Peavy and output transormer saturation

[teemuk]

So I think a significant part of his transtube circuitry is based on a misunderstanding of how transformers work.

What Peavey discusses in that document relates to their early 1980's "Saturation" circuitry and basically has nothing to do with modern TransTube -related designs. Infact, TransTube series amplifiers do not even employ similar bandwidth reduction at higher overdrive levels. I think Peavey pretty much abandonded the entire scheme after 1980's (and on that note, that particular scheme of supposed bandwidth reduction of a saturated OT).

Oh, and not to mention Peavey's idea was somewhat flawed from the start: Bandwidth reduction they discuss should apparently be a dynamic effect (as it's supposed to interact with signal voltage / frequency) . But the old Peavey "Saturation" scheme simply introduced a -fixed- amount of bandwidth reduction, ganged to the gain control dial. Effectively bandwidth is higher at lower gains, and reduced at higher gains, but there is no dynamic interaction at all whatsoever. Peavey's designers could have easily made the effect dynamically variable but for some reason they never did... Perhaps they found out that the dynamic effects had minor importance in comparison to major effect of bandwidth reduction before distortion alone.

Misunderstanding transformers... Well, I could partially agree on that. But Peavey designers should know better so I'm pretty sure they just chose a marketing approach and threw in one of those vaguely understood concepts that happens to sounds "sexy".... WOoooo.... "Transformer Saturation"..... Yet majority of people reading that article probably have no clue what that actually means in practice nor have even encountered it.

Anyone who has actually viewed output transformer saturation in an oscilloscope knows damn well it's one of those things that should be avoided almost at all costs. Distortion harmonics concernerd OT saturation is almost like very, very hard clipping, but instead of chopping off waveform tops random parts of the waveform are chopped off instead, and the signal gets manifestested with plenty of very high order distortion, odd and even. It will sound about as bad as poorly designed current limiter kicking in in a SS amp. Terrible! Avoid!

Luckily, avoiding is rather easy. It's pretty hard to saturate generic OT's with typical signal voltages, especially given typical signal frequencies. As said, you truly want to avoid OT saturation so this is no biggie, except maybe in bass amps that need to handle signals lower than about 80 Hz in frequency...
Just for the sake of reference, a typical modern high gain guitar amp may introduce a hi-pass filter centered around as high as 1 KHz (!!!!) just to avoid IMD. This pretty much even explains the key point of Peavey's old "Saturation" scheme (basically bandwidth reduction does marvels for IMD) but also makes existence of output transformer saturation quite dubious, given overall signal frequencies at the point of overdrive.

Case in point:

If the amp can put out an 80 Hz sine wave with nearly the rated power, then it appears that the relative power at about 80 Hz in these signals is too small to cause significant saturation.

So, if the amp can reproduce the lowest note of the guitar at its rated output power (usually at minimal % THD) then to get OT saturation you practically need to plug in a different signal source than your guitar. One that emits notes at frequencies lower than about 80 Hz. In my experience, the OT saturation typically start manifesting itself at about 30 - 50 Hz (largely depending on amp). That's no concern with frequency range of an electric guitar and even lesser concern with modern OD channels with highly restricted bandwidths.


Above is an oscilloscope capture illustrating OT saturation at 25 Hz signal with two different input levels. Bottom plot is the one with higher magnitude input signal, saturation correspondingly more severe. Note that @ 100Hz the reproduction would still be "clean". Yes, they are supposed to be sinusoidal waveforms but they happen to have this supposedly euphonic distortion from transformer saturation. You can just picture the harmonic distortion of this mess. Just think of the harmonic overtones created in comparison to generic soft clipping or even generic hard clipping. This will sound truly hideous, if not totally unusable in almost any musical context. And intermodulation will only turn things much, much worse.

For obvious reasons you can't really observe a potent trend to use underrated cr*ppy little output transformers... for that marvellous saturation tone. In fact, most people seem to praise transformers that are rather "transparent": low distortion, decent bandwidth and no saturation at typical signal frequencies / voltages. Take a vintage Gibson amp... People won't be hyping how great tone their pathetically tiny cheap output transformers have, more likely they'll be replacing the poor transformers with better ones. To improve tone.

There is also another explanation for bandwidth reduction during ovedrive that has nothing to do with transformer saturation: Simply, utilisation of negative feedback will increase bandwidth at the cost of gain reduction. If negative feedback loop becomes impaired because of clipping distortion the effect of bandwidth increase also becomes impaired. Basically, global NFB of the power amp stage can thus extent bandwidth of transformer coupling (basically even bandwidth of all kinds of circuits), but the effect is lost when the amplifier begins to clipping distort and global feedback loop is rendered ineffective. Unlike in true OT saturation, with this scheme the entire bandwidth (both lows and highs) actually is reduced and the distortion is also more "earpleasing" because it's not that of a saturating OT but that of the circuit just clipping in usual manner.

[Leo_Gnardo]

What Peavey discusses in that document relates to their early 1980's "Saturation" circuitry and basically has nothing to do with modern TransTube -related designs.

True 'nuff and in fact Peavey's early 80's "Saturation" knob has nothing to do with the output section much less transformer saturation. All it did is send preamp signal thru a standard fuzztone parallel reverse-direction diode pair. Wow. oops I mean (where's the tiny font?) wow. Well it does bring up a distortion which may or may not be to the players' taste. FWIW I've paralleled those silicon diodes with germanium for a smoother fuzz sound but aside from that, it's no big deal.

[teemuk]

Yes, Peavey's "Saturation" circuit is essentially nothing but a diode clipping stage, which's frequency response varies in gang with the gain control dial: Lower gain settings have higher bandwidth than higher gain settings.

But realistically speaking, Peavey did not invent applying EQ before distortion, and results and benefits of reducing both bass and treble before clipping must have certainly been known widely before Peavey went on to conceive and patent their "Saturation" circuit. So I'm essentially wondering what the proprietary part of that "Saturation" circuit actually is.

It's not like Peavey was or is the only manufacturer who makes a distortion circuit where gain control also happens to effect frequency response, most often by decreasing bandwidth towards higher gain settings. No. But then again, it's not any dynamic behaviour either.

Because the greatest irony is that the dynamic EQ'ing effect, which Peavey prominently discusses, is entirely lacking from their "Saturation" circuit. Variations in signal amplitude or frequency have no effect on frequency response of the circuit. No dynamic interaction whatsoever. Only thing that affects frequency response of that circuit is the gain control. You could crank the gain control of that circuit, and gently pick the strings ...and you would hear a bandwidth-reduced signal, not broader bandwidth reproduction you should hear with gentler picking touch IF aforementioned dynamic effects in EQ'ing were truly in play. So Peavey essentially left out the only feature, which could have made their circuit proprietary and different from other generic SS distortion circuits.

...And after happily leaving such dynamic effects totally unexploited for years Peavey suddenly sees need to revisit the issue in an article and pose as experts. Excuse me but I smell advertising, and not even very thinly veiled one.

[Enzo]

Suddenly?

[nickb]

When a guitar amp is overdriven, new frequencies are created that were not present in the signal that came from the guitar, or they were are such low levels that they were ignored. Some of these frequencies are below the the audio spectrum, they are artifacts of pick transients, note envelopes and asymmetrical clipping. Some of this was discussed here: http://music-electronics-forum.com/t37023/

Could these very low frequencies nudge the OT towards saturation in a way that causes compression of audible signals?

That's an interesting point. Sim for a PP stage fed with 100Vpp @ 100Hz and 10Vpp @ 20Hz shows no significant IMD products below 180Hz. The bias was 48V so this was pretty heavily overdriven with lots of blocking.


Blue is input, green is output.

[Mike Sulzer]

@Teemuk

There is also another explanation for bandwidth reduction during ovedrive that has nothing to do with transformer saturation: Simply, utilisation of negative feedback will increase bandwidth at the cost of gain reduction. If negative feedback loop becomes impaired because of clipping distortion the effect of bandwidth increase also becomes impaired. Basically, global NFB of the power amp stage can thus extent bandwidth of transformer coupling (basically even bandwidth of all kinds of circuits), but the effect is lost when the amplifier begins to clipping distort and global feedback loop is rendered ineffective. Unlike in true OT saturation, with this scheme the entire bandwidth (both lows and highs) actually is reduced and the distortion is also more "earpleasing" because it's not that of a saturating OT but that of the circuit just clipping in usual manner.

This is the one part of your post I do not agree with. With heavy clipping the amp slews from positive to negative saturation and back, and there is no linear gain at all and thus no bandwidth, although if the frequency is too high, the slew to clipping limit will not be achieved. With lighter clipping, the amp resumes linear operation between episodes of saturation, and the gain is normal during these periods of linearity.

[Mike Sulzer]

True 'nuff and in fact Peavey's early 80's "Saturation" knob has nothing to do with the output section much less transformer saturation. All it did is send preamp signal thru a standard fuzztone parallel reverse-direction diode pair. Wow. oops I mean (where's the tiny font?) wow. Well it does bring up a distortion which may or may not be to the players' taste. FWIW I've paralleled those silicon diodes with germanium for a smoother fuzz sound but aside from that, it's no big deal.

So let's examine the level of harmonics generated by different devices used differently. We want to examine the high current limit, since that is what matters for both positive and negative limits in a push pull circuit, and so we look at a single device and bias off center to make clipping on the high current side. First consider a small darlington power transistor used as an emitter follower. Global feedback is often blamed for sharp corners in clipping, but you cannot ignore the huge amount of feedback in an emitter follower all by itself. The attached plots show time domain and frequency domain. Notice the sharp corners and the gradual fall off of harmonics. The other circuit is a JFET low level common source stage of gain; no feedback, but also biased to one side to clip on the high current limit, which now is the bottom of the wave form. Note the rounded corners and the much faster fall off of harmonics. At one KHz (10th harmonic) there is nearly 40 db difference in relative levels! Nearly all SS amps use emitter followers, or two transistor equivalents. Nearly all tube amps use pentodes (not too different from FETs) as the output stage in a configuration with gain, and not much feedback. I think there is no reason to expect such circuits to sound anything like the same when clipping, as these plots suggest.

Dartd.pdf
Darfd.pdf
FETtd.pdf
FETfd.pdf

[J M Fahey]

Oh , I do agree with Teemuk on this point (as with most , if not all others ) , and you will too, once you look at the full picture.

You are referring to the electronic/active components satuirating and, true, once you have a squarewave it is a squarewave.

But what I see (have worked with that for decades, simulating it *without* transformers but adding the proper filtering at the approppriate points) is that in the full amp, you do have a "narrowband filter" after the clipping stage .
And which, contrary to what Peavey (and others) state, is quite stable in its own characteristics , specially NOT saturating by itself, at any frequency.

Now, consider an old cheap transformer, barely meeting frequency response limits **at low power** (many claim "x to y Hz bandwidth" ... at 1 W applied power) , definitely less (unspecified) at higher power.

So everything is nice anyway since the transformer is *inside* the NFB loop which evens things up ... as long as the amp has extra gain (not much in a tube amp which does not tolerate much NFB anyway) and "muscle" , but as soon as it clips, we are back to square zero.

So I believe there is a bandwidth limiting action in a regular/classic/oldstyle (pick one) transformer ... which was always there.

Magnetic saturation? .... "dynamic behaviour"? ..... never saw such thing as part of guitar sound.

What about the few non-NFB amps?

Well, they tend to be considered jangly/chimey which does not exactly suggest "rounded waveshapes", does it?

[R.G.]

Mike, I'm having trouble following your reply to Leo. How are your comments related to his?

Not quibbling, just lost.

[Mike Sulzer]

Mike, I'm having trouble following your reply to Leo. How are your comments related to his?

Not quibbling, just lost.

Not directly. After discussing differences in diode clipping, it seemed like a good time to introduce differences in clipping in active stages. That is what we really need to understand!

[nickb]

Teemuk - Thanks for the scope traces in post 16. Pretty jaw-dropping. +1

I only time I ever have looked at saturation was to do with power transformers, where things are a bit different.

For the benefit of others who might be following this, I captured the following today using a 60Hz 115V 3VA transformer driven by a variac and with a current probe in the primary. The output was also loaded to avoid the current waveform looking too distorted by the magnetizing current. The power here is 50Hz so it saturates that bit sooner as I raise the primary volts.

Normal Operation:



Saturating:


You can see that in this situation, with the source impedance being determined pretty much only by the DC resistance of the primary, as the core saturates the current jumps up due to the lower in inductance ( V = L . d(i)/d(t) so as inductance falls current goes up). Since the current can increase the secondary waveform shows much less distortion, looking remarkably like blocking distortion. For the tube amp example in #16, the source impedance is very high so the current cannot rise and so the output voltage collapses. That is why the two cases look so different.


Going back to Peavey, one thing that was a fairly common theme in the SS /Hybrid designs was the addition of blocking distortion, you see this in the cathode driven VTX Heritage right through to the Vypers. I guess Harltey either liked the sound of it or it was just blind pursuit of identical characteristics. I find blocking distortion a tad too buzzy for my taste.

[Mike Sulzer]

I do not think you and Teemuk are saying the same thing, but I might be confused. I think you are saying that if you have an output transformer with poor bandwidth that is extended by negative feedback around the transformer, then when clipping occurs (which happens before the transformer) and removes the feedback, then the response of the transformer acts as a low pass filter (or bandpass if the bass fall off is significant also). This is certainly true, although I doubt it is very signifiant because the guitar speaker falls rapidly at about 5 KHz, and transformers seem reasonably flat up to that frequency, and so I do not think much changes.

I understood Teemuk to be describing a more general effect for a wide range of circuits. But it is true that this effect can only matter for circuit elements after the clipping, but inside the feedback loop. I think the output transformer is the only one for which these conditions are met.

Oh , I do agree with Teemuk on this point (as with most , if not all others ) , and you will too, once you look at the full picture.

You are referring to the electronic/active components satuirating and, true, once you have a squarewave it is a squarewave.

But what I see (have worked with that for decades, simulating it *without* transformers but adding the proper filtering at the approppriate points) is that in the full amp, you do have a "narrowband filter" after the clipping stage .
And which, contrary to what Peavey (and others) state, is quite stable in its own characteristics , specially NOT saturating by itself, at any frequency.

Now, consider an old cheap transformer, barely meeting frequency response limits **at low power** (many claim "x to y Hz bandwidth" ... at 1 W applied power) , definitely less (unspecified) at higher power.

So everything is nice anyway since the transformer is *inside* the NFB loop which evens things up ... as long as the amp has extra gain (not much in a tube amp which does not tolerate much NFB anyway) and "muscle" , but as soon as it clips, we are back to square zero.

So I believe there is a bandwidth limiting action in a regular/classic/oldstyle (pick one) transformer ... which was always there.

Magnetic saturation? .... "dynamic behaviour"? ..... never saw such thing as part of guitar sound.

What about the few non-NFB amps?

Well, they tend to be considered jangly/chimey which does not exactly suggest "rounded waveshapes", does it?

[Malcolm Irving]

Trying to demonstrate OT saturation in a pentode amp can be a case of ‘chasing your tail’.
To saturate the transformer you need a low frequency and a high primary voltage. You can take the frequency as low as you like with a signal generator, but as the frequency gets lower the magnetising reactance of the transformer gets lower, loading down the pentode(s) so that the primary voltage collapses.