6AN8 & EL84/6BQ5 Project Amplifier
Trialing 100V-Line Speaker Transformers as Push-pull Valve Amplifier Output Transformers

January 2002

Update May 2003: a PP amp by Grant Wills, also using the Altronics transformers!


You've heard of the Saturday Night Special? Rude, crude, blunt and dangerous. Well this is the story of my Sunday Afternoon Special.

I have some of my best cogitations on the dunny. Either that or at two in the morning, lying awake and worrying that I can't get to sleep; busy day tomorrow! Either case is a good opportunity to stew on whether or not that 100V-line transformer I saw in the Altronics cattle-dog would work as a push-pull output transformer.  See, it's got all these primary taps, and there must be some symmetry there because their ratings, "15W", "10W", "5W", "2.5W", "1.25W" suggest a pattern.  Hmmmm. Power ratings in Watts represent a voltage2 term.  So I wonder what that'd work out to . . . either a loud rap at the door, or the Sandman, must've claimed my attention, but the thought wasn't lost.

The 100V-line speaker transformer, that this project was a vehicle for.  Follow the hyperlink to Altronics, and search on "M 1115" - remember to leave a space between the "M" and the firt "1" when you write M 1115 . . ..

Years later, and having this thought-experiment return to mind many times, I sat down with pen and paper. It worked out; the windings carried a symmetry about the "5W" tap on the primary. And what was more, a couple of the intermediary windings were symmetrical about this tap too - and just at the right point to be used for "Ultralinear" screen taps. What an extraordinary coincidence!  I wonder if 100V line transformers were originally old push-pull valve amplifier transformers used for another purpose - maybe I was re-inventing the wheel, only in reverse! Dunno. More cogitating on the dunny.

At about this time I also became interested in something Queensland valve-audio enthusiast Darryl Lock had done. That was to disassemble the E's and I's in little locally available push-pull transformers, and reassemble them co-incident rather than interleaved - thereby making the transformer suitable for single-ended use. He'd applied this in a guitar amp rebuild successfully. Cheep, albeit low power, single-ended transformers are normally made of unobtainium where I come from, so I admired Darryl's innovation and felt quite inspired to continue with my M 1115 transformer experiments.

Transformer Experiments

One Sunday afternoon, I tried hooking the mains via my variac, set to a very low AC voltage, to the "8 Ohm" secondary of this 100V-line transformer. I measured the voltage levels and phase relationships for each tap on the primary. Bingo! My calc's were quantitatively off, but the symmetry of windings and the order I'd predicted was definitely there. At least in theory these transformers had pretty close to the character of a push-pull valve amplifier output transformer, and 8 Ohms on the secondary would reflect as a 4 kOhm-ish load at the primary.  Could be something here . . . . .

There was only ever going to be one way to find out what would happen if they were used in this perverse fashion. Obviously I needed to lash up a little 10W or so push-pull valve amplifier, and see. I didn't want to invest either time or money; chances were that either the transformers wouldn't work at all, or if they did, part of the primary would get fried in the first five minutes and the whole project would expire summarily, or the sound out would be horrific. So a little project was born.

I know! I could use that little chassis I'd saved from a valve amp that'd been robbed of too many bits by a professional repairer, and eventually consigned to the jumbo-bin, only to be rescued and salvaged by yours truely. Now there's not much room there, so what to build?

Amplifier Notions

My initial thoughts were to build a conventional three stage amplifier using an EF86 pentode as voltage amp, 12AU7 or similar as concertina phase splitter, and a pair of EL84's in push-pull in the output. The EL84's would give about the 10W output that I expected was what my output transformers could handle. The concertina phase splitter would give the asymmetrical overload characteristic that I've come to prefer for its sonic qualities, and the EF86 pentode would provide all the gain I'd need, easily and with less fuss than achieving same from a triode based stage. Very typical of this topology is this Mullard circuit, (albeit with the style of phase splitter I mentioned).

The "Classic" or "MkIII" Dynaco circuit (and the "MkII" circuit) had what I was thinking of. But instead of the EF86 voltage amp then 12AU7 (etc.) phase splitter, a compound valve (a.k.a. a "compactron"), the 6AN8, was being used. This valve was therefore worth a closer look. The Heatkit W7M included the 6AN8 in this way too.

First, I sourced some data for the 6AN8. From Eric Barbour at the Vacuum Tube Valley website came;

"6AN8; small triode-pentode, used in the old Dynaco Mark 3 amplifier and a few others. Out of production, though old stock is still available. A 6GH8, 6U8 or similar can be substituted, by changing the pin connections . . ."

I found also that the 6AN8 was also used in a number of Univox guitar amp circuits; in the U1000, U1011, U1221 and U1226 as a spring reverb line-driver.

Detail Articles; Table of Contents

The following series of little articles details the various sections and bits of the thing.


Schematic of the project amp.  Click here for full size image.


Some shots of the Sunday Afternoon Special. The amp works really, really well, and sounds just fine; it ain't the classiest or finest amp ever, but it works, works well, and is certainly no shocker. A very typical valve amp sound, warm and rich and effortless to listen to.

Ah . . .and cheep . . .did I mention cheep . . .?

And, more recently, this from veteran enthusiast John Hunter, 27.02.03; "I found yesterday the part of your website on using 100V line transformers in valve amplifiers . . . and I thought *I* was the only one doing that! Yes, they do work very well, and the multi-tap primary ones are excellent for ultralinear."

And now back to cogitating on the dunny.
Yeah, OK, it looks like it's been in the wars, but none of us is no oil paintin', right? Looking down on the chassis from above; the mains transformer and power supply electro's at left, and the M1115 transformers, being used as output transformers, at left.
The chassis from above; the chassis itself had been recovered from an old public address amp, formerly robbed for parts by a repairer of such things. Underside; point-to-point wiring, using the valve socket tags, capacitor and large resistor terminals, and tagstrips.