Controller Troubleshooting

[hillzofvalp]

Hello. I have an EB-212-AS-1

I made the mistake of Misconmecting a phase to the wrong leg of one fet In a bank. I just replaced most of the fets (though they all seem to test fine), and now I am back to even resistance (~9.5K between battery leads and phases; 19K between phases).

To test them I just applied 7V VGS and checked for continuity . Then turned t hem off and diode tested them. I don't know why they seem fine..

First round of problems the wheel would hesitate in one direction. I heard a pop with only a 1K precharge resistor limiting current to controller. So then I discovered the Misconnection.. So I connected it the right way. Then the controller pulled power and smoked my precharge upon connecting battery leads (off position).

At some point something went pop.

Then I replaced that whole bank with new fets. Retested and now it wasn't pulling power upon connection to battery but there was resistance in wheel in both directions. (bear in mind it isn't a motor phase short). Some of the resistances between phases and leads and phases and phases were a bit different.

So now I have two banks of new fets and one bank of half new fets (only ordered 10). Hopefully now that the resistance is even I should be good.

Now when I turn it on I get gumminess in both directions. Seem s like the new FETs aren't turning off. Which are the gate resistors and what values should they be?

 

[hillzofvalp]

THe good news is that I found the problem area. I left a recharge resistor connected and one FET slowly melted the plastic shoulder washer and my 20W temp. recharge resistor smoked. This FET happens to be one of the old FETs, but it is also where I jumpered the gate from the mated pair to this one because the trace pulled off the board during resoldering. I'll give it another shot with a different FET (though the mated pair seems fine).

For some strange reason at this point in time (when everything went visibly thermal) the wheel jittered a little bit.

update: the fet that died was actually a new one in a bank of four new ones. If it was a resistor problem I would assume both FETs in the pair would have dissipated. Maybe static damage or a trace on on the back that was shorting?

update 2: with replaced FET the controller lurches when power button is pressed. Not enough to blow 5A fuse through 8Ohm resistor.I unplugged the phase I just worked on and it doesn't do anything on start. However, NOW the wheel has resistance even with the controller powered off (battery leads connected).

The gate resistors do not appear to be labeled with a number like the other black ones. When I pulled 3 of them off I could not get a reading off any of them, either meaning I'm destroying the resistor upon removal, they were destroyed during FET disruption, or I'm not measuring them correctly.

 

[hillzofvalp]

Here is an image of the controller with two gate resistors replaced. all of the gate-source readings are in the M-Ohm range (these fixed ones are now 10Ohms). I assume they were all damaged? please verify that these are the correct resistors to replace. at this point I'm very eager to have it functioning so I'm taking my chances.. already lost two of the original SMD resistors.

One pair is reading .43K GS with the gate resistor removed...?

 

[amberwolf]

Looks like an awful lot of little solder splashes all over, not only on the workbench, but on wires and the PCB. I think you might want to clean all of that up with a really good magnifying glass and tweezers or something, before continuing. I suspect you're getting shorts from those, and damaging things further becuase of it.

Not sure about anything else that might be wrong.

 

[hillzofvalp]

I will. Most of what you see is sloppy stuff on the wires that I don't clean in the midst of frustration. I'm particular about metal to metal shorts during testing. workbench stuff is from emptying solder sucker. Does it look like I replaced the right resistor?

update: everything is clean. I just want to make sure I'm not switching the wrong resistor.

and here is a picture showing the bad connections I made when I wasn't paying attention when I beefed up the traces (I've beefed up a different 12FET controller correctly). marked in red. Maybe that helps determine what was destroyed.

 

[scriewy]

O M G hillz this is so hilarious ! !

i thought to my self when you said "it was the easiest fix only 1 fet toast" what a lucky incident.

but seems you laughed to loud and the smoke GODS heard your joy through their ancient slumber !

now looks like all hell broke loose.

when i had similar issues i checked the transistors, the small 3 leged once like c1, there are usually 3-4 on each phase.

how did i check'm, in 2 ways i could with what i'v got, i tested firstly the working transistors with diode mode every leg to leg combo to see what is a normal reading (what ever the numbers are), if i couldn't determine squat, i desoldered them out and soldered 3 wires to transistor leads and stuck them in the NPN / PNP holes that our digital multi meters have, to get same readings as the working ones. this was a beach.

i replaced few suspect transistors and resistors on my 12 fet clyte, and i got it to limply work by spinning my motors very slowly and with much growling, it's sensored cont, i suspect maybe the halls in my motors are sleeping with the fishes, and i haven't updated my thread yet cos i didn't order halls for motor yet to test the cont and finally find out if i fixed it or not ! it's staled over 4 months now.


and wtf is wrong with your big cap lower left corner hanging off the board like that, suicidal ?

 

[hillzofvalp]

Lol. Nice post. The cap is an upgrade (5800mfd?) cap that gets soldered on with the battery leads cause it's so massive.

At this point I don't think trying to make this one work is worth my time, but I may mess with it some more. Even if I got it to work I don't think it would ever be as reliable again. At this moment I'm considering an 18FET controller or a new 12FET. Then I would have plenty of spare FETs and hopefully enough practice to not screw up another board.

Also to note is that my build is already sketchy enough as it is without running an extruded aluminum enclosure (heatsinnked to u channel instead). I'm not even sure it's work risking ruining an18Fet, especially in such a tight triangle enclosure