digitizing pneumatic piston position

[Technic tango]

Im trying to make return to center steering with pneumatic pistons due to not being able to use LAs or servos in a particular location. has anyone ever needed to "know" where their pneumatics are? I thought about using a spool of string attached to the end and a potentiometer on the axis of the spool to get a reading of where it is, but maybe there's a better way, with or without electronics. here's my driven, suspended, and hopefully, steered bogey I am trying to implement this on:bogey picture

[Andman]

The only way is to attach a motor to the turning point of the steered hub to read out the angle. I use that in an air suspend truck to adjust the heught. But the problem is the inaccuracy due to the internal resistance if the motor. I can post images once I'm at work.

[allanp]

I haven't needed to know their position to a greater accuracy than I can see with my own eyes. "In a bit more, out a bit more, perfect!"

But for return to center steering, you can put two cylinders back to back, to create one long cylinder that has a middle point when only one is fully extended and the other is fully retracted. But you might have to change your setup to fit that in.

I haven't tried your string fed feedback idea. Lego motors are stiff to back drive and Lego valves are stiff to move and with a noticeably large dead zone in the middle so it's a challenge but doable, but as you mentioned potentiometers, if you are talking commercial, non Lego potentiometers that move with no force at all, and then electrically driving the valve and taking the deadzone in the middle into account, that would be an interesting experiment, don't see why it wouldn't work. But if that's the idea then you could get tiny potentiometers and mount them right to the steering pivot point I guess.

Edited May 4, 2023 by allanp

[Technic tango]

51 minutes ago, allanp said:

I haven't needed to know their position to a greater accuracy than I can see with my own eyes. "In a bit more, out a bit more, perfect!"

But for return to center steering, you can put two cylinders back to back, to create one long cylinder that has a middle point when only one is fully extended and the other is fully retracted. But you might have to change your setup to fit that in.

I haven't tried your string fed feedback idea. Lego motors are stiff to back drive and Lego valves are stiff to move and with a noticeably large dead zone in the middle so it's a challenge but doable, but as you mentioned potentiometers, if you are talking commercial, non Lego potentiometers that move with no force at all, and then electrically driving the valve and taking the deadzone in the middle into account, that would be an interesting experiment, don't see why it wouldn't work. But if that's the idea then you could get tiny potentiometers and mount them right to the steering pivot point I guess.

I was also thinking about putting a hall sensor on the base and a magnet on the plunger but i dont know how the clashing magnetic fields here would work out. Ive seen game controllers having them fairly close together so maybe they just have magnets with a field perfectly fit for the distances between things

[m2fel]

Hey, not sure but looks like there are two problems with your set-up. When the suspension is compressed the pneumatic cylinder is kinked. The way it is build the 'black end' of the cylinder cannot move around the axle the way it needs when the suspension moves. If the movement was possible the whole cylinder would need to change its length when the suspension travels otherwise the the wheel will be turned (bumpsteer?). Maybe someone else can explain it better or even confirm my assumptions.

[allanp]

2 hours ago, Technic tango said:

I was also thinking about putting a hall sensor on the base and a magnet on the plunger but i dont know how the clashing magnetic fields here would work out. Ive seen game controllers having them fairly close together so maybe they just have magnets with a field perfectly fit for the distances between things

That could work, if you are worried about the magnets being too close maybe try a tiny rotary sensor on the steering pivot axis:

https://www.rls.si/eng/rm08-super-small-non-contact-rotary-encoder

The body of this one is 8mm.

Edited May 4, 2023 by allanp

[Technic tango]

5 hours ago, m2fel said:

Hey, not sure but looks like there are two problems with your set-up. When the suspension is compressed the pneumatic cylinder is kinked. The way it is build the 'black end' of the cylinder cannot move around the axle the way it needs when the suspension moves. If the movement was possible the whole cylinder would need to change its length when the suspension travels otherwise the the wheel will be turned (bumpsteer?). Maybe someone else can explain it better or even confirm my assumptions.

Theyre tatra suspension arms so it just needs to oscillate on the same axis as the springs. It is about.  stud off axis but i cannot see any visible bending or bump steer when manually compressing and steering

4 hours ago, allanp said:

That could work, if you are worried about the magnets being too close maybe try a tiny rotary sensor on the steering pivot axis:

https://www.rls.si/eng/rm08-super-small-non-contact-rotary-encoder

The body of this one is 8mm.

This looks super useful and way easier to implement, ill take a look at this

 

[Aerolight]

You really need two cylinders in series if you using pneumatics, pneumatics work with force unlike hydraulic which work with distance so the pressure would have to be constantly varying on both sides for stable steering without about 100 times a second adjustment it jitters all over the place and uses a ton of air. 

With two cylinders you can get a stable left right and centre for most mocs, if you can put up with 3 steering positions 

I managed to get a single cylinder system to work with a mechanical system that linked the steering servo and steering position driving the pneumatic switch, it worked well with a variable backlash but required an unrealistically large volume of air to work. I also tried linking two cylinders hydraulicly and driving one to drive the other but even a slight leak ruined the whole mechanism.

There is a third party hydraulic pump you can get and is the best solution if a bit expensive