Negative Feedback - an epiphany

[alerich]

I never realized what an impact negative feedback can have on a high gain amp until today. I installed a Soldano style depth mod in my Mig60. Test it out and as soon as I increase the depth control the amp got *way* louder. I mean with the depth control at minimum it sounded just like it always does and by 9:00 on the pot the amp was just roaring loud. Now I understand. The depth mod (just a 1MA pot withe a .0047 cap across it) is inserted between the NFB resistor and the OT tap. The NFB resistor on the Mig60 is 22K - fairly small for this style of circuit. My Marshall 2204 uses a 100K while my Soldano Hot Rod 50 uses 39K. I think this is why for the same relative setting on the 1M master volume control on each amp the Marshall is loudest followed by the Soldano and then the Mig60. I have always wondered why the Marshall is so much more freaking loud than the other two. The circuits are very similar otherwise and they are all 50 watt amps. I think this may be a large part of the reason.

One thing I don't get is that the Soldano with a 1M depth pot and a 39K NFB resistor doesn't exhibit the big volume boost when I turn the depth pot. It simply increases the low end as it is designed to do.The depth controls are wired in identical fashion. I think I'm gonna just dump the depth mod on the Mig60 and raise the value of the NFB resistor a little.

I know this is pretty basic stuff but it was a "light bulb" moment for me - might be for someone else so I thought I would share. I have disconnected the NFB on my Vibro Champ to get a little more out of the amp but I never considered the impact of NFB on a more powerful amp until today.

[PRR]

It's not the one resistor value. There's two resistors and the *ratio* is the main thing.

[alerich]

Yes and no. There are now two resistors (the existing fixed 22K NFB resistor and the variable 1M depth pot) but they are in series from the OT tap to the connection at the PI. It's not really a voltage divider. To the PI it's just one increasingly larger resistance. Part of it is bypassed by the .0047uf cap but that's that only place where the ratio of the two comes into play, as far as I can tell.

I dumped the depth mod. It just made the amp go insanely loud as you turned it up. The amp has plenty of bass, especially now that I increased the value of the NFB resistor from 22K to 47K. There is an marked difference in volume and grunt now. The amp was never anemic before but it was just never quite as loud as I thought it should be for a EL34 50W amp. I think I even asked you once about the strange bias situation. The amp has 640VDC on the plates and the schematic instructs you to set the bias so that you read 80mv across the 3.9 ohm sense resistors. That's 20ma x 640VDC = about 12.8 watts per tube. Never made sense to me and the amp was loud but just not as loud as I felt it should be. I even heated the bias up a bit with little impact on volume. Changing the NFB resistor to 47K erased all of that.

As if my tinnitus wasn't already bad enough.

[HotBluePlates]

Yes and no. There are now two resistors (the existing fixed 22K NFB resistor and the variable 1M depth pot) but they are in series from the OT tap to the connection at the PI. It's not really a voltage divider.

You're missing the 2 resistors PRR is talking about. The other resistor is between the long-tail inverter's "tail resistor" and ground. The typical presence control should either be this resistor or be in parallel with it.

The overall negative feedback (before NFB-reduction by way of a presence or depth control) is determined by the voltage at the OT secondary tap used (which depends on the actual tap selected and the output power), and is divided-down to a lower voltage by the voltage divider formed by the series negative feedback resistor and the resistor to ground.

The depth control can only make the amp louder by the amount negative feedback has already knocked-down amp sensitivity. Total power output doesn't really go up (clean output power should actually drop slightly, because the feedback reduces distortion for the portion of the amp it's wrapped around), but the output stage will require less drive signal to hit that maximum clean power. Said another way, feedback makes the output stage less sensitive, requiring more drive signal to hit the same output power.

If you look at the volume control setting and judge ear-loudness, then turning the depth up makes the amp seem louder because the same ear-loudness happens at a lower number on the volume knob. Really, it just distorts at a lower volume knob setting, which makes sense because the feedback was keeping the output stage cleaner.