BF Super Reverb aa763 Normal Channel -- how to increase headroom?

[TheTinMan]

I've learned some things in this thread, so it's not a total farce. A big part of electronics, as with any field I suppose, is learning what's possible and what's not possible. Up until recently I had really only done paint by numbers style of amp building with some debugging and general troubleshooting. I haven't done much experimenting or modding. In a big way I'm at the bottom of another curve. Joy.

First off, Mort started with a perfectly reasonable question which I understood just fine.

Most of the previous posters are assuming that the poweramp is breaking up before the onset of any preamp distortion. Why? I've only built one Super Reverb but the stock channel stayed pretty darn clean until it was really LOUD. I understand that the poweramp puts a limit on clean, undistorted sound volume. However, that doesn't automatically mean Mort's amp isn't starting to break up in the preamp before the poweramp starts letting go.

Also, no one has bothered to ask Mort what guitar his friend/customer has plugged into the amp (pickup type and output most specifically). "335" could cover a lot ground. Plug a high output humbucker into the Hi input of the Normal channel and I'm reasonably certain that the first two triodes start breaking up first. This is based on experiments in a Super Reverb derivative using a scope. Pedals up front further complicate the question.

The assumption poweramp distortion first may very well be correct, but we don't know without scoping the amp with the user's rig do we?

So, what could we suggest to get more clean headroom (higher output voltage swing without distortion) from the first two stages? This will let Mort figure out where the distortion starts first.

An LED in place of the cap/resistor on the first stage would provide effectively fixed bias on that triode. Assuming his cathode voltage there is about 2 volts, try LEDs that drop 2 and 2.5 volts.

Separate the RC cathode network for the second stage and experiment. In my SR build, I had a switch which added additional resistance between the RC network and ground to reduce the gain from that stage. Start with 470 or 680 ohms first perhaps. This really needs to be switched though to figure anything out.

What about different 12AX7s for those first two stages? Tube choice affects the breakup point in my experience.

Tone stack tweaks might help too if he has really hot input. A 68k slope resistor instead of 100k with increase insertion loss. It also will change the frequency shifting of the tone stack. Lowering the mid cap from .033 to .022 like the AB763 will change things too. Better or worse? I don't know because I don't have the amp in front of me, but even it the poweramp is distorting first we don't know what frequencies break up first. Maybe cutting mids (reduce 6.8k resistor to 4.7 or even 3.3) will do the trick. Maybe a .068 bass cap in place of the .1 could help if the distortion is from speakers with a bit of flabbiness in lower frequencies. From the soundclip, taming the bass a tad might be just the ticket. Instead of tone stack, dropping the first or first two cathode bypass caps might get the amp owner smiling. Nothing radical like a .68uf, but 4.7 or 6.8 could help. If it does, try it again on the second triode.

My main point is that there are some experiments Mort can try instead of throwing up his hands and telling the amp owner "it just is what it is. Tough." Also, our "instrument" is everything from whatever hits the strings to the speaker cone shaking around inside the cab. It's not just the amp in isolation.

BTW do we know what speakers and if there are more efficient alternatives?

One question for Mort: is the schematic drawn wrong for those first two filter caps or did you wire it as drawn? The two 350 volt caps are supposed to be wired in series with a 220k resistor in parallel with each cap. See : AB763 schematic

If wired as drawn, I would replace those caps because then they've been hit wth voltage WAY above their rating. However, I assume that you wired it correctly and just got the schematic fudged up.

Hope this helps,
Chip

[Malcolm Irving]

...
Most of the previous posters are assuming that the poweramp is breaking up before the onset of any preamp distortion. Why? ...

Let’s say the power tubes require an input swing of 50V to give the full rated output with 5% distortion at the onset of power stage clipping (i.e. clean). For that low level of distortion all the stages before the power tubes must either not be clipping, or also be just at the onset of clipping. A well designed amp would not have any of the pre-amp stages (in the clean channel) at the onset of clipping in the above circumstance. Why?
Because components (especially tubes) have a wide tolerance range, it would be very difficult to ensure that a pre-amp stage clips at exactly the same time as the power stage.
If we did try and design it that way, and the tolerances work against us, maybe a pre-amp stage clips first and we can then only deliver a swing of 40V at the power tube grids – which means the power output is reduced to 64% of its rating.

[TheTinMan]

Let’s say the power tubes require an input swing of 50V to give the full rated output with 5% distortion at the onset of power stage clipping (i.e. clean). For that low level of distortion all the stages before the power tubes must either not be clipping, or also be just at the onset of clipping. A well designed amp would not have any of the pre-amp stages (in the clean channel) at the onset of clipping in the above circumstance. Why?
Because components (especially tubes) have a wide tolerance range, it would be very difficult to ensure that a pre-amp stage clips at exactly the same time as the power stage.
If we did try and design it that way, and the tolerances work against us, maybe a pre-amp stage clips first and we can then only deliver a swing of 40V at the power tube grids – which means the power output is reduced to 64% of its rating.

I understand the design philosophy completely. However, if I were designing an amp today, I would have a preamp volume and a master volume so that I could balance the onset of clipping in both portions of the amp regardless of what guitar/pedal combination I put in front of the amp.

More importantly, think about the guitars Leo Fender & Co. had in mind when they were designing the 5E3. Telecasters and Stratocasters. IOW guitars with low output single coils. An amp designed just as you describe for these specific guitars might start clipping in the preamp well before the power amp if you plugged in a guitar with humbuckers. Like the 335 in the soundclip Mort posted. And like the Super Reverb I built and experimented with using an oscilloscope.

Some guys think that the "Lo" input was intended for Gibson humbucker guitars - not likely but it works by cutting the voltage swing right at the input jack.

Respectfully,
Chip

[Malcolm Irving]

.... Telecasters and Stratocasters. IOW guitars with low output single coils. An amp designed just as you describe for these specific guitars might start clipping in the preamp well before the power amp if you plugged in a guitar with humbuckers.

Sorry, that's not correct. As you increase the signal input, in a non-master volume clean amp, one stage or other in the amp will clip first. If that is a pre-amp stage, then the power stage is prevented from reaching it's rated clean output power.

While the amp is running clean, each stage is providing 'clean' gain. The same stage will clip first whether the input is single coil or hum-bucker.

[loudthud]

Even when you are playing "Clean", chances are that you are driving the output to clipping on pick transients. You can easily see this on an oscilloscope.

[Dave H]

Sorry, that's not correct. As you increase the signal input, in a non-master volume clean amp, one stage or other in the amp will clip first. If that is a pre-amp stage, then the power stage is prevented from reaching it's rated clean output power.

You are talking at cross purposes. I think he is saying that with a really high output humbucker the first stage will always be overdriven but the power amp can still amplify this distorted signal to its (the power amp's) full 'clean' output power (if that makes any sense).

[pdf64]

I agree that's possible.
However, I think it's unlikely to be the case with mort's amp / his customer's perception of it.
Otherwise the complaint would be more likely to be that the amp sounds overdriven even at very low volume settings, especially if it occurred with some guitars, eg 335 with humbuckers, and not with others, eg a strat with vintage type pickups.

Additionally, although the regular Fender input stage may theoretically be overdriven, or at least pushed well past its linear range, by high output pickups, it almost never seems to be a problem.

My guess is that it doesn't tend to be noticeable because it's only likely to happen momentarily, eg at the initial pick attack, and given all the other non linearities between the brain sending out the signal to pick and perceiving the result, its contribution may be fairly insignificant.

Bear in mind that the SF / BF TR have closely similar pre-amp operating points to the SR and are generally perceived to give very clean reproduction up to very high SPLs.

It just struck me that one way to mitigate input overload may be to change it from being dc coupled, and include a dc blocking cap, such that the time constant at the input grid was fairly long.
Then the regular bias shift mechanism would operate in response to input levels close to, and above, the normal operating point; acting to compress the signal.

[Malcolm Irving]

... You are talking at cross purposes. ....

Very probably!

Here is a simplified example to clarify what I mean.
Let’s say we have a clean amp with two preamp stages, a phase inverter and the power stage. The power stage requires 50V input swing to make it clip.
A technical definition of the ‘headroom of a stage’ is the input signal swing at the onset of clipping of the output of that stage. So, the headroom of this power stage is 50V, for example.
Let’s say that the earlier stages have the following characteristics:
Stage 1 gain x50 headroom 5V
Stage 2 gain x40 headroom 7V
PI gain x20 headroom 10V
To find the input signal that is needed to make the output stage clip: 50/(20x40x50) = 1.25mV
Check all the stage signal inputs under this condition:
Stage 1 signal input = 0.00125V
Stage 2 signal input = 0.00125 x 50 = 0.0625V
PI Stage signal input = 0.0625 x 40 = 2.5V
Power stage signal input = 2.5 x 20 = 50V
Notice that all the stages prior to the power stage are well within their headroom.
If we increase the input to the first stage to 5mV the PI will start to clip, but at this point the drive to the power stage will be 10 x 20 = 200V, so the power stage is now massively overdriven.
OK, the above example is very simplified, but I hope it helps to illustrate the principle.
A similar analysis can be applied to a real amplifier, but we need to take into account: different +ve and -ve headroom for stages which are not centre-biased, difference between small signal gain and gain achieved at clipping, attenuation due to tone stacks, resistive mixers, inter-stage attenuation, dynamic bias shifts, power supply sag, etc. etc.

[TheTinMan]

I still want to know if the first parallel filter caps are wired as intended or as drawn.

You are talking at cross purposes. I think he is saying that with a really high output humbucker the first stage will always be overdriven but the power amp can still amplify this distorted signal to its (the power amp's) full 'clean' output power (if that makes any sense).

First stage or second stage depending on volume setting - exactly.

Whereas with a lower peak to peak input from a Tele, the cumulative gain at each stage may be just right so that there is no significant distortion thru the preamp before the poweramp reaches maximum clean output.

Respectfully,
Chip

[mort]

I still want to know if the first parallel filter caps are wired as intended or as drawn.

They are wired as drawn. (?)

[Justin Thomas]

Mort,

That schem arrangement of those caps seems odd to me, too. It would seem the 220ks are in series with the individual caps instead of in parallel with them, and that the two R+C "units" are in parallel with each other? One cap has the resistor at the -ve end and the other at the +ve end, too. Either way, it would seem to me that as drawn you have a 200uF/350V cap there; not sure what effect the Rs have... but if it is that way, 350V is low for the amp' s B+... can you take a picture of the area, or are there wires going under the board?


I think the caps should be in series, + to -, with a resistor across the individual terminals of each one. Right now, the caps are in parallel, with the resistors in between them. It's like the other three caps, except as if you put a resistor between the grounded ends.

Not picking, the amp may work fine. Just curious, cuz I can't claim to know what would happen. You've built more than me.

Justin

[Dave H]

Sorry, that's not correct. As you increase the signal input, in a non-master volume clean amp, one stage or other in the amp will clip first. If that is a pre-amp stage, then the power stage is prevented from reaching it's rated clean output power.

The cross purposes I was alluding to are your analysing a hypothetical amplifier with no volume control and TheTinMan analysing a real guitar amp with a volume control positioned after the first stage (or later). There can be no disagreement here because you are not discussing the same circuit. You are both right

[mort]

Mort,

That schem arrangement of those caps seems odd to me, too. It would seem the 220ks are in series with the individual caps instead of in parallel with them, and that the two R+C "units" are in parallel with each other? One cap has the resistor at the -ve end and the other at the +ve end, too. Either way, it would seem to me that as drawn you have a 200uF/350V cap there; not sure what effect the Rs have... but if it is that way, 350V is low for the amp' s B+... can you take a picture of the area, or are there wires going under the board?


I think the caps should be in series, + to -, with a resistor across the individual terminals of each one. Right now, the caps are in parallel, with the resistors in between them. It's like the other three caps, except as if you put a resistor between the grounded ends.

Not picking, the amp may work fine. Just curious, cuz I can't claim to know what would happen. You've built more than me.

Justin

The person that was helping me early on with this project suggested it. The sum of the filtering should be 50uF at 700v I believe.

Here's his edits to my original rough drawing. He calls it a "totem" arrangement.

[Justin Thomas]

I may just be so used to seeing the traditional Fender way of drawing it that I can't wrap my head around something different... I'm open to correction on the matter!

Justin

Edit: okay, I think I get it now. It just doesn't look very "totem-poley" to me.

[Dave H]

Has the amp's output been measured and scoped?

Very good suggestion that. All this theory is fine but there is no substitute for measurements. Mort do you have a scope and sig gen?