Do I have to match voltage Gate-Source of mosfet in SMPS?

[Charoenchai]

I plan to build SMPS based on SG2325 and intend to use 8 IRFZ44 (4 per rail). Do I have to match voltage Gate-Source of thease mosfets for proper current sharing?

[1moreamp]

Originally posted by Charoenchai
I plan to build SMPS based on SG2325 and intend to use 8 IRFZ44 (4 per rail). Do I have to match voltage Gate-Source of thease mosfets for proper current sharing?

Good question.
I have not seen much work in that area. You can gain match if you like. I have seen Rockford Fosgate amps use 1 FET per conductor on there primary. This sort of bypass's the issue of gain matching the FETs for current sharing, as each FET has its own single load to operate into. PG does this also on there MS-2250 and MS-1000 (a dual 2250).

I have not seen much else ever done concerning the gain differences in each FET.
I believe this is mostly due to the fact that nonlinearity in not a real issue in a single frequency variable pulse width operated load. I know that I have seen very little matching effort in power supplies.
Of course car amps are built pretty cheap nowadays

[Charoenchai]

It should be SG3525 and not SG2325. Sorry for that.

[space]

What is probably more important than gain matching is RDSon matching. Differences in RDSon may lead to core saturation.

space

[Charoenchai]

Thanks for all comments. After searching for MosFets I have found IRFP2907( Id=209A, RDSon=4.5mOHM, Pd=470W),TO-247 package, which is very interesting. Can I use just 2 of this, only 1 per rail, to make sure that no current variation in each smaller MosFets. Poweramp section would be 4 channels of LM3886.

[1moreamp]

Originally posted by Charoenchai
Thanks for all comments. After searching for MosFets I have found IRFP2907( Id=209A, RDSon=4.5mOHM, Pd=470W),TO-247 package, which is very interesting. Can I use just 2 of this, only 1 per rail, to make sure that no current variation in each smaller MosFets. Poweramp section would be 4 channels of LM3886.

Your gate drivers won't accomadate such a change, you must re-engineer the drive circuitry for that large of a fet. the gate capacitance is way different and the turn on/ turn off will be way in left field. Probably just won't work from the get go.

Try replacing your Z-44 fets with IRFZ-48's they will lend higher reliability at no extra cost in circuitry change.

Your trying to build a very precise power supply for a rather rudamentry amplifier. I would go with the Z-48's and live with the improvement that generates they are rated at about 20 amps more power each, and in a array of 4 thats an extra 80 amps of switch headroom. your still dealing with the original toroid coil spec's so there is where your limited. Not the Fets...