Bike died mid ride

DRock153

100 µW
Joined
Jul 29, 2022
Messages
9
Location
Long Beach
So I replaced my Sabvoton 80a controller and ran my hall test and everything was fine. Took a test ride and all seemed good.

This morning I was going to ride to work and about a mile into the ride I turned it up to level 4. The motor started to jerk a little bit and then the bike completely shut off. It wouldn’t turn back on at all.

I unplugged the battery and plugged it back, the bike turned on. As soon as I hit the accelerator it shut off again.

Checked wires and everything seemed fine, went to try the unplug and replug again… this time I got a little sparky spark arc so I assume something is shorted. :/ left it unplugged and walked home.

I just got home after work and rechecked all the wiring. Found a single wire came out of one of the connectors for the motor. So I switched to the other connector. Double/triple checked all the other wires and plugs and found no issues. Went to plug the battery, heard a very light pop and sure enough, nothing turns on…
(Positive sides of battery plugs are black from arcing)

Am I completely screwed?

Please help

(72 volt Molicel, sabvoton 80a controller, 3000 watt rear hub motor, nb power display kit on a Super 73 RX)
 
When a system completely powers off, it's usually the BMS turning off it's output to protect the battery against damage that could lead to a fire (either within the battery or the devices connected to it), such as overcurrent (too many amps drawn by system) or overdischarge of low cells (problem in battery, or battery just empty). So generally it's not advisable to keep trying to repower a system that's shut itself off, until you first locate and fix the problem causing the shut off.



Sometimes a fuse makes a sound when it pops under severe stress, so if you're lucky it's just a fuse within the battery. If that's externally accessible, check that first...but:

The pop during the last plug in of battery is most likely the FETs in the BMS failing catastrophically from too-high instantaneous current draw by whatever is failed in the system. You'd have to open the battery to verify what actually failed--FETs that make a noise when they fail are usually visually obviously damaged (the noise is the case exploding or fracturing from vaporized FET-guts ;) ).

If it's blown you can probably replace just the FETs (all of them, not just the visibly damaged ones, preferably, as the others could be stressed if they didn't actually fail), and have the BMS working again. Or you can replace the whole BMS.



If the wire that had come out of the motor connector was for a hall signal wire, it could cause the controller to be out of sync with the motor phases during rotation, and that can cause current flow that can damage the controller itself (often blowing up FETs, which usually fail shorted--if it fails both top and bottom FETs in a bridge, it's a dead short across the battery positive and negative).

This isn't usually a visible failure, but sometimes it is. However a dead short is easy to test for with no battery connected to the system: set multimeter to 2kohms or 200ohms, and place it across the controller's battery input leads (with NO battery connected or it will damage your meter). If it's shorted it will read very low ohms about the same in both polarities (red meter to red controller and black lead to black controller and vice-versa). To be sure the motor isn't involved, disconnect that, too.

If it is blown, it's probably cheaper (and is definitely easier) to replace the whole controller, because it's not usually just the FETs that are damaged in these events--the gate drivers can also fail, and sometimes other things (and you won't know any of those until you replace the FETs unless you have an oscilloscope to test for signals in all the right places, and possibly a working unit to compare them to if all three phases were damaged leaving no comparison possible within the controller).
 
DRock153 said:
Ok I will test the controller it when I get home…

Does the FET inside the battery look like a fuse?
Sorry I don’t even know what FET stands for

Here's a BMS with blown FET on the right side of the picture; the ones on the left may also be failed but don't have obvious case damage like the right side ones do.
https://endless-sphere.com/forums/viewtopic.php?f=14&t=95853&p=1731334&hilit=%2Afet%2A+bms#p1731334
https://uploads.tapatalk-cdn.com/20220906/8d1320960e6e596aea0d143571df63a5.jpg
8d1320960e6e596aea0d143571df63a5[1].jpg


This one you can see the cracks from the edge where the FET blew out
https://endless-sphere.com/forums/viewtopic.php?f=14&t=116194&p=1715933&hilit=%2Afet%2A+bms#p1715933
https://endless-sphere.com/forums/download/file.php?id=316923
image_16863233[1].jpg

Another one with actual smoke damage near one of the FETs
https://endless-sphere.com/forums/viewtopic.php?f=12&t=110352&p=1703412&hilit=%2Afet%2A+bms#p1703412
https://endless-sphere.com/forums/download/file.php?id=312227
ANT_BMS[1].jpg

Another
https://endless-sphere.com/forums/viewtopic.php?f=12&t=110352&p=1703412&hilit=%2Afet%2A+bms#p1702448
https://endless-sphere.com/forums/download/file.php?id=311799
BMS-burnt%20area[1].jpg
 
Unfortunately that means it's extremely likely the FETs inside have failed shorted (and other damage like gate drivers could have occured).

If you feel adventurous and have money and time to potentially waste, you can test it to find out which FETs failed and replace them, using instructions at http://ebikes.ca in their Learn - Troubleshooting section.

It doesn't cover testing gate drivers, IIRC, however, and the only easy way to do that is with an oscilloscope once the controller will power on. :(

To just test which phase(s) failed, you can use the meter the same way you just did, except put whichever lead you used on controller + wire to connect to each controller phase connector in turn (without the motor or battery connected). Whichever phase(s) read zero or nearly that are failed on the "bottom" of their bridge.

Testing the "top" of the bridge you move the meter red lead back to controller + wire, and mvoe meter black to each phase connector in turn, same as above.
 
What was the previous failure, and what were the symptoms of it, and what happened just before it failed? Sometimes a pattern of failures can show up that indicates a root cause outside the failed device(s).
 
Last time it was a 30H error and the guy from the shop I bought my kit from had me send it back for him to test. He said it was just a “bad batch of controllers”

I posted about it here

https://endless-sphere.com/forums/viewtopic.php?f=49&t=117102&p=1726069#p1726069
 
Back
Top