AB763 NFB values

[tele_savales]

Say you have a single channel Ab763 w old Twin Reverb transformers, the OT has a 4 ohm tap and you are only running 2 6L6GC's into a single 8 ohm speaker, what should the NFB resistor and tail be as the schematic calls for 820R and 100R?

 

[sds1]

820R and 100R.

The speaker/tube configuration doesn't change the amount of signal at the 4Ω tap.

EDIT: wrong answer!

 

[tele_savales]

Thanks!

 

[NTC]

Umm...
1 watt into a 4 ohm load has a voltage of 2 volts. 1 watt into an 8 ohm load has a voltage of 2.82 v. At the feedback point, the voltage would be Vout (100)/(100 +820). For a 4 ohm output, your feedback voltage is 0.217 V. For an 8 ohm output, the feedback voltage is 0.307 v. Not much of a difference, but in relative terms you have more feedback. To get about the same value feedback voltage, you would have to change the 820 ohm to 1.2k ohms. Best bet is to use a 1k or 5k linear pot in series with the 820 ohms and dial in what you like - or try it with no feedback at all.

 

[sds1]

Yeah my answer doesn't make any sense because the signal voltage is measured across the load. So to say the load doesn't change anything about the NFB cannot possibly be correct.

Sorry @tele_savales for passing along bad info.

 

[NTC]

Yeah my answer doesn't make any sense because the signal voltage is measured across the load. So to say the load doesn't change anything about the NFB cannot possibly be correct.

Sorry @tele_savales for passing along bad info.

Don't be hard on yourself.

Playing Leo's advocate: EVERY AB763 derivative Fender amp in the 60's used an 820 ohm feedback resistor, including the Twin, Deluxe, and Super Reverb. All with different output impedances. Changing this value will affect high frequencies and distortion response slightly if using the value in my previous post.

 

[Ten Over]

Yeah my answer doesn't make any sense because the signal voltage is measured across the load. So to say the load doesn't change anything about the NFB cannot possibly be correct.

Sorry @tele_savales for passing along bad info.

I think your answer was correct. Two things changed: An eight Ohm load on a four Ohm secondary and only two output tubes. The eight Ohm load doubles the primary impedance which would double the open loop gain, but half the output tubes cuts the open loop gain in half. They wash and the open loop gain is unchanged so that the original NFB resistors are still appropriate.

 

[tele_savales]

I've been listening back to some recordings I made of the amp a few weeks ago playing a Strat w Mojotone noiseless pickups and I noticed sort of a harmonic buzz on on the bottom strings. Didnt seem to notice it as much playing my guitar w p90's thru it but it's still there. While at work I was thinking, maybe it's a nfb issue, maybe the speaker or baffle is buzzing, etc. I have switch on the back giving me 820, 1.5k, and 530 and it was in the center position during these recordings.

 

[tele_savales]

Thanks for the input, everyone

 

[tele_savales]

Also the amp has a cross line master volume which I think might be part of the weirdness. In general the tone is superb as you'd expect for an AB763 with a really good Weber speaker, except for like the E string on my strat and esp. on the neck pickup.

 

[Ten Over]

Also the amp has a cross line master volume which I think might be part of the weirdness. In general the tone is superb as you'd expect for an AB763 with a really good Weber speaker, except for like the E string on my strat and esp. on the neck pickup.

A cross line MV messes with the open loop gain, so it also messes with the NFB.

 

[sds1]

I think your answer was correct. Two things changed: An eight Ohm load on a four Ohm secondary and only two output tubes. The eight Ohm load doubles the primary impedance which would double the open loop gain, but half the output tubes cuts the open loop gain in half. They wash and the open loop gain is unchanged so that the original NFB resistors are still appropriate.

It's too late I already crossed it out.

I knew there was a "wash" going on there, but today thinking it was only relevant to the reflected primary.

I can't speak intelligently about open loop gain and how that factors in here, so if I happened to be correct in my earlier it was not due to me being well-informed.

Lucky guess?

 

[Ten Over]

Umm...
1 watt into a 4 ohm load has a voltage of 2 volts. 1 watt into an 8 ohm load has a voltage of 2.82 v. At the feedback point, the voltage would be Vout (100)/(100 +820). For a 4 ohm output, your feedback voltage is 0.217 V. For an 8 ohm output, the feedback voltage is 0.307 v. Not much of a difference, but in relative terms you have more feedback. To get about the same value feedback voltage, you would have to change the 820 ohm to 1.2k ohms. Best bet is to use a 1k or 5k linear pot in series with the 820 ohms and dial in what you like - or try it with no feedback at all.

He doesn't have an 8 Ohm tap. He has a 4 Ohm tap with an 8 Ohm load.

 

[Ten Over]

It's too late I already crossed it out.

I knew there was a "wash" going on there, but today thinking it was only relevant to the reflected primary.

I can't speak intelligently about open loop gain and how that factors in here, so if I happened to be correct in my earlier it was not due to me being well-informed.

Lucky guess?

This might help:


The voltage gain of the LTPI doesn't change. The gain of the output transformer doesn't change. The impedance of the OT primary does change and the number of power tubes does change. The voltage gain of the power tubes depends on the number of tubes, the transconductance, and the primary impedance. The number of tubes gets divided by two and the primary impedance gets multiplied by two, so the voltage gain of the power tubes does not change. Since all three factors in the open loop gain are unchanged, the open loop gain is unchanged.

 

[NTC]

This might help:
View attachment 1067709

The voltage gain of the LTPI doesn't change. The gain of the output transformer doesn't change. The impedance of the OT primary does change and the number of power tubes does change. The voltage gain of the power tubes depends on the number of tubes, the transconductance, and the primary impedance. The number of tubes gets divided by two and the primary impedance gets multiplied by two, so the voltage gain of the power tubes does not change. Since all three factors in the open loop gain are unchanged, the open loop gain is unchanged.

^ This is the BEST explanation. My math is correct - but does not take into account open loop gain or the resulting closed loop gain.

The long-tailed pi and output stage form what is basically and opamp with a low open loop gain.

 

[King Fan]

Amazing NFB analytics, you guys. I've been thinking about this, though:

I noticed sort of a harmonic buzz on on the bottom strings. Didnt seem to notice it as much playing my guitar w p90's thru it but it's still there. While at work I was thinking, maybe it's a nfb issue, maybe the speaker or baffle is buzzing, etc. I have switch on the back giving me 820, 1.5k, and 530 and it was in the center position during these recordings.

Heh, I'm not smart enough to consider NFB when I get a sympathetic / harmonic buzz (and yeah, it's usually on the low strings, of course). Part of me worries I've torn a cone or dislodged something in the voice coil gap; the rest of me worries it's not the speaker and I'll spend hours tapping and thumping and tightening and swapping cabs, etc. etc. (On a really good day, it's a rattling picture frame or speaker cable or something else extrinsic; I've seen a list of dozens of causes, but sadly most of them are in the amp).

So @tele_savales , does the buzz resolve when you change the NFB to 820? Am I right in thinking that even if NFB is involved, we're talking different amounts of bass energy, so that still means it could be any harmonic/sympathetic vibration driven by those frequencies? Not challenging your thinking, just trying to improve mine...

 

[tele_savales]

I haven't worked on it yet. Have a old Bassman on the bench at the moment. Part of the problem is I can't really run an amp thru it's paces in the apartment. But I will check the speaker and the cab- it looked like it had taken a fall down a flight of stairs when I got it and there was damage to the cab, and I pulled the particle board baffle and replaced it w a piece of what looked to be decent plywood and cut it for 1 speaker. I epoxied and nailed the new baffle in and added 1x2's here and there to stabilize it and the whole thing seems bombproof but there's definitely a chance there's some kind of sympathetic vibration.
Next day off is Wednesday, and the wife is at work so I'll dig in and update at that point.

 

[tele_savales]

This might help:
View attachment 1067709

The voltage gain of the LTPI doesn't change. The gain of the output transformer doesn't change. The impedance of the OT primary does change and the number of power tubes does change. The voltage gain of the power tubes depends on the number of tubes, the transconductance, and the primary impedance. The number of tubes gets divided by two and the primary impedance gets multiplied by two, so the voltage gain of the power tubes does not change. Since all three factors in the open loop gain are unchanged, the open loop gain is unchanged.

So standard nfb values unless my personal taste dictates otherwise?

 

[Ten Over]

So standard nfb values unless my personal taste dictates otherwise?

Yes.

I tried pulling two tubes and one speaker in a Twin Reverb years ago. The open loop gain was nearly identical to the stock setup. The closed loop gain was also nearly identical using the stock 820R/100R NFB resistors. So the amount of gain reduction from NFB is nearly identical to the stock setup.

 

[Badside]

Say you have a single channel Ab763 w old Twin Reverb transformers, the OT has a 4 ohm tap and you are only running 2 6L6GC's into a single 8 ohm speaker, what should the NFB resistor and tail be as the schematic calls for 820R and 100R?

The real answer is... The one that sounds best to you