New amp design

[Pdavis68]

So first, a bit of quick background. I have nearly completed an amp design that I got help with on diyaudio, working through the various stages of the design. This is my first build from scratch (I've done a bit of tube amp repair work and low voltage stuff). I've decided that my design that I did on diyaudio is too much for a first build. I still plan on building it, but with 2 channels, 3 pre-amp stages, reverb and tremolo (6 tubes in an SE amp), it just seems like I'm asking for trouble making it my first build (it's my dream amp, though, so I'm going to build it!)

So I took a step back and threw together something much simpler.

It's a straight-forward basic design, for the most part. Here are a few notes about it:

1> In the first stage preamp, the "squish" control is a sort of compression control. This is used in the AX84 blues preamp. I believe it's also discussed in Merlin Blencowe's "Designing Tube Preamps for Guitar and Bass" book, but I can't find it at the moment.
2> Instead of going with a tone stack, I stole the idea of the rotary selector with different cap values ala Matchless DC30 and a couple others I've seen. I threw in some values, but in reality, I'll probably go with a 3 or 4 position rotary switch and just find a few values that I like.
3> With the 2nd stage of the preamp, I've got the two paralleled 12AX7 halves... The reason to do it this way is I also want to play with another idea from the AX84 blues preamp. It has paralleled 12AU7, one biased with a green LED and another biased with a red LED. I want to try that with mine instead of the resistor bias in the schematic. It's supposed to "thicken" the tone. In fact, I was thinking of biasing one with a green LED and then having a DPDT switch that switches the other between a green LED (so they're biased the same) and a red LED (a "fat" switch). I also like the idea of being able to try out other 12A?7 tubes...

Otherwise, I think it's a fairly straight-forward design. I'm actually going to wire in an Octal and Noval for the power amp section so that I can try different power tubes for the output. I have an EL84 and wanted to hear what that's like. I also have a bunch of 6?6 (where ? = K, W, L, V, etc) and even some 5881s.

I was actually thinking about putting in a barrier terminal block for all the power tube connections. That way I could wire up either socket however I need to. Is there any problem with doing that?

Thanks for any comments.

BTW, most of the parts are already on order, so hopefully I haven't screwed up too bad :-)

[Pdavis68]

Oh, and one other note. Obviously missing is the power supply. I have a 230-0-230 Antek power transformer coming. I was going to go with 1N4007 as a full wave bridge rectifier... Seemed simpler than going with tube rectification.

[eschertron]

Why the values for R15 and R16? As grid leaks they could be much bigger. I'd think they will load the signal down too much as is.

I like your ideas for the tone selector and parallel gain stage. It will be interesting to hear your results. +1 for wanting to audition the sounds before declaring your design finished. There's nothing wrong with trying things out!

...and welcome to the forum

[pdf64]

V1 needs a grid leak path to ground, eg 1M.
I think that R16 and R15 will load the signal down to nothing? Not sure what function higher values, eg >1M, would serve though?
Instead of C12 bright cap, a tone control may be nice, eg as per the Harvard http://bmamps.com/Schematics/fender/...6g10_schem.pdf

[Pdavis68]

Oh... I wasn't thinking about loading. I was thinking about signal attenuation... I'll need about a 1:2 between R16 and R17 and about 1:10 between R15 and R6, right? Or am I thinking of that wrong? My thinking was that the 1st stage will have a gain in the 50s, I believe, so I'm looking at maybe 5-6ish volts out and I need that to go down to about 1.5 for the 2nd stage? And I've got about 90V coming out of the second stage and need to cut that down to about 8.5. Am I thinking about this right? This is an area I'm really shaky on... The whole interstage voltage attenuation thing..

The Valve Wizard site says about 40K grid stopper for a 12AX7, so should R16 be about 20K? Is that still too low? I could definitely use some help on this part.

[eschertron]

Oh... I wasn't thinking about loading. I was thinking about signal attenuation... I'll need about a 1:2 between R16 and R17 and about 1:10 between R15 and R6, right? Or am I thinking of that wrong?

The grid stoppers (R17, R6) can be ignored for the purposes of calculating attenuation. It's the grid leaks (R15, R16) and the preceding impedance (determined in your design by the position of the volume control) that will form the voltage divider. And you already have a variable divider with the two volume controls!

You probably want the grid leaks to be the max recommended value per the spec sheet. something like 1M for a 12A_7 tube, and 220k-ish for the EL84. They are there to provide a ground reference for the tube grid, and any other thing they do has a negative impact on your design (like the loading we're speaking about).

The Valve Wizard site says about 40K grid stopper for a 12AX7, so should R16 be about 20K? Is that still too low? I could definitely use some help on this part.

Remember grid stopper is NOT grid leak.

[nickb]

The screen grid of V1 should connect to the junction of C13 and R9. Not sold on the tone switch idea as it seems rather limited in functionality for so many components.

[eschertron]

As far as interstage attenuation goes, look at a Marshall preamp design (for example). They've taken your 1M vol pot and split it up, a 470k fixed resistor on top of a 500k pot. This limits the volume control's range to ~1/2 of the total available. I've used a 680k resistor above a 250k pot in some places for additional attenuation.

The MAX voltage available after the divider = signal before the divider * pot resistance / total resistance

When the vol pot is turned down, the signal avail = sig before * pot value above ground / total resistance

[Pdavis68]

Awesome. Now I get it. I'll make those changes along with adding a 5M grid leak to the input (that's what the champ uses for a 6SJ7).

[Pdavis68]

Doh! Yeah, not going to have much DC on that screen as-is... Thanks...

I'm open to changing the tone setup. I have a traditional FMV tone stack in my other amp design. I thought this would just be something new to try...

I'm really thinking of this amp as a playground to try stuff and get a better handle on amp design and building, more than anything else.

[Pdavis68]

Still not addressing the tone at this point, but I think I've gotten the other points addressed.

Here's the latest

Is this a bit closer to what I want?

[nickb]

R18 should be no more than 1 meg. Bigger also means noisier.
R15 & R16 are doing nothing useful other then modifying the law of the pots. If you want the change the attenuation make R19 or R20 bigger or put them across the pots rather than to the wiper.

I calculate the -3db points for the tone control (high pass) as about 10Hz, 20Hz, 48Hz, 70Hz, 212Hz, 272Hz - I think the bottom four are too low to be of much use.

[Pdavis68]

Removed R15 and R16 and set R18 to 1M. Thanks.

I hadn't actually calculated the frequencies on the tone. I had copied those values from an amp and meant to go through and calculate them at some point. But that actually raises another point... What would be a good range to shoot for with the -3db point? Obviously somewhere less than 48Hz as the highest. I just don't have enough experience with audio to know what would be too low for the cutoff and what would be too high.

[Enzo]

R18 in the pentode Champ was 5M because that tube was grid-leak biased. Yours is cathode bias. That is why the 5M is way too large for yours.

[nickb]

Removed R15 and R16 and set R18 to 1M. Thanks.

I hadn't actually calculated the frequencies on the tone. I had copied those values from an amp and meant to go through and calculate them at some point. But that actually raises another point... What would be a good range to shoot for with the -3db point? Obviously somewhere less than 48Hz as the highest. I just don't have enough experience with audio to know what would be too low for the cutoff and what would be too high.

I'd start with something like 150hz for the lowest and then x 1.5 for the next i.e. 150, 220, 330,...1100Hz. Just a WAG. Lot's to experiment with there. That is why I'd choose a continuously variable control, but, hey, it's your baby!

I really don't think the 'Squish' works to change the compression. It's just a variable low frequency gain reduction. You'd need something quite a bit more complex to do that.