Controllers for lightweight chairs

[Burgerman]

See this:
https://www.wheelchairdriver.com/gopro/bot.mp4

Its got 2x 4.5Ah Lipo batteries so 9Ah battery.
Its got a 20A per channel roboteq controller running the same script that my BM4 has by lenny, and thicker wires than your chair...

I am amazed that your chair works.

[emilevirus]

I'll see the current draw when I plug the PM to my computer. Curious to see how much current it draws. With current controller, it doesn't have enough power to get over small bumps. Now it sure will but will wires melt

[Burgerman]

Yes.

Do you understand why the current to motors is normally 3 or 4x higher than battery cables at stall?
And exctly the same as it reaches full motorvoltage (max pulsewidth) under load as you accelerate?

[biscuit]

What does stalled mean?

[Burgerman]

Drive your chair to a wall. Try and push it down. Motors get full power, but cant move! Tat motor is stalled. The controller feeds it max power at stall.

[emilevirus]

Got everything setup. It's actually drivable now and pops wheelies!

[Burgerman]

Good. Watch the motor current on the tiny wires!

What value did you set motor compensation to?
How did you determine that it is corect?

Quite literally everything depends completely on that being right. It is too non responsive under load (under different loads) to have any considtency of control if its set too low.
It is snatchy, may damage motors, and may even run away with no way to stop it either spinning in place or driving forwards or backwards or some combination, or occilating etc if set too high. And the correct setting is quite critical.

[emilevirus]

Good. Watch the motor current on the tiny wires!

What value did you set motor compensation to?
How did you determine that it is corect?

Quite literally everything depends completely on that being right. It is too non responsive under load (under different loads) to have any considtency of control if its set too low.
It is snatchy, may damage motors, and may even run away with no way to stop it either spinning in place or driving forwards or backwards or some combination, or occilating etc if set too high. And the correct setting is quite critical.

Ohh right, how do I determine that it's correct? I just left the old values that was on my old chair: 75 mOhms.
I checked a bit by blocking a wheel and seemed fine but not sure.

[Burgerman]

Well in the past, when swapping motors etc I just tested by feel.

I started with a value that was so low that it did completely nothing. Why? Because in order to "feel" it work you need to know how a chair responds without any when programmed in a normally responsive way. Try it turns in place, accelerating, decelerating, moving over a threshold, on a ramp. You will soon see why its super important! Even slowing down is drasically affected. Its ESSENTIAL to do this to gain a feel for things first

Then increase value in small steps and each step will appear to make no difference. Its only when you get close to the correct value that it suddenly begins to work. At this point a small change hs a big effect! Dont do large steps as if you go too far then you can have big issues...

At least in your case you know you can safely go to what you have set now.

Once you begin to feel it doing something and the chair feels much more responsive then keep increasing it and it will keep getting better. It will behave in a more controlled way regardless of load. The correct point is where for e.g it turns in place exactly the same "feel" on a thick carpet or a smooth tiled floor. Where at a FIXED low throttle (stick) position or set to speed 1 it doesent slow, OR SPEED UP as you meet a ramp regardless of angle. And where the response feels smooth still.

Go too far and it will get snatchy, uncomfortable. Go even further and it becomes almost impossible to control properly and very eratic. Go further still and you may get a runaway and broken legs. Or under a bus.

So thats what I do to get it correct.

[emilevirus]

Battery voltage is 26v and when I go forwards it drops to 23v. Does that mean the cells they used aren't enough C rating? Usually lithium shouldn't sag right?
Or wires are too small...

[Burgerman]

Well they sag a BIT... But it depends. What is the current at battery and motor. You measure battery voltage at the battery so wires dont make a difference. If you lose 3V in the wires you are going to see some smoke soon.

Which cells?

[emilevirus]

Don't know. I'd have to take them apart but they're chinese stuff so my guess is 2c 2600mah.
EDIT: found it online. LISHEN LR18650SK 2600

[Burgerman]

https://imrbatteries.com/products/lishe ... y-lr1865sk

Cheap, only 5A capable! 10 if you have a couple in series as you do. And thats maxed out.

So yes you will get some voltage drop and they wont last.

Lithium ions likes a nice easy life. Slow charge, slow discharge, keep between 80 and 20% full, and they will last forever. Dont, and they wont.

[emilevirus]

Yeah, sucks. They could have used newer 3500mah cells. I wonder if I can get my hands on the shell. I could make my own pack with Samsung cells.
https://movexmobility.com/product/batte ... ir-5-2-ah/

[Burgerman]

25.4V CHARGE? Means its 7S. And two strings in parallel. So 14 cells. And that powers the chair? There are bigger packs in my model helis!

Its only 5Ah. Why not 8 cells of lithium ion phosphate maybe 20Ah.

[emilevirus]

There are 2 packs in parallel so 10ah

[Burgerman]

Words are not as good as pictures. How many in parallel in total?

[emilevirus]

I mean there are 2x 5.2ah in parallel. They did it like that because airplanes. Batteries must be less than 200wh but 2 separated batteries is fine.
So basically one pack is 7s2p and I have two packs so it's 7s4p.

[Burgerman]

I see.

[emilevirus]

Idle consumption isn't bad after all... After 2 months of not using it battery voltage is still around 26V.

[faico_26]

Well in the past, when swapping motors etc I just tested by feel.

I started with a value that was so low that it did completely nothing. Why? Because in order to "feel" it work you need to know how a chair responds without any when programmed in a normally responsive way. Try it turns in place, accelerating, decelerating, moving over a threshold, on a ramp. You will soon see why its super important! Even slowing down is drasically affected. Its ESSENTIAL to do this to gain a feel for things first

Then increase value in small steps and each step will appear to make no difference. Its only when you get close to the correct value that it suddenly begins to work. At this point a small change hs a big effect! Dont do large steps as if you go too far then you can have big issues...

At least in your case you know you can safely go to what you have set now.

Once you begin to feel it doing something and the chair feels much more responsive then keep increasing it and it will keep getting better. It will behave in a more controlled way regardless of load. The correct point is where for e.g it turns in place exactly the same "feel" on a thick carpet or a smooth tiled floor. Where at a FIXED low throttle (stick) position or set to speed 1 it doesent slow, OR SPEED UP as you meet a ramp regardless of angle. And where the response feels smooth still.

Go too far and it will get snatchy, uncomfortable. Go even further and it becomes almost impossible to control properly and very eratic. Go further still and you may get a runaway and broken legs. Or under a bus.

So thats what I do to get it correct.

I'm modifying motor load compensation in my Q300R (4 pole motors) and it makes a huge difference on the carpet. It was set by default in 40 and following your instructions above I have increased it little by little until it makes an effect. Now i'm on 75 and it starts to be jerky. I have turn acceleration on 90 and turning speed on 25. Shoud low these settings or low the compensation? I'm worried to end under a bus if the chair starts running uncontrollably.

[Burgerman]

>>> I'm modifying motor load compensation in my Q300R (4 pole motors) and it makes a huge difference on the carpet.

Yes. And ramps, turns, and control generally.

>>>It was set by default in 40 and following your instructions above I have increased it little by little until it makes an effect. Now i'm on 75 and it starts to be jerky.

I am not surprised that far too high.

>>>I have turn acceleration on 90 and turning speed on 25.

You MUST set MINIMUM turn acceleration to the same. 90.
And you MUST set TURN DECELERATION as well as MINIMUM TURN DECELERATION also to 100.
Unless you do this it will:
a) steer like a hovercraft.
b) be impossible to determine when you have compensation set correctly.

Also. TORQUE adds a small amount of compensation (or a small boost in turn value) at VERY slow speeds only. So that should be used to help initiate a turn on carpet rather than having compensation set too high generally.

>>>Shoud low these settings or low the compensation? I'm worried to end under a bus if the chair starts running uncontrollably.

Lower compensation.

[faico_26]

Very clear, thanks BM. I will do as you say. For your reference, with the default 40 compensation the chair wasn’t able to turn on carpet or grass. Now it does, but too abruptly.

[Burgerman]

That could be because you have:
low MINIMUM forwards acceleration.
low MINIMUM reverse acceleration. Set THESE as high as possible.
low forward acceleration.
low reverse acceleration.
low reverse speed.
low MINIMUM turn speed
low minimum turn acceleration
low minimum turn deceleration. Set these all to maximum possible.
Low TORQUE setting. Does the same as higher compensation but only at VERY slow speeds. Set that higher instead of increasing compensation to initiate a turn.

[faico_26]

I tried all these settings as you say with 40 (default) compensation but when trying to turn on carpet with small joystick movement the motors try to move the chair but they can’t and I can hear electric sound from the motors trying to start moving with no results. So I have to apply full joystick with the result all we can imagine, sudden movement when the chair finally starts moving once the carpet resistance is over. The same with grass. This won’t happen with forward or backward movement. The higher resistance appears when trying to turn right or left.
When increased compensation this problem improved a lot. Now I will turn down compensation but keeping the ability to turn on carpet or grass.
The inability to climb curbs still is there. I am not surprised, 90amp (actually as you said in another post is 80amp with 10 seconds 90amp boost), a chair super heavy, probably the motors are pulling 100amp or more when facing high resistance as a curb. But I will check it with a clamp meter. Also it could be batteries going voltage down, they’re 64 AGM batteries.

[Burgerman]

Its all of those things. 40 may be too low. 60 may be too high. Set it where it works but isnt too jerky.

But it sounds like your biggest problem is that your chair is too nose heavy.

[faico_26]

I wouldn’t say it’s nose heavy, honestly. I am obsessed with having a proper rear COG. I could maybe move it a little bit back, I have to check it.
I have compensation in 60 instead of 65, it’s better now and it’s works on carpet. I tried 55 but the motors had problems moving the chair on carpet.
Now it’s not so jerky. Can I say it’s safe now? I mean, is there a risk that the chair runs by itself out of control? I didn’t feel any strange reaction.

[Burgerman]

If its smooth its safe.

[faico_26]

If its smooth its safe.

Thanks BM!

[Burgerman]

Here is what the R-Net HELP says:

COMPENSATION

This matches the Power Module to the motors used on the wheelchair. This is necessary to achieve optimal performance and control, especially on a gradient.

The programmable range is 0mOhms to 500mOhms in steps of 5mOhms.


First of all, check that the wheelchair is controllable and comfortable to drive on a smooth, level surface. If it is not, then the value of Compensation may be too high and should be reduced.

1. Then drive the wheelchair towards a gradient at a very low FIXED speed, perhaps 10%. There should be no reduction in speed as the wheelchair climbs the gradient. If the speed drops or the wheelchair stops, then the value of Compensation may be too low and should be increased.

2. As a final check, drive the wheelchair up a gradient at full speed and release the joystick. The wheelchair should come to a smooth, complete stop and should not rollback before the brakes are applied. If the wheelchair does roll-back, then try increasing the value of Compensation slightly. If the wheelchair appears to drive slightly up the gradient before the brakes are applied, then the value of Compensation should be reduced slightly.

All these tests should be conducted when the motors are cold and hot.

WARNING:

These tests should be conducted in an open space and a restraining device such as a set belt should always be used. PGDT accepts no liability for losses of any kind arising from failure to comply with this condition.

The wheelchair manufacturer is responsible for ensuring that the value of Compensation is correctly set for the motors being used on the wheelchair. This is particularly important if replacement motors are being sent out to upgrade a wheelchair already in the field. Failure to comply with this requirement may result in poor control characteristics, which in extreme instances can make a wheelchair uncontrollable and potentially unsafe. PGDT accepts no liability for losses of any kind arising from failure to comply with this condition.

The wheelchair manufacturer is responsible for always ensuring that any replacement motors supplied as spare parts are fully compatible with the originals that the system was programmed to match. Failure to comply with this requirement may result in poor control characteristics, which in extreme instances can make a wheelchair uncontrollable and potentially unsafe. PGDT accepts no liability for losses of any kind arising from failure to comply with this condition.

A control system must not be moved from one wheelchair type to another without first ensuring that the system is correctly programmed for the wheelchair it is intended to fit it to. Even if control systems have the same PGDT part number, they may not be programmed the same. Failure to ensure the system is correctly programmed for the wheelchair may result in a hazardous condition. PGDT accepts no liability for losses of any kind arising from failure to comply with this condition.