Capacitor bank for Castle controllers

[ferret]

I'm about to install A Phonix Edge HV controller, and want to fit a capacitor bank.

I've read the threads about adding a capacitor bank to reduce ripple voltage. A common recommendation is to fit 4 parallel 1000 micro-farad capacitors.
Is 1 4000 micro-farad cap better or worse than 4 1000 micro-farad ones?
Most of the threads say that the capacitance of the bank doesn't really matter, so why not fit for example a single 1 micro-farad capacitor?


Which of the following capacitors would be best?

http://www.ebay.com/itm/2-pcs-Panasonic-series-FC-1000uF-63V-radial-electrolytic-capacitors-LOW-ESR-/251441001786?hash=item3a8b0d353a:g:71QAAMXQkcZRj6Lf
http://www.ebay.com/itm/4x-1500uF-63V-Panasonic-FC-for-Hi-Fi-DIY-Low-ESR-Audio-Electrolytic-Capacitors-/351364704068?hash=item51cef7e744:g:CHwAAOSwymxVIpkk (3 parallel)
http://www.ebay.com/itm/5pcs-1000uF-63V-16x35-5mm-NCC-KY-63V1000uF-Low-ESR-Long-Life-Capacitor-/252427542441?hash=item3ac5da9fa9:g:64sAAOSwM4xXYPA2
http://www.ebay.com/itm/63v1000uf-63v-Electrolytic-Capacitor-LOW-ESR-4PCS-/181480189318?hash=item2a410fed86:g:xzYAAMXQVERSqHOC
http://www.ebay.com/itm/1x-8200uF-63V-Panasonic-AUDIO-for-Hi-Fi-Low-ESR-Audio-Electrolytic-Capacitors-/261842551909?hash=item3cf7084465:g:IQ0AAOSwstxVJLNN (single)

Thanks,
Avner.

 

[recumpence]

Honestly, the newer controllers really do not need them if you are using high c rate lipo cells.

 

[tahustvedt]

Are you going to extend the wires? That's when you need them. Smaller ones in parallel will have a lower total ESR. One big one will probably work too.

 

[tangentdave]

My setup uses almost 6ft of cable between the battery and Castle HV ESCs. Their logging software shows less than 1V ripple with 10Ga copper-clad aluminum cable. When I installed the Castle capacitor bank, the ripple voltage dropped to 0.5V. Castle recommends less than 10% ripple voltage, or less than 5V...

-dave

 

[bobc]

Inverter capacitor banks need enough Farads and enough ripple current capability. The output bridge also needs local high frequency capacitors (film or MLCC) to make a "commutation loop" to control switching transients.
Back to the capacitor bank...
How many Farads? To calculate:- how many amps are you pulling? what switching frequency? How much voltage ripple do you want on the DC bus? Throw some figures in -
40V DC bus
4kW load
20kHz prf
0.5V ripple
Worst case is around 50% modulation and you get 2 "active periods" per switching period; "active period" (in the PWM) is 50us/4 = 12.5us. delta V = I * t / C so you need C = 100 * 12.5u/0.5 = 2500uF. That's why 1uF probably isn't going to help much. The solution is usually aluminium electrolytic capacitors.
Ripple current is going to be at 40kHz (and its harmonics) and will be fairly close to a worst case of 0.5*IArms - 50amps in this case.
Finding a couple of thousand microfarads at 40V is pretty cheap and easy. Finding it with 50Arms capability at 40kHz is a bit trickier.
Frequently, sufficient current carrying capability can only be achieved by having too much capacitance, that's why folk quip that the capacitance doesn't matter - the bank size is dictated by its current requirements, not its capacitance

 

[ferret]

Thanks everybody.

I will be using 2 of these:
http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=74096 - 12s1p, 12,000 mAh, 10C constant 20 peak
in a backpack (connected by about 5 feet of 10 awg wire).
I don't know yet how much current I'll be pulling.

Bobc,
can you elaborate about the high frequency capacitors, what rating, were to wire them and how to gauge their effectivness?

Can anyone comment about the suitability of the capacitors from ebay that I linked?

Thanks again,
Avner.

 

[bobc]

I looked up an epcos brand electrolytic (now tdk)
http://en.tdk.eu/inf/20/30/db/aec_2015/B41456_B41458.pdf
The datasheet gives most of the information you need, but you have to do some deciphering..
I looked at the 2200uF size, that is the smallest diameter & you can see that has the lowest ESL & still works as a capacitor up to just over 10kHz. Note to properly judge the effects of ESR and ESL you have to do a proper simulation, which is fairly complex but doable & simulators (e.g. LTSPICE) are free off the web. You'll also note this thing has a current rating of 34Arms so it would possibly have reduced lifetime in our application (though the 50A ripple is worst case & possibly only occurs transiently - maybe it would be OK).
No way an EBAY ad will tell you what you need to know to do a design. You need to look at the original manufacturer's technical data. Also I'd expect ebay components to be counterfeit, the capacitance is likely to be near correct on purchase but degradation would be swift in a fake. You'd be better off going to one of the distributors (digikey, RS etc.)
Note failure in an Al can electrolytic is generally gradual reduction of C as the electrolyte evaporates (though they can fail spectacularly if you really over face them or wire them up backwards). So a dead one is like none at all, frankly most users wouldn't notice..........

 

[Gregory]

... were to wire them

As close to the MOSFETs as possible. In practice that would be on the battery wires immediately where they exit the ESC. This pic is from Castle. I'd try get closer than this. Beware the pre charge spark when connecting a battery to this!

Also back to the first post, multiple parallel caps will handle heat better than a single large one.

 

[spinningmagnets]

The three motor phase wires from the controller to the motor can be long, the two wires from the battery to the controller should be short and fat. The capacitors should be low "equivalent series resistance" Low ESR.

 

[macribs]

*Subscribed*

Let us know how the cap banks works for you.