AVA100 modular amp project

New: 26/09/06
Ver: 2nd draft 12:15 18/10/06
Rev: 22/1/08 (minor)

    Caution: (Oct '06) At this stage the AVA100-Series exists only on paper and is not a finialised design but a work in progress and as such may contain significant errors.

---

Such as...

22/1/08 - Minor detail; the 6GW8 isn't being currently manufactured. Oh dear. Assumption is the mother of all stuffups.

Never mind. If you happen to have a pair of 6GW8's, or 6V6's, or 6BM8's, or 6CM5's, or whatever, then the theory still applies and you can do your own worked examples (and send them in hopefully). So in the meantime where you see “6GW8” just think “my favorite pentode”.

Or how about a couple of 12AX7's as a quad in a push-pull parallel?

But for now it looks like the initial design will switch to 6L6's (or perhaps 6CA7's) as the reference circuit, simply because they are currently manufactured.

Experienced tinkerers may switch on their soldering irons at this point, but beginners looking at their first build would be advised to either wait for the dust to settle, or go look at (say) the established "AX" series.

---

The first actual incarnation will be the AVA101, so the AVA100 should be considered a mind-amp, a thought-experiment only, investigating the bits and moving them around before actually getting down and dirty with construction.

If you are already experienced, very keen, or just desperate, and want to have a go right away and don't mind a bit of risk and adventure, then jump right in. I'd love to hear your results and problems, but please don't complain if I have specified something that later turns out to be wrong, an expensive doorstop.

You want certainty? Wait for the design to be finished, or buy a name amp with a warranty.

But if you want to be part of a journey of discovery, if you want to learn a whole lot, if you're poor or broke (and what musician isn't?), or if you want to be in charge of what you are doing and not another hapless fashion victim or wood duck, then welcome aboard.

And if you are an old hand who would like to contribute your ideas, you're very welcome too.

This design makes no claim to magical tone. The initial aim is simply a small amp that works, but it is wide open to variation, expansion, and experimentation.

Building your own gear is very empowering and puts you in control, and if you are not satisfied you can rip into it and give it a rewiring it'll never forget. Similarly if it needs repair you have the circuit and already know where everything is and what it does.

The AVA100 specifically is intended to contain enough instructive information and grounding for you to use it as a design guide if you want to do something different. Questions, suggestions and corrections are most welcome. (Please put “AVA100” somewhere in the subject line)

Preface

Over time a number of people have written to me about an up-to-date design for an Aussie valve amp, pointing out the success of projects such as the AX84 and Spitfire. [link sites new window]

Silicon Chip responded dismissivly to one reader suggestion of a valve guitar amp design, saying such a project would end up “costing over $1000”.

But would it really?

I have already written favorably of their low distortion and high power solid-state amp designs, and if I were building a home Hi-Fi I'd be looking there.

But frankly the valve preamp debacle and outfall demonstrated that they don't have the first clue about guitar amplification; that amps ain't amps Sol.

---

Preamble

When I started messing with guitar amps in my teens there were few specialist parts and most were built using components common to the black and white all-valve TV's of the day, with a repair shop in just about every shopping strip. They may be exotic today, but they weren't back then.

Many a rig was build out of dead TV's and radiograms that had become unfashionable. They still could be, but now there is a better way.

Since then the valve tide has gone all the way out, only to come back in again in the past few years thanks to some extent to the Golden Eared Hi-Fi crowd.

A problem with current valve designs has been the transformers, the output transformer in particular, the valves themselves and bases, and associated high voltage capacitors and medium wattage resistors.

Component suppliers have become totally oriented to low-voltage high-current solid-state, but gradually some of the required components were creeping back into stocklists.

Then Paul Cambie made the happy discovery that cheap and available 100 Volt Line transformers made very acceptable output trannies, and the picture changed dramatically. He used 6CM5's but they seem equally applicable to several other type including 6V6 and 6GW8. Push-pull only at the moment I'm afraid.

There are still some questions to be resolved with a suitable power supply, but it is once again possible for the keen teenager to buy all the required components to build a guitar amp over the counter at reasonable prices and, as I and my friends did, again assemble a range of quite reasonable guitar amps.

In numbers, initial costings for a basic amp look around $A250, but there are also some expansions and offsets we'll explore later.

Since the 60's and 70's guitarists have also changed what they want from an amp, and today tend to be just as interested in a smallish amp with good tone as in the fire-breathing monsters of yore.

It is possible to build by stages, starting with a very basic amp and enhancing it as desired, so the design should be flexable and capable of producing good results even in the most minimal and inexpensive configuration, yet be capable of expansion by stages to a fully-featured version if desired, and open to creative variation to satisfy tastes.

---

The Design Process

In theory a design starts with the required specification and proceeds in linear steps to the working prototype. This spec may come from a boss, client, or your own desires.

In the real world the designer has to deal with something called parts availability. This normally means a snowstorm of catalogues and data sheets, going back and forward, making compromises between what you want and what you can get, and trying different arrangements. And that's what mainly drives this design and its variants - what we can get.

This design can make good use of scrounge and a well-fed junk box, but they are not required and it is assumed the average builder is starting off with nothing but a strong desire.

Designing something for yourself is one thing (you can make all sorts of dodgy compromises), but producing a design intended to be built by others introduces considerations of cost, future parts availability, and hopefully accessablity to home construction with minimal tools and skills.

The word design itself covers a wide range of sins, from requiring a whole new valve or transistor, to following the well-worn path of one of the typical specifications given in datasheets. This design is very much the latter and not intended to be adventurious or ground-breaking - quite the opposite.

I find spreadsheets like Visi-Calc, Lotus 1-2-3, As-Easy-As (DOS version) and Excel very useful for listing and comparing components such as transformers. This makes it quite simple to change a component and quickly see the impact on computed voltages, currents and powers, as well as costs.

So instead of it being a straight line the design process becomes recursive, in other words you go around in circles trying to find the best combination of all factors. There is seldom a Stop! sign and it is a matter of judgement when enough things have been considered and it's time to start drilling holes and soldering; to throw your ideas up against the Universe and see what sticks (old engineers' saying). But that is getting ahead of ourselves.

Pretty soon a set of project constraints emerge. One particular constraint on this project is that it can be constructed without having a big junk box or pile of scrounge to fall back on - all parts have to be purchasable from normal outlets at reasonable cost; no expensive specially imported or hand-wound output transformer, or the like, required.

---

Amp design constraints

In this case some of the most critical choices are almost made before you start.

Output transfomer

The key one is the output transformer, and since the choice is rather limited we can use it as a starting point.

This effectively rules out a single-ended Class-A design due to lack of a suitable transformer.

So our attention is directed towards a smallish basic push-pull design around 10 to 15 watts output.

Output valves

Another limitation is the choice of valves. If we decide to limit our choice to those valves still in production and available new, as a magazine would, then this boils down to 6CA7, 6V6, and 6GW8.

As it happens either the 6V6 or the 6GW8 are a reasonable fit to the available transformer choices, and both have been used in classic period amps. But with each 6GW8 output pentode you get half a 12AX7 free! Which is why it got used in a range of different smaller amps, Maton, Playmaster, Diason, Barclay and others.

Power supply

Thanks to SMPS's there are high voltage electrolytic capacitors available today that we could only dream of. Bigger capacity, low ESR/high ripple current rating, small size, and cheap; similarly rectifier diodes (free even).

Power transformers, however, are still a different matter.

We require, at a minimum, a heater voltage winding, and a high voltage winding. At this point (Oct'06) such an animal is generally unobtainium. If you do happen to have a suitable tranny you can certainly use it, but we'll try to keep luck out of the picture.

We can get a tranny that will run a few heaters, the trusty 2155 type, but the HT requires some wangling.

One answer, as used in the SC Mudlark Hi-Fi valve amp, is to use two trannies back-to-back.

You could also do what they did with the valve preamp and build an inverter, or more sensibly convert a computer SMPS to produce HT.

[article ref]

Another, lower noise and simpler, method is to use an available lowish voltage tranny and follow it with a voltage-multiplier rectifier. In this case this means a quadrupler (x4) driven by a 50-70Vrms trannie to give 250-300VDC at full load of 100-200mA.

Doubling was actually used widely in later amps, and particularly in those featuring 6CA7/EL34's I expect to find a voltage doubler.

Because of the needs of solid-state amps our main problem (and ironic paradox) here is choosing a tranny with a high enough voltage that isn't vastly bigger in wattage (weigh and cost) than required. This naturally prompts evil thoughts of an amp that can make full use of a 300VA (watt) tranny, but...

Following the theory of starting small and working up, we will start out with the most minimal practical amp and progress from there.

---

Introduction

Initial Working Specification

Object:

Design, produce and publish circuits and templates that can be used to construct simple to fully-featured guitar amps in any desired physical format.

Constraints:

To use only cheap, available over-the-counter parts.

To require minimal metal-bashing, no folds required, only drilling small holes.

Target:

The intelligent, perhaps poor and poorly-equiped in tools and skills, but motivated; such as teen band members or friends.

Scope:

Project stages from minimal no-controls “toaster”; through a basic single channel; to a full-featured small PM103-like; to...?

Valves:

6GW8's 6L6's and 12AX7's.

Output:

Around 13 watts.

Chassis:

Small-drill-only no-bend no-punch flat-chassis modules, steel, ali, or even wood-and-alfoil baseplate style construction. Dead computer power supply cases.

Physical:

To suit combo, or head and cab, and mounting upright, suspended, or on-edge, and with front panels on back, top or front.

Options:

Preamps/EQ, tremolo, reverb, stereo

---

Modules

* Power supply - heaters, HT, bias
Options: minimal, middi, and maximal

* PI and output - templates for 2x 9-pin valve holders, 2x Octals, and a line output trannie, nominal 10-13 watts rms per channel.
Options: two output transformers, 4/8/16 ohm output, triode/pentode/ultralinear selectable, NFB, different PI arrangements
Panel: minimal controls - volume and one tone

* Preamp - 1x 12AX7
Options: plain channel (EQ), tremolo channel, reverb drive and recovery

* Front panel
Options: inputs, volume, tone | treble, bass | treble, middle, bass, trem speed, trem depth, reverb, split/reverb select, master volume, footwitch socket

* Back panel
Power in, mains switch, external speaker socket, line out, Fx loop

---

The Case

Physical Formats

* small combo (like Maton/Barclay), cubic, single speaker, controls up-facing top-back

* medium combo (like PM102, Strauss Mouse, Chandler), 2 speakers stacked vertically or horizontally

* Head and cab, (bottom mount Marshall-like, or top mount Bassking-like)

Bass considerations.

---

Speakers

1x 10 or 12,
2x 8, 10 or 12,
3x 8 or 10,
4x 8
(1 or 2 tweeters)

---

Process

Phase 0 AVA100 Setting up, a place, tools, materials, design and choices

AVA100PA
Loadlines, power, distortion
Cathode circuit
Screen circuit
Phase Inverters
NFB

AVA100PSU
Worked examples for small, medium and maximal, bias, thermo-fan, solid-state.

AVA100PRE
A simple front-end with heaps of gain and radical single-knob tone control.

---

Phase 1 AVA101
Minimal PSU (Small heater tranny and HT filter caps, no bias components) 1st o/p stage 2x 6GW8's cathode bias
Panel: vol and top cut (or TSC Big Muff one knob?)

---

Phase 2 AVA102
Preamp 2x 12AX7 (provision for 2xEQ, 2xtrem, reverb)
Panel: vol, treb, bass (speed, depth, reverb, footswitch)

---

Phase 3 AVA103
Maximise PSU, (second heater tranny)
2nd o/p stage 2nd pre
Panel: vol, treb, bass, vol, treb, bass, speed, depth, split/chain (line i/o?)

---

Phase 4 AVA104
30 watt
4x 6GW8's (similar to AC30)

---

Phase 5 AVA105
40 watt
2x 6CA7's cathode bias

---

Phase 6 AVA106
55 watt
2x 6CA7's fixed bias

---

Aside...

Bounty

I happend to be browsing what else you could use the PlayStation series games machines for and I came across a site that pays bounties to people who produce projects. This may be old news, but it's news to me.

It seems to work like this; people propose a project, perhaps a different operating system, and other people who want this to happen paypal donations to a bounty. This bounty grows in speed and proportion to the interest and the first programmer to submit a working solution gets the bounty (and any following gratitude contributions). Contributors are listed from hundreds of dollars down to literally cents.

Will this evolve into the new Copyright?

<<<OzValveAmps