Motor stuttering after possible overheat

[cveen]

Hello,

Years ago I got a ready-to-roll kit from Grin. Over 15,000 very-satisfied km later I think we may have killed it by overheating taking 3 kids up to Grouse (local mountain - we've actually done it a bunch of times previously without incident).

Equipment:
* C4835-L10 controller
* V3 Cycle Analyst (with thumb throttle and regenerative braking)
* RH212_Std (Nice Continents rear hub) with Statorade and cooling fin kit
* 52.2V nominal battery voltage, 20A limit set on the controller

The Cycle Analyst was set up to start limiting power at 90C and cut out completely at 130C. Wife was biking the 3 kids up Grouse and noticed the loss of power, so waited a bit for it to cool off then continued without incident. Coming back down she noticed the regenerative braking wasn't working properly, and once it was time for some e-assist noticed that the power wasn't working either.

Now:
* Motor just stutters/clonks when throttled and doesn't provide much thrust. Cycle-analyst claims ~40W and that sounds about like what it might be doing.
* Motor just stutters/clonks when regenerative-braking applied and doesn't slow the bike down
* No noticeable voltage sag when throttling
* Speedometer correctly reports wheel speed when biking around manually
* Kinda clocks along and barely spins if the wheel isn't on the ground

I guess either the controller or something in the wheel are fried. Unfortunately, the (L-10?) connector between the two is seized.

Questions:
* Thoughts on what is the most likely problem?
* Were my thermal setting irresponsible?
* Suggestions for un-sticking the connector? My normal go-to for unsticking things is heat and/or oil, but depending on the plastic one or the other might not be a good idea.
* Once I get the connector apart I feel like measuring the resistance of the different phases might be informative?
* I guess I should take the hub apart and look to see if anything is obviously fried?
* Might it be possible (or a good idea) to get the same hub, and just replace the core, if that's where the problem is?

Thanks for any help!

 

[neptronix]

Ah, it must be the beginning of summer.

130C is the point where a lot of motors are going to fail. I would personally throttle at 80C ( and pull over for a while ) and cut at 105C because i don't like taking chances.

You probably burned the halls by overheating it.
If this happened there will be a smell inside the motor shortly after.
Or you may have shorted the phase by some of the insulation wearing off.
Open it up and see!

How steep is this hill, how long, and how much weight, on what size wheel?

I'd like to see your scenario on the ebikes.ca motor simulator.
Here is a link with what i'm assuming is your setup:
Motor Simulator - Tools

 

[amberwolf]

Most likely halls.

If you have the PR software and cable to connect to the ocmputer, you can use it to see the hall operation in i think the advanced tab.

if they are not cycling as expected you can change the pr to sensorless operation, no halls.

the motor magnets, if heat damaged, would cause the motor to spin faster than normal, rather than slower.

fets in the contorller bieng damaged usually would just error out and give you a blink code on the red light of the pr.

same with damaged windings or phase wires--they'd cause an overcurrent erorr and the pr would fault out.


IIRC the l10 is a screw-on connector, so you have to unthread the ring first gbefore you can unplug it. if you have salted winter raods it might've corroded shut; vinegar may help loosen that with a good soak but if the seals are compromised it can get into the conductors and it is an acid....

if the problem was a poor connection from oxidation/etc it's possible the connector has melted together, in this event the onnectcion problem itself is liekly causing the motor issue.

 

[cveen]

> How steep is this hill, how long, and how much weight, on what size wheel?

Let's say 150kg total of humans and 70kg of bicycle (it tops out my hanging-scale at 60kg, with a bit still left to heft). 20" wheel. The last stretch where it really gets cooking is 150m elevation in 2km, so 7.5% grade.

I plugged the numbers into the simulator, but I'm not sure how to interpret the results. I think the black line appears at what speed I'll get with full throttle and no human help? 15.8 mph and pulling 1300W? Seems a bit off... I'd say it probably goes about 20km/h, tops, on such a hill and will pull the whole 2000W that it's limited to.

> The connector

The metal screw-on collars of the connector still work beautifully - its the inner part that's stuck. I soaked it in WD 40 and followed the old advice "if brute force doesn't work, just apply more force". I figured separating the two was a necessary first step regardless - I'm not going to take the hub apart without separating it from the controller (although I did do the last wheel change leaving them plugged together). After some pretty aggressive pulling/wiggling it did come apart, with one of the female sockets still stuck on the male pin for one of the big current-carrying connections. So, that's one problem. Not sure if it was the original/only problem before, but certainly a problem now. Resistance is close to, but slighly above, the contact resistance of my multimeter probes (0.7 ish ohm, 0.3 ohm if I just hold the tips together)

Tomorrow I think I will take apart the hub itself and see what's to be seen in there.

 

[E-HP]

For future reference, while doing temperature testing on my steep test hill, i noticed that my motor would heat almost as quickly regenning down the hill as riding up. I hit 98C at the top climbing, and started descending before letting the motor cool enough and had to pull over pretty quickly as it rose over 100C. There’s no current rollback to protect the motor when the heat is from regen. It would be nice if it could disable regen due to overheating.

 

[cveen]

> For future reference, while doing temperature testing on my steep test hill, i noticed that my motor would heat almost as quickly regenning down the hill as riding up

Thanks for that - I had actually started wondering about the same thing.

> The connector

While sleeping I had the thought that, now that I pulled that socket out of the connector, I should plug it back in and see if the symptoms are the same. Now everything works. Leading hypothesis:

* Internal connection to the socket on the connector had been failing and possibly overheating for a while (it was stuck during the last tire change, months ago, and I decided to just leave it alone and change the tire with the wheel attached to the bike)
* Heading up to Grouse was the last straw and finally burned through the remaining strands of connection
* Plugging/unplugging the connector has managed to make things touch inside again, somehow
* Motor/controller are fine

So, the question is - can I buy a field-installable, waterproof version of this connector? I note that Grin carries an extension cord, which could be cut in half and spliced, but that seems pretty cheesy and I'd prefer more waterproofness. Seems that the manufacturer makes a variety of codings called L1019x... I think Grin uses A-coding from looking at the pictures? They do have a panel-mount version. I could do a small junction box with that on it pretty easily...