A Question For Fellow Amp Designers/Builders.....

[jrfrond]

What's your approach to amp design: "building block" with mods, ground-up fresh design, possibly with SPICE modeling, somewhere in between?

I'm guessing that it's the first one for many, since tube technology is old enough that the wheel doesn't require reinventing, and the tried-and-true designs, or variations thereof, are what is preferred by most players in the long-run.

[Howlin' Mad Mac]

Well I guess I use the "building block" method using tried & true designs from the past & attempting to incorporate some of the new or modern. I'd like to try Spice modeling, even downloaded the software, jus haven't quite figured out how to use it. Also still somewhat confused by some of the math needed for tube circuitry. They didn't teach us that stuff at electronics school because it was considered outdated (lol) so still trying to figure out a lot of things as I go along.

[diagrammatiks]

depends on what's being built.

if it's a simple circuit that's been done before in some way then building block is the way to go.

If not slide rule, paper, and pencil is the way to go.

Modeling and all that is a good way to generate numbers but it doesn't really tell you how something is going to sound, only if it's going to work or not.

If you want to get complicated with voltage references, current sources and sinks, totem pole circuits a spice simulation is just going to give you way to much information.

[Chuck H]

I'll go with building block, but... There's a lot to consider. If you want predictable results you can't just plug a "tweed" circuit into a Twin Reverb power supply. Plate voltages, rectifier type, time constants and such need to be considered when pairing design aspects. OTOH you can't know what a tweed type circuit will "sound" like at high voltages until you try it. Oh, wait, What OT primary Z to use??? The tweed circuit used 6600 for a pair of 6L6's but the Twin is more like 2100 for a quad??? You get the idea. Many things have been done that technically play on the edge of what one would expect to see in a tube amp. The building block method only works if you know what can and can't be paired. Oh... +mods. So that's "building block + mods" for me. Mods would include operational upgrades like elevated filament supplies, more uf main filtering, balancing gain stages, blah, blah... In the end you could hardly look at one of my amps and say it's using vintage circuits.

It takes time to learn what sounds like what, what you can pair and adjusting voltages, etc. In the end I think there's not much new under the sun WRT tube amps, But there's no limit to the variations.

[tubeswell]

I'm getting a lot of inspiration from published material now. But I know what things I like and which amps I like to get them from and its just a matter of finding the best way to combine them. Merlin's books have helped me to understand the issues of circuit design in greater depth in this regard, and when I sketch out circuit ideas now I do it in front of the computer and use the load line plotter programs freq formulas and other similar stuff.

(But don't ask me about spice modelling - is that like - how you cook your curry?)

[jrfrond]

So far, this is what I figured the answers would be. Many of us who still specialize in tubes have been at this for a long time and know what circuits sound like, how to lift them from one design, adapt them to another and try to make it work. Near-endless variations on familiar themes. Every once in awhile, you need to apply some science and/or math and numbers, but it's really in the empirical design process, debugging and tweaking (sometimes ENDLESS tweaking!), that make the designs fly. It has worked for many now-famous amp designers e.g. Randall Smith, Michael Soldano, etc. Even the original Marshalls were 5F6-A clones done a little differently, but yet became a legend in their own right.

I look at it this way: any well-versed, trained EE who knows tubes can design a simple audio amplifier with a given set of parameters, but it takes people with good ears and knowledge of what the player wants to create a guitar amp design that works, and sounds/feels good. They really should play the instrument as well.

Comments?

[Ronsonic]

The amps I've done have all been basically building-block derived. Practical design work usually is. Like you point out it's the dialing-in process that gets to be involved. That can get brutal, especially if you are trying to perfect a simple design.

[J M Fahey]

Personally I can design from the ground up. In fact, that's what I usually do.
What does *that* mean? :
> starting with a clean sheet of paper,
> figuring what I need to accomplish,
> making educated guesses as to how that could be achieved,
> sketching a block diagram as to have a general idea of what I'm getting into, sometimes it leads into a full re-thinking of the project, turning it inside out like glove or plain abandonong the idea, since under closer scrutiny it's not as good as it seemed to be before (yes, unfortunately that's also an option which should be considered)
You will see that up to now there's no SPICE involved, since still there are no schematics nor parts values to insert there.
> start filling each block with functional elements,
*here* some simulation can start.
Personally I never learnt to use them, (would love to, though) , but already have a quite clear idea as of what happens, meaning gain, frequency response, overload level, and even a rough idea of what overload waveform will get, at different levels.
> Protoboarding the building block (or its equivalent in tubes) provides the final answers.
Simulation is fine, an excellent tool, but the problem for me is that it's loaded with , for lack of a better word, the "standard" behaviour; while we overdrive things with 10X (or more) the maximum signal anybody in his right mind expects (maybe that says something about our mental condition) and all kind of "undocumented features" rise their collective ugly heads.
Also, as was stated before, we strive for a *sonic* effect, which although might be accurately depicted by a simulation or an oscilloscope waveform, can only be judged "by ear" so that's the Supreme Court for us.

That said, I find that the combination of 50 to 60 years old technology (those 50's amps *already* had a very good sound, are the benchmark even today) plus the fact that the exact same parts are used (12AX7, 6L6, etc.), mean that practically no *new* designs appear with any frequency, and what passes as the top of the top today is not much more than tweaking pre tweaked tweaked to death designs.
As in: 5F6 > English version (JTM45) > modernized (MarkII) > extra triode + master (JCM800) > 2 channels+tweaking (Soldano) > *sound* tweaking to death what's roughly the same (Rectifiers/AmpegVL/all the "high gain monsters" Diezel/Engl/Framus/Egnater/etc.)
As you see I do not participate at all in tube discussions, except some occasional answer trying to help some fellow who might have overlooked some small detail which could cause him problems.
As an example: how many burger variations can you design using the same old tired 6 elements Mc D and all others do?
As somebody posted: "since you can't be *that* different under the hood, go for the most outrageous front panel and cabinet design, not forgetting some weird logo" or something like that.
When I look at pictures from California Amp Show or similar events, I see exactly that.

Now, SS, that's a *much* wider and untapped field.
That's a real challenge, specially because it's not associated to Mojo at all.
Oh well.

[JoeM]

....<>
Now, SS, that's a *much* wider and untapped field.
That's a real challenge, specially because it's not associated to Mojo at all.
Oh well.

I've often wondered what amps today would be like if the timing of the vacuum tubes "demise" had occured before the beginning of the rock era and the early rock/blues/jazz players only had SS amps. I think today we'd have some very interesting designs.

[jrfrond]

It's not just tubes vs. SS anymore. Though all digital technology is SS, I place it in it's own category because we are now removed from analog, so it's Tubes vs. SS vs. DSP in my book.

The interesting thing is that there WERE some good, viable SS designs, and some of the same players that pissed on SS analog now think that DSP is the sh*t. I just think that DSP IS sh*t! I mean really, does ANYONE with a good pair of ears and hands REALLY believe that even the best modeling comes close to good tube sound and feel??? I don't. But what I think doesn't count in the real world, it's the players who buy into it and feel that it's "good enough".

Back when transistor designs were used in guitar amps, it was assumed that amplification was amplification. No one figured that tubes would have their own identity, and also be a much hardier design, and simpler to boot. This is why you really don't see too many SS guitar amp design gurus: less demand, and a bigger pain in the ass from a design standpoint.

[J M Fahey]

This is why you really don't see too many SS guitar amp design gurus:

Agree.
I don't see so many *Tube* design Gurus either
I mean, beyond self-appointed ones, that is. (which in my book doesn't count as one).

[JoeM]

...I mean, beyond self-appointed ones, that is. (which in my book doesn't count as one).

No shortage there!

[jrfrond]

No shortage there!

"Guru" is a title bestowed upon someone, not assumed by anyone, at least not in their right mind. After all, you know what happens when you ASSUME, right?

[J M Fahey]

at least not in their right mind.

I think you are touching a sensitive point here

[JoeM]

No one's ever accused me of being in my right mind!