Old 2xEL34 PA Amp - what should I build?

[SixStringing]

Thanks Nuke - sounds like you're onto something. There are a LOT of wires going into the Ouput Transformer (12 total - 6 'taps'?).

Most of those white wires are labeled w/ color codes (eg. RED YLW), but with all that tidy hand-tied wire bundling, it's hard to tell where everything is going. It's been a process of elimination for the most part. I am hopeful for an actual schematic, but hey - it's been a learning process.

So you think the 6AU6 circuit is acting as a limiter to the power section? I assume to reduce distortion - or maybe better worded as 'to reduce the ability to get to a distorted state'? Not necessarily what this guitarist wants to hear

One of the first mods I'll be making is to replace one of the XLR inputs with a 1/4" (the first mod being a 3-prong power cord). I may even be able to test with a guitar first - I think I can pull out one of the plug-in input transformers and insert a jumper wire to connect directly to the input stage.

And yes, that Reverb listing is the one. It was fairly local - we're in the same province.

[nuke]

It's possible to do ultra-linear two ways: By putting screen taps at 42% of the plate windings. So you'd get the following order:

Plate 1 / Screen 1 / Center Tap B+ / Screen 2 / Plate 2

And if there's cathode coupling winding:

Cathode 1 / Center Tap ground / Cathode 2

That would account for 7 wires on the primary.


Or, you can do the screens on their own windings with a center tap to the screen supply power node.

The transformer would then be:

Plates Winding:
Plate 1 / Center Tap B+ / Plate 2


Screens Winding:

Screen 1 / Center Tap B2+ / Screen 2


If you also had cathode coupling winding:

Cathode 1 / Center Tap ground / Cathode 2

That would account for 9 wires.

It's also possible to arrange the windows with the center taps brought out as separate wires.


On the input side, pop out one of the input transformers and take a 1/4" input jack on the 1-meg resistor. Yellow to the tip, blue to the sleeve and plug your guitar in there.

input-xfrmr.jpg

[SixStringing]

Well, thanks to Ingenium Canada (the Canadian Science and Technology Museum) I now have an actual schematic and manual!!!

I've attached a .pdf of the manual as well as a .jpg of the extracted schematic.

Hopefully someone can chime in on what's going on with the output transformer. And why so many B+ feeds.

Northern Electric PA-35A amplifier service manual AUDIO N8747 3006 1958-1964.jpg

[trobbins]

The output stage is likely a unity coupled cathode-anode configuration with a form or UL screen drive, and a form of grid coupling. I don't recall seeing all those schemes in the one output stage before.

You'd need to measure the turns ratios of the cathode and anode windings to see if they were unity. McIntosh brand are perhaps the most well known user of unity coupling scheme. Two aussie PA amps used the unity coupled scheme, but with cathode bias, and a fixed screen voltage relative to the cathode.

The negative bias is also used to suppress preamp stage hum by 'elevating' the heater voltage.

Local feedback is applied from the V5B post-MV triode gain stage, through the PI and driver stages, and including some portion of the output transformer signal via the grid winding. There is also local feedback by using the screen winding to supply the driver anode.

The winding phases on the schematic relate to the change in phasing need for grid, screen and anode.

Perhaps the screen winding is separate to avoid UL patent conflict, given the 1958 date.

The degree of interrelated feedback in the output stage would present a major risk to tinkering without lots of effort to confirm if any change has merit. Perhaps checking how well the output stage can be over-driven would be a good aim.

The various inputs likely provide a good way to make simple changes for alternate guitar inputs with different gain and overdrive designs, and tone adjustments. There are 4 gain control pots before the MV pot, which could allow various reconfigured signal paths for hi-gain overdrive control before going back into the tone control section. Input jacks along the signal path could allow different points of signal entry for different gain/overdrive options.

[nuke]

WOW!!!

I was not expecting so many windings. That takes the cake as the most thoroughly over-engineered output transformer in an audio amp that I've ever seen.

No less than 5 windings. I did not expect the 4th winding (the xxx-GRN winding). That's a new one for me, with the C- on the CT. Also a concertina phase-inverter, then pentodes to drive the grids.

All the additional primary windings on the OPT are for negative feedback purposes. You have the usual plate primary windings, then the winding for ultra-linear screens. The cathode coupled winding as mentioned in the thread provides negative feedback to the output tubes. The last one to the grids is a new one for me. Can't say I recall ever seeing that one.

Then there's negative feedback from the grid winding, all the way back to the cathode of V5b.

NFB gives the power amp low distortion, but I think the idea was to make it very stable to drive highly reactive loads, like widely spaced speakers on 70v distribution lines.

The power supply is complex partly to feed all those windings, plate, screen (ultra-linear), screen grids for the 6AU6, the PI and it's preamp stage in V5, two stages on intermediate preamp and mixing amps, then the three mic preamps and the phono preamp.

That was a very thorough engineering department!!! Probably got used to to doing military contracts or something before turning out PA amps.

Pop out one of those input transformers, they're just octal sockets. Put your guitar jack at pin-6 and the ground at pin 7 and see how she sounds!

[SixStringing]

I put in a 3-prong power cord, hooked up a guitar (pin 6/7 of one of the input transformer sockets) and took it for a short spin.

It's still very low/mid-low frequency heavy compared to a guitar amp. The 'speech filter' switch does a pretty good job of cutting down the lower frequencies to bring it closer to guitar territory.

It's very clean. And loud. There's a significant 'thumpiness' in the bass when you dig in to the bottom strings and barely any high-frequency 'sparkle'.

Turning it up (both input and master) makes it louder but still clean - no hint of breakup. Turn it up more, still louder and still clean.

Push it a bit more and it finally distorts, but not in a good way. It comes suddenly, instantly and all at once - like a gate. When the initial attack signal from the guitar fades a bit, instant 'off' of distortion and clean again. The distortion itself is also not the good kind - it's super raspy, almost like mixed with white-noise? Extreme clipping maybe??

The preamp amp section doesn't seem to be providing any distortion (probably by design), and I'm guessing that the common cathode/PI stage starts puking once pushed. All the volume controls are before the cathode follower/PI, so as a result I couldn't really drive the power section very hard to determine how (if?) it would distort.

So now what? I feel I'm in a bit of a chicken-egg situation. As a start, it probably makes sense to repurpose a couple input channels into one multi-gain stage with drive and volume controls and see what that sounds like. But at what point do I commit and replace ALL THOSE CAPACITORS before knowing what I can expect from the PI/power section?

/Dave

[TUBEDUDE]

The winding for each element tightly couples the tubes to the transformer for better efficiency and control of the tubes. Great in hifi to clean response up to the limit you may have just found. Most guitarists dislike the response of ultralinear or unity coupling, and prefer the gradual increase of distortion in strict pentode mode.
The response may be more to your liking if you use just the plate coil and leave the rest unused.

[nuke]

Yeah, agreed. They were trying to make that amp ultra-stable and clean. Being ultra-linear, lots of power too. I haven't looked at the EQ, but it is probably optimized for voice vs. music, except maybe the phonograph input.

Could do a Dumble-ish ODS-ish thing, with loads of clean power and two-stage distortion stage. Could be kinda interesting. Take advantage of that power and clean, with a smooth, creamy pre-amp distortion.

Lots of tube sockets and knobs available.

[trobbins]

Rearranging the preamp stages is likely the simplest effort - in a technical sense. You could even use a pre-out to check what any re-arrangements sound like.

Changing the output stages is more technically demanding. You could revert to simple pentode mode, but you'd need to confirm the turns ratios of the winding sections. It may be practical to shift the cathode winding into the anode winding in some manner, but that presumes the cathode windings have sufficient hv insulation, and the turns ratios and phasing make it easy to either series or parallel into the anode windings. Using just the anode windings would need the impedance ratios to be assessed. Having a lot of 'dark' copper is not such a concern for guitar output transformers, as they typically won't use global feedback, and coupling and bandwidth aren't such a concern, and the 70V winding part won't get used anyway. But you will need to make circuit decisions, and confirm the bias and screen operation are not going to quickly kill the valves. Lots of detail to consider imho.

[nuke]

If you want to revert to a more conventional output section, it would be very simple to omit the cathode winding and the screen winding, just tape them off and leave them unconnected.

Ultra-linear was used in the last generation of silver face Fender high-power amps. With 2x EL-34 in ultra-linear mode, you can get up to maybe 100 watts give or take. I'd expect a bit less, given how conservatively engineered it is.

You can measure the winding ratios just by putting a small AC voltage on the speaker side and measuring the voltage on the primary side. The voltage ratio is the turns ratio, and the impedance is the square of course.


I might be tempted to leave the power amp alone and see what I can get it to do. The dumble-esque thing is to rely mostly on preamp distortion and a relatively powerful (aka, clean) output stage into a very clean speaker (EV12).

If you want to stick to mostly pre-amp distortion with plenty of clean power, you could do it pretty easy. Lots of preamp sockets, and plenty of front panel knobs to play with. You could have an easy 3-channel guitar amp. Clean, dirty 1 and dirty 2, and re-do the phono preamp as a dumble-esque kinda thing.

[didit]

Hmmm --

I would encourage you to basically leave it as-is except to bring it fully back to working condition. It's a rare well preserved relic and, as has already been discussed, a remarkable work of engineering and manufacture.

It'd make fine vocal or keyboard/synth amp.

..

[SixStringing]

Thanks everyone for your info and feedback. I'm in no rush (this will likely be a fall/winter project), so I'll have a bit of time to research and learn.

I do agree that this thing would likely be best suited to a pre-amp playground with the multitude of sockets and knobs, and best to leave the output stage as-is for now.

/Dave