Hi, is it good practice to use " output " power transistors ( like 2sc5200 ) as pre drivers in high power amplifiers ?.
Would the high capacitance be a problem ? compared to " normal " drivers like MJE340 , BD140, etc. ?
How about enchanted VAS ( darlington ) with high power transistors ?.
- Bruno.
Would the high capacitance be a problem ? compared to " normal " drivers like MJE340 , BD140, etc. ?
How about enchanted VAS ( darlington ) with high power transistors ?.
- Bruno.
Nothing stops you from doing that, but I think you'll find that the smaller, faster VAS and driver devices make it easier to get good loop stability. You'll likely also find that their (usually) higher gain will give you better overall performance. In a CFP output stage were the drivers can sit on their own, smaller heat sink, you could potentially get better thermal stability with the physically smaller devices too.
Tom
Tom
In triple power amps like the PV-1.3k, a TO3 MJ15020 drove 5 parallel MJ15024. the MJ15020 had higher Ft than the MJ15024 - probably a key design feature to prevent cross conduction. 20 mhz vs 4 . The MJ15020 were driven by TO220 2sc 2sa something predrivers, later changed to MJE15031/32 .
Faster drivers in EF3 are preferred, but not absolutely required. The predrivers (or drivers in EF2) have to be fast. One can always use the smaller version like 0281 to drive a bank of 3281 or 21194. I’ve used 21194 to drive multiple 21194 at crazy power levels run at war volume, with fast TO-220’s for predrivers of course. Will it like being driven with a full power square wave at 200 KHz? Probably not but that’s simply not the use case. C5242’s do very well driving multiple C5200’s too, so fast driving fast works as well. Maybe even better, since the non dominant poles are all higher in frequency and further out of your way. People get the mistaken idea that too much speed or total gain in the output stage causes oscillations, but my experience says this is not the case. It can make it more sensitive to things like proper bypassing and layout (or lack thereof), but generally the global loop gets more stable when the outputs don’t eat into the phase margin.
Ive had to resort to heavy duty VAS’s before too, but run them in darlington. Now there is no need to do this at all at sane power/voltage levels. One simply does not need .0005% distortion figures on a 2000 watt amplifier, and premium devices capable of better are perfectly suited to 70, 80 volt rails.
Ive had to resort to heavy duty VAS’s before too, but run them in darlington. Now there is no need to do this at all at sane power/voltage levels. One simply does not need .0005% distortion figures on a 2000 watt amplifier, and premium devices capable of better are perfectly suited to 70, 80 volt rails.
I'd also question the sanity of using a $5 device when a $0.25 one would do just fine.
I agree that <1 ppm THD may not be needed, but on the other hand, why leave performance on the table?
Tom
I agree that <1 ppm THD may not be needed, but on the other hand, why leave performance on the table?
Tom
The question was more about curiosity, there are plenty of fast 20- 50W transistors that will do fine for drivers at even 1K output I think. Even the MJE15031/32 ,
@wg_ski Looking to build a 500W amp for a subwoofer. ( sound quality > efficiency ).
I don't know if I should get a cheap kit with IRS2092 picture down bellow, and change the output mosfets with something more powerfull , same'ish or lower Rds. maybe even the ic itself to be sure it's not " fake" .
Or build something from scratch with the IRS2092 , maybe some gate driver ic's from ir as well , A few Pioneer car amps with the ic I have found schematics.
or build a class AB with 4x or even 5/6 power transistors per rail ( maybe MJL3281A / MJL1302A ) .
needs to be 2 ohm stable .
I don't know if I should get a cheap kit with IRS2092 picture down bellow, and change the output mosfets with something more powerfull , same'ish or lower Rds. maybe even the ic itself to be sure it's not " fake" .
Or build something from scratch with the IRS2092 , maybe some gate driver ic's from ir as well , A few Pioneer car amps with the ic I have found schematics.
or build a class AB with 4x or even 5/6 power transistors per rail ( maybe MJL3281A / MJL1302A ) .
needs to be 2 ohm stable .
Attachments
500 W into what impedance? Most car amps are rated for 1 Ω or 2 Ω load to inflate the power numbers.
It's hard to beat Purifi on performance. Just saying... Assuming you're into Class D.
Tom
It's hard to beat Purifi on performance. Just saying... Assuming you're into Class D.
Tom
Other options would be tpa3255 or TAS5630 .
or maybe better / newer class D controllers like IRS2092.
I wouldn't mind going class AB , 500W isn't a lot of power.
or maybe better / newer class D controllers like IRS2092.
I wouldn't mind going class AB , 500W isn't a lot of power.
500 W into 2 Ω would require an output current of 15.8 A, RMS. There's no way those tiny screw terminals would survive that. Maybe for a few cycles at 20 Hz but not long-term.
Tom
@Nelson Pass These would be the advantages of using higher power transistors as drivers. Nowdays we have fast output transistors fT of 50 mhz or so , then saying using faster smaller transistors as drivers wouldn't be the case anymore . DIfference would be hFE and What about Capacitance of the transistors ? High power output transistors have 100 - 200p or so compared to a few pico or ten's of pico of smaller transistors
Changing the output mosfet's and heatsink of course. there are cheap high voltage , high current low Rds mosfets , I can " play " with the dead time on the IRS if I change the output transistors, right ?.
So is your question about output devices specific to this Class D module or are you asking more generically? I think since you posted in the Solid State forum most of us assumed that you were talking about Class A or AB amps, hence the comments about VAS, etc.
I'd be careful with random device swaps in a Class D amp. I've never designed one, so I'm not the best to ask about that, though. You'd probably get better help in the Class D forum if that's your intended application.
Tom
I'd be careful with random device swaps in a Class D amp. I've never designed one, so I'm not the best to ask about that, though. You'd probably get better help in the Class D forum if that's your intended application.
Tom
Yes the question was about transistors in Class AB , if using output power transistors as drivers / vas.
.
My mistake ! I did not make another post , and asked here for a 500W subwoofer amp if I should go class AB or class D, and the modifications to that class D module.
Yeah, I wouldn’t go swapping anything around on a cheap class D (or even AB) kit. Bigger/more hexfets have more capacitance to drive. That needs to be accounted for in the driver stage. Just as easy to start over. Your basic IRS2092 driving complementary drivers and three pair of IRFB4227 is already doing better than a lot of “pro” amplifiers these days and will easily do the deed. Especially if you put an actual heat sink on it, instead of a board full of SMDs and a fan blowing across the bare board.
500W into 2 ohms isn’t crazy for AB, although I might do H just to keep it cooler and because I can.
500W into 2 ohms isn’t crazy for AB, although I might do H just to keep it cooler and because I can.
500 into 2 class AB just takes a +/-58V real-world supply. Its just not that hard to do. Four pairs of ordinary audio outputs, driven by a 5th, and whatever TO-220 or TO-126 “driver“ pair that hasn’t been discontinued this week and is actually in stock. For a subwoofer you could build the whole output stage out of 10 2N3773’s, and MJE340 and a 350. Sure, you could use high speed flatpacks if they are easier to work with or easier to get. Standard 600VA 42-0-42 would take any normal duty cycle usage. Slap a blameless front end on it and it’s good to go. Or something more sophisticated if it floats your boat - that part doesn’t cost any real money anyway.
Thr ft of an output transistor will be 2 to 5 mhz at the currents used for biasing. The Hfe will also be lower there.