12 Fet Help Please - Resolved :)

[Danny Mayes]

Hi,

I built up a 12 Fet kit I got from Keywin.
Having a little issue in that i have full battery voltage where i should 5v for the halls.
I found out the painful way by smoking an optical board Next time i build one i will measure the voltage before hooking it up

Has anyone experienced this before and/or can anyone shed some light on the matter?

Any help appreciated.

Regards,

Danny

[amberwolf]

At a guess, if you have pack voltage instead of 5V at the halls, you probably also have it at all other 5V locations, likely including at the MCU, and if so I doubt it has survived. Same is possible for gate drivers and other things, which would run on the 12V. Since the 5V is derived from the 12V, then pack voltage must also be present there under one of the most likely circumstances (short or misassembly in the low-voltage section).

You'd want to verify the entire low-voltage power supply section's assembly and components, starting at the pack input connection and working your way to the 5V area.

[Danny Mayes]

Thanks amberwolf.

I had given up on getting a reply as I was waiting on an email notification but never got one Glad I checked in. Thanks for the support.

Makes sense. I wouldn't know where to start with trouble shooting the low voltage power supply section so I might discard the board and start again unless someone could point out something obvious...

Having said that, could I just solder up the fets (no heat sink), attach hall wires, throttle wires, power and phase wires and test it like this somehow before I build it up any further? To save my self the time if something ain't right from the start? I have a 16v 40amp power supply I can limit both voltage and current with.

Oh, I just realised something. I was limiting the current to 1 amp when I first fired it up. I used the above PS and slowley turned it up form 0v to 16. with it not working i then tried 50v from battery. Could the current limiting caused voltage to rise and cause the failure? I don't think so because the reg should have still dropped the voltage to 5v right? Thought I would ask though...

Thanks again for the guidance AW. Any more you or anyone can provide is much appreciated

D

[mdd0127]

Get out a meter and work the circuit. Start at the hall output and work your way back to the voltage regulation section. If you start getting battery voltage along the way with no components in between, you have shorted a trace somewhere. If you get all the way to the voltage regulator, then something is wrong before the regulator, which probably means it's toast and will need to be replaced. Then start at the voltage input and work your way forward, comparing your readings to what you think the circuit should be doing, based upon the components you come across on the way. I looked over your pics for a while and other than some flux left on the board and a couple of iffy solder joints on the bottom of the board on the upper right corner of the pic, everything looks great. I don't have a 12 fet anymore so this is going from memory though.

When I'm trouble shooting, I like to power things up, ground the meter, and follow circuit traces until I find something that doesn't make sense. You could have a diode out of place, a resistor in the wrong spot, or a bridged trace somewhere. Once I've verified that everything is in the right place, I usually start checking stuff at component level. Without having a working 12 fet board or a schematic in front of me, it's hard to troubleshoot this over the net though.

[Danny Mayes]

mdd0127, thanks.

I will do a little probing over the next few days following your advice and get back. I am not too confident in my ability to find anything meaningful but i will give it a go...

Reading a doc i found through google -

http://www.touzet.eu/divers/Infineon%20doc.doc (not sure who compiled this, but thanks)

Page 13 - "It is very easy for a user to set the LVC too low and blow the controller under a low voltage condition"

Maybe this is what happened as i started the controller at 1v and turned it up to 16v (thought it was a good idea at the time) trying to aviod explosions but may have just bits from the MCU from the start. I have no idea how a low voltage condition could blow the controller??? or how that would give me b+ voltage on the 5v rail. sounds more like a shorted trace the more i think of it. Anyhoot, i will prob around and see what i can find.

Thanks again mdd0127

D

[Danny Mayes]

OK, I found the issue.

It was 2 things. First was one of the fets was leaking voltage to the heat sink and then onto the bike frame. Second issue when I bolted the optical board to the bike the washer I used just touched on one of the legs of a sensor, passing 50v back through it (from the frame). Bad news is the optical board is fried; good news is the 12 fet seems to be working...
It was doing my head in for a bit because as I measured the voltage for the halls without the optical board plugged in I got 4.5v. when I plugged it in I got 50v An hour later the penny dropped.

Saga over and lessons learned thanks for the help guys

[DAND214]

Glad you found the problem.

when did you get that board? EB212 is now older board.

I have 6 EB3xx boards from Keywin in the last 4 months. # 12 fet and 3 6 fet. I also have a coupl eb812 boards but that is a while ago.

Great it's ok.

Dan

[Danny Mayes]

Hey DAND,

I got them from Keywin too. Must be 9 months or so ago now...



D

[dontsendbubbamail]

Doesn't look like your current shunts are soldered on the pad where you added copper. Might be something to check if that picture was after you finished soldering.

Bubba

[Danny Mayes]

Hey Bubba,

You are right, however they are soldered to the top of the board. I will drop some on there next time i got me iron out though...

Thanks for the heads up

D