Can't get R-Net with Bounder high speed package

[emilevirus]

I just took apart a Linx joystick and there's some kind of splashproofness unlike R Net. There's a gasket to prevent water from coming in. The other day I went to the store it was pouring rain. The CSJM2 started acting crazy, it was changing modes, profiles by itself. The next day, joystick wouldn't start at all. Hair dryer fixed it but that alone makes me consider Linx. Linx has its pros and cons. No on board programming. Must use a PC or a phone. Touchscreen with gloves, haven't tried yet but must be bad. Battery gauge can be adjusted for Lithium. Firmware can be updated and it's not cloud based.

[emilevirus]

Sooooo I switched to R-Net to try. Only thing left is compensation. On linx it's set to 195mohms. What would be a good starting point for R-Net. 90? I don't get why it's so high on linx.

[shirley_hkg]

Bounder motors are 700-800 rpm, which is very low , comparing to 3,200 other chairs'.

[Burgerman]

No sure.

On a bounder the motors are surprisingly weak. By which I mean pretty high impedance. But 190 sounds very very high.

As a general starting point its like this#;

Linix 35mOhm on 8mph 4 pole motors.
Linix or AMD, or AMT german 4 pole motors, for e.g both the 6 or 8 mph, are all 45mOhm. Many work best at 50mOhm.

6mph 2 pole many brand of motors are weaker, and are usually set to 70 up to 90 mOhm as stock.

The Linix 2 poles, used by Permobil for e.g are set to 65mOhm. So in between good 4 pole and 2 pole motors.

So 190 seems awfully high! Not sure that can be correct. I thought the bounders were around 70? Which is abot the same as a normal 2 pole???

If I am right and you set to 190 it will go into orbit. If its really 190 then they are junk.

[Burgerman]

RPM is not really much to do with it. Armature diameter controls RPM and Torque. That resistance measurement just determines how much current a motor pulls at a given voltage.

So a low impedance motor will draw the max current a controller can provide at a low voltage. So you get adequate torque to turn with a small voltage difference across the motors. It causes the compensation to work properly too. If impedance is high, the motors need a huge voltage to make any torque.

Just set compensation to 70.
Drive chair around. With turn acceleration and secelerations (and minimum ones) set to 100 each. So that you can feel the effect of compesation clearly. Then increase it 5mOhm at a time until YOU can feel it working, understand what its doing and then you will know when its too high. Go back a step.

Its easy enough.

Note. When setting any of the speed settings, or compensation etc in R-net you will find that the settings are very "clean" and accurate and only alter the parameter that you are adjusting and do so in a predictable way. Theres no overlap or affect on other parameters. Super predictable. They are very pure and seperated compared to the Dynamic systems. Takes way longer going in circles to get a dynamic system to respond how you want/expect based on the settings description alone..

[Yennek]

Sooooo I switched to R-Net to try. Only thing left is compensation. On linx it's set to 195mohms. What would be a good starting point for R-Net. 90? I don't get why it's so high on linx.

With the ST3 motors and R-Net I have my wife's Bounders compensation set to 150mohms. Any higher and it was uncontrollable.

[Burgerman]

150? Thats terrible. Those motors must be dismal. For several reasons.

[emilevirus]

I tried 195 but it's waaay too high. It wanted to move by itself. 150 is still too high. 110 seems better. It doesn't want to move by itself at least. It's annoying to fine tune though. Always have to unplug, open software, close, reopen etc.

[Burgerman]

OK I am bored let me tell you why.

REASON ONE.
When you use MAX turn speed, full stick, you get approximately a 10% difference in wheel speed to turn.
Or if you prefer, to turn in place quite fast, you only need one drive wheel to turn at 5% of your max speed FORWARDS. And the other one to turn at 5% of your max speed BACKWARDS. That means you will turn in place like a clock fairly fast as thats 10% speed difference between drive wheels. When watched from above. At a sensible turning rate.

So lets say you give it FULL left stick.
Your motors normally would give you say 10mph at FULL forwards speed (at 24V fed to each motor). But you are turning. Here full stick gives you only 5% of that 24V. So your "full LEFT" stick sends just 1.2V forwards to the right motor. And 1.2V backwards to the left motor. Does that make you turn? Well your motor is quite high impedance. 150mOhm or 0.15 Ohm...

If you apply 1.2V to a 0.15 Ohm circuit or motor thats just going the draw 8A at the motor. Is that enough? Not close! My own chairs take around 80 to 120A to turn in place. So you wont move. The joystick will, the chair will laugh at you. In fact you will need almost the full 24V battery voltage to turn. Why? Because with a 150 Ohm, (or 0.15 Ohm motor) you actually need 20V to get 120A. Not 1.2V. (ohms law).

Why then, does it turn? Because motor compensation SEES that 8A and it increases it. Its a positive feedback addition loop. So it adds some current to any current it sees (by increasing voltage way more than your full LEFT STICK actually sent). But its going to be very slow to ramp up all that way. And it may not even do so. You might top out at say 50A. So the motor really needs to pull MUCH more current at that 1.2V, that turn voltage, you commanded.

Remember VOLTS and RPM (or speed) are the same exact thing. These are completely tied together. Double the volts = double the RPMs or speed. Do you WANT to feed it 20V just to initiate a turn? Not really!
And remember that AMPS are directly tied to torque. Double the Amps, and you get double the torque, needed to initiate that turn. Theres 2 ways to do that. MOTOR resistance (impedance). Or Voltage applied. So you WANT low impedance motors. Those pull a lot of current and make a lot of torque, at LOW voltages. And they also make the compensation additive amount work much better. And because it doesent need to ramp up to a high voltage to initialise a turn its better controlled and faster to happen. (it steers properly)


Then theres REASON TWO.
You are rolling along or manoevering. Or accelerating. Doesent matter.
A LOW impedance motor draws adequate current (so loads of torque) at very low voltages or low speeds. So you might only need 6 volts to turn in plce or climb a threshold Instead of 24. Why is that important?
Because if you need 24V to initialise a turn on carpet. To roll up a hill. To accelerate across your kitchen. Then you are drawing FOUR TIMES as much battery current. Why?
Because 6V at the low impedance motor and get that full 100A needed to turn means 600 watts. (6V x 100A = 600W) And at the battery thats 600 watts divided by 24v or a 24AMP battery drain. But if that motor needed a full 24V to get 100A worth of current/torque (torque = Amps directly) then now you have 100A x 24V = 2400watts at the motor... And the same 100A / 24V 2400W is now drawn from the battery. 4X worse!

[Burgerman]

I tried 195 but it's waaay too high. It wanted to move by itself. 150 is still too high. 110 seems better. It doesn't want to move by itself at least. It's annoying to fine tune though. Always have to unplug, open software, close, reopen etc.

As I said...

But. Why are you shutting it all off etc? I fine tune this with a laptop on my knee. As I roll.

You also wont ever get compensation correct unless you set all the speed settings correctly inc setting turn accel, turn deceleration, min turn accel, min turn decel all to 100 to start with. Reduce turn SPEEDSS and min turn speeds as required. Set forward acceleration to 70.

Then you need to configure TORQUE correctly after you figure out compensation.

[Yennek]

And remember that AMPS are directly tied to torque. Double the Amps, and you get double the torque. Theres 2 ways to do that. MOTOR resistance (impedance). Or Voltage applied. So you WANT low impedance motors. Those make a lot of torque, at LOW voltages. And they also make the compensation additive amount work much better and because it doesent need to ramp up to a high voltage its better controlled and faster to happen.

Well yes and no. Torque is directly related to the magnetic flux. Magnetic flux is related to the current, but the structure of the motor also has a large impact on the flux as well. One way to change the flux is to change the magnetic strength of the stator magnets - this would increase the torque (but would decrease the maximum speed), if everything else is kept the same. The other way is to change the amount of flux generated in the armature. The amount of flux in the armature is proportional to both the current and the number of turns in a given winding (as well as a bunch of things like magnetic permeability of the core, and the cross-sectional area of the core). All things kept the same, an armature that had a coil of 10 turns flowing 10 amps, would generate the same flux as an armature with 100 turns flowing 1 amp, or a 1 turn armature with 100 amps.

So yes having a lower internal resistance does generate more flux if you keep the coils the same, but the coils are the main driver of the internal resistance, so they aren't really something you can vary independently. If you keep the physical layout of the armature the same, you can either fit more turns of a smaller wire, and so the impact of each amp of current goes up, but you need more voltage to get that current to move, or you can fit less turns of a bigger wire, so its easier to force the amps through the coil, but you need more amps to get the same torque.

For a given motor design, there are a number of tradeoffs as to what the "best" number of turns is. One factor is how tightly you can pack the turns of wire into the slots on the armature. with bigger wires, if they are round, you end up with more "open gaps" between them, but with smaller wires you may be able to pack them more efficiently but loose a higher percentage of the the space for insulation between the wires. I've seen some particularly high horsepower motors that used flat ribbon coils rather than wires.

It is hard to make a comparison between motors of different physical sizes and layouts. Size often has more to do with constant power handling capability, as well as peak torque. At a certain point, you just can't force more magnetic flux through a given size core. When that happens, the core saturates, and the linear relationship between current and torque flattens out, and the difference between that linear slope and the flattened slope is dissipated as heat in the core. There's also multiple ways to layout and commutate the armature coils as well, that can affect things such as the smoothness of the torque.

[Burgerman]

Whilst all true, I was generalising a lot to try and fit it all in a post!

But the problem remains. If you have a high impedance motor, in this case we decided on a 150mOhm for e.g. Then it still requires 20V approx (which is about what you actually get at STALL and a 24V lead battery due to battery chemistry and peukert etc . So to achieve 120A you need basically full battery voltage.

So the only way to get a turn in place, which sends around 10% pulsewidth difference to the motors (5% each + and -) is to add a huge amount of extra current via very tweaked load compensation. So you add your 10 percent difference across motors, and then we wait for the current to ramp up to whtever it takes. Does it TAKE 120A to turn in place. Yes sometimes. Measured. But remember that this can only occur on these motors at 0mph...

As you increase speed the motor generator effect reduces this current draw even further.

If we start with a 50mOhm motor, then it draws that 120A at 33% of battery voltage or a 33% pulsewidth. It ramps up faster as it actually has a starting current it can see and amplify. And it can still access peak torque as limited by the 120A controller only, not just at stall but while actually moving at some speed. So much greater control capability in motion too.

At half speed we only have 12V of voltage available. Which means the 120A of "torque" is still available. Where it falls to around half on a 150mA motor.
And of course battery current is higher for a given motor current than on a lower impedance motor. Yes generalisations. But also pretty easy to demonstrate regardless.

Regarding square windings, we dont see such things, or rare earth magnets etc on powerchair motors. I did see this in some fancy hobby motors and a tesla car motor. That stuff is great. But we dont get that! ££$$...

And yes I am generalising and simplifying for obvious reasons. Even so, general principles remain true here regardless.

[Burgerman]

Well yes and no. Torque is directly related to the magnetic flux. Magnetic flux is related to the current, but the structure of the motor also has a large impact on the flux as well. One way to change the flux is to change the magnetic strength of the stator magnets - this would increase the torque (but would decrease the maximum speed), if everything else is kept the same. The other way is to change the amount of flux generated in the armature.

But while this s also true, we then add a gearbox to give a free running speed of 6 or 8mph. Which basically changes this more torque/lower RPM (or the opposite) to give a very similar result...

So we have to compare thise things against designed road speed. Our 6mph or 8mph gearing. And we rely totally on load compensation. So we need a low enough impedance to allow that to work effectively as well without waiting for it to ramp up to 100% pulsewidth. So as to draw lower battery current at high motor current and quickly as needed. The controller is working as a better "electronic gearbox" on a low impedance motor which reduces battery current at lower pulsewidths while still having high torque. And yes generalisations/all else equal/etc. Which it mostly is though.

[emilevirus]

But. Why are you shutting it all off etc? I fine tune this with a laptop on my knee. As I roll.

How? I might have misunderstood the process then? I unplug usb, open OEM, set compensation, save file, plug usb, open dealer, flash the file. Is there an easier way?

[Burgerman]

I plug it all in, connect and just leave it that way. Even watch the motor current and amps as well as battery current etc as I drive around. Tweak as I go.

[emilevirus]

You must have OEM dongle then?

[Burgerman]

Yes. Never considered that.

[emilevirus]

Yeah you're lucky. It's annoying to mess with settings with that little trick. It's fine when you do it once but not trial & error.

[Yennek]

Dealer dongle user here. What I do is:
* Connect dongle to chair and laptop
* Turn on chair
* Open dealer version of the software, and let it connect, but don't close
* Open OEM version of the software
* Edit my existing OEM file and save, but don't close
* Switch to the dealer version and use it to write the OEM file.
* Repeat prior two steps as needed

Not quite as easy as being able to upload directly from the OEM software, but not terrible. If for some reason I have to disconnect the dongle from the USB, it seems pretty random which version of the software detects it first. If the OEM detects it first, then I do have to close and reopen it.

[emilevirus]

Yeah I'll try that. Would be less annoying than unplugging / plugging.

[Burgerman]

Before you start.
Save 10 clones of the setings file on your desktop with 4 or 5 higher and lower compensation settings.

So:
SETTINGS 45mOhm.r-net
SETTINGS 50mOhm.r-net
SETTINGS 55mOhm.r-net
etc...

Then just load as needed.

Delete the ones you dont need later.
etc etc...

[emilevirus]

I think 115 seems ok but I'll test outdoors soon. Best way to see is go uphills and release the joystick. It shouldn't roll back right?

[Burgerman]

Yes. And it should not slow or speed as you hit a ramp or slope. While you drive at say fixed joystick position at 1/4 of max speed.

Many things to test. But IF your programming is set up so that the chair drives in a linear fashion, and steers properly then its easy to feel when its too low, too high. You can end up at the right place and easily tell if its 5mOhm too high or low. By the way it drives and feels. Once there, the last thing to set up is torque.

If it needs 115 then those physically big motors are about as good as a set of 2 pole motors on a 90s chair. I am surprised. Compensation and torque isnt going to be hugely effective at low speeds. And will do almost nothing from slow speeds upwards. I doubt that it will ever reach that 120A limit other than at stall.

[emilevirus]

So a few things. R-Net can pull way more amps from the battery than Linx. Linx caps at 100A while I've seen R-Net peak at 140A. If going downhill, R-Net applies the brakes before the chair stops. No big deal for me but it'll wear out the brakes quicker for sure.
With Linx I could set it so that it'd always start the joystick at the lowest speed. Many people are using the joystick to move it out of the way and I don't want them to kill themselves as everything is set to 100%. Is it possible with R-Net? Like everytime it starts, it goes to profile 1 where acceleration is set to 1%.

[Burgerman]

Yes. 1%???
I set mine to 70. But have an EMPTY CHAIR profile where max speeds are limited to a crawl. Same with turn speeds. And compensation set to 10% less on that profile. And leave that enabled when I get out. A 4 year old can move my chairs around safely and easily.

R-Net can pull way more amps from the battery than Linx. Linx caps at 100A while I've seen R-Net peak at 140A. If going downhill, R-Net applies the brakes before the chair stops. No big deal for me but it'll wear out the brakes quicker for sure.

Both are a sure sign that compensation is working harder on the r-net. So you can match them.

[emilevirus]

Yes.

How? couldn't find it.
Carers are used to low acceleration so would rather be safe. It BARELY moves so can't kill anybody.

[Burgerman]

Set speeds and turn speeds to super slow. On a single profile regardless of high or low (1 to 5) speeds set.

Having a 1% acceleration just makes it less predictable and less proportional. If it goes exactly where they move the stick at stupidly slow speeds thats better for safety when moving it around empty.

Better still, teach tham to drive the thing. Low acceraraton just means it gets faster than they wanted over time. Same reason setting turn accelaration low makes it impossible to steer.

[emilevirus]

Ok but can I make it boot up at the lowest speed?

[Burgerman]

Not as far as I know. You can make it boot in many different modes like seating or programming or whatever.

My 1 speed is too slow for me to use even indoors. But I just set to 1 when I get out.

I also have a complete profile called EMPTY CHAIR where high and low speeds are all set to crawl... For any numpty to move my chair. Both ways make it super slow.

Personally I use 5 even indoors.