My first 2-channel amp - Please comment

[Gaz]

This is my first attempt at a two-channel amp, and I'd love to get any input before diving into the building process. One of my concerns is whether or not I 'need' the second gain control before the master volumes for the dirty channel. Thanks a lot for looking!

http://i56.tinypic.com/2e4xdzb.jpg

[imaradiostar]

It's just my own personal preference but I wouldn't want a gain stage after the dirty channel tone stack. I think the the gain channel will be better off without it.

You might take a look for a bogner shiva schematic- that seems to be the sonic territory you're headed toward.

Is your design based on the Kevin O'Conner design? I feel like I've seen something like this before. I've said this before but never tried it- if I were building a design along those lines I think I'd use the first half of a 12ax7 as the input tube and clean channel and the second 12ax7 as the dirty channel then funnel both into a mosfet source follower as a low-level effect loop output (with no additional gain before the loop). This would leave half a 12ax7 as gain recovery on the return side of the loop to drive the phase inverter. I know the dirty channel would have more than enough gain but I'm not sure about the clean side- I investigated lower loss tone stacks when I played with the idea. With this sort of arrangement your channel volumes become "send" level volumes and the master volume after the loop becomes the main volume control.

Have you considered a loop? I find I always like delay and chorus better post-distortion.

Those minor concerns aside, I think it looks like a solid design and I'd like to hear it.

jamie

[Gaz]

Thanks for the input, Jamie. The design is inspired by a preamp found in the Garnet book, which is again, inspired by KOC's design.

Why do you think the triode after the TS is unnecessary? - I'd like to know your reasoning. I think I found an okay schem of the Shiva, and I notice it has a CF, which I was trying to avoid. Do you think it'd benefit the OD tone more than the extra triode?

Thanks again for looking.

Alex

[imaradiostar]

Here's how I see it, I don't consider it gospel, just my interpretation.

First gain stage has about 35dB of gain. You'll lose between 10 and 25 dB in the Fender style tone stack. The switching and volume control will cause a little more drop. A mosfet source follower has a gain of approximately 1 so no loss there. This means a single triode would have enough gain to take care of the clean side up to the effect loop send. The error of many designs is having too high a send level. This arrangement seems like it would have an ideal level for common delay and chorus pedals.

On the dirty side you can assume the 3rd triode is going to produce hundreds of volts out so there will be more than enough level on the dirty side- you'll have to pad it down before the mosfet so the level matches the clean side. This is also why I wouldn't use a triode gain stage after the dirty tone stack.

So you'll have used 3 triodes and a mosfet before the effect send. This would leave a single triode to recover the low level effect return. A simple master volume would be placed between the recovery triode and the phase inverter.

The cathode follower is a personal choice- I like the way they sound but lots of great sounding circuits don't have them. I built a gain channel based on a SIB varidrive distortion pedal once (two triodes distorting, no CF) and I thought it sounded great too.

jamie

[Wilder Amplification]

I would seriously consider installing a 10M pulldown resistor on the relay that switches the PI input between the two masters to maintain the DC reference and prevent potential popping while switching.

[imaradiostar]

Found this in a folder on my computer. You can see there is a 470k resistor feeding the 100k volume pot. That's a substantial reduction, putting the distorted level closer to the clean level.



jamie

[Russ]

This may or may not be a problem, but since the input and output of the 2 triodes that get switched in and out of the high gain channel are in phase, and will be brought in close proximity to each other at the switch (DPDT (a)), you may experience squealing oscillation at high gain settings. Shielded cable not help here either, since its the close proximity at the switch that might be the source of oscillation's feedback.

Russ

[defaced]

Why do you think the triode after the TS is unnecessary? - I'd like to know your reasoning.

Don't forget that a LTPI does contribute gain. LTPI is almost the de facto phase inverter in high gain builds (and really guitar amps in general), and none of the schematics that I have seen use a gain stage after the tone stack, plate driven or cathode driven. Not that your design won't work, but if you haven't built it, don't put money on it - yet.

The 1u across teh lower resistor in the elevated heater circuit probably won't be enough to decouple. Try 22u to 47u.

Have you prototyped this circuit to see what it does?

[Gaz]

Well, this thread and some more research got me rethinking my untested design, and I changed things around a bit. Unfortunately, I had to add another 12AX7, but I really had no excuse to economize the way I did.

@ Defaced: The cap in the elevation circuit is .1uf, which has worked well IME.

BTW, Jamie, it's really interesting seeing the differences between KOC's original design and Garnet's variation.

Thanks for all the comments so far.

Here's the newer version:

http://i51.tinypic.com/34f14.png

[Wilder Amplification]

Well, this thread and some more research got me rethinking my untested design, and I changed things around a bit. Unfortunately, I had to add another 12AX7, but I really had no excuse to economize the way I did.

@ Defaced: The cap in the elevation circuit is .1uf, which has worked well IME.

BTW, Jamie, it's really interesting seeing the differences between KOC's original design and Garnet's variation.

Thanks for all the comments so far.

Here's the newer version:

http://i51.tinypic.com/34f14.png

Couple of questions -

Does your heater have a physical center tap on it?

Why so much filtering on the plate/screen filters?

Why the dual bias adjustment?

[defaced]

Jon, that filter arrangement is used on the SLO. It's no really alot, just more than most people are used to seeing.

Why the dual bias adjustment?

Which the way it's designed, will interact. Not sure how bad, but it will. There's a circuit on here that PaulP (I think that's who it was) and Enzo worked out that works really well. I've used it for both dual and quad bias arrangements.

Usually, people set up dual bias so they can use different tube mixes in the push-pull stage. Right now the way your dual bias is set up, you're just going to be able to unbalance the output stage. If that's all you want to do, I'd use a bias balance adjustment instead of two bias circuits like you have it set up now. If you want to be able to mix tubes, you need to redesign the PI to power tube coupling and add more bias feed resistors.

[Chuck H]

Only grazed the thread but I didn't see any mention of the switch for grounding the OD channel input being in front of the coupling cap. You'll want to move that .022 coupling cap feeding the OD channel ahead of the grounding switch.

[imaradiostar]

Only grazed the thread but I didn't see any mention of the switch for grounding the OD channel input being in front of the coupling cap. You'll want to move that .022 coupling cap feeding the OD channel ahead of the grounding switch.

What he said!!

Are you averse to effect loops? If you use delay at all you're probably going to want one.

I don't see the point of the parallel triode on the return side. I prefer the KOC design which is in line with your original idea of not having a cathode follower.

jamie

[Gaz]

Wilder:

Does your heater have a physical center tap on it?

Yes, but I prefer to use the hum balance/humdinger because it'd gotten rid of a pesky buzz before.

Why so much filtering on the plate/screen filters?

As defaced said, I think that amount of filter is really quite typical and should keep the amp a little tighter.

Why the dual bias adjustment?

So I can use two different pairs of tubes, or balance the power section when the tubes age asymmetrically.


Defaced:

Which the way it's designed, will interact. Not sure how bad, but it will. There's a circuit on here that PaulP (I think that's who it was) and Enzo worked out that works really well. I've used it for both dual and quad bias arrangements.

Usually, people set up dual bias so they can use different tube mixes in the push-pull stage. Right now the way your dual bias is set up, you're just going to be able to unbalance the output stage. If that's all you want to do, I'd use a bias balance adjustment instead of two bias circuits like you have it set up now. If you want to be able to mix tubes, you need to redesign the PI to power tube coupling and add more bias feed resistors.

You're right, and the interaction is pretty small - takes some juggling to get each pair balanced, however. I got the design from the AX84 "October" amp. I just want to be able to balance each half of the push-pull. Do you have a schem of the "bias balance" you mentioned? I've never done it, but even with the controls as they are, couldn't i use, say, one 6L6 and one El34 per side?


Chuck:

Only grazed the thread but I didn't see any mention of the switch for grounding the OD channel input being in front of the coupling cap. You'll want to move that .022 coupling cap feeding the OD channel ahead of the grounding switch.

Good catch! Thanks.


Jamie:

Are you averse to effect loops? If you use delay at all you're probably going to want one.

I don't see the point of the parallel triode on the return side. I prefer the KOC design which is in line with your original idea of not having a cathode follower.

Not really, I've just never built one before, and there's so many schematics out there. Is there a good one you can recommend thats simple and doesn't "tone suck?" I'd appreciate that.

There's really no "point" to having the parallel triode besides to perhaps make the clean channel a little more interesting. I didn't want to "waste" a triode with the addition of the CF to the other channel.

And Jamie, it's funny I changed the design because you mentioned the CF and the Shiva! I suppose I could just have the channels converge after the the clean channel's recovery stage, rather than before it, and having the channels overlap like in the KOC design. From your experience, do you think the lack of a CF will detriment the OD channel a lot?


Thanks a lot for all the input.

Alex

[imaradiostar]

I don't think the cathode follower is a detriment at all. It's just a different sound and the cathode follower helps to drive the tone stack harder. You'll notice in the KOC design the tone stack slope resistor is 100k- this is probably a good idea since it makes it easier for the triode plate to drive the tone stack. It makes the tone stack higher loss and weights it toward brighter/more treble. The cathode follower creates a tonal difference by drawing grid current on larger signals. The jury is out as to what sounds better- it's really a personal preference.

I was only assuming you wanted things simpler. If you don't mind the higher parts count you could build the Bogner design as found on the web. The layout is likely to be mildly complex but if you've built a few single channel amps it shouldn't be bad.

I'm just a fan of my simpler design because it allows you to use just 2 triodes and a high voltage mosfet to achieve a decent clean and dirty channel and a pedal-friendly effect loop. I have not built it yet so I don't know if it works correctly but there is really no reason why it shouldn't. If you'd like I can draw it up for you.

jamie