GoldenMotor.com Forum
General Category => General Discussions => Topic started by: Bikemad on March 31, 2011, 02:14:01 PM
-
Has anyone tried, or does anyone know if it's possible to connect two controllers in parallel to power a single motor?
I've been thinking how to reliably increase the power output from my internal controller in order to get some decent current out of it. (More than the 20Amps maximum that I currently have)
I'm convinced it should be possible to simply put another identical controller inside the hub and just connect all the wiring in parallel with the existing controller, so if anyone in the UK has a spare internal controller from a Pie, I might be interested!
I'd even considered putting another six (or 12) FETs on a separate heatsink plate, linking the power and the three phase wires directly, but with each of the gate pins being connected to the gate pins on the existing FETs. I think a shunt would still be needed on the ground connection to prevent too much current from going through the new FETs.
But I think a second controller would be the easiest option.
I'd appreciate some opinions on this if possible.
Alan
-
Hi buddy
Whereabouts in the uk are you buddy :)
Not sure about the double controllers, I think synchonisation might be an issue on them as whether the halls feeding to 2 lines would lead to issues...wouldnt need much de-sync the phases ??
Needs an educated answer from someone more educated me thinks..got a spare internal from the wheel I just converted to infineon......
-
Whereabouts in the uk are you buddy :)
...got a spare internal from the wheel I just converted to infineon......
Hi Ginge, I would have replied earlier, but unfortunately work has to come first.
Anyway, I'm located down in Sunny Devon, surrounded by far too many steep hills for my liking, hence the quest for more power! ;)
I am hoping to raise my current 20 Amp max to nearer 50 Amps (if all goes according to plan), which should give me some decent hill climbing power.
Alternatively, If things don't go according to plan, I can see an external controller moving much higher up my wanted list, but I am hoping I can keep things hidden away, so that it still looks as standard as possible.
If your old internal controller is going to be surplus to requirements, send me a PM with your asking price and I'll see if it comes anywhere near my ever diminishing budget.
Does your Infineon controller have a more responsive throttle than the GM controller, or does it still have a similar delay when you open the throttle quickly?
Alan
-
to be honest, the only time ive noticed any delay, is when the hall sensors failed and the controller was in sensorless mode.....then there was about 1 second delay....otherwise I never noticed
are you ok for citylink delivery ?, depot near etc.....or a different address for daytime delivery...
-
are you ok for citylink delivery ?, depot near etc.....or a different address for daytime delivery...
Ginge, whatever and whenever it's most convenient for you, there's absolutely no panic at this end. I'm in most days during the week as I work a late shift, and if I'm not in, my wife probably will be.
My dual controller theory is going to be put to the test! ;D
Thanks again Ginge,
Alan
P.S. Do you remember what the last programmed settings were on this controller?
-
hi mate
ill try get it to you early this week
48V low PAS, I think 70% regen....
I take it you want mounting plate as well....
regards
-
I take it you want mounting plate as well....
Yes please Ginge, as it would save me making another heatsink plate for it.
Alan
-
Did you try 2 controllers?Did it work?
I'm hoping it will fit without too much hassle, but it looks like it may have to wait until after Easter.
Once Easter has been and gone I will hopefully be able to find out whether or not it works, unfortunately it will be a bit more involved on my pie because I have a home-made spoke flange ring (http://goldenmotor.com/SMF/index.php?topic=1653.msg10177#msg10177) fitted over the cover on the speed controller side, which prevents the cover from being easily removed:
(http://goldenmotor.com/SMF/index.php?action=dlattach;topic=1653.0;attach=1731;image)
(http://goldenmotor.com/SMF/index.php?action=dlattach;topic=1653.0;attach=1735;image) (http://goldenmotor.com/SMF/index.php?topic=1653.msg9193#msg9193)
I will post the results of my proposed dual controller modification as soon as it's done, assuming I will be able to successfully fit it into the hub. I have attached a picture to show very roughly how it will hopefully be fitted and wired.
Alan
-
It's finally time for an update.(http://www.arhservices.co.uk/GoldenMotor/emoticons/7_small.GIF)
I have been very busy today performing some major surgery on my stator in order to get the second controller to fit nicely.
Here's the stator before modification:
(http://www.arhservices.co.uk/GoldenMotor/StandardStator.JPG) (http://www.arhservices.co.uk/GoldenMotor/StandardStator.JPG)
And this is where the second controller needs to go:
(http://www.arhservices.co.uk/GoldenMotor/Newlocation.JPG) (http://www.arhservices.co.uk/GoldenMotor/Newlocation.JPG)
But these reinforcement ribs are in the way and need to be removed first:
(http://www.arhservices.co.uk/GoldenMotor/Prepared.JPG) (http://www.arhservices.co.uk/GoldenMotor/Prepared.JPG)
There's no turning back now!
(http://www.arhservices.co.uk/GoldenMotor/Firstcut.JPG) (http://www.arhservices.co.uk/GoldenMotor/Firstcut.JPG)
And here is the final result after a lot of grinding (consuming 5 mini discs and a set of brushes in my drill):
(http://www.arhservices.co.uk/GoldenMotor/RibsRemoved.JPG) (http://www.arhservices.co.uk/GoldenMotor/RibsRemoved.JPG)
Fixing holes drilled and heatsink pop-riveted in place:
(http://www.arhservices.co.uk/GoldenMotor/Rivetedheatsink.JPG) (http://www.arhservices.co.uk/GoldenMotor/Rivetedheatsink.JPG)
Heat sink compound applied and second controller fitted:
(http://www.arhservices.co.uk/GoldenMotor/Bothcontrollersfitted.JPG) (http://www.arhservices.co.uk/GoldenMotor/Bothcontrollersfitted.JPG)
And here's a view from the other side:
(http://www.arhservices.co.uk/GoldenMotor/RearView.JPG) (http://www.arhservices.co.uk/GoldenMotor/RearView.JPG)
It's looking very good so far, with just the wiring to be completed when I can find some more spare time, and then I should find out whether it works as hoped (double the power) ;D, or simply blows both controllers (zero power) :'(
I should hopefully know one way or the other within the next few days!
(http://www.arhservices.co.uk/GoldenMotor/emoticons/fingerscrossed.GIF)
Alan
-
Fingers crossed :D.
Don't forget to duplicate your ebrake, in case you're in cruise etc
regards
-
Nice Stuff Alan! Very brave of you!!
Gary
-
Very interesting, let see if it goes.
-
Yesterday, I was finally able to complete the wiring, I'd like to say that the testing was a complete success and that it all went according to plan, but I can't, because it didn't. :(
Here are a couple of pictures of the completed wiring, which actually took much longer to do than I had expected:
(http://www.arhservices.co.uk/GoldenMotor/Wiring.JPG) (http://www.arhservices.co.uk/GoldenMotor/Wiring.JPG)
(http://www.arhservices.co.uk/GoldenMotor/Wiring2.JPG) (http://www.arhservices.co.uk/GoldenMotor/Wiring2.JPG)
I also added an additional loop of wire to each of the shunts to try and achieve nearer 25Amps max from each controller:
(http://www.arhservices.co.uk/GoldenMotor/Shunt.JPG) (http://www.arhservices.co.uk/GoldenMotor/Shunt.JPG)
With the wiring completed, the wheel was quickly reassembled and refitted to the bike ready for testing. I eagerly connected the battery, producing a very loud crack accompanied by the anticipated bigger spark. But I had not expected this to be immediately followed by three beeps coming from the hub, indicating throttle failure. I quickly rechecked all external connections and even tried a spare throttle unit, but still three beeps. I then tried pressing the cruise button and the wheel instantly spun into life , which made me feel a bit more optimistic.
I carried out some simple checks with the voltmeter and discovered that I only had 0.5V instead of 5V coming from the throttle supply at the mini connector.
Unfortunately there was no way around it, the wheel had to come off again so I could access the internal wiring.
I removed the side cover and pulled the stator out for easier access and mounted it in the bench vice and reconnected the wires to the bike which was underneath it.
This time I heard just 2 beeps (hall sensor failure) so I grabbed the voltmeter again and started checking voltages again. All three of the hall sensor outputs were reading around 4.6V, so I grabbed a magnet and tried moving it past each of the sensors, but the reading was still the same. I turned the magnet around and repeated the same process. This time, each of the sensor outputs went to 0V as the magnet came near to the respective sensor. So I deduced that the hall sensor failure warning had only occurred because the stator was not surrounded by the magnets.
Then I focused on finding the cause of the low throttle voltage, and a quick check with the meter showed 4.9V on the 5V throttle supply on both circuit boards. ??? So I rechecked the voltage at the external throttle connector, but for some strange reason this was also showing 4.9V. ::)
Unable to find anything wrong with the wiring inside the hub, I reassembled the wheel and decided I'd test it in the vice first. Connected the wires and no beeps whatsoever, so I anxiously tried twisting the throttle and the wheel started to run properly. ;D
I then refitted it to the bike again so I could take it for a quick spin, as I was taping the wires to the frame to keep them away from the tyre, I heard the dreaded three beeps again ??? but at least I now knew what was causing the fault - a poor connection in the main junction block on the Pie's harness! Not wanting to strip this apart, I moved the block slightly so it was pushing slightly onto the cable instead of pulling, which seemed to cure the problem.
The moment of truth had finally arrived and it was time to test out all my hard work. I took the bike for a quick spin around the back garden and it definitely seemed to pull me up the slope much easier than before. I checked my watt meter, and this confirmed my thoughts by showing the highest reading I've ever had of 1340Watts, but still not quite as much as I had hoped for.
I continued to test the motor and made it work as hard as I could in an effort to get a higher reading, but this did not happen. The motor did eventually get slightly warmer than it had before, but definitely not hot. After a few more tests, I noticed that the throttle would not work unless the bike was moving, and the noise from the motor had become slightly more harsh. I disconnected the battery and then reconnected it, only to hear two beeps (hall sensor failure) :( but as I was running out of time, I had to put the bike away and go to work.
Today, I decided to try the bike again so I could see what it was like from cold, but unfortunately I still received the two beeps. I pedalled the bike and twisted the throttle and received a slight nudge from the motor, followed by the dreaded sound of five beeps and a wheel which is now lumpy to turn, indicating the death of a controller, or more likely both of them knowing my luck!
I don't have the enthusiasm to start pulling it all apart again now, so the post mortem will have to be done at a later date.
But at least this experiment has provided an answer to my original question, "Can Two controllers be used to power a single motor?" , the answer is yes, but unfortunately I would not recommend it because it probably will not work for very long! ;)
I think an external controller conversion may well be happening if both controllers are found to be dead.
Alan
-
May be when you have finished that little job. You could take the wye apart so you get six ends and wire it as two motors with two controller look forward to hearing how it went tomorrow ;D ;D
-
aw man....I had such hopes......that little board did miles :(
listen, if you decide to go external, ive got aplphawire ecowire ill send you some meters for nowt....as a gift after the controller ( all red coloured though, although I got round this with coloured cable ties)
i regularly pull well over 1700 watts of my infineon controller on one pie ( confirmed by very happy cheshire cat andrew and webby)
If you need any help ( not that I think you would) , gimme a shout ;)
-
Hope you can solve the problems and make it work properly soon.
I just remember that we made something similar on a big airplane (over 6000W power) we run a double controller set up, not like yours but similar thing... and in order to make it work properly we fired up the controllers by turn first one, wait for the beeps and then second one. Do not know if you can make it but may be it helps.
I wanted to post a picture but they are too big :'(
-
hey mate
Download fotosizer from here
http://www.fotosizer.com/
you can make your pictures smaller to attach here :D
-
listen, if you decide to go external, I've got alphawire ecowire ill send you some meters for nowt....as a gift after the controller
Thanks for the kind offer Ginge, I may well take you up on it if both controllers turn out to be dead.
I'm still not sure whether the second controller was actually making any difference at all, perhaps it could have died when I initially connected the battery, and the 25% power increase may just have been the result of the shunt modification alone, as I still feel the second controller should have given nearer 100% increase if it was working correctly.
I also noticed that only one of the controllers beeped (the original one) to falsely indicate a hall sensor fault when I had the wheel apart, but this might be because it's an earlier circuit board and the fault warning criteria might be slightly different.
Prior to the failure of the hall sensor/s, it was running very nicely indeed, so perhaps both controllers trying to run in sensorless mode at the same time has damaged some of the FETs because the hall sensors were no longer able to synchronise them?
I'm still puzzled as to what actually caused the hall sensor/s to fail, as I don't think the hub was anywhere near hot enough for the sensors to have become heat damaged as the hub was barely warm to the touch. Perhaps the extra load of operating two controllers was just too much for the hall sensors to withstand?
Anyway, the novelty of walking the dog on foot has already worn off after just one day, so I'm going to throw a standard wheel back on for the time being and remind myself of how difficult pedalling up steep hills unaided actually is! ;D
Alan
-
To be honest, it does sound like somehow there were some voltage instability, which shouldnt really be with either the shunt mod or the dual cotrollers.....
I wonder if the fets had a problem due to voltage coming back from the opposite controllers phase firing, if they were slighty out of sync ??....
Well, the controller sent had done many miles without hitch, as had yours, both were tried and tested, so I guess we sorta know the outcome......shame though
Wire up an external harness, get an infineon of some description on there...apart from the cruise which takes a little getting used too, you wont be dissapointed...and not that much more than an external GM controller
-
I've just read the whole of this thread, such a great idea and a very brave modification. Gutted for you Alan, all that grinding too! >:(
-
hey mate
Download fotosizer from here
http://www.fotosizer.com/
you can make your pictures smaller to attach here :D
Thank you but I use Mac and that program is not for OS.
I have to look for something similar thought...
-
The wheel was removed and dismantled today and revealed some debris loose inside the hub accompanied by a strong smell of burnt components. :'(
Here's what the three small capacitors on the rear of my original controller look like:
(http://www.arhservices.co.uk/GoldenMotor/Goodcaps.JPG)
and this is all that remains of the same capacitors on Ginge's (second) controller:
(http://www.arhservices.co.uk/GoldenMotor/Blowncaps.JPG)
And this capacitor has also split its end open on Ginge's controller:
(http://www.arhservices.co.uk/GoldenMotor/SplitCap.JPG)
I unsoldered the interconnecting wires and removed the second controller and then powered up the original controller with the main harness left disconnected. I was expecting to hear the same five beeps as before, but for some strange reason I heard four beeps, indicating a brake fault. ??? So I hastily connected the main harness to eliminate throttle and brake faults and powered up again. This time it had reduced to two beeps (hall sensor failure), so I grabbed a magnet and voltmeter and started testing the three hall sensor outputs and was soon able to confirm that all three sensors have been killed. :'(
I'm now pretty hopeful that my original controller is still working, but I'm tempted to throw the wheel back together just to confirm it still works in sensorless mode, just to put my mind at ease.
I will post an update when I've either tried it as it is, or when I've replaced the sensors and tested it 100%.
Ginge, I have not yet ruled out the possibility of converting the wiring and using an external controller, so please don't use up all of your alphawire wire just yet. ;)
Alan
-
I was going to throw the wheel back together and test as it was, but I thought it might be interesting to increase the thickness of my additional shunt, so I soldered on another thin wire loop and then decided to solder them both together to give an even larger cross sectional area.
I reassembled the wheel and put it securely in the bench vice and connected up a 7cell 5Ah LiPo pack measuring just under 29.4V.
After the two beeps, I spun the wheel slightly and twisted the throttle and she burst into life running as smooth as ever. I tested the throttle, cruise and the regen functions and all are working fine. ;D
Now the really exciting bit; I put on a pair of heavy duty gloves and decided I'd do a stall test to see how much current it would draw in sensorless mode with the improved shunt mod. I opened the throttle and grabbed the tyre to slow the wheel down as much as possible, but I could not quite manage to stop it completely. There was a strong smell of burning rubber from the friction of the tyre against the leather gloves and then I noticed the reading on the Watt meter:
(http://www.arhservices.co.uk/GoldenMotor/54Amps.JPG)
(http://www.arhservices.co.uk/GoldenMotor/1458Watta.JPG)
I bet you can imagine how I'm feeling right now? (http://www.arhservices.co.uk/GoldenMotor/emoticons/7_small.GIF)
(http://www.arhservices.co.uk/GoldenMotor/emoticons/6_small.GIF)(http://www.arhservices.co.uk/GoldenMotor/emoticons/everyday_dance_230608.GIF)(http://www.arhservices.co.uk/GoldenMotor/emoticons/everyday_clap_230608.GIF)(http://www.arhservices.co.uk/GoldenMotor/emoticons/everyday_thumb_up_230608.GIF)
If it draws 1458 Watts on 7 cells it should be at least 2916 Watts (3.9HP) on 14 cells, but I don't know how long the controller will last at 55 Amps, I'm pretty sure something is bound to expire (probably in a big way) sooner or later with this amount of power.
I just couldn't resist trying it:
(http://www.arhservices.co.uk/GoldenMotor/2929Watts.JPG)
(http://www.arhservices.co.uk/GoldenMotor/5641Amps.JPG)
... and it still lives!
I wonder if I can manage 4HP?(http://www.arhservices.co.uk/GoldenMotor/emoticons/everyday_think_230608.GIF)
I'm still not sure whether the second controller was actually making any difference at all, perhaps it could have died when I initially connected the battery, and the 25% power increase may just have been the result of the shunt modification alone, as I still feel the second controller should have given nearer 100% increase if it was working correctly.
I think my initial feelings may well have been correct in this instance, but I'm still puzzled as to why all three hall sensors failed.
Alan
-
I feel really guilty for some reason...i know the controller was good, covered nearly 700 miles.......but yet im still feeling guilty...honest guv, it was alright when I took it out.....
Still for a capacitor to blow that bad, in a very short space of time, I think one controller dumped all its volts through the phase interconnect into the other, and if they were in sensorless mode, could explain the lack of sync....
No worries on the alphawire alan, ill reserve some for you :D
I think once you get on your pie at that power, you will want it to become a permanent mod......maybe the controller will last that long......personally if I were you, before you seal it up, maybe a good time to add a LOT of heatsink metal to take the heat off the fets
(still feeling guilty :-[ )
-
I feel really guilty for some reason...i know the controller was good, covered nearly 700 miles.......but yet im still feeling guilty...honest guv, it was alright when I took it out.....
(still feeling guilty :-[ )
There's no reason why you should feel bad at all, I knew I was taking a big risk by doing what had not been done before, and I half expected something to go wrong, but I just had to do it to test my theory. Anyway, I'm very pleased with the 56.41 Amps maximum that it now draws, and the 2929.9 Watts of power being consumed really does feel a lot when trying to stop it by hand! I could see my tyre being quite literally worn away within my gloved hands! ;D
I don't suppose there are many people stupid enough to try a burnout with their bare hands, as I think bits of skin would have been removed instead of bits of rubber, if it wasn't for the heavy duty gloves I was wearing. :o
Disclaimer
Remember kids, this sort of testing should only be carried out by a fully trained and highly qualified idiot, and in a properly controlled environment, so "Please don't be tempted to try this at home!"
I initially thought about videoing it, but I do not want to be responsible for any loss of limbs!
It will be interesting to see how new hall sensors survive with just one controller again, I measured around 2mA per sensor to pull it down to 0V, so that's around 4mA per sensor when running both controllers. I wonder what the maximum current is likely to be on the GM hall sensors? Hopefully the Honeywell ones will be better. ;)
I still think it may have been more successful if both controllers had been exactly the same model number, and programmed with identical parameters.
Alan