AX84 P1x Build - "Stiff" Sounding Output

[bw1985]

Hi all,

Many years ago I built an AX84 p1x amp that uses the single octal output tube. I've since rebuilt the preamp section of the amp several times. I've tried the original design, and then a higher gain preamp and now have rebuilt it again using the front end of the 1974x / 18W design.

Everything in the preamp section sounds great. I've got the right flavor/sound of the 1974x, but the output feels very "stiff" or almost solid state like to me. I've had this same kind of sound with every preamp I've tried on this thing.

I'm wondering where I can improve the design. Should I swap the filter caps to lower values (they are currently all 47uF)? Should I change the OT (currently using the Hammond 125ESE)? I'm willing to spend more cash on an OT if it makes a world of difference. Not sure if I've ever been a huge fan of Hammond OTs in general.

I would appreciate any advice or circuit/component tweaks anyone could recommend to make this thing sound a bit better! I would prefer to not change to a tube rectifier if possible.

Link to the schematic on Archive.org:
https://web.archive.org/web/20160531...P1x_101004.pdf

Thanks!

[g1]

Does it have a tube rectifier? If not, why not just add a sag resistor after the rectifiers, simple and effective.
Edit: I see there is already a 100R between the rectifiers and B+1, you could try adding another between there and the OT, or try increasing the value of the 100R, see what you like.

[Helmholtz]

Does it have a tube rectifier? If not, why not just add a sag resistor after the rectifiers, simple and effective.

Schematic shows a 100 Ohm sag resistor. But it won't have much dynamic effect with a class A output stage as average power/current consumption is more or less constant.

[loudthud]

I'd knock those 47uF caps down to 22uF.

[Dave H]

Perhaps it's the EL34? You could try a 6V6 without having to do too many mods. Increase the pi filter resistor R2 to 1k? to reduce B+ voltage and keep the 6V6 dissipation to less than 14W and set the OT impedance to 5k.

[Helmholtz]

I'd knock those 47uF caps down to 22uF.

Single ended amps require better plate and screen supply filtering than PP amps.

[Chuck H]

Once again Helmholtz in onto the real problem. It's a class A, single ended output. It's the nature of the design to NOT sag. No amount of sag resistors or tube rectification can change the fact that the single power tube has to operate continually somewhat near the max of it's useful current for conducting audio. The same for soft, quiet playing as for hard, loud playing. Which means there's no where to sag to, basically. All the audio current the power amp will use is already being used when it's just sitting there amplifying nothing. In fact it's not uncommon for current to decrease with signal conduction in class A, single ended amps. I think the guitar player community at large has the opposite expectation of "class A" relative to the media hype.

[bw1985]

Thanks for all of your replies, they are much appreciated!

I suppose changing the output tube to a 6V6 could be the best option here. I did try the amp with one, but didn't make the necessary changes to the circuit for it to be biased properly. It sounded better but not perfect. I've built other single ended amps (more of a Champ style circuit) and didn't really have this issue. Maybe I'll take some notes from the power supply in that amp.

I'll take Dave H's advice and make those circuit changes soon.

[Chuck H]

I've built other single ended amps (more of a Champ style circuit) and didn't really have this issue. Maybe I'll take some notes from the power supply in that amp.

Perhaps there were properties other than "sag" that were giving you the favorable results you're after then. Because the situation is very much as Helmholtz said. You may be chasing the wrong solution because class A, single ended power amps don't really sag.

Can you provide schematics? One for a single ended amp you like the behavior of and then this one which you don't. We may be able to identify some other parameter that is overtly different about them.

My guess (only a guess) is that it may have something to do with how much drive you have to the power tube in either deign. A single ended amp will clip more asymmetrical with more drive. Perhaps if you decreased preamp clipping and increase power tube drive for more asymmetry would put you in the zone? So don't be fooled into thinking the 6V6 is offering more "sag" if it works out. It could just be that the 6V6 requires more grid drive to reach the same playing volume levels as the EL34 does now.

[bw1985]

Perhaps there were properties other than "sag" that were giving you the favorable results you're after then. Because the situation is very much as Helmholtz said. You may be chasing the wrong solution because class A, single ended power amps don't really sag.

Can you provide schematics? One for a single ended amp you like the behavior of and then this one which you don't. We may be able to identify some other parameter that is overtly different about them.

My guess (only a guess) is that it may have something to do with how much drive you have to the power tube in either deign. A single ended amp will clip more asymmetrical with more drive. Perhaps if you decreased preamp clipping and increase power tube drive for more asymmetry would put you in the zone? So don't be fooled into thinking the 6V6 is offering more "sag" if it works out. It could just be that the 6V6 requires more grid drive to reach the same playing volume levels as the EL34 does now.

The amp that I built was based on this old Valco schematic. It's similar to a Champ I believe but has tons of gain. I built it using an old Bogen CHA-10 that I had sitting around. I am however using a solid state rectifier as I found it worked well for this amp.

I also swapped the 6V6 for a 6L6 in that amp and it sounded awesome because it actually tamed the gain levels a bit and gave me slight headroom improvement.

[Mick Bailey]

I've experimented with sag in SE amps and the results can be quite good but adds to the complexity. The idea came from an old Gibson amp schematic and I implemented it with a vactrol driven off the OT secondary, just like an optical compressor. There are a number of issues with this simple approach, the main one being that there's a sweet-spot for operation due to the rather narrow operating voltage for the LED side of the vactrol. It needs a rectified voltage that's trimmed to give the turn-on, and voltage limiting so as not to destroy it (and set the upper operating level), though in low-powered amps this is perhaps less likely. In practice I could get a really nice sag at a particular volume level, but as soon as this was changed the sag circuit needed adjusting to find the new sweet spot. At the time I though it would have some merit in that the sag could be made more or less extreme by adjusting the time factor of the LED circuit. At the most extreme the entire amp behaved like a compressor set to maximum squash.

[Chuck H]

Wow! That's as simple as it gets, huh! It's using grid leak bias for stage one and no cathode bypass on stage two, but still, not much loss with the tone control and basically a shit ton of drive to the power tube grid with what amounts to two series 12ax7 stages.

You mention having changed the preamp in the current project a few times. To be honest I don't know what you may have going for the preamp in the amp in question so I can't comment on what might be mitigating grid drive there. BUT...

Notice that in the schematic you provided that the screen node is fed through a whopping 6.8k and that may have an affect on how the screen grid behaves even in a single ended amp. I've read that under clipping conditions that this can create a sort of sag in the form of current transfer between the plate and screen circuits, but understanding this is beyond my pay grade Maybe someone else will chime in on this.

So in the amp you like there pretty much a straight shot and high grid drive to a power tube which has the screen fed through a higher than typical resistance. This is probably different from the amp in question (no schematic).

Probably also worth noting is that in the amp you like the power tube response of doesn't have a bypass cap on the power tube cathode circuit.

You might try bumping the screen node resistance in the power supply (and reduce the preamp node resistance so you don't lose the preamp tone you have) and then lift the bypass cap from the power tube cathode. See if that makes a difference for you.

[bw1985]

Wow! That's as simple as it gets, huh! It's using grid leak bias for stage one and no cathode bypass on stage two, but still, not much loss with the tone control and basically a shit ton of drive to the power tube grid with what amounts to two series 12ax7 stages.

You mention having changed the preamp in the current project a few times. To be honest I don't know what you may have going for the preamp in the amp in question so I can't comment on what might be mitigating grid drive there. BUT...

Notice that in the schematic you provided that the screen node is fed through a whopping 6.8k and that may have an affect on how the screen grid behaves even in a single ended amp. I've read that under clipping conditions that this can create a sort of sag in the form of current transfer between the plate and screen circuits, but understanding this is beyond my pay grade Maybe someone else will chime in on this.

So in the amp you like there pretty much a straight shot and high grid drive to a power tube which has the screen fed through a higher than typical resistance. This is probably different from the amp in question (no schematic).

Probably also worth noting is that in the amp you like the power tube response of doesn't have a bypass cap on the power tube cathode circuit.

You might try bumping the screen node resistance in the power supply (and reduce the preamp node resistance so you don't lose the preamp tone you have) and then lift the bypass cap from the power tube cathode. See if that makes a difference for you.

Thanks for the reply and the information. There is a lot to think about here, and definitely a few changes I could make. For reference, here is the schematic of the preamp I am using:


However, I am using a 220k resistor to ground instead of 470k after the 0.01uF coupling cap.

[The Dude]

I've experimented with sag in SE amps and the results can be quite good but adds to the complexity. The idea came from an old Gibson amp schematic and I implemented it with a vactrol driven off the OT secondary, just like an optical compressor. There are a number of issues with this simple approach, the main one being that there's a sweet-spot for operation due to the rather narrow operating voltage for the LED side of the vactrol. It needs a rectified voltage that's trimmed to give the turn-on, and voltage limiting so as not to destroy it (and set the upper operating level), though in low-powered amps this is perhaps less likely. In practice I could get a really nice sag at a particular volume level, but as soon as this was changed the sag circuit needed adjusting to find the new sweet spot. At the time I though it would have some merit in that the sag could be made more or less extreme by adjusting the time factor of the LED circuit. At the most extreme the entire amp behaved like a compressor set to maximum squash.

I wish I could like that twice! I've never considered that...... and it's brilliant!

[Chuck H]

I've experimented with sag in SE amps and the results can be quite good but adds to the complexity. The idea came from an old Gibson amp schematic and I implemented it with a vactrol driven off the OT secondary, just like an optical compressor. There are a number of issues with this simple approach, the main one being that there's a sweet-spot for operation due to the rather narrow operating voltage for the LED side of the vactrol. It needs a rectified voltage that's trimmed to give the turn-on, and voltage limiting so as not to destroy it (and set the upper operating level), though in low-powered amps this is perhaps less likely. In practice I could get a really nice sag at a particular volume level, but as soon as this was changed the sag circuit needed adjusting to find the new sweet spot. At the time I though it would have some merit in that the sag could be made more or less extreme by adjusting the time factor of the LED circuit. At the most extreme the entire amp behaved like a compressor set to maximum squash.

Love it! I've contemplated a way to get controllable compression (of a desirable nature hopefully) out of an overall design without the need to build a "compressor" into the circuit. Since you seem to be "on the path" as it were, my idea was to use a voltage divider off the top of the power tube cathode circuit (in a cathode biased amp with a full AC bypass) to feed voltage to an earlier preamp cathode (also AC bypassed). That way when voltage rises at the power tube cathode the drive signal is analogously reduced from the preamp. I haven't tried it because I'm doing other tricks with the cathode circuits in my cathode biased amps. But here it is for the world to see now. In the "public domain" as it were now. As is your idea. Just report back if you try it I'd love to know how it works out.