Hiwatt dr103 output impedance

[SteveRube]

Hi,
As the title suggests, I’m trying to determine the output impedance of the 8ohm tap on the dr103. I don’t have an amp to measure so I’m just going off the schematic online. Hiwatts seem to have a lot of nfb (10k/2.2k basic with an extra shunt to get rid of some of the treble feedback). Does this reduce the output impedance as compared to say a Fender?
cheers
Steve

[Helmholtz]

Output impedance can be calculated from schematic values, but it's quite involved and requires a lot more data than just the NFB divider ratio.
Actual "amount" of NFB depends on divider attenuation times open loop gain, which varies between amps.

To calculate output impedance you need to know or calculate no-load gain and open loop impedance first.
This requires quite a bit of data (like power tube rp and gm, OT impedance ratio, PI forward and NFB gain, NFB divider and maybe PI tail input impedance).

I once did the full calc for 3 different amps (JTM45, JTM50 and Fender Super Reverb).
Took me quite some time.

There's an article about NFB and output impedance on Aiken's page. It's a bit unclear at times.
I used different literature (mainly Zollner and Kuehnel).

[The Dude]

I would also think that NFB cutting some extra high frequency wouldn't effect impedance except at those frequencies that are attenuated, since impedance is frequency dependent.

[Helmholtz]

I would also think that NFB cutting some extra high frequency wouldn't effect impedance except at those frequencies that are attenuated, since impedance is frequency dependent.

Well frequency dependent NFB makes output impedance frequency dependent.

[Helmholtz]

Schematic would be useful.

[The Dude]

Well frequency dependent NFB makes output impedance frequency dependent.

That was precisely my point,.....and that it should not have much effect on rated impedance at frequencies not "adjusted" by NFB. In other words and for example, you have an OT rated at 8 ohms. You have an NFB circuit that cuts (for example) everything above 5kHz. The only impedance change would be to those frequencies above 5kHz. Frequencies below should still be roughly 8 ohms. And, of course, this is only an example. It's more likely to be a slope than a hard cut.

[Helmholtz]

That was precisely my point,.....and that it should not have much effect on rated impedance at frequencies not "adjusted" by NFB. In other words and for example, you have an OT rated at 8 ohms. You have an NFB circuit that cuts (for example) everything above 5kHz. The only impedance change would be to those frequencies above 5kHz. Frequencies below should still be roughly 8 ohms. And, of course, this is only an example. It's more likely to be a slope than a hard cut.

I think you misunderstood the OP.
Rated output impedance of the amp is known.
But actual output impedance (or source impedance, which affects speaker damping) at the 8 Ohm output depends on amp circuitry and can vary from much lower (with lots of NFB) to much higher than 8 Ohm without NFB.
Amps without NFB can have a source impedance greater than 100 Ohm at the 8 Ohm tap.
Result is little to no speaker damping causing a boost at the bass resonance as well as at high frequencies.

[The Dude]

I saw it as a 2 part question.

1) The question you just addressed.
2) The question of whether HF filtering in the NFB has further effect on impedance.

I was only addressing the 2nd part of the question. But onward, I'm certainly not up for starting a pissing match here.

[Helmholtz]

Rated impedance is fixed by definition, just as rated mains voltage.

[SteveRube]

Thanks for all the replies. I was asked for a schematic, I’m using this one for preamp
https://hiwatt.org/Schematics/DR_Pre4Input_v1a.pdf
and this for pa
https://hiwatt.org/Schematics/DR_100wOutput.pdf
am I right in thinking that a cut to the feedback at high frequencies (10n in series with 470) will increase the output impedance at higher frequencies, but which is in series with an inductive load already much higher than 8ohms (probably above 30ohm reactance) at say 10khz?
Steve

[Helmholtz]

am I right in thinking that a cut to the feedback at high frequencies (10n in series with 470) will increase the output impedance at higher frequencies, but which is in series with an inductive load already much higher than 8ohms (probably above 30ohm reactance) at say 10khz?
Steve

Yes, lower NFB at higher frequencies increases output impedance (and even more so gain) at high frequencies.
Also speaker impedance rises to around 30 Ohm at 10kHz.

Amp output impedance and speaker impedance form an AC voltage divider.

The 10n+470R to ground is part of the tricky presence circuit of this amp, which employs both positive and negative feedback.
It is essential that negative feedback dominates at all frequencies.

[SteveRube]

which employs both positive and negative feedback.

I had no idea there was positive feedback. Could you explain that a little more? I had always assumed that the presence control was essentially like an eq, heavily loading the stage before the cathode follower. It’s only when the presence control gets up to its last 10% or so that it starts shunting the treble nfb.

[Helmholtz]

I had no idea there was positive feedback. Could you explain that a little more? I had always assumed that the presence control was essentially like an eq, heavily loading the stage before the cathode follower.

Main NFB path is NFB signal after divider (10k/ 2.2k for low frequencies) via 100n coupling cap into right side PI grid.
The NFB signal is in phase with the PI input and the right side PI grid is the inverting input of the power amp.
Via the presence circuit a variable part of the higher frequency content of the NFB signal is fed to the grid of the CF.before the PI.
As a CF in non-inverting, the NFB signal which gets to the PI input is in phase with the normal signal causing positive NFB.
There's also a low pass loading effect on the plate signal before the CF by the presence circuit.
But to achieve this, it wouldn't have been necessary to connect to the NFB signal.