Exploded Kelly Controller

[Farfle]

I fired up my Kelly controller for the first time today, and I puffed it on the first run-up. The controller went bang right after my friend applied a tiny bit of throttle, the motor stuttered, and then the controller emitted a large amount of black smoke. What I don't get, is that the motor checks out fine on the Lyen phase tester, it spins smoothly, and when it is spun by hand, all three lights light up and cycle left-to-right, and the hall lights all turn on and cycle in a similar direction. The other thing, is that when it exploded it left a 10 amp fuse untouched :? . On the positive side however, I have had such horrible luck with controllers lately that I decided to film the setup, and the first plugin in hopes that I could catch why it was blowing up if it decided to. The first video got a bit screwed up when I chopped the curses out of it, where the screen goes black, its pointing out the 10A fuse I had inline to the B+ terminal.


[youtube]UWcPZBZl3lk[/youtube]


[youtube]QPFm_-oCyng[/youtube]

 

[neptronix]

Was this controller getting quite hot before it went bang?
I've heard a few reports of these controllers being very overrated in terms of amps/watts, some folks are only getting 1/2 the rated power O_O

 

[Farfle]

Was this controller getting quite hot before it went bang?
I've heard a few reports of these controllers being very overrated in terms of amps/watts, some folks are only getting 1/2 the rated power O_O

It was pretty warm after it exploded, but it had never been run before and even if it was only good for 75a (150a rated controller), it *should* run a 5304 under no-load just fine.

 

[neptronix]

sorry my bad on the reading comprehension part. lol

 

[Ricky_nz]

oh no, not good.
Looked like it didn't even manage to go through one hall switching cycle before the popcorn was ready
Either an internal fault in the controller or a hard short of phases that the controller didn't manage to control the current in.

 

[Farfle]

Hmmm... Ill make a video later testing it with the Lyen phase tester. hopefully it should still be working, unless something changed while it was sitting on the bike. Meanwhile, have sent an email to Kelly, hopefully they can be of some assistance.

 

[liveforphysics]

That's a FET pass-through event. I've had it with RC controllers a couple times.

Even a 1amp fuse doesn't help those, the energy in the caps alone is enough to send all the caps to hell in a pass-through.

 

[Farfle]

hmm, so the amperage flowing thru the FETs coming from the caps is enough to pop them? and what causes this? It keeps circulating back to a shorted phase, but the motor spins freely by hand, and the motor generates AC voltage across all the phases :? .

 

[Ricky_nz]

All it takes is one dry joint in the wrong place in the controller to cause both top and bottom FETS to come on at the same time and its all over.

Blame lead free solder...

 

[liveforphysics]

hmm, so the amperage flowing thru the FETs coming from the caps is enough to pop them? and what causes this? It keeps circulating back to a shorted phase, but the motor spins freely by hand, and the motor generates AC voltage across all the phases :? .

No, a pass through doesn't need a shorted phase at all, doesn't even need a motor to be hooked up. A pass-through event is when the controller tries to send the same phase leg to be both positive and negative at the same time, internally. This results in much smoke and plasma.

 

[liveforphysics]

All it takes is one dry joint in the wrong place in the controller to cause both top and bottom FETS to come on at the same time and its all over.

Blame lead free solder...

Lead free solder really is the worst idea EVER. I can't even imagine how much more e-waste the world has, and how many millions of products get tossed out in a week because of lead-free solder.

 

[Ricky_nz]

Lead free solder really is the worst idea EVER. I can't even imagine how much more e-waste the world has, and how many millions of products get tossed out in a week because of lead-free solder.

Especially when combined with poor quality control processes.

 

[amberwolf]

Well, at least it is probably repairable, but you'd have to find out why it did that, because it probably did turn on both legs of a phase. If so, then I doubt it would be software causing it, but probably a hardware defect, like bad solder (as mentioned above), short from solder bridge or a wire or component leg not properly inserted into the hole, etc.

(as an example, a long time ago, before I got into motor controllers, I was disassembling and salvaging a car audio amp that blew up on first power-on, and found that what was left of one of the FETs still had the gate leg not even inserted into the hole--it was squished up underneath, and touching something else. So the gate was probably turned fully-on, rather than just a little here and there, and the leg probably also passed more voltage/current into the gate then it was meant to handle).

Assuming a single manufacturing defect, it's pretty likely that the other two phases are not damaged, but the one that failed likely has no intact FETs (the rattle is probably what's left of them, and may include little metal balls that you'll need to find and remove, that could've stuck to various things inside while they were still white-hot).

 

[Farfle]

Ill probably send it back to kelly, hopefully they warrantee it...

 

[Doctorbass]

Sorry to see that you got a problem with your kelly

I have the KBL 220A 120V that Method sold me a year ago.

I installed it on my 5303 with 75% power setting and 24s lipo in the program. I get up to 12kW continuous during the acceleration ramp and also it maitain 6-9kW at 100-110km/h on the road.

I really love that controller compared to the 18 fets!.. good power, still no risk of blowing work all the time, and.. AND.. you can se the previous program setting!!!.. compared to teh infineon that dont!

I plan on upgrading to 30s lipo for my next speed record.. i'm still not decided on IF or not i'll put the current to 100%

The controller is set for TORQUE ( i can also choose speed and BOTH)

My C-A is perfectly calibrated and i used two different REFERENCE shunt i have to validate the current and also my calibrated FLUKE 337 DC clamp meter. I record easy 239A burst on the C-A after the 2-3 first sec of acceleration wich correspond to the (phase current) too.

My voltage drop to 95V from fresh charged 20C 15Ah lipo.

Controller rarely get warm...

Doc

 

[Farfle]

I got the 150 amp 120v kelly controller. Its programming software is sooo much nicer, but Its only rated for 75 amps continuous which is all that I really use. Only complaint I have is that its really big! And that it exploded, but that's just a side note

 

[Farfle]

Nope, it was used with a 5304, I ended up returning the controller because it was so large, expensive and complex. A lyen 18-fet outperforms it, and puts $300 back in your pocket.

 

[Andje]

when you say a lyens outperforms it can you be more specific?
i have a lyens 18 fet paired with a 5304 too, and i'm happy with the performance i see from the 18 fet. It does get hot during use, but seems to peak even though the motor continues to heat. Regen also seems to generate most of the heat in the controller.
Anyway I had thought about going to 120v much along the lines of DB using a 120v kelly. I'm just interested in your on the bike feelings between the two at 100v...

 

[Farfle]

well, on the bike the Kelly did have a few low-speed start up issues, very mild but caused some grumbling if pegged from a stop. But set at 125a, which was what I had my 18-fet set to, The Kelly did get very warm, much warmer than the lyen controller did at that same amperage. Overall the 150 amp Kelly is a waste of space, but the 250 would be worth it, as its ub an identical case. Best programming software that i have ever seen however.

 

[Doctorbass]

I am using on my Giant the old 220A 120V Kelly controlelr of the KBL serie ( now replaced with the 250A)

I bought it from Methods why roasted his X5 with it .. On my X5 ( 5303) i have a temp meter so i know the exact temp of the winding all the time.

Everything worl well.

i'm using 30s lipo ( 126VDC right of the charger) and max current set to 75%. The controller temp stay under 100F adn i easy draw 15-20kW with it.

I am very satisfied with it!!!

Remember.. the 5303 IS what make my infineon 18 fets controlelr to blow! it dont like the high phase current !

on the other hand The kelly is perfect. compare to all problem i had with many 18 feys controllers

Doc

 

[oatnet]

The first Kelly controller on my x5tracycle blew after a 3,000+ miles of 88v92a, and Kelly sent me a new one, delivered in 2 days with no questions asked. Shinyballs had a similar expereince. Both are posted here:

http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=3415&start=45#p85865

-JD

x5tracycle Build, and Upgrade, threads
http://www.endless-sphere.com/forums/viewtopic.php?f=6&t=3514
http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=10201

.

.