Torturing an OT what's the limit?

[nashvillebill]

Here's the background: I've picked up an old Heathit AA-50 stereo integrated amp. Each side has a pair of 7591's running 420 volts B+ and 360V on screens. Because the amp has cosmetic damage, it was not well constructed, all caps are ancient, and several tubes are cracked, it's hardly worth salvaging. I'm going to re-purpose the PT and OT's which appear to be very stout iron.

Now, while it's tempting to just build a guitar amp using the same OT and same 7591's in the same P-P, looking at the hefty OT makes me wonder...how much abuse could the OT take if I used a pair of 6L6's rather than the 7591's? (Plenty of heater current from the PT since it was powering a stereo integrated amp). Or pushed it even further...use SS rectification instead of the GZ34/5AR4 to raise B+ up AND run the 6l6's... But rather than just experiment, I'd kinda like to learn what I'm doing before I do it.

In other words, what REALLY determines the power limit of an OT? For an audio amp, would it be the power at which the manufacturer feels it saturates and becomes non-linear? As opposed to a guitar amp in which the designer may wish to saturate the OT? Assuming of course we have not exceeded the breakdown voltage of the transformer winding insulation.

As always, thanks for edumacating me!

[tedmich]

some people like the new JJ 7591's and the older (real) 7591s have a ton of fans for guitar. The old Ampeg Rockets were 7591 based and many have converted them to 6V6GT I believe. The OT should be like 7.6K at the primary, perfect for EL84, but a bit high for most bigger bottles, although they can work. OTs are quite robust if not run FULL on, but radically higher voltages can kill them, especially if old.

[R.G.]

In other words, what REALLY determines the power limit of an OT? For an audio amp, would it be the power at which the manufacturer feels it saturates and becomes non-linear? As opposed to a guitar amp in which the designer may wish to saturate the OT? Assuming of course we have not exceeded the breakdown voltage of the transformer winding insulation.

The power limit of an OT - or any transformer at all - is REALLY determined by (1) how hot it gets, because this can cause the insulation on the wires to break down and cause an internal short, and (2) the maximum voltage across the closest approach of two current carrying wires inside the transformer under a voltage spike or overheat condition. That is - the temperature rating of the insulation and whether there is a voltage spike or mechanical vibration that wears through insulation.

Transformer makers worry about the internal hot spot temperature. The iron and copper will work at temperatures where smoking-hot motor oil is a good coolant. But when the insulation fails, the electrical arc that forms will burn the copper through, and the transformer dies. The limit is temperature versus the rated insulation temperature.

Which is not what you wanted to hear. Saturation doesn't matter as long as the current is limited so the resulting higher current doesn't overheat the insulation. It won't necessarily sound good there, but saturation isn't death. Overheating is.

The way the pros tell temperature is to run the trannie at the anticipated max conditions for about six or eight hours so it gets fully hot - it takes a long time to stabilize that much mass thermally - and measure the change of the resistance of the copper winding wire. Copper resistance changes by 2.7 nanoblems per degree, or something like that.

[nashvillebill]

That makes sense RG, I typically tend to think only about the PT temperature and not really much about the OT getting hot. So...I guess I stick to the orginal 7591's, which as tedmich pointed out, have become kind of a fad.

I don't think I'll be using the original 7199 phase inverter, though....even though I have a pair of them, they seem to be unobtainable today. Guess I'll be studying Merlin's book some more, as I'd like more drive than the Ampeg Rockets.

[loudthud]

I had one of those amps back in the 80's. Re-capped it, installed a solid state rectifier, cut the chassis down and rack mounted it. Kept the original 7591/7199 circuit. I used it to drive the horns in a PA system. It always sounded real good, sort of mellowed out the horns. Never heard it clip no matter how loud we pushed the system. IIRC it only made about 22W per channel.

[Steve Conner]

Edit: whoops, I got my tubes mixed up.

[Chuck H]

Also. consider that hi fi OT's are intended to reproduce a full frequency range. Not of much consequence re: the HF but huge demands with WRT the LF. A guitar amp won't demand nearly the same current in the LF. Hammond, for example, actually told me that their off the shelf models can be used at double the published spec when used in guitar amps. I have an amp that I pretty much run wide open all the time. I built it a little over four ears ago. It uses a Hammond 1608 (spec-10watts) with a pair of el84's at 360V biased at 100%. The OT only ever gets warm. Never hot.

I also have a pair of old Fisher OT's that were pulled from a "The Fisher" 500B. Each OT ran a pair of 7591's. One of these OT's is in an old Traynor chassis operating a pair of 6l6's at 460Vp with no trouble and sounds fabulous. The other is on the shelf awaiting a project.

The original OT spec for the Trainwreck Express (a pair of el34's @ 420Vp) was a Stancor (A-3801?) with a spec of 30 watts. I've never heard anyone complain about an OT failure in those amps and they obviously live dimed since those amps are sort of one trick ponies.

So. I can't say WRT your OT. But my past experience is that if it looks robust and was intended for 25 or 30 watts in a hi fi amp, it should work fine (even cruise) at 40W or 50W in a guitar amp.

OTOH 7591's are said to sound great. Similar to el34's. Some say better. I've never built with them.

JM2C

[Wombaticus]

Chuck makes a good point. So if you do decide to try out some bigger bottles, consider trying the KT-66 instead of 6L6s. KT-66 tubes like much higher plate loads than other big bottles. If I recall correctly, JTM45 OTs were in the range of 6.6k - 8k depending on what manufacturer Marshall was using at the moment....

[imaradiostar]

I would be wary of using KT66 with a PT originally intended for 7591's only because of the increased filament current draw. If you're building a guitar amp and use the stereo PT with only one of the OT's you'll be fine, of course. I love the way KT66's sound- the Beano album is one of my favorites.

You can't go wrong with JJ 6v6s or the EH 7591a. I've used both and while some would argue that neither are "vintage correct" they both have a lot of usable tones and can handle the voltages and plate impedances you're likely to see with your transformers.

Jamie

[Wombaticus]

I would be wary of using KT66 with a PT originally intended for 7591's only because of the increased filament current draw. If you're building a guitar amp and use the stereo PT with only one of the OT's you'll be fine, of course. I love the way KT66's sound- the Beano album is one of my favorites.

True, but the OP suggested in the first post that there was ample heater current.

[nashvillebill]

Exactly, I was going to use the original stereo PT but only one of the OT's. So plenty of heater current.

Great replies guys, I really appreciate the thoughtful responses.

[loudthud]

Heater current:

KT66 1.27A
6L6GC 0.9A
7591A 0.8A

7591A data is from an old RCA data sheet. Can't find one for EH. JJ says 0.8A

[Wombaticus]

Here's another fun thought. Consider running the two OTs in parallel, a la the earliest JTM45/100's. Marshall was initially unable to source a 100W OT, so just paralleled a couple of smaller ones. An expensive hack, but it worked. This solves the power handling conundrum *and* the big-bottle impedance matching issue in one fell swoop.

[Gingertube]

Guys,
Re-use of HiFi'ish Output Trannies for a Guitar Amp is something that falls into the cracks in my knowledge.
I'm one of those guys who take the philosophy that there is no shame in being ignorant, only in remaining ignorant.
So some questions!!!

Can I just say we have a power bandwidth, an if I reduce the bandwidth I can push more power through it in a direct linear relationship?

Example: A Hammond 1608 rated at 10W 30Hz to 30kHz in Ultralinear for HiFI.
For Guitar Amp I don't need anytghing below 60Hz so I can push 20 Watts out of it - YES/NO?

In addition it is my limited understanding that trannies have a max idle current spec and if you exceed it then you get a primary inductance reduction even before you hit saturation, that is going to drop the bottom end off a bit anyway.

Is there any theory/formulae to apply or is it in the just try it and see realm?

My only direct experience is use of a cheap Chinese HiFi Tranny. It is an 8K Raa with UL taps designed for 10W from a pair of EL84 in Ultralinear. It sounds just lovely in a PP 6V6 (Pentode Mode) Guitar Amp with around 15 Watts output.

Cheers,
Ian

[R.G.]

Can I just say we have a power bandwidth, an if I reduce the bandwidth I can push more power through it in a direct linear relationship?

Close, but not exactly.

Since the power handling ability of a transformer is really a thermal one, you have to look at what causes the heat. The two major causes of heat in a transformer are resistive losses in windings and core losses from eddy currents. Resistive losses go up with current. Core losses go up with frequency, regardless of the power through the wires. The edge of saturation, which causes primary current to skyrocket, goes up with frequency too. So a 100W transformer at 60Hz can handle roughly twice that at 120 Hz because the voltage it can withstand at 120Hz without saturating is doubled. This is what makes for tiny switching power supplies.

Example: A Hammond 1608 rated at 10W 30Hz to 30kHz in Ultralinear for HiFI.
For Guitar Amp I don't need anytghing below 60Hz so I can push 20 Watts out of it - YES/NO?

Yes - with the proviso that you tinker the load impedance to take 1.4 times the voltage at the same currents through the wires. This doubles the load power (V squared divided by R) while keeping the same currents (V divided by R). Of course, you have to have twice the voltage on the primary to put through it to get this operation.

In addition it is my limited understanding that trannies have a max idle current spec and if you exceed it then you get a primary inductance reduction even before you hit saturation, that is going to drop the bottom end off a bit anyway.

It's almost that true, but complex. Magnetic saturation is reached by the flux density getting to X. This can be expressed as a volt-time integral for signal, or ampere-turns product; They're two ways of saying the same thing. In an SE transformer, it's absolutely true; the idle current needs to sit near the middle of the BH curve so it can swing equal(ish) amounts above and below the bias point. In class-A push-pull, it's kind of true(ish) except that the idle currents cancel magnetically in the core, so there is no magnetic limit on them. There are other limits, but not core saturation ones.

In class AB pushpull, it's more special case. There are equal and opposite idle currents at zero signal, and those should come close to cancelling. The volt*time integral on signals adds to or subtracts from the static idle currents in terms of flux density in the core. As long as the instantaneous sum of idle current times turns plus volt*time integral for both halves of the winding are below the start of saturation, you're good.

Is there any theory/formulae to apply or is it in the just try it and see realm?

Yes; lots of theory and formulae. Too many to achieve closed-form solutions in most cases. Transformer designers' most precious possessions are the accumulated test results on their past designs. This tells them how their designs really worked in a situations with more equations than can be cleanly solved.

I wish I had a simpler answer for you, but those are the facts.