P-Channel high side gate drive for 150Vbus BLDC motor drive

Apr 16, 2007 4 Replies

Gents,



I've seen several app notes showing using complementery N/P channel fets to drive a brushless DC motor with a low bus voltage (say 12 - 80VDC).



I have an application that has a 150Vbus. Is there any reason I could not use a 250 - 400 watt P-channel with appropriate currnet rating?



Id the appnote stop at 80 volts because the P-channel devices of the time were wimpy and not desired?



I know I could use all n-channels but the floating gate drive is getting to be problematic due to the long duty cycles. It can be done and I'm investigating a small footprint solution.



My problem is that the application is at very high temperature (>150C) and the standard "easy" answers don't work. (IR2110 or optocouplers). The next step is gate drive driving a carrier through the transformer and demodulating it on the other side.



Seems like if it could work, a P-channel high side drive would be a better choice.



What are your opinions on complememtary motor drive for a 150Vbus brushless motor driver.


I'm sure you mean volt, not watt.

A quick look at Farnell's site (which isn't the best-stocked, but has all the jellybean stuff) shows a single 250V P-Channel MOSFET, and that's a TO92 device. After that they havbe only 100V parts.

I don't know what makes it so hard to make decent P-channel power MOSFETs. Maybe it would be possible today, but there's no market because everybody has settled for N-channel only with floating high-side drivers.

Go bipolar? Oh... just checked. Not much luck there either.

My hunch is that even if there are high-power, high-voltage PMOS or PNP devices out there, they might be single-sourced and expensive. Personally, I'd definetely opt for all-NMOS and somehow wiggle my way around the other problems.

robert

High power, high voltage N-channel fets are cheaper and more available.

Drive is easy: use a cheap potted dc-dc converter to get floating power, then an optoisolator and a chunky gate driver; maybe $12 for the whole thing.

John

This approach is a bit more complex, but popular and very inexpensive.

There are some reasonably-good 400 to 500V p-channel power MOSFETs, but as you say, they aren't as good as n-channel (e.g. a poorer Rds-on * price, etc., figure-of-merit). But if the O.P. feels the need... He may think p-channel-drive circuits are more simple, but after he's added the ability to drive high currents into the p-MOSFET gates, I dunno.

Yup, I'm still using lots of the 300-watt Ixys parts you suggested. They're on the heatsink beast whose photo I posted to abse last week.

But it seems to me that there wouldn't be much difference in gate drive complexity. You wouldn't want to capacitively couple directly into a p-channel gate, not with the source that far off ground. I guess if you had multiple p-fets working off the + rail, h-bridges maybe, you could share the dc-dc converter.

Oh, did you ever used the drift step-recovery gadgets I sent? The project is dead on my end (ie, all the lawsuits are pretty much settled) so I don't consider the stuff to be secret any more.

John

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required