be sufficient to move the plate in the first place but since the angle of attack on the flap changes, the sp gets less relevant at highter opening angles so in theorie a fan with medium sp and high airflow can outperform a high sp fan in the normal test but your version of the test would fix that problem. however it would be best set the flap to a set of fixed positions and then take the measurements at those positions. otherwise it would be not possible to take precise measurements at low air restriction scenarios. also the flap should be inside of the tube and not at the end to get better readings( as @evil_me also stated in his comment)
@bxe_prototyp0118 yeah man thats exactly what i was thinking aye, have specific opening points to reference against then use the flow meter and force gauge/point scales to make a graph of pressure vs flow for each fan. And you could also test the static pressure at a set flow rate of say 25cfm buy adjusting the flap until your at 25cfm then take your pressure measurement
this still "measures" air flow through an obstacle (your grid infill). the radial scale could be improved/extended with a linear one so small changes at higher lift result in more noticable results. static pressure would have to be measured before the obstacle - u-shaped tube with water in it, like other suggested.
That's the thing. It does not directly measure static pressure, but it can be inferred through this test which fan would have higher/lower static pressure. It's good enough for me, it's just a youtube channel for entertainment purposes.
Airflow through an obstacle is what's relevant for a cooling fan anyway. It is the air that flows through the radiator/fins that carries the heat away.
The way is made it's not at all linear making it impossible to compare beyond bigger/smaller. Maybe using more of a lid that lifts straight with a linear spring of some sort.
edit: to ckarify, I mean not leaving an opening, which would allow a better pressure reading without airflow/turbulence having an effect.
Basically a closed chamber with a clear tube full of water and you measure the distance the water travels ;) Genius!
No that's not how it works the water measurement for pressure is how much water that pressure can "support", so you'd have to either add/remove water or widen/tighten the tube to achieve a balance.
Well, if you were wanting to measure a certain ...pressure... that a fan could ... Statically... support.
@@nschubach What's your point?
Hi, I think your solution is pretty good for what you're trying to do.
While you were explaining your idea of how to measure static pressure, I visualized a U shaped pipe with the bottom filled water and one end sealed with what you already built so that the air coming out of the fan goes into the pipe. That way the air coming out of the fan would push the water up the open end of the pipe and it should be easy to measure how high the water is pushed. That said, I think it would take some experimenting to find the best diameter of the pipe and it would need to be tall enough so the water doesn't overflow out of the open end. I hope the description makes sense.
Love your videos, will see you in the next one.
Yes with carburators I use tubes filled with mercury
As I understand it, that's literally why the measurement in the spec sheet refers to water -- that is THE way to measure it in high precision.
The issue with that is that if the Noctua can only do about 2.3 mm H2O, how are you supposed to accurately measure the differences between the fans? You'd be measuring tenths or hundredths of a millimeter. You could try a lighter fluid like mineral oil for around 15-20% better accuracy, or set the measuring tube at an angle as long as you have the space and the tube is rigid and stable enough, which can be an issue.
Another option would be to control the "ambient" pressure holding the liquid down. You could pull a slight vacuum on the "output" side, for example use a plastic hose and clamping it with some air between the end and the liquid. Basically a reverse water barometer I guess.
I like last season's "blowing through a radiator" type of method better, static pressure and air volume are two unserapable parts of a fan's performance as you always have to overcome some kind of resistence in a real life situation, and air is what ultimately takes away the heat so volume matters too. But the concept is nice, even though numbers on the gauge barely mean anything, this device can be useful in comparing fans, maybe add this to last season's test suite. The lower half of the gauge is basically unreachable using ordinary pc fans, as the angle got wider, the contact area will get smaller, and more airflow will just pass by the plate without applying any force or just a little force.
In theory, could get a digital reading with a magnet and a Hall effect sensor. Not sure about repeatability and consistency, though Hall effect sensors are used to great success in flight sticks.
Came here to say the same thing. Digital read out would be much better and should provide greater resolution for those tie breakers.
This crossed by mind as well. With such minor differences possible, a visual gauge will introduce a lot of error.
Since you asked, A larger reinforcing grid on the door to lighten it would allow more movement, making the gauge more accurate.
I love this concept. It reminds me so much of the wind tunnel the Wright brothers used to design their propellers on their aircraft. Balance weighted and simple, they used the box to design a propeller with an efficiency close to today's standards.
I also have read that orthogrid (rectangles) are more efficient for stiffness to weight.
A lot of people advised to use a U shaped manometer. My two cents to this idea is to use a fluid which has the lowest density as possible. Oil is less dense than water. This can increase the sensitivity of the device.
There's always isopentane but you need to keep the liquid under 80F, or it will boil. Also flammable 💥💥💥
@@knibknibknib Haha, good one! I don't think a liquid with high vapor capabilty (eg alcohol, pentane) would be a good choice, though.
Hi automotive engineer here, I love your mechanical gage design idea! But I can see a few issues with your measurement. The door flap will not open linearly to the gage. There are two things at play. The door wants to come down, but the force at which it does is related to the angle. The higher it goes, the more the center of gravity is pushed beyond center, and Fg increases exponentially. Also, as the flap opens, the area of the air that can escape is increasing exponentially, too. Your scale drawn is linear, so there could be alignment there. It's hard to calibrate, though. To eliminate the escaping air, you could enclose the flap with like 0.5m of plexiglass. For the angled gravity, you could try a counter balance so that this is perfectly able to rotate freely. I hope this helps.
Interesting idea. Maybe have the flap fully enclosed and sealed with some kind of rubber bracket (or equivalent) on the side facing the fan, then hooked up to a weight scale. Like a food/baking one should be sensitive enough. There's even entirely mechanical ones that you could repurpose. 🙂
Hey, i am using a Sensirion sdp810 differential pressure sensor for a project. it measures 500Pa, communicates over i2c to build a data acquisition system with a microcontroller, it is a very good resolution sensor and not that expensive (35 bucks I guess). but I also like the flap measurement system.
This sensor looks awesome for the price.
May need a bit more than 500pa as that's only up to 2" WC.
My main concern is that the scale may not be quite linear but more logarithmical in nature since the opening gets larger for a larger scale reading, which alleviates some of the additional pressure.This implies that small differences in the scale readings with regards to high performer fans actually means a greater pressure difference than expected
I love the idea and suggestions. You can also clearly see the curve slowing as the fan reaches the top. Makes everything so easy to see.
As for the manometer a single tube filled with water (or oil for increased displacement) making a U bend and with open ends before and after the load (different thicknesses/style of radiator) would be enough to show the pressure. With a camera and a macro lens against a precision scale you could get quite precise measurement. If you want to you could use some sort of iris with variable aperture (like in a lens) as the load/resistance to create a Pressure/Velocity graph.
I think this will work for comparison. But the actual number will be meaningless. You should be able to tell if one is better than another, but not really accurately by what percentage.
The resolution might be too low, getting above 10 or so seems nearly impossible, most fans will probably be between 0 and 5. I imagine a lot of fans will be at the same number. You might need a bigger dial with a longer indicator at the top to get a more accurate reading.
love the progression of your testing and the tools you're making to test this!
A key thing to keep in mind is that the air pressure and air flow are related and not necessarily in a linear way. All of your tests so far, including this one, are testing some point on the pressure-flow curve using different resistances.
The open tunnel was about as low of resistance as you could get, which is basically the maximum airflow - an interesting number, but not really reflective of a practical application.
The radiator applied a considerable resistance, which makes it a much more practical working example. Measuring the airflow with the radiator tells us something about a point a long that pressure-flow curve, where the pressure determines the airflow through the resistance (the radiator). If we knew more about the radiator, we could calculate the total power of the fan as a combination of the pressure and flow developed.
What you are aiming for now is pure static pressure with zero flow. In order to get that, you need it to actually be zero flow, so it has to be into some kind of closed orifice. The classic water tube is one way to measure the static pressure at the outlet. Perhaps you could direct that pressure into some other mechanism that would indicate the pressure, such as gauge using a diaphragm and spring. Which, at the end of the day, is just an analog pressure gauge.
I looked at low pressure digital manometers for last season and the costs were prohibitive. Even more annoying is that lots of gauges will mention the resolution of their display, but not the minimum pressure or the accuracy of the measurement itself. I bought one that I thought could be useful and it just barely tipped the scale off of 0 when applying the full static pressure of a 50mm blower for a 3D printer nozzle, so that's useless. Models that seemed useful were at least a few hundred dollars, which might be worth it for running a UA-cam channel, but certainly wasn't worth it for me to play around with making my own fan blades and nozzles.
However you manage to measure the static pressure, I think it's important to judge a fan just by a single characteristic. A fan that maximizes peak air flow at zero resistance will have terrible static pressure and vice versa. That's why airplanes use different motor and prop configurations depending on whether they are meant for level flight (airflow at high air speed) or 3D use (static pressure for hovering). That's why all those compressor designs did so poorly for maximum airflow in the wind tunnel test.
Another consideration is that fans have a stall speed. If you apply too much resistance and the airflow through them decreases, the behaviour of the airflow over the wing changes and you get a completely different behaviour. Knowing the static pressure at zero flow and the max flow at zero static pressure tells us what the ends of the spectrum look like, but nothing in the middle.
What I would love to see is an actual sweep of the pressure-flow curve. What kind of resistance can it handle? What use cases is this fan good for? Does it fall apart at the slightest resistance or can it handle a clogged radiator and still function decently? I think you'll find that the Noctua is good specifically because it maintains a solid flow even up to relatively high pressures.
What about the wind tunnel but with a variable grid in the middle, like you might have in a floor vent (I forget the name of that grate design). By varying the cross-sectional area of that part of the tunnel, you control the vast majority of the resistance along the path, and therefore can see how pressure and flow vary with that resistance. Sweeping this through several points from fully open to fully closed should give you a good idea of that pressure-flow curve. Even if you can't measure the static pressure that builds up behind the grid, we know that, below the maximum air flow, the air flow is determined by the resistance of the grid and the pressure behind it, and the grid is the same for all fans tested, so the pressure is simply a function of that flow.
Bonus points if the settings for the grid can be matched against common situations like a typical computer case and a radiator to give an idea of how that fan would work in those everyday situations.
As for scoring them, it's not immediately obvious how to get a single number from this, but a common metric in similar situations is to take the area under the curve. By taking the area under the curve, we measure how well it works across all situations, rather than just one. With the same area, you could balance it to be really good at static pressure or really good at max airflow. But if you manage to improve both, that will show up as an increase in the area, making it an overall better fan.
You should add a geared indicator so you can have more discrimination between intervals relative to flap displacement angle or a longer indicator tip so you can include more gradations between presently available marks. Your ability to measure differences will likely suffer without something like this.
Don't know if this would work but. What about using another fan at the end of the pipe and reading the volage generated from air flow.
This is a good idea. It just needs some refinement.
1) The flap needs to be enclosed in a shroud significantly longer than it is. Right now, you have a lot of spillage that is lowering your results.
2) You need to calculate the moment of inertia of your flap. The effect of static pressure on it is not going to be linear. It is going to vary significantly with the angle.
3) You need to determine the amount of resistance your flow guides impart. Alternatively, if you put an extended shroud on the outlet, they may not be necessary.
4) I would create a second test. Turn your apparatus vertically and create a slightly outwardly sloping shroud. Put a weighted plug at the end that just barely seals where it contacts the apparatus. The static pressure will push up the plug until the pressure maintained under it equals it's weight. The plug should have sloping walls that match the shroud walls. If you know the weight and the elevation change, you can calculate the force applied (F=ma), and from that, the pressure (P=F/A). This would be a simpler design to calibrate, I think. However, it is still not linear because the gap around the edges opens as a function of height since they both slope.
Fellow mechanical engineer here. Seeing lots of people suggest a U-tube manometer, but I'm not convinced that'd work very well for this application since the difference in pressure between the two sides of the fan is so low, not to mention any static pressure would likely be overshadowed by the venturi effect resulting in a device that essentially measures velocity anyway.
The biggest problem with this machine at the moment is that it's not measuring static pressure at all -- it's measuring a quantity related to the stagnation pressure, which is a function of both static pressure and velocity (it's not quite the stagnation pressure since the air doesn't quite come to a complete stop when passing the flap). If you wanted to accurately measure static pressure you need to make the velocity of the air zero, i.e. have a fan blow into (or out of) a sealed container and measure the difference between the pressure within the container and atmospheric pressure somewhere far away from the fan. The only practical way I can think of for doing this with such low pressures is an actual manometer, unfortunately.
That being said, I still think that this is an interesting way to gauge the effectiveness of a PC fan and could definitely be used in the next season, albeit with a change in what you claim the device is measuring. Several other people in the comments have left great ideas about how to improve the repeatability and resolution of the device so I won't repeat them all here.
You could make a tunnel where the fan sits in the middle of it and there is a pitot tube both at the inlet and outlet side. The difference in the pitot tube would give us the pressure difference. 😊
I like this idea. Probably depends on how good a pitot tube you could buy for the small reading.
A pitot tube measures stagnation pressure, providing flow velocity, not static pressure.
No need for a pitot tube, what we want is a static port.
Like the idea, much nicer to see this than a number 😀
Also, how does it look with smoke pushed through it 🤔😁
Im an aerospace engineer- i can say that this should do alright, but it’s technically still a dynamic pressure meter/ flow meter- flaps like this have actually been used on early aircraft to measure speed. It’s still a solid idea though.
This is a really cool idea, it's a great way to visualize it, i'm in favor of this! cheers man!
Hi there, my suggestion is to balance your door/flap so that the T-30 just about maxes the gauge out... because no normal 120mm fan running on an AX12-25 is going to go that high. That should also bring the stock AX12-25 closer to the middle of the scale.
I like the box idea, just wish it had more resolution, but I don't know how that could be done. Maybe put a long needle on the end of the gauge and make the gauge circumference bigger.
Im also curious on how the "flow straightener" impacts the ability of the fan to lift the door. Can you try one without the flow straightener and one with? That would be easy to design and print.
Yes! He definitely needs to make the arm to flap ratio a lot larger right now it's maybe 2:3, with the arm shorter than the flap. I think shooting for 2:1 (with the arm double the length of the flap) would give him a much more precise measurement.
Maybe some sort of constant scaling on the force on the flap? Not sure how he can implement that but maybe a small motor that can react to the force applied on it depending on the angle of the flap?
He could select a low-angle, easily accessible for most fans to reach, and gradually add weight to the door in small increments until it can no longer stay open, recording the weight at that point.
I would strongly suggest to print at least 1 or 2 alternative flaps with more weight and probably also oversize the length by a good margin. The static pressure creates a force against your flap and you basically only measure that force by seeing how far it pushes your weight up. The current problem with this is that the greater the angle of the flap the lower the amount of air that actually reaches it and applies force. So you have a system that for e.g. linear gains in static pressure only measures sublinear changes. With your concept I don't see a solution how to circumvent this entirely but if your flap is longer it means that even at an angle the 'whole' air stream will still reach it. And if your flap has more weight the applied force will move it less, meaning your angles stay within a smaller interval and have less impact on the force balance.
If you are interested in a different concept I would suggest to use a spring with a low spring constant and place that inside your tunnel with a flap that moves collinear with an increasing air stream instead of at a varying angle.
Good job. I like the simplicity and the fact we can make our own.
It would be interesting to see the difference of an unimpeded box (without the grid pattern you have to clean up the airflow). I understand why it's there, just curious if it's necessary or not.
not a bad setup and simplicity is king.
while you might have issues keeping it level, maybe find some ultralight linear bearings on some guide rods, blowing air up against gravity holding the weight down would provide a slightly more linear pressure reading. this way its always blowing against the full surface area rather than a diminishing surface.
EXCELLENT! Now for the micro improvements over time as with any first build.
I've seen suggestions but as I have no idea about this stuff I can't evaluate them. So just want to say, this refresh is a really good idea, always making adjustments to the testing methods, maybe it wouldn't be necessary but for sure it makes the series even more interesting! I'm so happy about this!
I like it, I wonder if it will be sensitive enough to identify small increments. What was the difference in static pressure measured between the A12 and the Winner of last year in CFM?
How about a load cell attached to the end of a free swinging door with slots around it for air to pass through. Should give you a linear readout of pressure on the door for all wind speeds.
E4 mafia to E4 Mafia. You can take this concept linear! Use some springs, and make it push the grid open from the flat side and the flat side only. Making it push the grid open on one side only completely delegitimizes your solution!
Great idea, even better execution!
cool test rig, I like the idea of using it along with other testing to create a more complete picture.
I think the angles measured could be not granular enough to differentiate between different fans and you might end up with lots of entries having the "same" result. And even reading the correct value seems pretty prone to error.
Maybe a thinner, longer extension of the indicator might increase granularity?
Oh nice measurement concept, I love it!
I actually made wind tunnels to test samples in conditions at work. My dad is retired NASA and put me in contact with the head of the wind tunnels at his base. He made two very helpful suggestions to ensure I had laminar flow at my sample port. I needed to use streamliners in the tube. He suggested jumbo straws. Next I needed 6x the diameter in tunnel length from the streamliners to the sample port and at least 2x after it. It worked perfectly. I get nice even and non turbulent flow at my test port. In your case your test port is your flap. So add a tube after and lengthen you tube before and add straws equal in length to the diameter. I used 6" PVC for my tubing.
Hey man! Remember psi is per square inch! I think a better way would be to see how many grams of thrust is produced then dividing that by the square inch size of your push plate. This giving you thrust per square inch. Also don’t flip the flap, the flap should be pushed parallel to the thrust produced.
Would be awesome to start a contest on submitted Impeller blades for a 3D-printed water pump and test self-priming (height), psi, vertical lift, and GPM.
Could still be PC-related if used for CPU cooling.
EDIT: yes, using a static pressure test would be cool, and other tests.
The water test mentioned might be a better test tho.
Interesting idea. If you split the flow straightener to leave a little gap in the middle and then 3D print an averaging ring duct around the sides with ports all around the edge. This can then be tapped and piped off to a simple inclined U-tube manometer to get very accurate static pressure measurements.
I tried to google a term for the averaging ring port but I can't find out whether there's actually a proper term for it. It's bascially a way of minimizing the dynamic pressure variations over a duct section by sampling a number of points around the perimeter of a duct and merging them all together into a single dead-headed measurement port. An inclined water manometer is just like this, in case you've not tripped over them before: vlabs.iitkgp.ac.in/mv/exp1/index.html
I like your idea. I would love to see how the results look as the season goes on
Man I've missed the fan showdown!
Can you use known values from certain fans to calibrate your scale?
thanks for doing these tests! lots of cool ideas in the comment section, hopefully you can incorporate some of them if they sound legitimate. the engineering is deep!!
I think what I like the most about this channel is the way people think outside the box and im all for it.
I love the idea of using this for the next season! I think the numbering should be modified though, where the gap between units is less as the numbers get higher. Here is my reasoning: To get a fan to go from a 2 to a 4 is pretty easy, where if you were to try to get a fan to go from a 14 to a 16, it will take much more airflow to do so. Therefore, the units should be closer together as they get larger, to help differentiate the top competitors more easily without getting into decimals/fractions. In any event, I would love this to be an equal factor in determining the best fan, along with the traditional methods used prior. Thanks for the awesome content!
You're so smart bro. Very inspirational. I love the if I can create it I will mind set. As well as the open sourcefulness. ❤
Man, ur one innovative cookie! Keep it up bud!
You could make a rectangular box out of glass/plexiglass so the fan can be placed at the bottom blowing air up. Then inside the chamber would be one of the little parachute guys we all played with as kids. The winning fan is the one that makes him fly the highest.
Like several comments have said, a u shaped of tube filled with water or other liquid i.e. alcohol, can buy used as a manometer, tilting the board the tube is mounted to will increase the resolution and you can use trigonometry to find the actual results, or just use it as is to compare fans without getting a perfect number
The wind tunnel is lovely, and those are some nice high quality prints.
At work, we use Dwyer 477AV manometers for low pressure gas measurements, but they are a bit pricey at ranges of 20" w.c. or less (which you probably want the 4" or maybe even 1" models for the highest level of accuracy). I have found Superior Sensor Technology has some low pressure sensors which can communicate either I2C or SPI with good accuracy specifications. They even have some evaluation kits that you can use to hook up via USB to a PC for taking measurements (such as EK04-HV-SPI or EK04-HV-I2C). I wouldn't recommend using an actual water column manometer, because getting a good one that is in the 0"-1" w.c. range can be quite expensive and you have to periodically fill those back up with special fluids.
I think that if you (mostly) close the sides of the door, even with a transparent panel or something like that, the increase of the area where the air can escape would be slower when the panel opens and you might be able to get more granular results than you're getting now where I fear the fans might end up all too clustered on similar value. Closing the bottom too might help if the sides are not enough.
Same thought. As it is now, the air seemed to be moving down the direction of the flap. Closing the sides and bottom longer if not the same length of the flap could drive the air straight.
Alternative solution I have in mind is to make a vertical tube with the fan at the base. Along the wall of the tube there is a thin slit cut. Place a plastic ball in the tube and the air pressure will lift the ball. the higher the ball floats, the more air gets pushed out the slit in the tube. Eventually it will reach an equilibrium where the ball is effectively held in the air and you can measure that point along the slit. The adjustment points would be the weight of the ball and the thickness of the slit.
I think that it would also visually look cool and be duplicatable.
I was thinking about using the definition of the static pressure itself: mm H2O
My idea was a sealed chamber with a hole that a fan can be mounted onto it with a decent seal.
another opening is pointed downward then curl back up like a U that is filled with water.
The pressure will be measured by the difference in height between the 2 half of the U pipe.
I had another thought just now. Instead of seeing how well it can blow, you can check to see how well it can suck. Take your box and turn it on its end. Put a lid on the bottom that it can pull up on. Add weights until it falls off. m*g/A is your static pressure.
It may also work the other way. Turn your box on the other end. Blow into the box. Add more than enough weight to the lid to hold it down. Remove weight until the lid doesn't stay any more. When balanced, the static pressure is m*g/A.
In either case, you no longer have to be concerned with turbulence.
Mount a laser pointer to the flap so you can project it long distances and get a more precise reading. You could use a double mirror to extend the beam length in a tight space if needed. Just make sure it stays calibrated if you do that.
Interesting concept to show an apples-to-apples comparison between fans. For the purposes of what you're doing, it's a really neat solution. For more precise measurements, you want to measure static pressure at zero airflow, so you'd need a sealed chamber and a manometer. Most PC fans are in the
Another thought is still using a smaller door for more resolution but require a minimum open angle while Adding some type of small weights to the door on the standard a12x25 to reduce to this minimum angle.
This way you can judge each fan by the amount of weight required to open the door required distance.
This will definitely restrict the outlet volume to the sane size on each fan while giving an idea of the fans ability to deal with static pressure. It wouldn’t matter if you even used paper clips as long as the increments are small enough to get good resolution in the weights.
It’s the simplest way to go build this without extra complexity.
My very first thought was to try to build something similar to the air flow meters found in 80s cars like the Miata and MR2. Very cool!
One flaw of this is that your flap angle is nonlinear due to the change in orifice size, and the fan P-Q curves of the fans are nonlinear as well. So each fan would be operating in a different performance regime. A fan pushing more air might see less restriction so it would show as having a higher static pressure, which might not be accurately. Essentially, comparing the results might not be strictly apples-to-apples.
I see a lot of people suggesting this already, but U-tube-based manometry is probably the right tool here. Electronic manometers are expensive because they're compact, but old-school water-and-tube based manometry is simple and can be tuned for all sorts of pressure ranges. It could also look cool as hell mounted on a wall with colored fluid and some LEDs.
My suggestions:
As others said- you'll need some sort of tube/walls around the flap so the air doesn't just get blown downward. It would improve resolution quite a bit.
-Maybe add some sort of gasket around the fan? Construction paper should be fine. It would help eliminate pressure losses due to blow-by (slight loss makes a difference with low values)
-The indicator doesn't show well on camera. Maybe bevel it with the outside corner lower? I'd guess 45 degrees instead of 90 would look good.
It's a neat idea. I'm impressed with your numbering- they're still very readable at that small of a size
To the Author, did you realise that the first airflow meters for fuel injection engines in the mid '80's used this 'trapdoor' arrangement? They were restrictive but accurate and I love you for using this method to upgrade your test results 😅 love your channel and will stay tuned to all your future posts😀
Love it! Use a high vis protractor to get accurate angle of deflection and you'll be GOLDEN!
To the Author, i think your low rpm test reshiems are fine and don't listen to haters/critical responses, i thought your channel when I first saw it was a little underpowered 😢 but now I see that your low speed tests are very representative of how they would act at high speed (10x's or more) so happy to see your video's and hope to see more from you, i am a sheetmetal worker/engineer that builds my own jet engines from old turbos but have learnt more from your fan testes! Keep up the good work and look forward to more from you 😀😀😀
I think pushing the "door" open, thus letting air out, muddies the pure pressure measurement. There will be some airflow aspect to the measurement as air is escaping. Instead a closed box with a pressure sensor (scale) at the end would be better.
A theoretical tiny fan may be able to produce a lot of static pressure, but may not be able to move enough air volume to keep the large hole of the door open.
This leads to the question: why should we measure static pressure for PC applications. I think the measurement using the radiator leads to the most realistic grading of the fans. Still, I like this experiment.
I would recommend measuring flow and pressure separately, and graph both over the full available range of restriction, from a minimum restriction represented by the grid alone to completely "shut" static pressure. These graphs, charted with RPM and Noise measurements would be a complete characterization of each fan that is basically the "end game" of testing for all fans.
Use a manometer and a mass air flow sensor system combined.
I prefer your inhouse engineered testing solutions. I think it is what makes your channel so great! That being said, you do run into issues with airflow exiting the sides of the flap once you reach a certain point. Your data point grouping will be very very close. But a mechanical visual result is way more entertaining than numbers on a 200$ device.
I think what you might be interested in is a "Characteristic pump curve". It shows the power transmitted to the fluid against the amount of fluid that is pumped. For both you can use many different variables. Like head, static pressure or specific power and flow rate, volumetric flow rate or mass flow rate.
This setup is fine IMO. I see people getting all hung up on the accuracy of the measurement and whatnot, but all you really need is a repeatable A/B comparison. The number you get doesn't need to correlate to any actual measurement to see which is better in this context.
I think exploring in the direction of static pressure is a natural direction for this sweet experiment.
I used a digital kitchen scale(4" x 4") with a plastic bag made into a tube tape sealed to it, sealed that to the bottom of a 120mm wind tunnel. Got repeatable results matching Noctua's P-Q graph. Another guy on reddit used a cork floating on a sealed water dish with an indicator off the side of that, got solid results too. Some pics on print-ables --> GPU/CPU Fan, the PSF (Psychedelic Super Funkitron )
I have worked in instrumentation, I have a cheap solution, get a centering Ruler in mm(18$ Amazon). mount it vertically ( Magnets Work well). Get 2 acrylic tubes and some hose and make a U Shape and make it plumb, with a little Hose on both ends( long enough to connect to a test) fil your tubes about half Way and center your adjustable magnet center ruler at the bottom parabola of the water line. you now have a working monometer. Hose clamps some tubing at home depot and some Magnets and you can put it on your fridge. Pro tip use yellow with a hint of green food coloring to see it easier. Use a Protractor for your Angle gauge with pivot at center. I have Worked on Forced and Induced Draft Boilers. Also Pins Will have less inertia than bearings at small scale. This is why Hot Wheels don't have bearings. Hot Wheels hobbyist have cheap wire fore the Axles that work Very Well It comes strait. Get one end hot and insert into plastic Leave a pilot hole before the hot insert. I have also worked In Micro Switch manufacturing. Your prints are Missing Pins as pivots instead of plastic posts( I wanna See A pinned double opposing Prop, one with Fans/ Bearings). That is all it would take for you to have a Pro level product. Hot Wheel or Lego lube is magic on plastic.
Step 1: Make an air tunnel with the ability to drop-in meshes of various blockage percentage in the middle. On the upwind side, put a nozzle for a tube. Connect the nozzle to a u-shaped tube, and put water in the U so we can measure pressure using the same units as the manufacturers (mmH2O). Put an anemometer at the end of the wind tunnel. Now you have the ability to create PQ curves for each fan, at whatever resolution you have the patience to test.
Step 2: Modify the wind tunnel so you can stick off-the-shelf radiators in it instead of meshes. Now you're drawing the PQ curves for the radiators.
Step 3: Plot both sets of curves on the same graph.
Step 4: Become the only 3rd party database of actual fan performance. Go mad with power as gamer's nexus bows to your test methods and I guess.......Profit?
I like this design. I think it's a very good way to compare fans
cool elegant result
Pretty good Mk. I static pressure gauge. Can't wait to see the tests, and possibly a Mk. II when what improvements can be made become more apparent.
This will work, but I have another idea.
Reuse the wind tunnel, put a radiator behind the fan and a solid panel 10-15mm in front of it to simulate a restricted airflow case. Then use dual manometers, use water tanks with clear plastic tubing marked at different heights to measure the pressure. You can also use your anemometer at the end.
I like the pressure machine. Great work!
My recommendation would be to extend the length of your indicator arm, which is serving as a counterweight by another 50% or so. You would have a longer travel distance at the tip, resulting in a higher fidelity look at the score if you print a longer arc scoring path to match. If you are concerned about the center of mass or the damper not closing, then simply print a small extended rod at the very bottom to move the center of mass further out from the chord line of the damper.
A separate take on how to measure this would be to keep your damper but add multiple small rods/arms near the base that you can hang small weights on. These would be printed of course in fractions of grams and used to determine the weight needed to keep the damper closed even with the fan blowing. The weight required to close the damper becomes the score, and the fan requiring the most weight has the highest static pressure.
What a great contraption bro ! This is soooooo coool. I want to see a ton of fans tested ! Can you test some of the lian li fans and some more phanteks ?
This works for the job. As a few have pointed out, making the arm at top a little longer while adjusting the weight value for the counter arm to get finer measurement. Kinda like adding a .1 to .9 to the scale for the deciding numbers. New fan designs and even some of the older fans that would show promise to test again can be in this group.
This is pretty cool. I see a bunch of comments talking about non linear measurements on it which make sense, but this scale can be logarithmic similar to Decibels. Neat concept. Curious to see how this plays out this season.
A relatively simple way to increase the resolution of a water manometer is to use an inclined scale. It won't give you a digital output, but with appropriate graduations along the length of the manometer tube you will still achieve measureable results out of it. These can be purchased, but can also be DIY'd fairly easily. If DIYing if you want transferable results it is important to use a liquid with the correct specific gravity. Water can be used, although it should be distilled or deionized, and its temperature also effects its density.
Interesting. These new tests make the show even more interesting.
In a wind tunnel using the Bernoulli effect, the manometer can be a simple acrylic u-shaped tube full of water. That is what they use to figure out pressure. It doesn't need to be super fancy