Kelly KLS reverse polarity damage-malfunction

Ca7

10 µW
Joined
Aug 12, 2022
Messages
6
So, I managed to IQ over 9K when I connect the battery with reversed polarity(series XT60 cable, thought I can't mess up with XT's lo) to my Kelly KLS-7230s, by some absolute miracle, the thing still works, but the motor makes a low frequency rumble at low speeds, I assume caps are shot, they self discharge rather quickly(~60sec? Not sure), also controller now needs more volts than before to turn on(before LEDs started blinking ~12v now ~30v), so there is probably more damage, any ideas?
 
Can you open it up and look inside? I don't know if those are potted. Also, can you measure the current when it's on?

The voltage issue sounds like possibly a voltage regulator problem. From the battery, there will be a regulator for around 12v and one for around 5v. One of these might be sick. You can measure the 5v on the hall sensors or the throttle in most cases (not exactly sure for a KLS).

If the caps went bad, they usually swell up or explode. 60 seconds to self-discharge is probably close to normal.
 
THe most common things for reverse polarity damage are

--the phase FETs (which may not outright fail but can be damaged from the extreme overcurrent and die heating thru the body diodes)

--the capacitors (from outgassing / electrolyte damage from the extreme overcurrent and internal heating; sometimes lead damage if they aren't mounted body right to the PCB with very short leads, sometimes they can desolder from the PCB partly or completely from the heat). In most cases the damage to caps, leads, or solder is visibly obvious, but sometimes it just changes the characteristics (ESR, capacitance, etc), sometimes enough to make some definite change to system behavior, sometimes not.

--the first stage of the LVPS (low voltage power supply) that supplies the 15/12/5v etc to the rest of the system. That might just affect the output from it vs the input voltage, or decrease it's efficiency, etc., or other things.

--Sometimes further downstream stages can be damaged if the first stage outright fails. That is unlikley in your case since the controller still functions (which it would not do if the LVPS first stage was destroyed).

--the shunt, if it uses a soldered-in shunt wire (the Kelly I have here has an external bolted-on shunt, as well as external bolted-on fuse; yours is probably soldered in as either a set of manganin wires or SMT shunts), can desolder from the heat, sometimes completely, sometimes partially. One controller I got as a known-damaged unit from a reversal actually had it's shunts burned open.
 
Resistance of Y B phases to ground is 54k G phase is 108k, idk what normal values are, some images https://imgur.com/a/QKecrqf I didn't really see or smell anything, I've since put it back together, 5v measures at 5.13V and 12 at 10.8V the current idle current at 28V(bench PSU) reads 0.04/5A. When buying it there was an option for external shunt, but I didn't get one.

On my cheapo china controller, the caps not sure but probably loose ~5V per day
 
That all sounds pretty nominal.

Have you tried running it at full throttle? Just wondering if maybe one phase quit working and is giving you the rumbling sound.
It's remotely possible that all the main caps failed in some manner, but they typically show signs of it. One crude test is to disconnect the controller from the battery and quickly short the controller battery wires and see how big the spark is. With good caps, it should make a pretty big spark. If you have a meter that can measure capacitance, you could remove caps from the board and test them. Caps are pretty cheap, you could just simply replace all of them, but this may not fix the problem.
 
Ca7 said:
Resistance of Y B phases to ground is 54k G phase is 108k,

They should all read about the same value. If one phase reads twice what either of the others do, there's a chance that half the FETs in the one different phase have failed or been damaged, or unsoldered, etc.

Or that the other two phases have FETs damaged so they are lower resistance than usual, but it would be unlikely for them to both fail identically....so it makes more sense that the 108k phase is the problem.

5v measures at 5.13V and 12 at 10.8V
5v sounds ok. 12v sounds pretty low--it's usually used for the gate drive of the FETs, and if it's not high enough, then depending on the design it might cause low gate charge current which might not turn FETs on fast enough or all the way, and cause incorrect phase drive. Since it's usually the first stage of the LVPS, it's the most likely to be damaged by reversal, depending on design.

On my cheapo china controller, the caps not sure but probably loose ~5V per day
That could just be that there's no safety drain resistor on those, but that there is in the Kelly (I don't know for either design). Some systems put a high resistance across the main caps so they will drain relatively quickly when no power is applied, so that servicing presents little risk.
 
Thought that was a possibility with the resistor and caps, ah curses, maybe it would be a good idea to send it to someone to fix?(EU) I think the low voltage small components stuff is beyond me.
 
Depends on how they designed it. Good clear well-lit (direct sunlight is best) high-resolution pics of everything with it opened up, both sides of the PCB, might show us enough of the insides to suggest specific test points. No guarantees, but it might be as simple as replacing a TO220-format LM317 or 7812, as far as the 12v supply goes. Some image examples of those:
https://www.google.com/search?&q=TO220-format+LM317+or+7812,&tbm=isch

Replacing FETs, if theyr'e damaged, can be harder, because it's harder to heat the pins up quickly enough without damaging anything else, even with TO220 cases. If they're any of the SMD/SMT formats, it can be even harder.
 
Back
Top