How would pure power tube distortion sounds like?

[thinkingchicken]

Hi, I heard someone said Trainwreck Express produced 100% pure power tube distortion. Is this true? Can someone explain how the power tubes even begin to clip first without the aid of any components at all? I think it is impossible for power tubes to distort on their own, thus pure power tube distortion which is purely come from isolated power tubes and without any aid from other components isn't true.

[wpaulvogel]

A Trainwreck Express is purely power tube distortion. If you’ve ever played one, you’ll notice that there’s not much reason to turn the volume control much past 12:00. I’ve always used Glen K.’s recommendation of basically setting everything at noon. I play Les Paul guitars and they have enough output to drive the amp with everything at noon. It’s also VERY overdriven at these settings. Alright back to the question.
If you send a 100mV signal into the amp, the power tubes begin to clip at about 9:00 on the volume control. At 12:00 the power tubes are putting out square wave with some neat asymmetrical form and some anomalies and this sounds Awesome. While this is occurring, if you scope the phase inverter, it’s clean sine. Any previous stages are also the same with just lower amplitude. That’s pure power tube distortion.
You can turn up further on the volume control but to me the output never gets better sounding.
Just in case you’re wondering, a Marshall is the same way, but the output waveform doesn’t get the interesting forms of a Trainwreck. The cascading gain without attenuation creates the power tube overdrive.

[thinkingchicken]

A Trainwreck Express is purely power tube distortion. If you’ve ever played one, you’ll notice that there’s not much reason to turn the volume control much past 12:00. I’ve always used Glen K.’s recommendation of basically setting everything at noon. I play Les Paul guitars and they have enough output to drive the amp with everything at noon. It’s also VERY overdriven at these settings. Alright back to the question.
If you send a 100mV signal into the amp, the power tubes begin to clip at about 9:00 on the volume control. At 12:00 the power tubes are putting out square wave with some neat asymmetrical form and some anomalies and this sounds Awesome. While this is occurring, if you scope the phase inverter, it’s clean sine. Any previous stages are also the same with just lower amplitude. That’s pure power tube distortion.
You can turn up further on the volume control but to me the output never gets better sounding.
Just in case you’re wondering, a Marshall is the same way, but the output waveform doesn’t get the interesting forms of a Trainwreck. The cascading gain without attenuation creates the power tube overdrive.

According to this link http://home.polstra.com/amps/wreck1/scope/, it looks like at the maximum volume the power tubes, phase inverter and preamp tubes especially the cold-clipper are all clipped although the power tube clipped the hardest. So, it is not a pure power tube distortion.

[wpaulvogel]

You didn’t read what I wrote. At 12:00 volume setting only the power tubes are clipping. Buy a Trainwreck or build one and draw your own conclusions. The amps really loose the wonderful tones when the other stages begin to clip.

[pdf64]

It could be argued that in AB1 (or any of the non grid current '1' modes), it's the preceding stage to the power tube that clips, not the power tube itself.
Such that the power tube clips the preceding stage, rather than the power tube itself being unable to be pushed any further.
As the plate current can almost always be increased by beefing up the preceding stage's output capability and allowing operation in a grid current '2' mode.

[thinkingchicken]

It could be argued that in AB1 (or any of the non grid current '1' modes), it's the preceding stage to the power tube that clips, not the power tube itself.
Such that the power tube clips the preceding stage, rather than the power tube itself being unable to be pushed any further.
As the plate current can almost always be increased by beefing up the preceding stage's output capability and allowing operation in a grid current '2' mode.

Sorry, do you mean even what was shown in this link http://home.polstra.com/amps/wreck1/scope/, what are actually distorting and clipping are not power tubes but output transformers, speaker output, etc.? If you look at the first pictures, the speaker output clipped first and it is the most hard clipped followed by the phase inverter and then the third preamp stage. Trainwreck amps are push & pull class AB.

[katopan]

Zen poem.... Is clipping in a distortion pedal occurring at the diodes or the stage driving them?

If you go digging I posted a heap of waveform captures years ago that show much more than the polstra ones. Every stage in the amp at a full range of volume settings, with commentary on what was happening and why the Express has the magic. In the magic zone the cold clipper is not overdriven itself, but the 10K cathode resistor is about gain control. The PI operation is all about coupling cap charge up to shift the PI operating point under power grid clipping conditions and this is what causes the uneven 2nd harmonic rich distortion out to the speaker, not the cold clipper. This PI behavior is also working with the power screen sag to no just provide compression, but actually cause a hybrid clipping where the narrower/higher side of the output waveform is pure power grid clipping, but the wider/lower side of the output waveform is actually PI grid clipping operating underneath the threshold of that side's power grid clipping.

Do a search and then you'll have some waveforms to look at for yourself. Lastly, please don't make the common mistake thinking that the clipped waveform on the plates of one stage to be overdrive of that stage. You have to be able to properly identify what is that stage grid clipping, what is that stage cutoff and what is grid clipping of the follow stage.

[thinkingchicken]

Zen poem.... Is clipping in a distortion pedal occurring at the diodes or the stage driving them?

If you go digging I posted a heap of waveform captures years ago that show much more than the polstra ones. Every stage in the amp at a full range of volume settings, with commentary on what was happening and why the Express has the magic. In the magic zone the cold clipper is not overdriven itself, but the 10K cathode resistor is about gain control. The PI operation is all about coupling cap charge up to shift the PI operating point under power grid clipping conditions and this is what causes the uneven 2nd harmonic rich distortion out to the speaker, not the cold clipper. This PI behavior is also working with the power screen sag to no just provide compression, but actually cause a hybrid clipping where the narrower/higher side of the output waveform is pure power grid clipping, but the wider/lower side of the output waveform is actually PI grid clipping operating underneath the threshold of that side's power grid clipping.

Do a search and then you'll have some waveforms to look at for yourself. Lastly, please don't make the common mistake thinking that the clipped waveform on the plates of one stage to be overdrive of that stage. You have to be able to properly identify what is that stage grid clipping, what is that stage cutoff and what is grid clipping of the follow stage.

Sorry I'm a newbie to this thing but what's your interpretation about the scope images of this Trainwreck Express waveforms http://home.polstra.com/amps/wreck1/scope/? Where did the distorted sound mainly come from? Is it actually the phase inverter, output tubes or what.

[pdf64]

It could be argued that in AB1 (or any of the non grid current '1' modes), it's the preceding stage to the power tube that clips, not the power tube itself.
Such that the power tube clips the preceding stage, rather than the power tube itself being unable to be pushed any further.
As the plate current can almost always be increased by beefing up the preceding stage's output capability and allowing operation in a grid current '2' mode.

Sorry, do you mean even what was shown in this link http://home.polstra.com/amps/wreck1/scope/, what are actually distorting and clipping are not power tubes but output transformers, speaker output, etc.? If you look at the first pictures, the speaker output clipped first and it is the most hard clipped followed by the phase inverter and then the third preamp stage. Trainwreck amps are push & pull class AB.

Given sensible control settings, the earliest clipping point in an AB1 amp is, by definition, the power tube control grids. ie the diode effect created when the signal peak at the control grid overcomes its bias and attempts to raise the control grid's potential positive with respect to its cathode.
AB2 operation allows the control grid to be driven linearly in the positive grid / grid current region, thereby achieving higher plate currents.
In AB1, allowing the power tubes' grid-cathode to be the only clipping point would typically result in bias shift, leading to blocking distortion, at fairly low levels of overdrive, ie it probably wouldn't sound very good.

[thinkingchicken]

Given sensible control settings, the earliest clipping point in an AB1 amp is, by definition, the power tube control grids. ie the diode effect created when the signal peak at the control grid overcomes its bias and attempts to raise the control grid's potential positive with respect to its cathode.
AB2 operation allows the control grid to be driven linearly in the positive grid / grid current region, thereby achieving higher plate currents.
In AB1, allowing the power tubes' grid-cathode to be the only clipping point would typically result in bias shift, leading to blocking distortion, at fairly low levels of overdrive, ie it probably wouldn't sound very good.

So at the maximum volume especially on Trainwreck Express amps, is it only the power tubes that are producing the distorted sound or some combinations of power tubes, phase inverter, etc.? I wonder how would a pure output tube grid clipping sound like.

[pdf64]

...So at the maximum volume especially on Trainwreck Express amps, is it only the power tubes that are producing the distorted sound or some combinations of power tubes, phase inverter, etc.?

Dunno how many ways there are to phrase the same same thing.
Given reasonable control settings, for any non-grid current amp, the earliest (ie as signal level is increased from 0) point to clip will almost always be the power tube grid/s.
Therefore at very light degrees of overdrive, only power tube grid clipping will be heard. As the degree of overdrive increases, stage/s before that will also overdrive. But the power tube grids must be the first point as otherwise the amp would not be able to reach its full AB1 power output.
Katopan's thread demonstrates in great detail the precise characteristics and mechanisms of how that works with the Express.

...I wonder how would a pure output tube grid clipping sound like.

Almost any tube amp can be adjusted to achieve that. However, whether by accident or design intent, to avoid excessive bias shift (leading to blocking distortion), anything more than a minimal degree of overdrive will result in stage/s before that also overdriving.

[thinkingchicken]

...So at the maximum volume especially on Trainwreck Express amps, is it only the power tubes that are producing the distorted sound or some combinations of power tubes, phase inverter, etc.?

Dunno how many ways there are to phrase the same same thing.
Given reasonable control settings, for any non-grid current amp, the earliest (ie as signal level is increased from 0) point to clip will almost always be the power tube grid/s.
Therefore at very light degrees of overdrive, only power tube grid clipping will be heard. As the degree of overdrive increases, stage/s before that will also overdrive. But the power tube grids must be the first point as otherwise the amp would not be able to reach its full AB1 power output.
Katopan's thread demonstrates in great detail the precise characteristics and mechanisms of how that works with the Express.

...I wonder how would a pure output tube grid clipping sound like.

Almost any tube amp can be adjusted to achieve that. However, whether by accident or design intent, to avoid excessive bias shift (leading to blocking distortion), anything more than a minimal degree of overdrive will result in stage/s before that also overdriving.

That's why I ask Katopan in his own thread regarding what produced the overall distortion at the maximum volume. Also, his explanations are quite complicated for a newbie like me.

To your second statement, to achieve that pure output tube grid clipping sound do you mean by lowering the preamp volume and increasing the master volume? The phase inverter is there, so it is the whole power amp producing the distorted sound.

[katopan]

There's a lot to talk about here, so it all depends on how deep you want to go. The Express is the most complicated mechanism of overdrive you'll find in any amp and they way it works it like none other. It took me a few months to work it all out and the thread you reference https://ampgarage.com/forum/viewtopic.php?f=1&t=15694 shows that still not everyone agreed with what I found.

Many, many traditional push-pull amp designs (particular non-master volume types) will let you hear "pure power valve distortion". I am a self confessed power valve overdrive junkie by the way. That's the reason why I never build an amp without VVR. Combine that with real pentodes, with their additional screen compression effect compared to beam tetrodes, and you have what makes for my favourite type of amps. So the absolute benchmark for power valve distortion is the Marshall 18 Watt. It runs totally clean single stage preamp channels, into a PI that only overdrives once the EL84s are so overdriven it doesn't matter anymore. You can measure when the PI starts to overdrive but you can't hear it in the output tone. It's pure EL84 distortion at all settings with any guitar. JCM800s (non-master or with the master turned right up) also do it well, but above a point the PI clips which you can hear, then depending on the model the preamp clips, which you can hear. It's then a blended distortion. Master volume models allow you to set the mix of preamp to power stage distortion to taste. In my view running them with just preamp distortion (actually any amp, but that's just me) is blah and I might as well have a distortion pedal. Power overdrive is where it's at! But this isn't only a Marshall thing. With the right settings and not overdriving too far, I've tested and proven that power stage overdrive dominates or is the only component of the distortion in amps from different Fender models including all series of Bassmans, Blues Junior, Deluxe Reverb, to various Laney amps, and many others. The power distortion tone and just as importantly the dynamic 'feel' depends on which power valves, OT impedance, power supply sag response, screen resistor sizing and then other secondary things.

The short answer to "what produces the distortion in an Express at full volume" is that there is clipping in the power stage, the PI and the 3rd cold clipper. Because of the unique operation, clipping in the 3rd cold clipper stage actually does the opposite to normal/expected, and rebalances the power clipping to be more even and the output to the speaker moves back closer to being a symmetrical square wave. See in the Polstra photos Fig 4 to Fig 5, he comments that the output is not much different, but actually you can see it's starting to even up again. My scope shots show if you keep going past that point the mark-space ratio becomes even again, it looses the steeper slope on the bottom swing and +ve and -ve magnitude become balanced again, ie. back to pure square wave. It's not at its best at that setting and looses both the tone and the feel of what makes an Express a great amp. At a basic level, the output stage clips first and continues to clip as you turn it up more, so it's a major contributor all the time.

AB1 amps can be configured to overdrive very, very heavily in the power section before the PI clips (18 Watt is an example). But many aren't and in that case, you can only overdrive the power stage so far before PI or even preamp clipping kicks in. You'll generally hear that transition point with a change in the tone and see it in the output waveform.

If you want to know more about why the Express is so special, it's all in that thread. But it's not an easy mechanism to get your head around, so feel free to ask any questions.

[thinkingchicken]

There's a lot to talk about here, so it all depends on how deep you want to go. The Express is the most complicated mechanism of overdrive you'll find in any amp and they way it works it like none other. It took me a few months to work it all out and the thread you reference https://ampgarage.com/forum/viewtopic.php?f=1&t=15694 shows that still not everyone agreed with what I found.

Many, many traditional push-pull amp designs (particular non-master volume types) will let you hear "pure power valve distortion". I am a self confessed power valve overdrive junkie by the way. That's the reason why I never build an amp without VVR. Combine that with real pentodes, with their additional screen compression effect compared to beam tetrodes, and you have what makes for my favourite type of amps. So the absolute benchmark for power valve distortion is the Marshall 18 Watt. It runs totally clean single stage preamp channels, into a PI that only overdrives once the EL84s are so overdriven it doesn't matter anymore. You can measure when the PI starts to overdrive but you can't hear it in the output tone. It's pure EL84 distortion at all settings with any guitar. JCM800s (non-master or with the master turned right up) also do it well, but above a point the PI clips which you can hear, then depending on the model the preamp clips, which you can hear. It's then a blended distortion. Master volume models allow you to set the mix of preamp to power stage distortion to taste. In my view running them with just preamp distortion (actually any amp, but that's just me) is blah and I might as well have a distortion pedal. Power overdrive is where it's at! But this isn't only a Marshall thing. With the right settings and not overdriving too far, I've tested and proven that power stage overdrive dominates or is the only component of the distortion in amps from different Fender models including all series of Bassmans, Blues Junior, Deluxe Reverb, to various Laney amps, and many others. The power distortion tone and just as importantly the dynamic 'feel' depends on which power valves, OT impedance, power supply sag response, screen resistor sizing and then other secondary things.

The short answer to "what produces the distortion in an Express at full volume" is that there is clipping in the power stage, the PI and the 3rd cold clipper. Because of the unique operation, clipping in the 3rd cold clipper stage actually does the opposite to normal/expected, and rebalances the power clipping to be more even and the output to the speaker moves back closer to being a symmetrical square wave. See in the Polstra photos Fig 4 to Fig 5, he comments that the output is not much different, but actually you can see it's starting to even up again. My scope shots show if you keep going past that point the mark-space ratio becomes even again, it looses the steeper slope on the bottom swing and +ve and -ve magnitude become balanced again, ie. back to pure square wave. It's not at its best at that setting and looses both the tone and the feel of what makes an Express a great amp. At a basic level, the output stage clips first and continues to clip as you turn it up more, so it's a major contributor all the time.

AB1 amps can be configured to overdrive very, very heavily in the power section before the PI clips (18 Watt is an example). But many aren't and in that case, you can only overdrive the power stage so far before PI or even preamp clipping kicks in. You'll generally hear that transition point with a change in the tone and see it in the output waveform.

If you want to know more about why the Express is so special, it's all in that thread. But it's not an easy mechanism to get your head around, so feel free to ask any questions.

Thank you for the detailed response Katopan! So if I'm not mistaken and if I understand you correctly, compared to Trainwreck Express amps and their clones, the older NMV amps such as Marshall 18 watt and Plexis produced purely power tube distortion at any volume while TW Express amps produced blended distortion sounds when the volume pass 12 o'clock. Correct?

Am I correct when I said what makes Trainwreck Express amps sound a little bit higher in gain and thus more distorted, especially when we cranked them pass 12 o'clock, than the older non-master volume amps is because of the blended distortion from the power tubes, PI and 3rd stage preamp? So in order to make a NMV amp to sounds higher in gain and more distorted, we will need to make other components not just the power tubes inside the amp to start clipping too?

What puzzles me is square wave distortion sounds produced by solid-state instruments such as Fuzz Face and other fuzz pedals sound different from those square wave sounds from Marshall 18 watt. Marshall 18 watt produced a lower gain sounding square wave distortion sounds while the Fuzz Face produced a louder, harsher, higher gain square wave distortion sounds with longer sustain although the circuit topology of Fuzz Face is much simpler.

[katopan]

Thank you for the detailed response Katopan! So if I'm not mistaken and if I understand you correctly, compared to Trainwreck Express amps and their clones, the older NMV amps such as Marshall 18 watt and Plexis produced purely power tube distortion at any volume while TW Express amps produced blended distortion sounds when the volume pass 12 o'clock. Correct?

Yes, but with the caveat that the Express blended distortion is in a way that is like no other amp. When the Express is in its sweet zone, the top half of the output waveform is pure EL34 power stage clipping and the bottom half is pure PI clipping underneath the power grid clipping threshold. That's why as raised in the other thread that when overdriven, one power valve in the Express is working really hard and the other isn't. That's because one is overdriven into grid clipping and the other is not. If you overdrive a Plexi far enough, the PI distortion overlays on the power valve distortion on both sides of the waveform.

Am I correct when I said what makes Trainwreck Express amps sound a little bit higher in gain and thus more distorted, especially when we cranked them pass 12 o'clock, than the older non-master volume amps is because of the blended distortion from the power tubes, PI and 3rd stage preamp? So in order to make a NMV amp to sounds higher in gain and more distorted, we will need to make other components not just the power tubes inside the amp to start clipping too?

The Express amp topology has a higher gain (as in voltage signal multiplication) between the input and the power grids than say a Plexi. So it can overdrive the power stage more (longer time in class AB1 grid clipping per cycle). That's nothing to do with the blended distortion. To get a Plexi to sound more distorted you need to drive it harder (upfront boost) or modify it to have more gain between the input jack and the power grids. But the amp topology is such so that will only go so far before other stages (PI and then the preamp) will overdrive. That will make more distortion of course, but blended rather than pure power stage.

What puzzles me is square wave distortion sounds produced by solid-state instruments such as Fuzz Face and other fuzz pedals sound different from those square wave sounds from Marshall 18 watt. Marshall 18 watt produced a lower gain sounding square wave distortion sounds while the Fuzz Face produced a louder, harsher, higher gain square wave distortion sounds with longer sustain although the circuit topology of Fuzz Face is much simpler.

It's all about the clipping curve or transfer characteristic. As I said before my favourite amps use real pentodes and have them set up for a fair amount of screen compression leading into clipping. This is different to compression like from a compressor, which is like a form of averaging automatic gain control. Screen compression acts every cycle without an averaging effect, and softens the transition into grid clipping which can be seen in the waveform. Beam tetrodes don't draw the same screen current leading up to and in grid clipping, so they don't have as much screen compression effect. A Fuzz Face has none. When comparing square waves, you've got to zoom in on the corners and see how sharp/soft they are. That equates to a difference in harmonic content that doesn't look like much but makes a big difference in tone. Actual tone shaping of the amp or pedal circuit before and after the clipping is another factor on top of this as well that can change the tone completely.