Highside Mosfet keeps blowing

[Analogrider]

Hi,

My wife purchased an Envo Lynx 20 ebike directly from Envodrive.com and it keeps blowing the high side (positive) mosfet on the yellow phase wire.

The bike uses a Kunteng KT36SVPRCD-CBWF06L controller. Specs in photo below:


The motor is advertised as 500watt but based on the controller seems more likely to be 350?
Battery is 36V 12.8Ah.

It's blown the same mosfet on 2 (probably 3) controllers now. I didn't get a chance to confirm on the first controller, as it was replaced under warranty but on the 2 subsequent controllers I used the troubleshooting method outlined on the Grin website and traced it to the highside mosfet on the yellow phase wire. I salvaged a good mosfet from one of the controllers, tested it, replaced the bad one in the other controller, and it work's fine, so that the rest of the circuit seems to be ok, it's just that one mosfet.

Every time it's happened basically the same way: from a standstill, with no one actually sitting on the bike, the throttle is engaged to the maximum setting. Basically, we were using the throttle to try assist getting the bike up our front steps. The first time just on the bare stairs, the second on a ramp, and the 3rd time my wife caught the throttle on the virgina creeper that grows on our fence.

All 3 controllers were used for a few days before blowing, under load, for a dozen or more kilometers of riding. They were primarily used only via PAS but the throttle had been tested.

After the second controller blew, I modified the settings to limit current to 75% of the maximum. (C5 set to 5 on the LCD3 display). It still blew.

It seems to me that this shouldn't overload or overheat the mosfet? Is it the sudden inrush of current that is causing this? Or is there a bad component or short in the motor or motor cable on the yellow phase wire?

I'm new to ebikes, but know a bit about electronics and have been a home (traditional) bike mechanic for over 10 years.

This forum has been a great firehose of information getting me up to speed on ebike tech and I greatly appreciate it. I look forward to your insight, let me know if you need any additional details or photos. Thanks.

 

[neptronix]

Prebuilt bikes coming with absolute garbage controllers ( and worse stock tunes ) is nothing new.
You may have a phase short ( maybe mild/intermittent ) in the motor causing this, it's strange for them to die so soon. KTs are usually of OK quality.

It sounds like you did everything possible to try to mitigate the problem. I'd consider moving to a higher quality controller like a baserunner. It as least has diagnostics onboard that would allow you see if the problem is with the motor. A VESC ( a lot more advanced ) would also show you some signs if the motor is the issue. Both would likely shut down when encountering this short.. whereas a lower quality controller won't detect it and keep putting power in.

 

[dgc]

Although I have changed to mid-drives now I still have a couple of e-bikes with the KT controllers I have used them for years without any problem, yes, it could be a motor fault in the windings but you seem to be over driving the controller as I would say it a 250 watt unit true the controllers are usual rated at 7 amps continuous and 15 amps peak. I would at least be in the 11 - 12 amps continuous and 22 - 25 amps peak range. One bike has a Bafang 36v 350w / 48v 500w rear hub drive on which I run with a KT 12 amp cont 25 amp peak controller which I run on a 36v battery. Next question I would look at is what is the position of the controller, is it enclosed? as this will not help with frying the controller, but what is going thro' my mind at the moment is that if you are from a stand still trying to go up a steep stairs what sort of angle are they and the start loading on the controller may be too much and it does not have chance to shutdown, doe's the throttle work OK with normal riding?. That why I went mid-drive, drop down to a 46T cog on the back of my e-mtb and go straight up a 45 deg bank, but that is not from a standing start. Still cannot work out why the yellow phase goes unless it a weak point either in the controller or motor.

 

[amberwolf]

It's blown the same mosfet on 2 (probably 3) controllers now. I didn't get a chance to confirm on the first controller, as it was replaced under warranty but on the 2 subsequent controllers I used the troubleshooting method outlined on the Grin website and traced it to the highside mosfet on the yellow phase wire. I salvaged a good mosfet from one of the controllers, tested it, replaced the bad one in the other controller, and it work's fine, so that the rest of the circuit seems to be ok, it's just that one mosfet.

Every time it's happened basically the same way: from a standstill, with no one actually sitting on the bike, the throttle is engaged to the maximum setting. Basically, we were using the throttle to try assist getting the bike up our front steps. The first time just on the bare stairs, the second on a ramp, and the 3rd time my wife caught the throttle on the virgina creeper that grows on our fence.

All 3 controllers were used for a few days before blowing, under load, for a dozen or more kilometers of riding. They were primarily used only via PAS but the throttle had been tested.

The most likely cause is a wiring fault in the cable from controller to motor, usualy at the point it enters the motor, where the cable jacket is damaged and allowing a phse short to the axle when the cable moves or vibrates in the right way.

But it an also be inside the unbroken but visibly scuffed/etc jacket where the insulation between two phases or a phase and a ground wire are boht damaged and allow them to short together under the right conditions.

It can also be inside the motor, with a winding on the lamination stacks too tight at a corner and insulation cut thru enough to allow a partial short between them, etc.

Less likely it can be a poor connection from controller to motor on that failed phase, allowing arcing between the poor points. That can be at the connector pins, or at the back of them with a poor crimp or solder joint, or a broken wire at the back of the contact or anywhere inside the cable.


All of those can be intermittent types of failures, and especially the winding/lamination short can be one that requires a higher voltage to short, so only under higher power outputs (commanding a higher voltage across the phases). These are hard to detect with a plain multimeter because they may not read any differently at the very low voltage of a DMM, and instead need a visual inspection to find, or a "megger" or "hipot" type meter.


Not really an answer, just a (few) starting point(s).

 

[dgc]

I would test with a good old fashion megger.

 

[docw009]

KT controller will max out at their rated current, which is 17A in this case. On a fresh battery, you could get close to 700W on a hill if you had a KT display that showed watts, but I don't think your unit shows watts,.

Amberwolf's comment about a short has merit. but it's tough to damage the 9 pin motor cables mechanically, Users do melt them, but not with a small 17A controller. The phase wires are pretty low impedance, so it will be hard to see any difference checking resistance. They're also connected together. Maybe check for a short to ground. That's the center pin in the 9 pin cable.


I run throttle on my 20A and 25A KT controllers all the time. It's part of the pre-ride check, tilt bike on kickstand, and spin motor to make sure it works. Today, on my FWD bike, I had the tire spinning on the wet pavement going uphill. You're saying the bike blows the MOSFET when the wheel is in the air? It shouldn't. That is actually low current,

That brings up another thought though. Although this is a commercial ebike, with factory connectors, is there any chance of a miswired motor connector? With 3 phase wires and 3 position sensors, there are 36 permutations, but only a few will spin the motor, and one or more are false positives that will draw big currents, Have you ridden the big long enough to determine if it gives normal response? It doesn't growl or make noise?