Scoping a Dumble

[LeftyStrat]

I'm by no means "Master O' The Scope,' it seems I keep finding ways to use it every day, but I can give an example of my explorations with my Express.

I have a Weber Mass with a line out. I set this up for full load, with no speaker attached, and I plug a 1/4 inch jack with no housing into the line out. This way I can easily clip a probe on the output of the amp.

I feed a low level sin wave into the amp, and look at each stage by attaching the probe after each coupling cap, to avoid having to deal with DC offsets.

I have a dual scope, so I usually have one trace on the output and move the other around.

I adjust the amplitude of the sin (you could just as easily use the gain on the amp), until I see the start of waveform flattening, i.e. clipping. Then check each stage to see if it is clipping.

What's interesting in the Express, is that the output clips first, followed by the PI. But then something really interesting starts happening. The third gain stage clips, but only on half of the waveform. Since this stage is AC coupled to the next, it actually shifts the crossover points of the waveform, and you can see that the almost square wave on the output looks a little like PWM, altering the duty cycle.

So Kenny wasn't using the third stage so much as a clipper, but as more of a "waveform animator."

At least those are my conclusions. Nonetheless, I feel like I understand the circuit better, and can add this knowledge to my bag of tricks.

[Structo]

Cool!

Thanks for the tips!

Keep'em coming.

[ayan]

Hi,

You mention that "the third stage clips, but only half of the waveform." That's what common cathode preamp stages do, at some point the grid volatge becomes too large and larger than the cathode voltgage and the tube loses its bias. The stage always clips the positive cycle at the input, so clipping will appear at the bottom of the waveform at the output, since these are inverting stages.

There's a great explanation about how things work here:

http://www.valvewizard.co.uk/

You can look at the "Triode Stage Link" which has some great information on this subject.

Cheers,

Gil

I'm by no means "Master O' The Scope,' it seems I keep finding ways to use it every day, but I can give an example of my explorations with my Express.

I have a Weber Mass with a line out. I set this up for full load, with no speaker attached, and I plug a 1/4 inch jack with no housing into the line out. This way I can easily clip a probe on the output of the amp.

I feed a low level sin wave into the amp, and look at each stage by attaching the probe after each coupling cap, to avoid having to deal with DC offsets.

I have a dual scope, so I usually have one trace on the output and move the other around.

I adjust the amplitude of the sin (you could just as easily use the gain on the amp), until I see the start of waveform flattening, i.e. clipping. Then check each stage to see if it is clipping.

What's interesting in the Express, is that the output clips first, followed by the PI. But then something really interesting starts happening. The third gain stage clips, but only on half of the waveform. Since this stage is AC coupled to the next, it actually shifts the crossover points of the waveform, and you can see that the almost square wave on the output looks a little like PWM, altering the duty cycle.

So Kenny wasn't using the third stage so much as a clipper, but as more of a "waveform animator."

At least those are my conclusions. Nonetheless, I feel like I understand the circuit better, and can add this knowledge to my bag of tricks.

[katopan]

LeftyStrat, I'm very happy to read your brief but to the point description. It is exactly what I also see. In fact some time ago I posted a heap of very detailed waveforms of all this and was a bit surprised at the lack of response and agreement. Well, each to their own. But you can learn a lot with this sort of testing, especially as Jelle said if you keep in the back of your mind what the amp sounds like at those settings. You can form links in your head of what test waveforms are doing visually with what you hear with a guitar plugged in. Then thinking through design changes and concepts in your head, you can think of response curves, imagine what test waveform that would generate, and then relate that to how you know it's going to sound.

[LeftyStrat]

Hi,

You mention that "the third stage clips, but only half of the waveform." That's what common cathode preamp stages do, at some point the grid volatge becomes too large and larger than the cathode voltgage and the tube loses its bias. The stage always clips the positive cycle at the input, so clipping will appear at the bottom of the waveform at the output, since these are inverting stages.

There's a great explanation about how things work here:

http://www.valvewizard.co.uk/

You can look at the "Triode Stage Link" which has some great information on this subject.

Cheers,

Gil

I guess I used a bad choice of wording. I tend to think of clipping as squashing of the top or bottom of the waveform, though you're right, clipping is what happens on one side and grid current limiting is what happens on the other.

Normally we bias in the middle of these two so that both halves of the waveform are affected. Ken's cold bias insures that only one side of the waveform is affected.

Also the lack of a bypass cap tends to soften the clipping and sounds much more like compression.

I once tried using that stage in a Marshall amp on the second triode. It increased sustain, but you lost a lot of the Marshall crunch from the 820R/.68 combo.

[LeftyStrat]

LeftyStrat, I'm very happy to read your brief but to the point description. It is exactly what I also see. In fact some time ago I posted a heap of very detailed waveforms of all this and was a bit surprised at the lack of response and agreement. Well, each to their own. But you can learn a lot with this sort of testing, especially as Jelle said if you keep in the back of your mind what the amp sounds like at those settings. You can form links in your head of what test waveforms are doing visually with what you hear with a guitar plugged in. Then thinking through design changes and concepts in your head, you can think of response curves, imagine what test waveform that would generate, and then relate that to how you know it's going to sound.

Thanks, and I completely agree. As I mentioned in the previous post, I tried that cold biased stage in a Plexi circuit, and while it noticeably increased sustain, you lost the Marshall crunch that comes from compressing the other side of the waveform from grid current limiting. Which again you can clearly see on a scope.

But like you and Jelle say, being able to equate what you hear with what you see contributes greatly to your understanding. It's not enough for me to build an amp that sounds good. I want to know why it sounds good.

[ayan]

I guess I used a bad choice of wording. I tend to think of clipping as squashing of the top or bottom of the waveform, though you're right, clipping is what happens on one side and grid current limiting is what happens on the other.

Normally we bias in the middle of these two so that both halves of the waveform are affected. Ken's cold bias insures that only one side of the waveform is affected.

I guess we'll have to disagree here. In a "regular common cathode stage"(i.e., designed a la Dumble), to clip the other half of the waveform you'd have to hit you voltage rails out the output... By the time that would happen, you'd would have long lost the "positive" half of the waveform at the input and the sound would be -- I think -- less than dersirable. If you current starve the tube using a large (usually unbypassed) cathode resistor -- a la Soldano, some Marshalls, Mesa Rectos, etc. -- then you'd have an entirely different situation and sound. Not many amps are designed that way, certainly no Dumble I've ever heard of. In Dumbles, you will see some clipping on one side, then the next stage will clip the other side, etc.

Cheers,

Gil

[LeftyStrat]

If you current starve the tube using a large (usually unbypassed) cathode resistor -- a la Soldano, some Marshalls, Mesa Rectos, etc. -- then you'd have an entirely different situation and sound.

My quote was about scoping the Trainwreck Express, which uses the same technique as you described above.

I haven't had a chance to scope a Dumble clone, as I haven't built one yet.

[ayan]

If you current starve the tube using a large (usually unbypassed) cathode resistor -- a la Soldano, some Marshalls, Mesa Rectos, etc. -- then you'd have an entirely different situation and sound.

My quote was about scoping the Trainwreck Express, which uses the same technique as you described above.

I haven't had a chance to scope a Dumble clone, as I haven't built one yet.

Gotcha, we're in synch then. Dumbles are very "predicatable" when looking at them with a scope; clipping occurs gently, one side at a time.

Cheers,

Gil