Is a Class D amp appropriate if you have an external feedback loop?

[Circuitsoft]

So, a year or so ago, I got this idea about trying to build a high-power tube-sound instrument amp by driving a low-power tube amp into a transistor amp, then feeding the output current back through to the tube amp, like in the attached schematic.

Would a Class-D amplifier have the response necessary to make this work, or do I need to go for Class-AB or -B?

Thanks,
- Alex

[Pano]

Well you're not the only one to have this idea, it gets kicked around a lot.
And it does actually work, after a fashion. You can get some of the tube amp sound by using it as what is essentially a line stage or preamp. You may not like the results, but that's a different story. You might even end up with too much tube flavor.

I'm not quite sure why you would want the the feedback loop, but maybe someone else here could comment.

[Circuitsoft]

Perhaps I should elaborate more. The purpose of the external feedback loop is to give the tube amp the same load it would have if it were driving speakers. By drawing a proportional amount of current from the tube amp that is being drawn from the transistor amp, I simulate the load of the speakers exactly to the tube amp.

[Pano]

Yeah, OK.

But really you just need a load resistor for the tube amp. Use whatever, 4, 8, 16 ohms. The load and the taps you use will determine the sound of the amp, the harmonic structure. Play around with different loads to find the sound you like. You'll only be driving about 2-3 volts out of the tube amp, so no worries.

[Circuitsoft]

A speaker is not just a resistor. The impedance of a speaker system changes drastically with frequency, as does the effective output impedance of the tube amp. The characteristic sound of a tube amp is partially defined by the interaction between the speaker cabinet, output transformer, and whatever feedback there may be in the amp.

Being that a speaker is a reactive device, only an active feedback such as this will properly pass on the load.

So, the root question of this thread is: Will the output voltage of a Class-D amp track the input voltage fast enough for a loop such as this to effectively represent the impedance and phase of the load?

[CLS]

Hey, I'm thinking of this, too.

Not for the 'tube' sound, but trying to tune the system Q with the much higher than normal output impedance, which is very useful for low Q woofers in OB.

I'm not familiar with class D (T) circuits With some shallow studies on the datasheets and manuals, it seems the half bridged (single ended) design is much easier to mod. Now I'm playing with Amp 6 which is full bridged thus floated outputs. So a differetial amp is essential to isolate both ends (of feedback loop). I'd guess UcD amps are much easier to play with -- single ended (grounded) output, and a negative input, how convenient!

Another idea, is it possible to use a line level transformer to get the differential signal (a part of the output current) and feedback to the input (which is marked negative in Amp 6)? Primary is connected to both ends of the series resistor at output, and secondary is connected to the input (via a resistor) and the bias pin. Will this work?

[Circuitsoft]

I believe that's done, in reverse, on some hifi amps to compensate for resistance in the speaker wires.

That's different from what I'm doing, though. You want the power amp fed-back from a virtual point as if there were a resistor in series with the speaker. I think I explained my intent well in the previous post. If not, ask further questions and I'll try to explain better.

[Circuitsoft]

I believe that's done, in reverse, on some hifi amps to compensate for resistance in the speaker wires.

For isolating output current measurements from input, just use a balanced-input opamp circuit. That's what U1A and R2-5 do in my circuit above. Ignore R1, and the output of U1A will be the voltage across R6 using the other end of R2 as a reference.

That's different from what I'm doing, though. You want the power amp fed-back from a virtual point as if there were a resistor in series with the speaker. I think I explained my intent well in the previous post. If not, ask further questions and I'll try to explain better.

[Pano]

Originally Posted by Circuitsoft

A speaker is not just a resistor. The impedance of a speaker system changes drastically with frequency,

Yeah, OK. But IMO you are way over thinking this. Sure, you could come up with some active or passive network to imitate the load or sound of the speaker connected to the amp, but why? Will it really be better? I doubt it.

A simple load on the amp and some EQ if you want it should get you 99% of the way there. Why not at lest start out simple to hear if you like it or not?

[raintalk]

Use this to dummy load the tube amp
Speaker load
Take the output off of R1 through to feed the output amp.
As mentioned, there are already commercial amps that do just this.

Better yet, put a speaker in an isolation box with a mic in it and feed the mic to the class-d through a mic preamp. You really need crappy speakers to old mics to make this sound authentic.