coolskywizard
Posts: 1
Joined: Sat Jun 01, 2019 5:46 am

Self-managing race car air suspension

Sat Jun 01, 2019 6:49 am

Alright guys, I'm new to this and hope I'm in the right spot.
A friend recently sold me a Pi 3 Model B. I've been trying to figure out what to do with it, then it hit me. :idea: I want to try and build an "affordable" system that can actively monitor and adjust air pressure in each of the 4 air-shocks using height as the reference measurement. The plan is to use these shocks https://www.airliftperformance.com/product/82607/ and this encoder https://www.adafruit.com/product/377 with an arm instead of a dial in order to measure height.

I am having a tough time finding pressure sensors for this application though. The shocks are rated at 100psi so I'll only need a sensor that goes a touch above that, and reads very accurately and frequently.

I am also curious if/how I could use my Pi to operate four separate solenoids in order to provide air to the shocks? And what solenoids I could use?

And in the end I would also like to be able to use this as a recording system so I could plot the high and low pressure points on a graph.

I was also planning on using python to program this being how simple it seems.

So please share any ideas you might have, I'm all ears... err eyes I guess :lol: Cheers

User avatar
omegaman477
Posts: 148
Joined: Tue Feb 28, 2017 1:13 pm
Location: Sydney, Australia

Re: Self-managing race car air suspension

Sat Jun 01, 2019 11:42 pm

This is the sort of pressure transducer I would use: https://au.element14.com/honeywell/px3a ... transducer
It produces a 0-5VDC voltage proportional to the pressure sensed. You would then need to pass that signal through a AD convertor. There are several 4 channel ADC hats available for the RPI. I would overrate the sensor by at least 200% as I am sure you will get some fairly high pressure shocks, and these sorts of sensor do not cope well with being exposed to over limits.

The rotary encoder you suggest is ok, but I would be concerned about its ruggedness. There are industrial versions available. A cars suspension system would be the most hostile place on earth :-)

You should also sense the temperature of each shock, as with temperature rise, you will get a increase in absolute pressure. The sensors are only partially compensated, not for the extremes of a car n all weather conditions.
..the only thing worse than a stupid question is a question not asked.

PhatFil
Posts: 1368
Joined: Thu Apr 13, 2017 3:55 pm
Location: Oxford UK

Re: Self-managing race car air suspension

Sun Jun 02, 2019 4:33 am

interesting project..
any of the 1/8" gas solenoid valves ebay spat out at me looked suitable if they publish pressure limits in bar or mpa rather than psi, google should provide a conversion easily enough..

in my limited experience most gas valves are suitable to control source pressures within the range of most regulated sources and my co2 regs can supply way more than 100psi not that i have much use for co2 at that pressure ;)

a solenoid valve will draw a significant amperage (in pi terms) and most suitable options will likely draw @12v not 3.3v so to switch on and off you will probably devote a single gpio pin to controlling a mosfet or relay to actually switch the load current to each of the valves. so 4 x gpio output to solenoid control..

keep us posted on your findings plans and progress please.

PiGraham
Posts: 3604
Joined: Fri Jun 07, 2013 12:37 pm
Location: Waterlooville

Re: Self-managing race car air suspension

Sun Jun 02, 2019 8:24 am

coolskywizard wrote:
Sat Jun 01, 2019 6:49 am
I want to try and build an "affordable" system that can actively monitor and adjust air pressure in each of the 4 air-shocks using height as the reference measurement. The plan is to use these shocks https://www.airliftperformance.com/product/82607/ and this encoder https://www.adafruit.com/product/377 with an arm instead of a dial in order to measure height.
Is this a real world project intended to control a real race car at racing speeds?

If it is the here's a not of caution. What's the worst that could happen if something goes wrong with hardware or software in this control system?
I should think there is the possibility of destabilising the car, and an unstable race car can be deadly to driver and others. Can you incorporate safety devices that prevent anything nasty happening?

You really need a "hard-realtime system" with guarantees on latency and watchdogs. A RPi running Linux can't do that.

User avatar
omegaman477
Posts: 148
Joined: Tue Feb 28, 2017 1:13 pm
Location: Sydney, Australia

Re: Self-managing race car air suspension

Sat Jun 08, 2019 10:53 am

PiGraham wrote:
Sun Jun 02, 2019 8:24 am
coolskywizard wrote:
Sat Jun 01, 2019 6:49 am
I want to try and build an "affordable" system that can actively monitor and adjust air pressure in each of the 4 air-shocks using height as the reference measurement. The plan is to use these shocks https://www.airliftperformance.com/product/82607/ and this encoder https://www.adafruit.com/product/377 with an arm instead of a dial in order to measure height.
Is this a real world project intended to control a real race car at racing speeds?

If it is the here's a not of caution. What's the worst that could happen if something goes wrong with hardware or software in this control system?
I should think there is the possibility of destabilising the car, and an unstable race car can be deadly to driver and others. Can you incorporate safety devices that prevent anything nasty happening?

You really need a "hard-realtime system" with guarantees on latency and watchdogs. A RPi running Linux can't do that.
@PiGraham raises a very good point. One which we see often on this forum. My following post is more generally directed, to anyone considering mission critical applications.

The RPI is really not appropriate for industrial "mission critical applications" where its unforseen failure could place life at risk. If your hardware design was such that a complete failure of the RPI meant that the suspension system would simply lock and hold a limp home mode then your fine. I wont discuss why an RPI is not the best choice here, that's a different and far more indepth subject.

A good designer thinks about what will happen when everything works, a great designer thinks about what will happen when everything fails. Please consider reliability and fault tolerance in your design.

I draw on the Auto industries path towards full Electronic Braking systems. Even in todays most advanced drive by wire cars (Tesla, Mercedes E series etc) there is still a basic mechanical backup linkage between the pedal and the brake caliper. Simply put, the best car networking, software, computers, servos and sensors are all useless of there is no power. No power no brakes.

But as we have seen in aviation, full fly by wire is a reality, but even then the pilot can manually pump the ground brakes in case of a total hydraulic and electrical failure.
..the only thing worse than a stupid question is a question not asked.

GettingGreyer
Posts: 5
Joined: Mon May 20, 2019 10:15 pm

Re: Self-managing race car air suspension

Mon Jun 10, 2019 1:50 pm

Looked at this 4-5 years ago...

Perhaps reverse engineering something like this?

https://www.ridetech.com/products/air-s ... ol-system/

Here's something that might be interesting as well.
https://www.ebay.com/itm/Air-Suspension ... -Navigator

Look to the industrial space for precision. Reach out to a local pneumatics firm to get vendors/parts. There are all sorts of 5-way valves that you can fire with relays (not much power at all) as well as digitally-fired/wireless/CANN...

As to the comments about "appropriateness of technology"? Pi is a platform to facilitate experimentation. Start here, prove out some use cases and pivot to specialized. If you try it in a race, perhaps figure out a way to keep the thing cool and pot the crap out of everything!

User avatar
Imperf3kt
Posts: 2775
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: Self-managing race car air suspension

Mon Jun 10, 2019 11:31 pm

GettingGreyer wrote:
Mon Jun 10, 2019 1:50 pm

As to the comments about "appropriateness of technology"? Pi is a platform to facilitate experimentation. Start here, prove out some use cases and pivot to specialized. If you try it in a race, perhaps figure out a way to keep the thing cool and pot the crap out of everything!
I think you missed the point.
The raspberry Pi running Raspbian is not a real time operating system. There may be delays on response as the CPU is busy doing another job.

In other words, if sensor A calls the CPU and says EMERGENCY, RELEASE THE PRESSURE RIGHT NOW! the CPU is going fo say wait your turn and that could potentially lead to a loss of stability in the cars suspension and inevitably, a crash, the very thing the system is supposed to help prevent.

Don't use a tool just because it's convenient, use a tool because it's appropriate for the job, 'experiment" or not
55:55:44:44:4C
52:4C:52:42:41

User avatar
omegaman477
Posts: 148
Joined: Tue Feb 28, 2017 1:13 pm
Location: Sydney, Australia

Re: Self-managing race car air suspension

Fri Jun 14, 2019 2:32 pm

GettingGreyer wrote:
Mon Jun 10, 2019 1:50 pm
If you try it in a race, perhaps figure out a way to keep the thing cool and pot the crap out of everything!
While potting is a sound method to protect against mechanical shock, moisture and dust. It does not help heat dissipation, in some cases it will work in reverse and make the problem worse.

Solve your heat dissipation issues first with a passive solution (heat pipes etc) then POT.

There are some new thermally conductive but dielectric potting compounds on the market but they all perform poorly with chip on PCB style dissipation problems.

My best method is to thermally bond a copper bar to the chips, and then both that heat sink/pipe) to the external metal case.

Potting is a bitch of you want to repair (or swap the SD card).
..the only thing worse than a stupid question is a question not asked.

Return to “Automation, sensing and robotics”