150v 35amp FDP2532 mosfets X-Lite Controller

[fechter]

It would work well in the v2 controllers, at least physically. Would the drivers be able to handle doubled FETs all around in the v2 units?

Not sure. I think they would be OK, but can't say for sure without testing...

 

[keywin]

Hi,very happy to know everybody ,i from China,my English too bad.
i also tested the FPD2532 before on month,but it is not good,because the FPD2532 internal resistance too high,most bad is the internal resistance very sensitivity for the temperature,so will squander a great lot battery power,i think use FPD2532 for controller is not feasible.
i hope master advise,thanks.

 

[tron]

I think keywin is right about the fets. I believe E-bikes site had a problem with fets inside a batch of Crystalytes that had a high internal resistance that caused them to fail. Anyone on this, correct me if I'm wrong. If I'm wrong it still makes sense that an efficient system would have low internal resistance.

 

[The7]

One other thought: Since there are always two high side switches on at any given time, against one low side switch, you could double up on the low side FETs only. This will drop the heating. Since the low side FETs switch at a much lower frequency, the effect of slower switching speed will be minimal.

It does seem that the high side FETs are switching at the PWM frequency (say about 17kHz) and the low side FETs are switching at the frequency of the 3-phase output (say about 400 Hz at max speed).

If so, the switching power loss of the high side FETs will be many times higher than that of the low side FETs. However, the resistive (conduction) power loss of the high side FETS is about 1/2 of that low side because there aree two on each of the high side.

The total heat generated by FETs = Switching power loss + Resistive (Conduction) power loss.

 

[lazarus2405]

i think use FPD2532 for controller is not feasible.

Except that Stevo's experiences show that it is feasible. He isn't having serious problems. The controller gets warmer, but it seems manageable. Just add some fins and mount the controller where it can get plenty of airflow.

I'm not saying that it is ideal. 14mOhms on resistance is not great, but it's the only option to run those higher voltages.

 

[fechter]

I think the conduction losses are much greater than the switching losses, especially when you get up to speed, the switching losses are very low since the PWM is 100% on.

With 4110's, I would expect the switching losses to be about the same, but the conduction losses are lower due to lower on resistance. In actual testing, the 4110 controller runs much, much cooler than the stock 4710's

14 mohm sucks. It would be OK for 20 amps. This is the same on resistance as the stock Crystalyte 4710s. If you have enough cooling, you could run higher current, but heating is definitely an issue.

 

[solarbbq2003]

if doubling up fdp fets works, gives an on resistance of about 7mohm comared to 3.5 for 4110, looks pretty good
not worrying about outer case, if you put second set of fets at right angles to originals ( same plane as the board) would be an easy connect and allow for heatsinking pretty easily too ( wouldn't get in the case of course!)

 

[fechter]

if doubling up fdp fets works, gives an on resistance of about 7mohm comared to 3.5 for 4110, looks pretty good
not worrying about outer case, if you put second set of fets at right angles to originals ( same plane as the board) would be an easy connect and allow for heatsinking pretty easily too ( wouldn't get in the case of course!)

Right. If you put enough of them in parallel, you can get the on resistance down to any level you want. You might need to add additional gate drivers beyond a point.

I think you could piggyback an extra FET under the ones there using an aluminum spacer between the heatsink tabs. You would need to raise the board up a few millimeters, and put a spacer on the regulator transistor. I think it would still fit in the case (analog controllers).