PINNED - Roboteq Controller - developing for powerchairs

[rcbook]

You really want to get a Spektrum reciever like the AR500 that failsafes to null output on loss of signal.

No I did not think of the failsafe’s to null output on loss of signal. OO stupide me
But the fact is I still loss the signal not safe

[Burgerman]

Ok I think I know what you are doing.
You are connecting the joy stick directly to the controller and then using the scripting to control
Or tame down the jumping of the motors at low speed with the compensation Right?
I could really use your help. Can you get me the scripting?

Yes, email me. Burgerman@ntlworld.com

Compensation gives full torque at small stick deflection. So chair slows, and climbs thresholds, turns at LOW speeds etc predictably and with better "power".

Turn rate, reverse speed is all adjustable inside the script. Like a mobility controller.

Some player’s need the dead band open up more than others because of lack
Of muscle like my son we have dead band set at 15% with only 30% total through movement of
Joystick would make full speed on the chair. If the dead band is to close and he gets hit or hits a bump in the road as you know any slight movement of the joy stick causes the chair to react right now.

This is all adjustable, and theres 3 smoothing exponential curves too? And throws are now adjustable in script so not a problem!

Like you I have all the p@g factory level programming. We find the r-net can be set up to do what
It’s not supposed to do like react fast turn in a dime set power levels up so on
But it still does not compare to the roboteq.

Tell me about it! It just goes! Esp on 45v lithium... But it makes control EASIER and way more predictable.

And yes I have been in rc airplanes for 38 years it was easy using rc gear and you are right it’s not safe
But I am not good at all with scripting or open source codding. Back in the day when I first
Started using 2550 I connected the joystick directly to the controller and it was way to jumpy
And when we lost one leg /joy stick wire broke OMG hold on for a wicked ride. Yes it hurt lol
When I finely stopped ran to something that did not move or break down. Of course after that
A kill switch was installed for safety.

That is also settable as a safety (out of limit range) so it stops?

I hope I am not asking for too much I have spent the last 15 years working with the soccer teams
Across the US In my little world I am the go to man.
Thanks for the quick response
Robert

Email me, discuss here later! This is Lennys script, a member of my forum.

Yes, email me. Burgerman@ntlworld.com

[LROBBINS]

Deadband, exponential (or logarithmic, but I can't imagine a use for that) curving, out-of-range detection (that will invoke a safety stop if a joystick wire breaks) and many other things can be set directly in the Roborun configuration page - they don't need to be scripted. The Roboteq user's manual explains how to do all of those things. If you have not downloaded the current versions of the software, firmware and manual, you should. PWM output joysticks are available and would have better noise tolerance than analog ones. The Roboteq can handle either.
Ciao,
Lenny

[Burgerman]

PWM output joysticks are available and would have better noise tolerance than analog ones. The Roboteq can handle either.

The only prob is that I really want and use the RC input often. Otherwise I would use a pulse width joystick myself. But to do so means you used up the RC input...

[cdb0ewm]

Working on getting my understanding of the Roboteq controller I've read all the posting and getting started with the manual.

Is there a script available that includes the latest and greatest work of forum members combined in to a complete working program?

Thanks for all the time and effort by everyone

[ex-Gooserider]

I think the latest script is linked to one of the messages earlier in this thread, at least that's my recollection...

Since BM is still building on the BM3, it hasn't had much active development lately, as it seems to be working as well as can be determined by bench testing - further work needs to wait until he is in a position to get some actual stick time in the chair...

In addition, Lenny has started a really ambitious project to develop a complete open-hardware / free software CANbus control system. When done he thinks it should be at least as good as any of the commercial systems for safety, while supporting all the adjustability we want, the higher voltage support that BM (and eventually some of the rest of us) want, and the complex control needs of his daughter or other folks with severe limitations... It sounds like it will be in a ready to test state soon, and will solve a lot of the problems we currently have in figuring out how to interface with the Roboteq safely.

ex-Gooserider

[LROBBINS]

ex-Gooserider mostly has correctly summarized the state of my activities. The last version of the script that John will be testing is somewhere in this thread. Until it is actually running on a Roboteq with motors attached there's little more I could do.

Yes, I am working on a full open-source CANbus control system. I have not, however, programmed it for higher voltages and the DC-DC converters powering each node are also not suitable for >30 V. Only a few numbers would need to be changed for higher voltage use, and a central DC-DC converter would need to be added to provide 24V (or 12V as John intends to use), to go to higher voltages, but the latter would be needed for lights, brakes etc. in any case. I will eventually, however, be testing on a 24V system. I have one more CAN node to prototype, (containing a display, SD card for data logging, and a real-time-clock, none of which I've ever programmed before) and if things continue to go well, I will then plunk down the money to buy a Roboteq and get it delivered to Italy. Even after that there'll be quite a bit of detail design to do: mounting a main contacter for the Roboteq, relays or MOSFETs for lights & actuators, boxes to case things etc. I am a total amateur at this, learning as I go, and I have no "staff" to help me, so the time frame is decidedly long - I've been at this about a year so far - and I offer no guarantees that I'll actually succeed.

Ciao,
Lenny

[Burgerman]

If you email me I will send you the latest version that is working for me here on the limited testing I have done. But it seems good to go.

[PowerHungry]

Hi
Can Roboteq be used with odyssey batteries as well?

[PowerHungry]

Scrap that question. Would i be better off getting the roboteq instead of pilot plus for the 2 odyssey batteries and 6mph motors?
Thanks.

[LROBBINS]

Given the current state of things, only if:

(1) you like to experiment and build things, and have the skills to (at a minimum) add main power contactor, actuator circuitry (if needed), lighting circuitry (if wanted), brake circuitry and so on.
(2) you don't need any specialty controls
(3) you accept the risk of using an analog or pulse joystick and the somewhat untested fail-safe and safe-fail properties of the Roboteq.
(4) you can't get an OEM-level programmer for the P&G or Dynamic controller you chose
(5) you are likely, in the future, go to LiPO4 and/or higher voltages (even if that is to gain range and battery life rather than speed)

With a 24V lead system, all you will gain with the Roboteq is higher maximum current. At 6 mph, unless you and your chair are very heavy, or you are playing power-chair sports, that probably won't buy you much.

Ciao,
Lenny

[PowerHungry]

Given the current state of things, only if:

(1) you like to experiment and build things, and have the skills to (at a minimum) add main power contactor, actuator circuitry (if needed), lighting circuitry (if wanted), brake circuitry and so on.
(2) you don't need any specialty controls
(3) you accept the risk of using an analog or pulse joystick and the somewhat untested fail-safe and safe-fail properties of the Roboteq.
(4) you can't get an OEM-level programmer for the P&G or Dynamic controller you chose
(5) you are likely, in the future, go to LiPO4 and/or higher voltages (even if that is to gain range and battery life rather than speed)

With a 24V lead system, all you will gain with the Roboteq is higher maximum current. At 6 mph, unless you and your chair are very heavy, or you are playing power-chair sports, that probably won't buy you much.

Ciao,
Lenny

Hi Lenny and thanks for reply.
This is my first time building a chair and i am learning as I go along. I do plan to switch to lithiums later on but when I understand how the charging system and correct setup works . I do not weigh much at all and not really a sporty person. I was reading through the posts and see a script has been made for roboteq so that had me thinking why not get roboteq instead as it seems cheaper than pilot plus.

[LROBBINS]

If all you need is a basic two-piece system: joystick + power module, you might be able to find a lightly used pilot plus (even if you have to buy a used chair you don't intend to use) for less than what a Roboteq + homemade joystick pod + contactor etc. would cost; there might not even be much difference with a new pilot +.
Ciao,
Lenny

[PowerHungry]

Is this the p&g controller I need?

http://www.ebay.co.uk/itm/Penny-and-Gil ... 41749412cf

[Burgerman]

Yes, there are 3. That's 80 amps. There are also 50 and 100 Amp units. The 100 amp one is obviously best. Try to get the connecting cable with any one you buy, as its stupidly priced.

[cdb0ewm]

Thank you

Sent you an email yesterday. Did you get it

[Burgerman]

Yep. Replied!

[Burgerman]

Apparently ROBOTEQs torque mode, requires no wires even though its closed loop...

Does this do the same thing as motor compensation?

See new PDF file...

http://www.roboteq.com/index.php/robote ... #resources

[LROBBINS]

No, torque mode essentially does the OPPOSITE of motor compensation. When a wheel contacts an obstacle, load rises, current rises, torque rises, but the wheel slows down because of the load. Torque mode reduces PWM thus reducing current thus keeping torque constant - the wheel will slow down even more. In other words, torque mode is useful if you want constant force, but we want constant speed. Hence, we have to INCREASE PWM and current when there's an obstacle, and that's what motor compensation does. Motor compensation also doesn't need any extra wires, but it is certainly not closed loop and can't be made closed loop because the feedback is positive.

For both, some extra circuitry is needed if you need either constant torque or constant speed at low PWM. That's because the Roboteq measures only battery amps and then calculates motor amps as Battery Amps X PWM - if PWM is only a few percent, that becomes a rather inaccurate measure. That's why Roboteq says you may have to add motor current sensors and read them as analog inputs instead of using the calculated motor amps. Because motor compensation is most needed at low speeds, I suspect that doing it with the calculated values won't work too well. It would be nice if you proved this wrong as mounting the Allegro high current sensors on my boards proved to be one of the trickier parts of the whole project. They have some very unusual solder pad requirements to dissipate the heat generated even in a 100 micro ohm shunt at the maximum over-current rating of 1200 Amps for 1 minute (that comes to 144 watts!). We'll never see those kinds of currents, but the physical layout of the leads, and hence the solder pads, are designed for that.

Ciao,
Lenny

[Burgerman]

Well we will get to see if your method works soon. I hope so!

External current sensors adds too much complication. It SEEMED pretty much ok via RC without any compensation. Maybe because at 45v the motors behave as if they have half the impedance.

Been wiring today...
http://www.wheelchairdriver.com/BM3-con ... trics1.jpg

MORE
http://www.wheelchairdriver.com/BM3-construction/

[LROBBINS]

I also have the impression that you rarely go so slowly that accurate compensation at slow speeds will be an issue.
Ciao,
Lenny

[Burgerman]

It depends how accurate you mean at low currents.

Look at it like this. If a motor can pull 150 amps at full stick stalled, at 24v, it will pull 150Amps at HALF stick on double this voltage.

So at 1/4 stick that's 75Amps! Or almost 40 amps at 1/8th stick. That's enough to start a turn.

I doubt less than 20 amps will make it turn even slowly. What do you call low current? These are tall geared 4 pole 8.5mph motors. I think calculating it from battery amps will be adequate.

Anyway we will soon see.

How confident are you that your compensation code will work? I know it works for turn rates, reverse speed etc.

[LROBBINS]

I'd be more confident if you bench tested it with motors attached before trusting your bod to it, but it should (FLW?) be OK. Just be sure to set a very low milliohm value to start with. If you set it to 0, that's no compensation so the chair should behave the same way as without the script unless there is a gross logic error, so maybe you should do that first to make sure I didn't thoroughly screw it up. Also, DO NOT set the Roborun to autorun the script until you've done lots of testing.

Motor compensation really isn't of particular importance for turning, or at least no more for turning than for going straight. Imagine that you are in hour house moving very slowly forward and there's a screwdriver that you didn't see in front of one wheel. Motor compensation should help you go straight without needing a joystick correction even if you are only going at 1% of maximum speed. It's in that situation that PWM% X BatteryAmps might be questionable - the compensation boost needed may be relatively large compared to the base current at that speed, say a doubling to 2% PWM but NOT to 1.5% or 2.5% of PWM, but that's a very small fraction change compared to 100% PWM. We'll see, or rather you'll see, whether BATT AMPS is good enough for that 1/2 of 1% difference.

Ciao,
Lenny

[Burgerman]

If I disable motor compensation or set it really low, on a stock powerchair it becomes very odd to drive, and will not turn.

It wont turn because you may have say 40 percent turn rate set at full stick, in a high speed setting.

So you get about 5 volts to a motor. That may not be adequate to initiate a turn. The compensation gives it a kick! Without it, no go...

Worse, it doesn't accelerate in a straight line until you give it a handful... And then it runs on and doesn't slow down when you release the stick. You have to become fast at using minus stick to decelerate. Trust me its not easy!

The Roboteq SHOULD be the same. But doesn't seem too bad (at least by RC in very limited tests).

[LROBBINS]

When Rachi's chair first arrived here with it's poor quality 300 mohm Fracmo motors, it had compensation set at 40 mohm. Its handling was awful, but not quite as bad as what you describe. Totally unsuitable for head switch driving, however. Re-set to 260 it started to behave decently, but, being far too puny, it just ran out of steam with the slightest obstacle or slope - compensation only works when you're not already at controller maximum output. With the Roboteq and no delays it will probably behave better than you expect even with no compensation. In any case, running it without the script and then with the script with compensation zeroed would be a just check for gross programming errors - I doubt you'd want to drive a chair programmed that way.
Ciao,
Lenny

[Burgerman]

40 Mohms is about right for the 4 pole motors on my F55.

Try turning it completely off. Its horrible!

[Williamclark77]

How would I go about connecting a limit switch? Physically connecting for now. I will have to get yaw's help programming later. Basically, what I am needing to do is connect a push button toggle switch to the 25 pin connector of my Roboteq HBL2360 controller. I would like to be able to push the button and slow the chair down for indoor use. I'm unsure of which pins I should connect it to or if they are not designed to just accept a flat on/off voltage command, ie - I wire the switch to connect a + pin to an input pin and write the script to, for example, decrease forward speed to 50% and acceleration to 80 rpm when said pin receives voltage. Clear as mud?

The switch in question that I want to use:





The pinout for my connector. Pins 1, 14, 15, and 16 are used to my rc connections. Pins 13, 23, 24, and 25 are used by my analog joystick.

[Burgerman]

Much simpler to switch in a resistor at the analog joystick. So that full stick has say 20 to 80 percent output on turn/speed.

But you can do it your way. Any pin is map able directly in software, and usable for in/out digital or analog, frequency, duty cycle, pulse width etc as you wish. So you could use it to do any number of things.

[Williamclark77]

Thanks!

I was kind of worried that doing a resistor would decrease the resolution and make it, what's the word, spongy feeling when driving. I really don't know. Never done that.

There would be no damage to the controller if I wired my switch to, for example, pins 14 and 21? I wanted to do it this way so it would be programmable and for the physical packaging of the wiring. I already have the looms and hardware there. Adding two more wires would be quick and easy.

[LROBBINS]

Roboteq digital input pins can be wired as either pull-up (between the pin and 5V) or pull-down (between the pin and GND). See pages 44 and 45 of the Roboteq user's manual. To use as pull down you must add a resistor between the digital input pin and 5V. If used as pull up, a resistor (between the digital input pin and ground) is optional, but would reduce noise sensitivity.
Ciao,
Lenny