Actuator hardware - switching times

[james-portman]

Hi,

Not so much an issue but I was having a look at electronic solenoids as well as pneumatic like you are using (Festo 525150?).

I did some napkin maths and I think even at 1000 rpm, the time for a stroke in a four stroke engine is about 30ms e.g. the intake stroke.
From the Festo website it looks like the pneumatic valve you are using takes about 8ms to switch on, 4ms to switch off.
Some of the linear electronic ones had vaguely similar numbers to that.

Looking at higher RPM this might become an issue, 3000rpm for example I think would be 10ms for a stroke, but you have 12ms worth of valve opening time.

Also just in general, if you ignore the opening and closing times in the code then you will gradually retard the valve timing as a consequence, as RPM increases.

I just wondered how far you have gone into this type of calculation, I might be wrong with my maths or missing something, just interested in general, it's a really nice project that you are doing!

** After writing all of that I just saw that Festo also make some faster switching versions called MHJ, with more like 1ms switching time, that could make this type of system more feasible, they are expensive though - https://uk.rs-online.com/web/p/pneumatic-solenoid-pilot-operated-control-valves/1752145/

Thanks

[whitfijs-jw]

Looking over some freely available research papers it seems the advantage of using pneumatic valves is simplicity of the control system with the main limitation being exactly what you've pointed out -- valve actuation delay:

http://kth.diva-portal.org/smash/get/diva2:542744/FULLTEXT01

1ms valve response time would definitely help, but the total delay between the rising edge of the control signal to valve opening would have to be characterized to determine the effective RPM range -- valve position sensors are needed. This doesn't take into account things like soft closing of the valve to prevent premature wear on the valves and valves seats -- if we care about that kind of thing.

More recent research with electro-mechanical actuation requires considerably more complex control systems (at least this configuration), but it seems that it wouldn't be limited to the same issues that as the pneumatic system.

https://www.researchgate.net/profile/Zlatina_Dimitrova/publication/330801103_Robust_control_for_an_electromagnetic_actuator_for_a_camless_engine/links/5c62aed9299bf1d14cc074de/Robust-control-for-an-electromagnetic-actuator-for-a-camless-engine.pdf

It would be a struggle to implement this type of control on an Arduino. The sample period they used to resolve valve positions at 6000 RPM in the paper above was 50 usec.

This was from a pretty cursory literature search and I don't have access to a lot of the papers that someone at a college/university would have access to through their institution. Maybe there is a pneumatic configuration that is not dependent on the air valve opening times.

[james-portman]

Just a quick note about the speed of arduinos etc, there are things like the Teensy 4.1 (and others) which way faster than an Arduino mega
https://www.pjrc.com/store/teensy41.html