I have a prospective project where I'd like to make an optimal choice of complementary P & N channel power mosfets, probably @ 200V breakdown with 2 pair for each up to 200W/Channel 8 ohm in the 16 to 24 A range in a TO-247AC type package plus I'm hoping for the best short term SOA window without the Crss being too nasty? I'm planning to drive these two pair with what amounts to a class A push pull 12AU7 & considering bootstrapping. Would 150 volt MOSFETs be cutting it too fine for a +/- 65 volt supply?
So far, N channel parts by IRC & P channels by IXYS are the most suitable I've come across.
Also, I was thinking of making a temperature tracking bias network for the MOSFETs with a NPN transistor mounted to the output device heatsink as a 'zener' with an adjustable band gap shunt reference in series with the emitter to be able to trim the bias out over a temperature range. Comments on this?
TIA -
So far, N channel parts by IRC & P channels by IXYS are the most suitable I've come across.
Also, I was thinking of making a temperature tracking bias network for the MOSFETs with a NPN transistor mounted to the output device heatsink as a 'zener' with an adjustable band gap shunt reference in series with the emitter to be able to trim the bias out over a temperature range. Comments on this?
TIA -
For power 200 W into 8 Ohms you will need minimal two ( rather three ) fets with dissipation 150 - 200 W each and Uds 200 V. Each of this fets have input capacity approximately minimaly 2500 pF, so summary input capacity will be very high. ECC 82 ( 12 AU 7 ) have Ri 7,7 kOhm, which is too high for this using - you get amp, which will be " very lazy " and which will have big distortion on higher frequecys. If you can use tube VAS, go on power ones, but don't wait any miracles 
.
.Hi, Upupa -
That's a good point, which is one reason that I'm giving primary consideration to only two pair instead of three, although if I find a complementary pair that has particularly low Crss at a lower maximum drain current, I can always change my mind.
Actually, my calculations indicate that I should be just about able to avoid slew limiting out to 20khz since I plan to operate the output devices as followers with an expected average Crss of about 400 - 500pf per each of four power mosfets. I also had the experience, a while back, of building a somewhat similar amp using a long tailed pair 12AT7 driver with only pullup resistors to drive the gates (5mA Ip/each side), and the thing slewed at 15V/uS at the zero crossing, which was significantly better than I expected. You might be wondering how drove power mosfets with a setup like that, but I did, and both halves of the totem pole were N-Channel, at that. (It's my very own driver/output topology and probably patentable, so I don't want to be too specific about how I made it work.)
But, why go with a 12AU7 totem pole driver, you might ask? Because if that works out I think I can do the input *and* driver of a 200 watt amp with a single compactron tube. Kewl.
That's a good point, which is one reason that I'm giving primary consideration to only two pair instead of three, although if I find a complementary pair that has particularly low Crss at a lower maximum drain current, I can always change my mind.
Actually, my calculations indicate that I should be just about able to avoid slew limiting out to 20khz since I plan to operate the output devices as followers with an expected average Crss of about 400 - 500pf per each of four power mosfets. I also had the experience, a while back, of building a somewhat similar amp using a long tailed pair 12AT7 driver with only pullup resistors to drive the gates (5mA Ip/each side), and the thing slewed at 15V/uS at the zero crossing, which was significantly better than I expected. You might be wondering how drove power mosfets with a setup like that, but I did, and both halves of the totem pole were N-Channel, at that. (It's my very own driver/output topology and probably patentable, so I don't want to be too specific about how I made it work.)
But, why go with a 12AU7 totem pole driver, you might ask? Because if that works out I think I can do the input *and* driver of a 200 watt amp with a single compactron tube. Kewl.
Slew limiting
It is not easy find fets with low input capacity and great Pd. Nearest are lateral types of BUZ 900 family or " historical " Hitachi's or their new " clones " in TO 247 case. If you can limit frequency characteristic for 20 kHz, 5 mA will be enough, but distortion will be very rising at this frequency ( look at calculation from Erno Borbelly in The Audio Amateur ). My recomedation is to use any " predriver " beside tube ( some lateral type such as 2 SK 213 ), which have input capacity much lower and becouse they are lateral, have better temperature coeficient of Ids.
It is not easy find fets with low input capacity and great Pd. Nearest are lateral types of BUZ 900 family or " historical " Hitachi's or their new " clones " in TO 247 case. If you can limit frequency characteristic for 20 kHz, 5 mA will be enough, but distortion will be very rising at this frequency ( look at calculation from Erno Borbelly in The Audio Amateur ). My recomedation is to use any " predriver " beside tube ( some lateral type such as 2 SK 213 ), which have input capacity much lower and becouse they are lateral, have better temperature coeficient of Ids.
- Status
- Not open for further replies.