OTL amps for guitar duty

Status
Not open for further replies.
Not forgetting of course that sag created by the tube rectifier.
I agree totally that "sag" is very much a part of guitar amps designed for at least some genres of guitar music.

But I think "sag" is more like building an accidental compressor into the circuit. Not quite the same as creating THD because of a nonlinearity in the transfer function.

Put another way, THD-generating mechanisms involve a gain that changes instantaneously, varying rapidly during each individual half-cycle of (guitar) audio. Meanwhile, sag is a much slower gain variation, with gain changing rather slowly over a period of many cycles of audio, so it is the audio envelope that is "squished", rather than individual half-cycles of the waveform.

-Gnobuddy
 
I have heard this many times, and I do believe it's quite plausible. The logic being, "tone" actually is a misnomer for "timbre" in this context, and anything that creates roughly 10% or more harmonic distortion has the potential to change the timbre.

There are only a few components that can potentially create that much THD in a guitar amp, and those include the actual valves, the speaker itself, and the output transformer.

So it's certainly plausible that the output transformer contributes to guitar amp "tone". But it bothers me that I've never heard of anyone actually doing an experiment to prove it.

Oh, we have plenty of people saying that their guitar amp sounded better after they put in a $200 Mercury Magnetics transformer or something like that. Trouble is, there is no AB comparison possible, and human psychology dictates that you usually will think something is better if you just spent $200 on it. So these sorts of claims more or less fall into the "that could possibly be true, but it could also just be the placebo effect" category.

In principle, it doesn't seem too hard to come up with an experiment to test if guitar amp output transformers actually contribute to the amps timbre. Say, drive a guitar speaker with increasing amounts of sine-wave 500 Hz power, first from a low-distortion solid-state amp direct-coupled to the speaker, and then from the same amp via a 1:1 ratio output transformer made with the same-size iron as the actual valve amp OT. Or even two back-to-back OTs, one stepping the voltage up, the other stepping it back down, for an overall 1:1 ratio (less losses).

If you hear significantly different timbre via the transformers, the hypothesis is proved. If you don't, well, it turns out to be yet another amp-guru myth. Either way, we all learn something useful!

-Gnobuddy

An output transformer can have quite an effect, as depending on the design, might be limited on the bottom and or top. Phase shift accompanies the cutoff frequency and is as a rule of thumb going to be present to about 10X or 1/10th the cutoff, depending on which pole we're talking about. Guitar OPTs cut off fairly high (20-40Hz), so its plausible that phase shift can exist to 200-400Hz or higher. The ear translates broad spectrum phase shift like this into tonality, so its easy to surmise that's part of the 'sound'.

The thing is, in an OTL, you just impose similar bandwidth limits, like smaller coupling caps...


Not forgetting of course that sag created by the tube rectifier.

Rectifier sag does not have to translate to sag in the power supply. Much depends on the design of the supply; you can engineer sag using semiconductors as well (I prefer HEXFREDs as they tend to be lower noise similar to a tube rectifier).
 
I have been a professional guitar amp manufacturer for some years ( and still I'm building them for friends or myself).
Hi fi and guitar amps share just the fact that they use the same kind of electronic components, and nothing else.
In a guitar amp, generally , you are searching for some distortion ( also on clean sounds) in order to enrich the harmonic content of the sound that, otherwise is quite sterile and poor. Think that the guitar loudspeakers, that are another fundamental part in the soundchain of any guitarist, have their boxes made to resonate, and if you make the dead and stiff, the sound will be " dead and stiff". So really, there a not great contact point between hifi and guitar amps. A different perspective could be the amplification for Bass guitars or for acoustic guitars and basse, where there is some request of "hifi" treatment of the signal.

regarding otl as a guitar amp: I know there has been a commercial attempt with a 6c33 amp ( it was presented at a namm show a few years back) but as far as I can see it's been not succesful. Not even mentioned as a curiosity among the guitar guys, neither in the Hi End of guitar amps builder....
 
Last edited:
Not that an otl cannot sound decently for a guitar, but there are some basic facts that are to be considered whe you are designing a guitar amp in my opinion.
1) guitarists are sticky to some sounds ( and circuit) topologies (fender, marshall etc)
2) the approach in making a very good guitar amp starts not from the sound, but from the feeling that the entire amplification chain is giving you. Starting from the feeling you have under your strings related to the sound you hear. That's where a good amp builder is different form a normal one. What you need to do is "translating" what a musician wants to feel in a physical amplifier. Technically, this is where you play a lot with power supply, sag and feedback, that eventually have an impact in the timbre. Balancing all these elements is what makes for a great guitar amp.
3) replacement of tubes anywhere you are. Sourcing an el34 is much easier than any otl tube. This means that if you step in a music shop anywhere in the world , you'll find always an el34 or a 6l6 , or some el84.

These are some basic points you might want to consider when professionally designing an amp.
OTOS, If you are making an amp for your personal use that will sit in your practice room, forever, you can use any kind of topology ,style, or inspiration, that once the sound will be pleasing at your ears, it will always be a perfect amp.

edit I forgot to point out a fact: a guitar amp, needs to satisfy the player first, and only in a second moment the listener....
 
Last edited:
3) replacement of tubes anywhere you are. Sourcing an el34 is much easier than any otl tube. This means that if you step in a music shop anywhere in the world , you'll find always an el34 or a 6l6 , or some el84.

This is actually the big deal. The other points are all doable.

You are coupling the bad traits of SS amps with the cost and complication of tube amp ... which to boot will not sound like a TUBE amp anyway.

This statement is outright false. OTLs most definitely sound like tube amps!

As far as bad traits of solid state??

Hard to say what that might be- soft clipping? -that's a tube thing.

Greater bandwidth? - that's a design thing and has nothing to do with tube or transistor.

Parts availability?- the tubes we use are 12AX7s, 6SN7s and the output tubes which are 6AS7Gs. They are available cheaply, so it would not be much expense to include a spare set if actually selling the amp.
 
Guitar OPTs cut off fairly high (20-40Hz),
A normal six-string guitar in standard tuning has no fundamental frequencies below 82.4 Hz. A 40 Hz lower cutoff frequency would be a complete waste of effort.

Because of the (range of) physical locations at which the strings are picked (fairly close to one end), usually there isn't much output at the fundamental frequency. There is much more output at the second harmonic, so most of the guitar's output is actually above 165 Hz!

Leo Fenders open-back guitar cabs have lousy (acoustic) low-frequency response, too, particularly in the physically smaller amps. Bass frequencies from the front and back of the speaker just cancel out. Those amps still work well enough that tens of thousands of guitarists love Fender amps, though!

As for the effect of phase-shift, guitar amps are not subtle things, and I have my doubts about the audibility of transformer phase-shifts. But if transformer core saturation and hysteresis is enough to generate several percent THD, that could be audible, though there doesn't seem to be any definitive evidence for this (yet).

The thing is, in an OTL, you just impose similar bandwidth limits, like smaller coupling caps...
Indeed. Wee little caps was part of the Marshall designs even fifty years ago!

-Gnobuddy
 
But if transformer core saturation and hysteresis is enough to generate several percent THD, that could be audible, though there doesn't seem to be any definitive evidence for this (yet).

OPT Transformer misbehaviour is indeed a part of the sound of certain amps - the presence knob in the feedback loop utilises this in many amp designs.

Somewhere there's some screen shots of plate voltages of what happens as BOTH valves cut off under overload - lots of spiking without the feedback.

Rob Robinette also notes that saturation provides compression and a less sterile sound (see his voicing an amplifier page)
 
OPT Transformer misbehaviour is indeed a part of the sound of certain amps
So I've heard, many times, just like the stories about Yeti sightings. :)

Certainly there could be an undiscovered mammal out in the wild. But the Yeti stories are problematic because after all these decades of claimed sightings, there is still no incontrovertible evidence; not even one little strand of fur containing DNA that matches no known animal.

That one little strand of fur would be solid proof. But, year after year, decade after decade, that solid evidence never actually shows up.

In this case the story (about transformer saturation providing audible change to guitar amp timbre) is plausible, certainly much more so than Yeti sightings.

But it bothers me that after years of reading claims about audible output transformer distortion in guitar amps, I still haven't found any solid evidence yet - for example, a couple of distortion spectra showing a 10% THD increase resulting from passing through a guitar output transformer operated under typical guitar amp voltages and currents, or at the very minimum, a couple of sound clips with different-sized output transformers installed.

the presence knob in the feedback loop utilises this in many amp designs.
The presence knob usually removes negative feedback above a certain frequency, causing a substantial change to the frequency response of the amp (specifically, it boosts a range of mid and upper frequencies by several dB).

So the presence control causes a dramatic and certainly very audible change in EQ. That being the case, how can we know if we're also hearing a (presumably subtler) nonlinear distortion from the transformer buried under the huge EQ change?

Rob Robinette also notes that saturation provides compression and a less sterile sound
Another possible Yeti sighting. But still no DNA! :D

There is no doubt that there are massive amounts of saturation occurring in the valves. There is also no doubt that at high enough power levels, some loudspeakers also saturate and add some compression.

So is the saturation we hear from a guitar amp caused by the valves, the valves and speaker, or the valves, speaker, and output transformer? No way to tell, if they're all mixed into the stone-soup like that.

Maybe it's just the nature of guitarists and guitar amp gurus, to not consider doing an actual experiment, and instead to operate on faith and belief and legend. Maybe they really are correct about audible transformer saturation distortion. I don't know. I want to see that DNA test.

The trouble is that faith and beliefs and legends are wrong a lot more often than they're right. So, who knows if this particular one is wrong or right? It's repeated a lot, sure, but that doesn't constitute evidence, only hearsay.

-Gnobuddy
 
Another possible Yeti sighting. But still no DNA! :D
I've only measured OPTs in hifi applications so I've no Yeti sightings. Too busy measuring how deep the snow was:)

Grant Wills (of "lamington" amplifier fame) would be the person to talk to. He's just spent some time comparing different SE OTPs for guitar amplifiers for his "SOLO" amp - see here for some background.

He's the person most likely to have a shaggy pelt on his lounge floor :D
 
I completely agree with @Gnobuddy about output transformers in guitar amps.

The usual transformer equivalent circuit is very accurate for these transformers at audio frequencies. The only nonlinear (saturable) part of that model is the shunt magnetizing inductance. This comes into play only at the lowest frequencies, because the impedance of that shunt inductance is high otherwise. Even at those low frequencies, the nonlinear part is in shunt, and can only affect the signal by loading the output valves (in parallel with the proper load reflected from the speaker).

The factor that puts the shunt magnetizing inductance into saturation is the current through it. But that depends on the time integral of the voltage across it (which is the OT primary voltage). Again it is only at low frequencies that the time integral of voltage can become large, but at those low frequencies the output valves are being heavily loaded by the magnetising reactance and may not be able to provide sufficient output voltage to cause saturation. This is particularly true for pentode output stages which have a high output impedance.

Output transformers also have inter-winding capacitances and leakage inductances, but these are linear effects and change the ‘equalisation’ of the guitar tone without introducing any further distortion.
 
Last edited:
I thought I saw a yeti on YouTube. It appeared to be very partial to Ibanez guitars. I don't know what kind of amplification it was using because my first instinct was to run...

I saw something on (some?) transformer capacitance being non-constant by virtue of being dependent upon voltage level. That was also disturbing but I was left with a curiosity about how that could ever be true. I was reluctant to accept it.

I don't remember where I saw it; possibly in a book or chapter on transformer winding, or a transformer manufacturer's website. It was a source I felt knew far more than I did, but I didn't want to believe...

Then there was this other video...

https://youtu.be/oH-yoJKSp60
 
Last edited:
With respect to OPT distortion, although a little over my head, it doesn't make sense that a guitar OPT doesn't distort more. They are half the size, and can have massive quality shortcuts such as non-interleaved windings and even windings on separate bobbins.

I came across this by Crowhurst in Audio, March 1957, p. 22:
 

Attachments

  • Crowhurst quote.jpg
    Crowhurst quote.jpg
    105.7 KB · Views: 168
32 ohm loudspeaker for OTL

I didn't want to read all 73 pages on this amp... I was thinking about a guitar amp with less costly speakers. I noted the availability of greater power output with higher impedance loads, but finding a 32-50 ohm voice coil seems like a big waste of time. ....

Hi there m: re post #1, Eminence B810, 32 ohm, 10-inch diameter, 150watts guitar loudspeaker is available from US Speaker for about $70. ...regards, Michael
 
Last edited:
Status
Not open for further replies.