Express Output Waveform Shots

[katopan]

PM sent.

[ve3vxo]

Hi Craig

I found your Thrifty Wreck design. Am I correct that you are talking about using that, but replacing V3,V4 and V5 with two 6F4P ? Would it be correct to assume that the split plate loads for the PI are to be retained? Also I found a datasheet but not an English one. Do you have a suggestion about how to choose the operating point for these tubes?

Joe

[jazbo8]

Gingertube (a very experienced EE and amp designer) posted awhile back on MEF wrt to the Trainwreck mojo, he highlighted the grid leak resistor for the third stage, and the third gain stage itself, there was no mention of the screen grid resistor or the PI clipping - perhaps those were viewed as secondary effects. Just more food for thought...

[katopan]

Thanks Jazbo. For some reason I couldn't get that link to work when you first posted this. But now it worked and I've had a read through. I have a huge amount of respect for Gingertube and enjoy his posts.

I said it earlier in this thread. The assymetry of the output waveform is backwards/upside down to what is explained by thinking of the 3rd stage as a cold clipper. It's that anomoly that got me to investigate this further in the first place. Otherwise I would've just written it off as all 3rd stage assymetrical clipping like many have before and moved on. But it turned out to be a lot more interesting than that.

[thinkingchicken]

Just to finalise it, here is my description of how the V4 grid backing off part way through clipping to become PI clipping dominent, creating the mixed mode distortion:

The shift in duty cycle means the positive side of the V4 grid widens and the positive side of the V5 grid narrows, even though the amplitudes are similar. On the power grid side of the coupling caps these signals rebalance themselves around the power grid bias point. The V4 side with its wider positive peak moves downward while the V5 side with its narrower positive peak moves upward. When the duty cycle is symmetrical the PI power node hasn't dropped a heap yet (from power screen current being drawn through the 1K screen power chain resistor) and the PI clipping is well above the power grid clipping. So the output is power grid clipping dominant on both sides with equal headroom, plus the smoothing compression effect of standard screen compression.

When the asymmetry happens the PI power node voltage has dropped (from the power screens drawing current) which by itself lowers the PI output headroom, still above the power grid clipping threshold. The output peaks of the PI plates are reasonably flat topped, while the power grid leak resistors have a dropping slope because of the PI to power grid coupling caps charging up. You then add in the shift from asymmetry and the PI pin 6 plate side drops below the V4 grid clipping threshold during the last half with that dropping slope, moving to PI clipping dominant with less headroom. But the V5 side shifts upward and therefore stays above V5 grid clipping, leaving it as power grid clipping dominant.

How would pure power grid (power valves for example EL34) clipping distortion sound like? Are there any audio examples of them at middle and maximum volume? At the maximum volume on Trainwreck Express, which one of them dominated the distortion sound?

[pdf64]

Y

How would pure power grid (power valves for example EL34) clipping distortion sound like? Are there any audio examples of them at middle and maximum volume? At the maximum volume on Trainwreck Express, which one of them dominated the distortion sound?

You've already started a thread focussed on such matters, why post basic generic queries here too?

[thinkingchicken]

Y

How would pure power grid (power valves for example EL34) clipping distortion sound like? Are there any audio examples of them at middle and maximum volume? At the maximum volume on Trainwreck Express, which one of them dominated the distortion sound?

You've already started a thread focussed on such matters, why post basic generic queries here too?

Katopan seems to be the most experienced regarding this matter so that's why I asked him. But nevermind, lets get to the other thread.

[rooster]

Katopan, alright, just wanted to say hi and tell you that your work is much appreciated here. (I just reread all 14 pages.)

[katopan]

Katopan, alright, just wanted to say hi and tell you that your work is much appreciated here. (I just reread all 14 pages.)

Thanks Rooster, I appreciate the feedback. I put a lot of time into this original investigation. So many other builds and other bench testing I want to do, but just haven't had the time in the last few years.

[Colossal]

I will second Rooster's comment. Excellent work and a good read! I find myself revisiting that thread periodically.

[katopan]

I will second Rooster's comment. Excellent work and a good read! I find myself revisiting that thread periodically.

Thanks Dave.