Need help diagnosing motor problem.

[drutledge]

I have a 9C motor that has stopped turning properly. Sometimes it will make a very faint 'Ting' sound and not turn at all. Sometimes it will turn (usually if helped) but any sort of load will stop it (like just a hand gripping it). I'm thinking hall wires? How to diagnose if it's in the motor, controller or a connection in-between? This problem seems to have been intermittent for a short while before but now constant. Infineon controller with a CA.FWIW.

For what it's worth, this is the same setup that I lost my regen capability on while still having the brake cutoff ability. (Still haven't fixed that) Probably not related but thought I'd mention it.

Thanks.

 

[dnmun]

does it seem like each move is little bumpy steps? if the phase wires are shorted then the motor will resist turning and you feel it as each magnet impedes the movement until it has passed a critical point then it jumps to the next resistance, one jumpy jump at a time. test the mosfet with the diode tester from the phase wires and the two power terminals on the controller.

for the hi side mosfets the drain is attached to the B+ and for the low side, the drain is attached to the phase wires. measure source drain with the diode tester both ways for each 3 phase wires and each power lead. 12 measurements.

 

[drutledge]

The wheel does make a sort of ratchet sound when moving. This testing method seems a bit beyond my abilities/comprehension. If the phase wires are shorted wouldn't a simple continuity test work?

 

[Icewrench]

You can turn the wheel by hand to feel for heavy cogging.
If it is hard to turn then disconnect the phase wires and turn it again.
If it is now easy to turn there just might be a problem with the controller.

 

[John in CR]

Intermittent could be a bad connection. Unplug that hall plug and check for spades and sockets not shiny and new looking. Something as simple as sliding it in and out could fix it till you clean them up well, if you're lucky enough to be all it is. A breaking wire is possible, though I have seen halls develop an intermittent short that progressively got worse and the cause was insulation deformed from heat and virtually impossible to find.

 

[dogman dan]

Intermittent definitely points at a bad connection, such as the classic one where a wire has had it's contact back out of the plug housing a bit.

Or it could be a cut wire, which can happen without actually cutting the outer sheath. Usually at the hub, but I've had a wire rub on a tire cut a wire once.

The easy way to test stuff is with a motor controller throttle tester. Many sources, lyens, ebay, etc. But again, a hall sensor failed in the motor would not do so intermittently. So I think its your wires (90%)
or (10%)could be the controller itself.

Shorted phases is easy. you'd feel a LOT more cogging resistance than before when they were not shorted. Often the short is in the controller, so try unplugging the wheel and see if the resistance to turning changes.

 

[drutledge]

So after much hassle, I eliminated the motor itself and the aprx. 9 feet of wiring leading to it (it's on a large rickshaw) as the likely problem leaving only the controller. I removed the controller and noticed that it would 'thunk' when tilted. Not a normal sound for a controller to make! I opened it up and found that the heat sink and whatever is attached to it on the right side (looking from the switched end, Infineon 35A) had broken free. The left side heat sink is screwed to the case but not the right. I'm sure this is the problem and would explain why it was intermittent for a short while. The weight of the heat sink probably flexed and finally broke the three connectors like a paperclip. The controller only has about 500 miles of riding on it according to the CA and, as I said before, it is in a rickshaw so it's not like I've been jumping off of ramps with it.

I bought this from Justin at Grin. I'm hoping he will stand behind it since it seems like this is not something that should have happened. I'd think that there should be at least one screw on that side to prevent this. Any thought as to if this was also the cause of my earlier problem of no regen but having brake cutoff? I have no idea what piece of electronics it attached to that heat sink.

View attachment 1

 

[dogman dan]

That doesn't look right to me. An awfull lot of weight hanging on just the fet legs.

 

[drutledge]

I took the end cover off to see what was going on inside. The first photo is the problem part.

 

[dogman dan]

Yeah, sorry about that first reply, I just edited it after I noticed the second pic.

First time I've seen that, a second heat sink on the other side.

 

[dnmun]

looks like the voltage regulator but you did no read out the part number on it for us. are there any screw holes in the side to hold that heat sink in place?

your pcb may and likely does have a bumper underneath the pcb to hold up the outer side. nothing unusual about the way it is built. all the controllers are like that.

but you need to tell us what the part number is on that device so we can clue you in on how to fix it.

 

[drutledge]

There are no screws on that side to hold it to the case. Or any bumper to hold it up. I think the thermal paste would normally just stick it to the side. Maybe a sharp jar in shipping knocked it loose and from there it just slowly flexed itself to death. I took it apart further this morning and everything else looks ok inside. No scorch marks. The part number is TL 783C, a High Voltage Adjustable Regulator, according to a google search.

I've got an email in to them about the problem. I'm hoping they'll offer to fix it or replace it, my soldering skills are no great shakes, though this looks like one of the easier ones to do. Preventing it from happening again is the question.

 

[dnmun]

from mouser: http://www.mouser.com/ProductDetail/Texas-Instruments/TL783CKCSE3/?qs=sGAEpiMZZMtUqDgmOWBjgHdj12KM6gWG7%2fD7nPrnouU%3d

ebay: http://www.ebay.com/sch/i.html?_odkw=TL793&_osacat=0&_from=R40&_trksid=p2045573.m570.l1313&_nkw=TL783&_sacat=0

and this is for TL783 kc: http://www.ebay.com/sch/i.html?_sacat=0&_nkw=tl+783+kc&_frs=1

this is the data sheet: http://www.ti.com/lit/ds/symlink/tl783.pdf

cheapest shipping for small orders from mouser is now $4.99 but you may be able to talk them into sending it in a jetpack envelope for $2-3.

this is why ebay gets so much business. shipping versus free shipping.

 

[drutledge]

Just got a reply from Adam at Grin, they said they'll either send me the part or give me an RMA to return it for repair. Evidently they had a small batch of controllers with way too big heat sinks attached to them. They suggested either doing away with it all together if using <48V or just cutting it down to size to avoid a repeat performance.

 

[ambroseliao]

I just bought one of these controllers and saw your gigantic heat sink. Here's an image of the one in mine.

 

[drutledge]

That seems a bit more reasonable!

 

[dnmun]

since they are designed to eat a lotta voltage they will make a lotta heat.

you can tell the difference between the two because ambrose's has the voltage regulator grounded to that heat sink with kapton tape on the backside.

on your heat sink the kapton tape is under the regulator to keep it insulated from the heat sink. if the regulator touches the case then it can short out.

since yours is insulated, you can drill a hole through the case into the heat sink and attach it to the case to keep it from moving around.

check continuity between the middle leg and the heat sink when you reassemble it.