DIY amplifier: OPA 1611 + Bipolar transistor

Hello,
I am not an electronic engineer and my english is poor.
Sorry for my mistakes.

I work on this amplifier for a few month with MicroCap 12.
Here is the schematic, distortion with 1Watt output and phase diagram.

What is your opinion and your comments?

The power supply will be a SMPS from HYPEX + - 64V
 

Attachments

  • schema.jpg
    schema.jpg
    191.9 KB · Views: 679
  • Distortion-1W.jpg
    Distortion-1W.jpg
    250.1 KB · Views: 575
  • Phase.jpg
    Phase.jpg
    340.9 KB · Views: 536
I want a quiescent current of 100 mA by adjusting X11.
I know that the power dissipation will be around 100W, each way, without signal.
I have two big aluminium dissipators (9kg each).
Do you see a possible problem?
 
Yeah, I’ve always had the best result if the predrivers were run cooler with less temperature swing than the drivers/outputs. And for best thermal tracking and output transistor power sharing, only ONE driver transistor, not multiples. Consider MJW0281/0302.
 
Hello WG_SKY
If I understand, you perfer one pair of MJW0281/0302 to drive the 2 x pairs of MJL4281/4302 instead of two pairs of MJE15030/15031 ?
And the reason is to be sure that all the power transistors are drived by the same signal !
Two pairs of transistors, even paired, do not give the same result ?
 
The two separate 1503x may not run at exactly the same *temperature*. Then there will be a mismatch in bias between the groups of outputs. It’s already hard enough keeping a long bank of the same outputs close enough to each other in temperature, and TO-220’s are notoriously difficult to heat sink correctly. You don’t want these imbalances growing and potentially getting out of control.
 
AFAIR: With a bipolar emitter-follower output stage the optimal bias is one thermal voltage (Vt) across the emitter resistors.

Vt = kT/q

k = 1.38E-23 J/K (Bolzman's constant)
T = absolute temperature in K (so 300 K = 27 ºC)
q = 1.602E-19 C (elementary charge)

So you're looking for about 26 mV at room temp. I can't for the life of me remember if it's Vt across each emitter resistor of if it's Vt across both emitter resistors. I think it's across both.

Douglas Self writes quite a bit about it in his Audio Power Amplifier Design book.

The design reminds me of an old Linear Tech (or ADI) app note.

Tom
 
Last edited:
I Have a lot of MJL4281/4302.
Can I use it for Driver, instead of MJE15030/31 ?

With MC12 simulator, one pair of MJL4281/4302 is better than two pairs of MJE15030/31.

Yes, you can use the same driver as the outputs. I do it all the time with MJL21194’s. I only suggested the 0281 because it is “big enough” to replace a pair of 15030.

Don’t believe everything a simulator tells you, though. I doubt you would be able to tell a difference in he sound. It *will* give you less trouble in and actual amplifier that doesn’t have all perfectly matched transistors, and are all mounted to ever so slightly different thermal environments.
 
So you're looking for about 26 mV at room temp. I can't for the life of me remember if it's Vt across each emitter resistor of if it's Vt across both emitter resistors. I think it's across both.

Douglas Self writes quite a bit about it in his Audio Power Amplifier Design book.

The design reminds me of an old Linear Tech (or ADI) app note.

Tom

The app note is attached in Post #3.

Most of the time, with amps this big you end up running them under biased anyway. 26 mV across a tenth or two gets you a *lot* of heat with than many output transistors. Big *** EF3’s tend to get very low crossover distortion even with less than optimum bias. If you do plan plan on running optimum bias, plan accordingly for the heat. If you plan on sizing the heat sink for sustained high power operation it will be big enough. If what was planned was for “normal” listening levels only - optimum bias at +/-64V supplies, with 8 output pairs per channel, it might surprise you how much you’re dissipating at idle.