Pneumatic Clutch

[joekid11]

Has anyone attempted to design a pneumatic clutch actuation system for their car? I ordered a piston from Bimba and I plan to use it to pull the clutch cable. But I want to control the flow rate coming out of the piston with a miniature proportional valve. I want to know if anyone has tried using a similar system and if it has worked? Thanks.

Joe Santore
The Cooper Union

[joekid11]

Has anyone attempted to design a pneumatic clutch actuation system for their car? I ordered a piston from Bimba and I plan to use it to pull the clutch cable. But I want to control the flow rate coming out of the piston with a miniature proportional valve. I want to know if anyone has tried using a similar system and if it has worked? Thanks.

Joe Santore
The Cooper Union

[Superfast Matt McCoy]

Controlling the flow out of the clutch cylinder sounds good for launches, but the flowrate would need to change for different track/tire friction. Not only different tracks, but tire and surface temps.

We used a pneumatic clutch for an auto clutch on downshift, but we also had a manual clutch for launches and such. I recommend this. There are some teams that use just a pneumatic clutch with a flow control. Kansas seemed to have pretty good luck with theirs.

I would mount the cylinder directly to the clutch lever on the engine, rather than the cable. If not, make sure the cylinder bottoms out before the cable stops pulling.

[Wes Johnson]

MST has a similar system that you are talking about. They pretty much have to launch the car full tilt every time they want to move it. We ran a similar clutch setup to Matt's with reasonable success (well at least with respect to the clutch).

-WJ

[Wesley]

Will the valve be servo controlled? I ask because if you have a slow clutch release, you won't get
your best shift times.

I'm unsure if there are anything more than binary solenoids available for air that aren't prohibitively expensive/heavy.

You could run two solenoids, one with a valve. Heavier though. You could also (if you could find a variable electronic valve) just run a voltage decay circuit and adjust it with a pot for track adjustment.

[Erich Ohlde]

advice from the pro (without trying to sound too much like a prick http://fsae.com/groupee_common/emoticons/icon_wink.gif)

on the kansas car we run a bimba cylinder with two solenoids. one that controls air into the cylinder and one that controls air out. the 'exhaust' solenoid has a straight dump to atmo and a flow controlled needle valve outlet. for launching we use the flow control and for regular down-shifts we use the straight dump. it works very very well. I think wisconsin uses a dental solenoid (for nitrous oxide in the dr. office) that has a pulse width modulated control to vary the speed of the bleed off.

[joekid11]

I was thinking about a very similar idea Erich. Except that the Bimba cylinder has just one three-way solenoid and I will control the 'exhaust' with flow control. I plan to do a lot of testing and I will post my results here in a couple of weeks if any of you are interested. Thanks for the insight.

Joe Santore
Cooper Union

[Erich Ohlde]

did you get the bimba cylinder with the integrated solenoid? we went away from those a few years ago because we have more options with external solenoids.

[Seb@NLR]

we use humphry 310 valves. We have about 500 in stock if you need any

[Julian Choquette]

We've been running an electro pneumatic system since 2006. We have a single action pneumatic cylinder on the clutch with a 3/2 12V solenoid valve. For the shifts, we have a double action cylinder with two 3/2 12V valves. On the release side of the clutch cylinder is a flow control valve. This allows us to launch electro pneumatically( although, not very cleanly).

In 2007, I worked on a clutch and shift controller based on a microcontroller. It was all time based and worked OK for shifts. As for the clutch, we replaced the 3/2 with a 0-10 bar pressure regulator. Even with the regulator, we couldn't precisely control the clutch position because the force is not very linear. Also, the pressure regulator couldn't handle the electrical noise in our car (mostly clicking relays and ignition coils). So during Michigan, we had to revert back to our 2006 system because the car wouldn't clutch sometimes because the pressure regulator wasn't responding.

This year, we will be attempting a version 2 of the shifter/clutch controller based on closed loop feedback. Basically, a potentiometer will be fixed on both shifter and clutch barrels to measure their position. Then, by actuating a 3/3 valve using a microcontroller, we'll be able to position reliably the clutch or shifter.

At the same time, we will remeasure the required loads and try to optimize our air cylinder packaging.

Good luck with your project!

[Dave Cook]

For clutch control my school tried the 3 way valve with a restriced exhaust in the past, but there is a significant time delay issue that no one is mentioning here. When you set the exhaust restriction you are controlling the air release to get a smooth and somewhat slow clutch lever travel over the short distance of travel between the friction point and what little overtravel you can adjust in. The monkey wrench in this concept is the fact that you use more pressure to disengage the clutch than is needed to just hold the clutch open. So when you first deactivate the clutch solenoid to cause the clutch to engage, the clutch lever doesn't start moving until the pressure in the cylinder has gone from the system set pressure to the 'holding' pressure of the clutch cylinder. Once enough air has exhausted from the chamber (through a small restrictor now) the clutch lever starts its travel and the clutch engagement process starts.

This effect was most noticable not when launching, but when downshifting and having to wait on the clutch while driving. We eventually removed this system and continued using the hand clutch for launching the car. A hand clutch on the side of the car can be very light, this was a few years ago and it was a cable clutch on a Honda F3

Dave Cook

If incorporated into an airshifter system you may need two regulators, one for shift and one for the clutch or play around with the clutch cylinder bore size until the system pressure closely matches the 'holding' pressure.

[joekid11]

I see what you're saying about the time delay, but I am not attaching the piston directly to the clutch pin. The gxr engine comes standard with geometrical clutch actuation through a cable, so I plan to pull the cable to get more "play" with the clutch.

So, I bought a PWM controlled proportional valve and am going to attach it to the exhaust of the solenoids you gave me. With the use of a microcontroller and a linear pot, I'll be able to use a feedback system on the location of the piston and be able to variably adjust the flow electronically each time so that the acceleration of the piston does not exceed a certain value and bogs down the engine.

I plan to test it within the next week, but if I notice a time delay on the clutch when running it on the engine on the dyno then i'll let you know. Did you test the clutch system on an engine on a dyno because I know the clutch acts differently when its cold compared to when its hot and that may have been causing an unexpected time delay?

[Dave Cook]

Joe

I think you are going to find the behavior of this system to be very nonlinear. When you first start venting the cylinder you will see no motion for a time as the pressure drops from the system pressure to the pressure needed to hold the clutch open. Once these pressures cross you will have an increasing rate of motion that will require at rapid change in the pulse width that could be hard to do smoothly.

Its not impossible, but I also remember how many of the Formula 1 teams continued getting this wrong when they did launch control. It will be interesting to see how you do. I'd keep a hand clutch design for launching the car in my back pocket just in case.

Dave

[DJ Sanghera]

hey Joe,
I've been working on a similar system as you. We have a needle valve on our clutch to control release speed, but I'm not sure how it will hold up in testing. How is the pwm control valve working for you so far? Any success? Where did you find this valve, and what kind of valve is this exactly?
Thanks.

[joekid11]

Unfortunately, I have not had a chance to test the system yet. Hopefully, I will get to test it over the next few weeks in coordination with the launch control on the car. I haven't even done preliminary tests yet, but I plan to start them soon. I'll let you know how they go. I bought the pwm valve from kelly pneumatics the part number is kpi-vp-10-25-A0-V. Please let me know how it works out for you, I'm interested in comparing results.

[DJ Sanghera]

I looked at the part. It looks like they sell a driver board for it aswell. Do you know if you can just input a direct 0-5 volt signal via PWM straight to the valve? Or do you need to input the PWM to the driver board and then go from the driver board to the valve? Looks like a good valve with a 3-5 ms response time.
Thanks for the help!

[phil lord]

this is the first year we are running a air shifter. does anyone have any advise for us? right now we are trying to get the air shifter to put the tranny in neutral.. we are using a 3/4 bimba cylinder with parker 3 way valves.

we would love to test a pneumatic clutch setup, but what size bimba cylinder are you guys all using? i was thinking another 3/4 cylinder, but single acting with one more of the same 12v, 3-way parker soleniod valves.

[aerowerks]

I believe we (UAH in 2004) were the first to use a pneumatic clutch actuation in FSAE competition (please correct me if I'm wrong). I designed the system for no clutch - power cut up shift, clutched full flow dump down shift, and restricted flow launch. The flow restriction was controlled by simple needle valve on the outlet of a solenoid. The system worked OK but suffered from large delay on launch, and the clutch engagement was rather abrupt.