Crystallite Controller Capacitor Failures and possible fix

[Puppyjump]

I've read of issues elsewhere on this site with bike kits using Crystallite controllers.
In regards to reports of capacitor failures, I think we all have seen the intense, loud, spark that happens when the controller is connected to the battery. Is it possible that the current surge associated with the spark is causing the caps to fail by stressing them with the inrush current? This current surge is especially bad with SLA battery packs that have no BMS to limit battery current.

If this is the case that battery inrush current causes damage to the caps, then I have a solution: When the controller is connected to the battery, it should be done first with a resistor in series with one of the leads. A 5 or 10 Ohm works nice. After allowing the controller caps to charge up softly thru the resistor for a few seconds, then it can be removed and the wires connected directly as the controller caps hold a charge for several minutes (assuming the controller switch is OFF).

The best way to do this is with a SPST battery disconnect switch in one of the controller leads, probably the + lead. Then, a NO push-button momentary switch can be connected across the SPST disconnect switch with the resistor in series so that when the push-button switch is pressed, the controller is connected to the battery pack thru the current limiting resistor.

The bike power-up sequence would then be operated in this sequence:

First, insure the controller switch is OFF.
Next press the push-button switch for a few seconds.
Then, close (turn on) the controller battery disconnect switch.
Finally turn on the controller power switch and go for a ride.

After the ride, to prevent the controller from slowly draining the battery pack, open (turn-off) the disconnect switch.

In this manner, the controller caps will never see a current surge associated with an instantaneous connection to the battery.

 

[northernmike]

The R/C controller folks have been doing this for a while. Check the non-hubmotor drives section. :wink:

 

[potatonet]

thats a good idea,

probably what caused my cap to blow...

 

[fechter]

The inrush current when connecting the batteries is probably less likely to cause cap failure than overheating during operation. Vibration and breaking of the leads seems to be a common failure mode also.

I use a precharge circuit on most of my controllers and have still seen capacitors explode.

 

[dnmun]

do caps blow because they get hot inside from the leakage current?

i assume they leak more for higher voltage spikes and also for higher frequency because the field is strongest for a larger proportion of the period with higher frequencies?

but the heat builds up in the layers and gets trapped?

 

[potatonet]

maybe the cap blew and then my fet blew

or my fet blew and my cap blew at the same time, the thing is I didnt even hit the gas that much...

 

[Puppyjump]

The inrush current when connecting the batteries is probably less likely to cause cap failure than overheating during operation. Vibration and breaking of the leads seems to be a common failure mode also.

I use a precharge circuit on most of my controllers and have still seen capacitors explode.

Bummer. Oh well, it was a gallant effort.

 

[fechter]

There are probably several failure modes, but under high motor load, the capacitors are seeing some pretty high current for extended periods of time which will cause them to heat up. The lower the ESR, the less heating they have. Once a cap fails and goes open, the FETs will immediately blow in most cases. If you have a very low resistance connection to the batteries and the batteries have a low ESR, they take some of the current off the caps.

Capacitor current seems to be the highest at low duty cycles, like when you are just beginning to accelerate. Most of mine that blew went when just taking off from a stop.

 

[potatonet]

that is exactly when mine broke!

and what is ESR?

I have low resistance line to my lifepo4 battery

 

[rhitee05]

what is ESR?

ESR = equivalent series resistance

Real capacitors are not ideal components (nothing is). A more accurate model of a capacitor adds a resistor (ESR) and an inductor (ESL) in series. The values for those vary depending on the material and construction type of the capacitor. For duty like a motor controller, the capacitors are used to smooth the voltage as the FETs cycle on and off. This means they alternately sink and source fairly large currents with each PWM cycle - which is called ripple current. All that current flowing through the ESR causes efficiency loss and the wasted energy heats up the cap. Lower ESR makes for better efficiency, more stable output, and less heat. As fecher pointed out, a low-resistance/low-inductance connection to the batteries also puts lower stress on the caps because it means the batteries can source/sink part of that current themselves.

Electrolytics are cheap and high-capacitance, but generally have high ESR. You can find more specialized ones with lower ESR. Tantalum caps have nice low ESR, but are more expensive and you'll pay through the teeth for high-voltage ones. Ceramic caps also have low ESR but lower capacitance. Generally, the best approach is to use several types and values in parallel. Tantalums or ceramic in parallel with electrolytics can still give good performance without breaking the bank.

 

[potatonet]

I may put the capacitor I blew up here later and ask for the proper one I think it was the one labeled SM36 or something like that, no capacitance value on it.

 

[StevenR]

I had several C'lyte controllers fail due to capacitors breaking leads. C'lyte now straps the cap to the board with a zip tie like they should of all along. The hole was always there and was obviously for just this purpose. One controller blew up bad but the other seems to work. If I replace the 1000uF 100V, what kind should I use? Is there benefit in oversizing the cap? Is a higher voltage rating usefull?

 

[rhitee05]

If I replace the 1000uF 100V, what kind should I use? Is there benefit in oversizing the cap? Is a higher voltage rating usefull?

At the risk of over-simplifying, you can't go too far wrong by increasing the capacitor size. It will increase the inrush current when you first plug the battery in, though. Higher voltage rating isn't useful unless you plan to seriously increase the voltage in your system. A 100V cap is more than adequate for a 48V setup.

 

[Roy]

What a bizarre co-incidence! Something blew yesterday while starting to accelerate from 15mph (leisurely) after the bike had been fine for months and parked for a week. It was slightly warm outside. Symptom: virtually no power output. The wheel starts to move with throttle but then immediately goes "neutral" Readout voltage on my Drainbrain, seems too low and shifts around for no apparent reason. I'm guessing it's the controller because I don't see how a hub motor fault gives this behavior. CL 20 amp, 72v controller running at 60v. Pulling it apart later today....

 

[potatonet]

thermal cutoff?

mine did that too, just turned it off then back on, waited a few mins, was fine.

is the controller encased in something?

 

[Roy]

What a bizarre co-incidence! Something blew yesterday while starting to accelerate from 15mph (leisurely) after the bike had been fine for months and parked for a week. It was slightly warm outside. Symptom: virtually no power output. The wheel starts to move with throttle but then immediately goes "neutral" Readout voltage on my Drainbrain, seems too low and shifts around for no apparent reason. I'm guessing it's the controller because I don't see how a hub motor fault gives this behavior. CL 20 amp, 72v controller running at 60v. Pulling it apart later today....

Guessed wrong. One of the 5 Sonic Batteries had died. wouldn't pass current and only about .5 volts under load. Swapped in an older half-good B&B and now all is running again. What a relief, but what is up with the battery I wonder. Anybody else have sudden failure of an SLA?

Now with a mixed batch of batteries, 'am thinking of adding a tiny voltmeter and rotary switch to see all the battery voltages under load.

Back on the road