EL84 OPT

[Joel]

I know, I know!  Another output transformer thread....

I'm looking at doing a 15W version of a Matchless DC30 (only 2 x EL84's), and I have an old 5.8k to 8ohm output transformer (calculated).

The EL84 datasheet examples say 8k primary.  My investigations of 15W-ish EL84 powered amps show;
 * Marshall 18W amps use 8.4k primaries, and
 * Classic tone and Hammond replacement 'vintage' spec transformers for Vox AC15 specify 6k and 6.2k primaries respectively.  (Even though some of the AC15 schematics I've read spec 8k primaries).

Question for the guru's:  Is my 5.8k transformer OK for an EL84 powered Matchless-style amp?  Would you?

[Bieworm]

If you like OPT distortion a lot the 5.8k would do the job.
That's closer to kt88 spec than el84.

[sluckey]

I'd use it. Easy to change if you don't like it.

[pdf64]

I wouldn’t, because Matchless tend to idle hot anyway. And lower impedances tend to have a lower conduction angle, so with cathode bias, even hotter bias would be needed to bring the conduction angle back up and so limit excessive bias shift at high power outputs.
But you may have an old PT that’ll only put 250V on the EL84, so need to put some numbers on the other variables and see how the load lines pan out.

[DummyLoad]

I know, I know!  Another output transformer thread....

I'm looking at doing a 15W version of a Matchless DC30 (only 2 x EL84's), and I have an old 5.8k to 8ohm output transformer (calculated).

The EL84 datasheet examples say 8k primary.  My investigations of 15W-ish EL84 powered amps show;
 * Marshall 18W amps use 8.4k primaries, and
 * Classic tone and Hammond replacement 'vintage' spec transformers for Vox AC15 specify 6k and 6.2k primaries respectively.  (Even though some of the AC15 schematics I've read spec 8k primaries).

Question for the guru's:  Is my 5.8k transformer OK for an EL84 powered Matchless-style amp?  Would you?

maybe use it with 2 pair. output would be < 30W and depend on B+, etc.. what's was the OT original installation? power rating? 

--pete

[Joel]

Thanks for the input everyone.

I should have mentioned that it looks like a 15W transformer.  It's about the right size for 15W.  I don't know where it came from.  I got it in a box of parts somewhere along the line.  It's been sitting in my parts drawer for a long time.  The only markings on it are GTC.1194 and a hand written 6V6.  It would probably suit 6V6's better,   but a 6V6 powered DC30-ish amp probably wouldn't sound the same.

[HotBluePlates]

... Is my 5.8k transformer OK for an EL84 powered Matchless-style amp?  Would you?

You're only asking half the question. The other piece of the question is what power supply voltage will you have.  The two go together to set a limit on output tube plate current (or at least a limit before the tube runs into distortion).

Knowing the supply voltage that will be present gives you a chance to check pdf64's concern that the output tubes will over-heat during part of the signal cycle.  If the supply voltage is low enough, it won't be an issue.

... The EL84 datasheet examples say 8k primary. 

They do?  This Philips EL84 data sheet does say 8kΩ, but also a bunch of different things from 3.5kΩ up to 10kΩ, and all around same/similar supply voltages.

You can actually-design, or you can copy an existing design.  If you do the latter, you'll find it hard to use scrounge-parts unless they happen to have the exact same specs as the original-part.

[DummyLoad]

couple of 6L6G/A/B/C in P-P running class A₁ would work with that OT - it's a closer match than for 6V6 or EL84.



250-270V to anodes - Rk ~120R, should make about 12-15W. 6L6s' would be loafing and should last a long time.


--pete

[DummyLoad]

a pair of 6L6G/A/B/C in P-P running class A1 would work with that OT - it's a closer match than for 6V6 or EL84.

250-270V to anodes - Rk ~120R, should make about 12-15W. 6L6s' would be loafing and should last a long time.

--pete

[PRR]

...This Philips EL84 data sheet does say 8kΩ, but also a bunch of different things from 3.5kΩ up to 10kΩ, and all around same/similar supply voltages...

Most of those are single-ended. (DC30 is push-pull, right?) The four push-pull pentode cases are all 8k; the couple of push-pull triodes say 10k.

[AmberB]

A pair of EL-34s with lower plate voltage would probably work with that OPT as well...

[Backwoods Joe]

Try what you have. Study the tube data. Make sure you don't go over the voltage & current ratings. "Rumor has it" Matchless used output transformers with lower primary windings. Don't know that for a fact but that is supposed to be part of the Matchless sound...

[Joel]

I've done a bit of research and drawing.  I used http://www.valvewizard.co.uk/pp.html and https://robrobinette.com/Drawing_Tube_Load_Lines.htm as dummies guides.  Note: a lot of articles about drawing load lines talk about a thing called the "knee" yet no one bothers to explain what this "knee" is.  What is the knee?

The attached picture is what I came up with using 300V anode voltage with my 5.8k output transformer.  It looks alot like Mr Robs 5E3 6V6 example on his website so I have to assume that having segments of your load lines over the max dissipation curve is OK.

The challenge now is finding a power transformer that will give me that low a B+ voltage with the required current capacity.

[DummyLoad]

The challenge now is finding a power transformer that will give me that low a B+ voltage with the required current capacity.

that can be these days.

most of what you find off the shelf puts B+ at well over 300V+. one approach would be to use a suitable EL84 P-P PT with a 5Y3GT rectifier. that should get things down around 275V and keep 6L6 in class A₁ operation. the preamps will be running at 250V or lower, depending on your power rail design. use larger filters with lower values for the series dropping resistors - e.g., use 3.3K-4.7K instead 10K-22K as commonly used. i ran a PSUD-II sim on a hammond 270ex and attached the results. datasheet for 6L6GB class A₁ also attached, and i'd suggest the use of the cathode bias plan. you could use fixed bias, but you'd need to use the early marshall bias PS with the hammond PT. another PT choice is the Hammond 261M6 - simulation is about 285V with SS bridge. either get you very close to where you'd want to be B+ wise with a low-volt 6L6.   

--pete

[PRR]

...What is the knee?...

[HotBluePlates]

Most of those are single-ended. (DC30 is push-pull, right?) ...

You're correct of course, though that's not so much my point.  I'm pointing out that if "max power output" is not a requirement, a range of different loads are acceptable pending a check that the tubes won't overheat with the range of probable drive signals.


I had a mental-block about transformer loads for a lotta years, on facts that should have been "obvious" after a week of reading.  Like many on forums, I didn't see the tube/transformer relationship as being very-simple "Volts, Resistance, Current" and tended to think in the terms I now see others repeat:  "the EL84 wants 8kΩ."

Folks fall into this trap and think there's one correct load, regardless of the supply voltage or class of operation, or even whether the builder needs/wants max output power.

On those other forums, I've also seen near-universal need for a way to deliver less power to the speakers: guys are chasing every combination of attenuators, reactive loads, master volumes, re-amp solutions, load boxes because they can't use their amp's native volume-before-distortion.


So Joel wants to use the junk-box transformer.  Can he?  Probably, as long as "all the watts" isn't a goal, and maybe the supply voltage is brought down from stock (though I'd have to spend more time investigating the design than I really have to devote).

I'm thinking of it somewhat like putting Prius wheels a pickup:  it's less capable, but as long as the drivetrain isn't scraping the ground Joel can use the Prius-wheels he's got, and he might be okay with the outcome.  If Joel wants a full-capability truck, he shouldn't bind himself to the sets of wheels sitting in the garage.

... The attached picture is what I came up with using 300V anode voltage with my 5.8k output transformer.  ...

The challenge now is finding a power transformer that will give me that low a B+ voltage with the required current capacity.

Kudos on diving into load lines!

~300v is where my vintage AC30s and AC10 sit.  They use "proper load" transformers.

Now let's be mentally-flexible and ask, "What about 200-250v?"  Consider re-working your loadlines with an eye moving the B+ down into that range, where it no longer strays above the dissipation curve (though a brief excursion above the curve is not a big deal in practice).

You will learn some stuff from this:

     -  "A good OT load" is whatever seems to work as a balance between the tube's (plate current) capabilities and the supply voltage to be used.

     -  You'll have to go explore power transformer options to see if "the bright idea" is supported by available parts.

     -  The "cost savings" of using junk-box parts can often be overwhelmed by the expense of changing other parts of the amp-design, or in the time burden of re-thinking the amp-design.

[Joel]

Thanks for the reply Mr. HotBluePlates.

If we start dropping the B+ too far we begin changing the operation of the amp as a whole too much.  I must say I've learned a few things about power amp design thanks to you fine gents.  Whilst I can use my junker OPT, the trade-off is essentially redesigning the whole power supply.  I'm OK with that, and I think I could roll the power transformer myself.  But my goal is an amp that is close to the Matchless DC30 sound/feel.  If I go ahead and use my OPT and redesign the output section and power supply I'll be making a different amp. 

I'll just buy myself the Marshall 18W iron (which seems to be used on all the 15W clone versions of the DC30 these days) and save the junker OPT for another 6V6 project.  Of course, this means I'll have to build more amps... 1st world problems hey?

Thanks again.  I do appreciate the responses and mentoring attitude displayed on this forum.

[Backwoods Joe]

If you want to be close to the Matchless sound... MojoTone has the Vox AC-15 iron. By the way the OT primary is 6K.

[HotBluePlates]

... MojoTone has the Vox AC-15 iron. By the way the OT primary is 6K.

That seems wrong to me.

Firstly because the AC15 schematic says "8kΩ".

Secondly because I've measured my 1965 AC10 at ~8.1-8.2kΩ.  Best I can tell, the AC10 had the same output stage as the AC15, except for the presence of plate resistors on the EL84s to waste a little output power.

[Joel]

This is what originally had me confused.  The various "vintage" repro output transformers made by Mojotone, Classictone, Hammond, etc all have 6k primaries.  The associated "vintage" repro power transformers all are 290V or 300V CT.  Yet, at the same time, older AC15 schematics which actually specify the OPT primary impedance say 8k.

According to my interpretation of load lines I drew, when using the "vintage" repro power and output transformers the output tubes should be burning up in short order. 

The various amp kits available online tend to use Marshall 18W repro transformers (290V CT power, and 8.4k OPT primary) which is a good pairing.  The Trinity kit is a good example, yet even it has a switch to go from 8k to 5k OPT primary windings...  Again should burn those output tubes right up according to the load lines. 

So there must be some mysterious mojo magic going on, or I'm missing something when drawing my load lines and they pass over max dissipation.  Or the 6k value is in error and everyone has blindly just kept making "vintage" spec output transformers without ever actually measuring many of them.

[AmberB]

There is another tube that you might consider for that 5.8K output transformer.  the 7591A is listed in my RCA tube manual as using a 6.6k OT with either a 350 volt B+ and fixed bias for 30 watts, or 450 volt and fixed bias for 45 watts, which is probably more than you want.  With 450 volts and cathode biased with a 200 ohm bias resistor, the output is listed at 28 watts.  I'm guessing the output would be even less with 350 volts B+ and cathode bias.

[PRR]

Tube amps do not "burn up" for 25% lower load impedance. We don't play them THAT hard that long. And self-bias amps tend to de-bias when worked on low Z loads.

The 7591 is a significantly bigger tube than any mentioned in this thread. More like a high-gain 6L6 (which certainly pulls 6K loads). And 7591 specifically died in the 1980s; I gave-away an amp when I realized they were not making any more and NOS was scarce and overprice. There "are" new-made 7591 today but they are just 6L6 shimmed a little closer than good sense suggests.

[HotBluePlates]

This is what originally had me confused.  The various "vintage" repro output transformers made by Mojotone, Classictone, Hammond, etc all have 6k primaries.  The associated "vintage" repro power transformers all are 290V or 300V CT.  Yet, at the same time, older AC15 schematics which actually specify the OPT primary impedance say 8k. ...

I've noticed similar, but with respect to power transformers:  The power transformers in many of my vintage amps have higher winding resistance than found in some modern "reproductions" which means their output voltage & resulting d.c. volts will be lower than the modern part.  My 1964 Deluxe Reverb comes in a bit under 400vdc on the 6V6s with 120vac at the wall outlet (and 120vac at the wall gives 6.3vac on the heaters).