french version here : http://wiki.electrolab.fr/Projets:Perso:2012:controleurfauteuil
We are trying here to find a solution to our most important technical issue in powerchair football, "over heating", or the fact that after a while, the chairs starts to slow down and this has a huge impact on their way to play the game.
The hypothesis we want to check for now is that the "over heating" phenomenon is due to a temerature rise in the controler , that triggers a "safety mode", which slows the chair down. This has to be verified.
We are currently using 400W motors, with 2X80A controlers. Electronics are made by Dynamic control a New zealand brand owned by an american disability products company, Invacare.
- 2 mosfets half bridges ceach composed with 8 transistors. - current measure by "DC" current sniffer
[To be completed]
In situation controler behaviour analysis
The first idea we had was to observe the behaviour of the chair in play, to understand what is really going on.
That would mean to concomitamently measure physical data la battery voltage, the battery current and the motor controler temerature.
a thermocouple is available but the acquisition module we plan on using also accepts PT100 and RTD. http://www.omega.com/prodinfo/rtd.html We can also use a silicium sensor like the KTY81-121 with R25 = 990 Ohms and which varies from 700 Ohms between 20 and 100°C Tmax = 150°C : a simple spreadsheet can then convert it into temperature.
200A Shunt available
The acquisition module we plab on using has numerous channels (20 maybe more, it os possible to do simultenaous measures more thorough than the measures we mentioned earlier. For exemple we can do several temperature measures on the controler, the motors, PWM equivalent voltage, etc.
In situation Acquisition
the acquisition module power supply is using a transformer and works on 230V: we can use a small inverter 150VA to recreate the 230V on the chair: But our inverter only does square, so not so easy to do.
- either it works with the square 230V (= see if the transformer does not over heat, (knowing that a trial will last at least 10 min)
- Either we filter the saure, but doing a 50Hz filter that cuts fast enough towards 70Hz is a real challenge
- Either we find a small sinus inverter
- Either, we limit to 2 channels (battery voltage, controler temperature), in that case a portable scope on battery should do the trick.
Because an ascillo only measures voltage, we need to use a thermocouple converter-> voltage/current or PT100 -> voltage/current [available] to ease the temerature data recording. Exemple of another voltage temperature converter with its sniffer: http://www.gsamart.com/webdata/prodimg/FLU_e0007903_lg.jpg
motor controller possible eviolution
a few unanswered questions for now::
- Is it really the controler that causes the overheating? (the answer must come from in situation audit)
- if the controler is indeed the cause of the issue, what can we modify in order to increase its tolerance?
- We have several 60A controlers, is it possible to "boost" them towards 80A (or more!), that would mean huge savings for us?
Answer from powerchair specialists
This project was posted on a forum specialized in powerchairs, and below are a couple of answer full of insight:
I also suggest that your friends download the DX System manual, and all other DX and DX2 manuals as well. They will, for example, already find that temp foldback (1) does respond to power module temperature, and (2) does reduce output. In other words, some of their questions are already answered. Moreover, they will discover that the foldback parameters are programmable and, by default, are pretty conservative. A slight increase in the temp for foldback could give a lot more playing time before foldback without actually risking the controller. Undoubtedly, JoeC will come along to tell you what he uses on his wife's soccer chair. He will also probably describe how he's used some small fans to improve heat dissipation. With the manuals, an OEM programming dongle and a PC you will at least know what temp has been programmed; even the cheapest commercial grade components have a temp limit of 80oC, and better components have temp limits quite a bit higher. So a foldback setting of say 60o is certainly giving you a huge safety margin. You probably should have some buffer as the temp sensor isn't necessarily at the hottest spot in the module, but a 20oC margin, assuming cheap components, is surely more than's needed to protect the module. Of course, you may find that avoiding PM foldback also leads to the motors getting rather hot - JoeC should again be able to bring the sound voice of experience to bear.
Lastly, there was some mention here that the difference between 60A and 90A PM's is just some on-board connections or passive components. I've never seen confirmation of this, though I suspect that it makes sense for Dynamic to use the same MOSFETs in all PM variants. If that's true, perhaps your friends can discover what needs to be changed.
To answer simply, here are replies to your questions:
1- Yes, it really is the temperature of the controller that causes the chair performance to be decreased. Keep the controller cool, and it will not reduce its output.
2- What to do? Reprogram it to a higher value, and it will be able to work harder. I have reprogrammed these controllers to a maximum temperature of 85C, from the stock value of (I think) 60C. I have used fans blowing on the controllers, and this helps. I have mounted the controller on a flat aluminum plate, and this helped a little, but not as much as fans, reprogramming, and moving the controller to a space where it can get free air flow.
3- Can you convert 60A DX2 to 90A? Probably not. As the design sophistication and manufacturing control is improved, they are less likely to overbuild the controllers, and more likely to have diagnostic self checking at power-on that will detect and object to any tweaks. If they are older than DX2, say DX or older, you stand a better chance. Even DX might be too clever to do it as simply as connecting a jumper.
All of the measurements you've suggested are helpful for getting a really good idea of what's going on with the chair. The temperature things can be spot-checked with a thermometer (infrared, or a taped on thermocouple), since it does not change so quickly that you can't stop playing and get the temperature checked. It's also fairly well understood (by me, at least) exactly what's going on with the temperature, and I tell you that further investigation won't give you much insight.
What you *should* look at more is voltage. Look at motor voltage, battery voltage, cable loss voltages, anywhere that voltage drops. You'll quickly see that shorter fatter cables and high quality batteries are the way to go.
Another crucial thing to understand, if you want to get so technical as to do current measurements and bring out oscilloscopes, is to get a good understanding of the difference between battery current and motor current. Once you understand that you can have 80 amps coming out of the batteries and 120 amps going to each motor simultaneously, you will see the importance of good motors with low resistance.
Once you get your power module reprogrammed for higher temperature, your wires thickened, good batteries installed, and the chair balanced for good traction, THEN (but maybe sooner) you are likely to start causing excessive motor heat that can degrade performance due to higher resistance, brush damage, and loss of magnet strength. If you can comfortably hold your hand on the motor, then the motor is nowhere near hot enough to be the limiting factor. I have seen motors get hot enough after all my mods to cause burns on your skin, if you were able to hold your hand in place for long.