Amp design exercise, mentorship needed...

[taj]

With apologies to the mods, I originally posted this in the Instruments and amps forum, since it's for guitar ultimately, but there just isn't much traffic there, and I got very little joy. All the tube amp expertise is here.

I'm trying to learn how to design an amp, so this is a project intended mainly to drag me through the process and calculations as a learning experience. The only design goals; small practice/recording amp, SE using a 6DQ6B since I have a couple dozen NOS ones in my junk box. I'm basing its topology on Bob Richard's Bluesman amp, but using different tubes, again, just to force me through the calculation process.

Here's a loadline I've created for the output. I graphically scaled the curves up to G2=175V instead of 150V, so they aren't exact, but should be really close.

Your advice is appreciated. (this is the 2nd stab at it. Thanks Michael K. for the first response.)

..Todd

[rsumperl]

Perhaps these might help.

[taj]

Hi rsumperl,

Well, not really. I did see that (and others), but I won't learn much copying a different design, or even just the operating points, so I purposely avoided looking too closely.

..Todd

[Miles Prower]

Originally Posted by taj

Your advice is appreciated. (this is the 2nd stab at it. Thanks Michael K. for the first response.)

The 6DQ6 is not at all suitable for SE use. The most linear part of the characteristic lies well within red plate territory. If you take another look at that loadline, it strays out of the saturation region at a very low Vgk, just barely above Vgk= -30V. That's an open invitation to screen grid poofage, as the screen current will increase tremendously once you leave the saturation region.

If you're gonna use those 6DQ6s, best design a PP final. If you want to stay with SE, then E-Bay 'em for types more suited to SE, Class A operation.

See this thread for design walk-through I did.

[taj]

Hi Miles,

In fact your walk-through is one of a handful I have printed out and re-read regularly, hoping one day for it to suddenly sink in and make sense. It's starting to now that I'm applying it. Thanks for writing it.

I'll take your word for it about the 6DQ6B being unsuitable for SE use. I just have a bunch kicking around, and the output power is about right in SE mode for a little practice amp, and I didn't want to buy anything for this. I'll see what else I have kicking around.

[rummaging...] Ooo.., I have a couple used 7189s and a pair of little companion SE output transformers from a donor console stereo. I'm sure they'll be fine for this. And I'll have spare set if I destroy the first attempt (or love it so much that I want a second one.)

Miles, I'd love it if you could elaborate on why the 6DQ6 isn't suitable; I don't know the term "saturation region" but it sounds like you are suggesting the operating point and load line are too close to bottom (0 mA) cut-off area. Also, I don't understand why the screen current acts as you say.

And for guitar amp purposes, how much should I be concerned about using a tube's most linear region? For hi-fi that's a no-brainer, but for guitar I may be wanting to purposely avoid it. I'm not really sure how the guitar amp guys approach that. But I do know that achieving that perfect distortion spectrum is an art-form I'd like to learn.

..Todd

[Miles Prower]

Originally Posted by taj

Miles, I'd love it if you could elaborate on why the 6DQ6 isn't suitable; I don't know the term "saturation region" but it sounds like you are suggesting the operating point and load line are too close to bottom (0 mA) cut-off area. Also, I don't understand why the screen current acts as you say.

Look again at the plate characteristic you attached at the top of the thread. Everything to the right of the blue line is the saturation region. This is where the plate current becomes (more or less) independent of Vpk. It is also where screen current remains manageable. Straying outside of that drives Vpk to such low values that the plate becomes ineffective as an electron attractor, sending the screen current skyrocketing to possibly destructive currents. It's especially bad for any guitar amp design, given how axmen tend to abuse amps.

For any SE amp, the plate bias current should be about half the max signal plate current. The 6DQ6 and other TV horizontal deflection finals were designed to pull some big currents at Vpk= 0. For the 6DQ6, that's about 350mA. so you'd need a plate bias of 350/2= 175mA. At Vpp= 300V, that means that Pd= (300)(0.175)= 52.5W. The type is rated for Pd(max)= 18W. Granted, the TV horizontal deflection types are rated very conservatively, given the demanding nature of the intended service, and can stand some spec-busting for much less demanding audio amplification, but not that much. It'll red plate badly.

And for guitar amp purposes, how much should I be concerned about using a tube's most linear region? For hi-fi that's a no-brainer, but for guitar I may be wanting to purposely avoid it. I'm not really sure how the guitar amp guys approach that. But I do know that achieving that perfect distortion spectrum is an art-form I'd like to learn.

You wouldn't necessarily be so concerned about linear operation for guitar amp useage. However, you still need to consider staying within the type's saturation region so's you don't poof the screens. The only other alternative might be to use it as a pseudotriode, and keep the Vpp under the 220Vdc max specification for the screen voltage, unless you can determine if you can bust that spec without making the screens glow. Some of the TV HD types are more tolerant of this than others.

[taj]

Okay, thanks Miles. I understand most of that. Regarding straying outside the saturation zone, you mean if the output voltage swings too much to the left (negative from bias point). But wouldn't that be the case in every tube load line? I'm not getting why this is more of a concern with this tube or load line. I guess there's something I'm still not seeing.

I'm going to tackle this with a 7189 instead.

..Todd

[Miles Prower]

Originally Posted by taj

Okay, thanks Miles. I understand most of that. Regarding straying outside the saturation zone, you mean if the output voltage swings too much to the left (negative from bias point). But wouldn't that be the case in every tube load line?

Not if the loadline intersects the Vgk= 0 line. If that's the case, then you'll have grid clipping at the same point where the loadline leaves the saturation region, and the nasty sounding distortion will let you know you've turned the volume up too high. (It's something you need to watch out for with RF amps, as mismatched loads can cause the same problem, and high powered RF pents often incorporate screen grid over current protection.)

I'm not getting why this is more of a concern with this tube or load line. I guess there's something I'm still not seeing.

The original loadline leaves the saturation region at way below Vgk= 0, and so you won't have clipping distortion to let you know you're getting into possibly destructive territory.

[taj]

Okay, Finally got it. Thanks.

..Todd

[sreten]

Originally Posted by taj

Hi rsumperl,

Well, not really. I did see that (and others), but I won't learn much copying
a different design, or even just the operating points, so I purposely avoided
looking too closely.

..Todd

Hi,

Analysing existing designs, good and bad, are an
excellent way of checking your analysis skills.

If you can't do that, you can't design anything.

rgds, sreten.