help sabvoton 72200 meltdown!

[ridethelightning]

I have got a sabvoton 72200 +qs 205 50h v3 to try out on a friends bike(first sabvoton for me), i have programmed it and finally got a good grasp of the software operations. i finally let him take it home, and on testing, he said the throttle seemed unresponsive so he tried to get it to go by giving it a full twist. it was then that flame and sparks came from the controller!

we have not opened it yet to see but it sounds to me like the fets have probably blown.
i had it running fine for several day before this, and ran up 15km from testing it on the stand.

im suspecting that the halls test results have somehow not been saved and he was trying to spin it up with no combination configured.

-has anyone had these issues with sabvoton?

-are they easy/good to deal with for ordering parts?

any info greatly appreciated.
Cheers,
RTL

[Jonno]

Hi

I had this happen on one of my sabvoton 72200s.
I traced the problem back to where the cable exits the motor and the brake disc had cut into it. The short to earth will only occur when the disc retaining bolts pass over the cable. There is very little gap for the cable to pass through hence the problem.

Unfortunately the controller was written off

Jonno

[Manny]

just open it up! than we can see what is dead.

replacement parts are probably best sourced from digi-key, mouser or similar

good luck

[ridethelightning]

Hi

I had this happen on one of my sabvoton 72200s.
I traced the problem back to where the cable exits the motor and the brake disc had cut into it. The short to earth will only occur when the disc retaining bolts pass over the cable. There is very little gap for the cable to pass through hence the problem.

Unfortunately the controller was written off

Jonno

well in my case there was no disc fitted yet i dont think(the bike is at my friends house and i dont think he put them on yet)
but ill definately be keeping an eye out to see if something like that happened.

the qs205 has a pretty long axle shoulder too so i doubt there is any friction there.
ill be seeing the chargrill tomorrow so ill try to post some pics then.
thanks for the info.

[ridethelightning]

so i went to see the fried controller today.
on careful inspection, the guy had fired it up with only one, maybe 1 and a half, of the xt 150 phase plugs connected
i think they had become disconnected when he moved the bike to his place with a van, and as the bike is still only partly built, the phase cables were loose on the swingarm and the pedals(no chain yet) probably spun and knocked the cables pulling the plugs.

so basically he tried to give it full throttle with only one phase plug connected, the other was partly pulled out when i saw it and the third was loose altogether.

on opening the controller one of the banks of fets is obviously where the damage hit, 10 fets corresponding to one phase wire.
(there are 30 fets altogether)
would it be ok to replace the 10 broken fets or is it essential to replace them all to make sure they are consistent?

the pcb and other components look to be in pretty good shape still, just soot flared over them.
pics comn'

[ridethelightning]

so looks like the svmc 72200 uses Techcode TDM31026 fets-
http://www.techcodesemi.com/datasheet/TDM31026.pdf
but not easy to find on ebay.
iv emailed techcodesemi but since its chinese holiday, might take a while.
anyone know where else to find them?

[amberwolf]

Well, if only one phase leg was connected, no current could flow, so nothing could fail.

What is more likely is that the other two unconnected phases touched each other and shorted together, or both were touching the same piece of metal somewhere on the bike, shorting together. (or one or both were touching exposed metal on the phase that was connected, if any was exposed).

A third possibility is the partly-connected phase, if there was no electrical contact just mechanical, and hte gap was small enough, it oculd've arced across the contacts, generating a lot of RF whcih travels on the wires tehre and ends up doing unexpected things inside the controller to the FETs (and sometimes gate drivers). So the two "connected" phases could have current flow thru them but with a lot of RF content from the arcing. How or if that would affect the unconnected phase I don't know.


As for which to replace, any that test bad for sure. Generally if one in a phase is bad, so are likely the rest in that phase, so 10 of 30 makes sense.

If shipping is really cheap, you could just get the 10 you need, and then order more if it still doesn't work.

If shipping is expensive, you're probably better off getting all 30 at once whether you need them or not.

[alec]

Replace all the FETS, is what I would suggest.

[luc37]

Hi,
The same thing happened to me.
The hall angle was good and everything was runing grate for a few days. I did a 100km before it explode.
For no reason the rear wheel lock up when i was going down a small hill and all the high side fet's exploded and burn.

I have no explanation for this....

I think Sabvoton controller are really really bad !!!
I dont recommend them it is the third one that burn out with me for no reason.

Did you managed to repair yours ?

[ridethelightning]

nup. luckily it wasnt mine
wasted postage sent back for warranty and was a rightoff.
the guy who owned it got a mobipus 72200 to replace it. much better controller.
btw i found the sabvoton fets for sale(techcode) aprox us.50c each..says it all really

[Mick48]

Sorry to hear your unfortunate experiences with your Sabvoton 72200. Moreover, it sounded like the issues/meltdown you had was due to human error, not controller fault.
I have had my Sabvoton 72200 controller now for almost one year and all I can say it is one of the most reliable,quiet, and powerful controllers I have ever had. Also, it rarely gets warm even if I run the full 150-200 amps through it.