Quark fried, blown capacitor...

[zeropointbug]

Originally Posted by BrianG

ZPB: Just remember that caps in series calculate like resistors in parallel: Total capacitance = 1 / ( 1/C1 + 1/C2 +1/C3 )
So, if you get three 330uF caps and put them in series, the total capacitance is only 110uF, but the voltage rating is the sum of the voltage ratings of all three.

AAngel: Yeah, sometimes buying in quantity can save a bunch of $$$. So, if anyone wants to get some low-ESR 220uF/330uF 35v 105*F caps, I'm in for a group buy for 10-12 of them (plus shipping of course).

I know that, but if I get the 35v 330uF ones, I will use them in parallel. I was talking about the Novak trans-cap in series, which I don't really like doing, BUT IT'S SO DANG TEMPTING! :027:

What do you think about the trans-cap in series? three of them.

I also wouldn't mind getting a bunch of caps as well Brian. Have you found any radial type caps, all I have found is surface mount.

[zeropointbug]

wait, don't you mean resistors in series as well? That's not a very good comparison still, IMO. :030:

[Serum]

Nope, with resistors it's just a simple sum.. (in series that is)

Here; this offers some help;

Resistors in series;
http://www.physchem.co.za/Current%20...ors.htm#Series
Parallel;
http://www.physchem.co.za/Current%20...s.htm#Parallel

Capacitors in series;
http://www.play-hookey.com/dc_theory...apacitors.html
Capacitors parallel;
http://www.play-hookey.com/dc_theory...apacitors.html

[zeropointbug]

I know, resistors in parallel, you divide the Ohm value by how many you have, series, you multiply, but in each case, the power dissipation goes up by the total number you have.

is that what you mean? power dissipation?

I really like this trans-cap idea, I think it's worth it, $45 for three, and possibly the best protection for a $280 controller, and prob better performance. IMO

[Serum]

You have any idea why capacitors are on controllers? not to protect the controller. They protect the batteries from the fast pulses the BL controller uses to spin the rotor.

Batteries will get a high internal resistance on pulse loads.. (on both discharge and charge) (throttle/brake) voltage will drop, batteries will get hot. They just can't deliver the juice on the high frequency the controller works on. Not talking about acceleration peaks but the switching frequency. Not to mention the pollution it will cause in the HF region..

Best way to protect the controller is to use trancil diodes on the motor side.

[zeropointbug]

Yes, they are to protect the controller, it would probably run cooler if you have more capacity, low-esr of course, slightly more power. Yes, they also protect the batts as well, they simply can't deliver the short pulses.

[Serum]

Read my post again, i am saying the opposite.

Why would they protect the controller?

[zeropointbug]

I meant protect the controller from heat from switching, better caps will provide faster switching? Is that right? And I thought they protect whole system from noise, spikes, voltage ripple?

[BrianG]

Those Trancil diodes seem like they might help to clamp the back-emf pulses from getting to a really high value, especially with high-inductance motors. I just wonder how much power they can absorb. It's hard to see the underside of the board, but I don't think either Quark or the MM use them...

Personally, the only thing I'd want to do is add a little more capacitance on the input so the existing caps don't have to work as hard.

It's too bad these controllers pull so much current in a way, because I think a low turn coil on the battery lead would also help supply a bit of current boost as well as filter the ripple from the batteries. That in conjunction with the input caps would create an LC filter. Unfortunately, the wire would have to be quite thick (or even use Litz wound coil) and may even need a ferrite core to concentrate the flux, but then the controller would start getting kinda heavy.

[Serum]

I've seen them which are capable of killing 1500 watt spikes. And you can use them parralel to gain more power.

[zeropointbug]

Originally Posted by Serum

I've seen them which are capable of killing 1500 watt spikes. And you can use them parralel to gain more power.

What kind of voltage are talking here? Where have you seen them, on a controller, or just the part?

[Serum]

It's a transil diode instead of trancil. My bad.

http://www.st.com/stonline/products/...re/an/3568.pdf

they are available in several voltages (a LARGE variaty, from 6-600volt or so; found an 18V type, which can withstand 1500 watts during 1 milisecond.

[Serum]

Good thing about these transils is they work bidirectional.. (at least, that's what i thought)

http://www.ortodoxism.ro/datasheets/...onics/8694.pdf

here, the 25V type seems to be interesting;
SM2T27A

I don't know if their capacity has got negative influence on the EMF signal though..

[BrianG]

I thought about the effects of the needed feedback for the controller, but I figure it would clean up the signal if anything. If the signal is not a perfect spike (noisy on the top), then these would effectively shop that noise off - kinda like how FM modulation works in RF. However, because there would be no actual "peak", the signal might resemble a square wave more than a spike...

[zeropointbug]

Cool, would that part be something like what the Novak trans-cap uses then? In the little pic, they look very similar.

Wouldn't 25v not be enough though? Wouldn't we need something at least like 30 volts or more?