Does 5E3 have crossover distortion?

[lowell]

I'm curious now, learning that cathode biased tubes can create crossover distortion caused by bias shifts when the tube is working harder. I have a 5E3 amp that does not sound bad at all, in fact it sounds incredible when cranked to "11." I don't have the amp in my possession right now so I cannot scope it. Can someone tell me if the 5E3 goes into c-over distortion at any point? If so, how does it still sound so good?

[MWJB]

Very probably, depends on how hot its biased, many push-pull amps will show a cross-over notch on the output at overdriven volumes. But here's the pertinent question - If it sounds "incredible cranked to 11" and I assume that you're getting satisfactory cleans at nominal rated output, why are you worrying about it having a particular kind of distortion? If it sounds good, it IS good.

I would be worrying about EXCESSIVE crossover distortion if you had a raspy, thin, fuzzed-out tone at all volumes and couldn't shake it. This is unlikely unless you have very low voltage with the stock circuit, or a MUCH larger cathode resistor than 250/270ohms.

Use the scope for isolating & troubleshooting problems that you have with the tone, rather than scaring yourself into believing you have a "problem" when you actually like the tone & performance is probably typical.

[lowell]

MWJB,
I'm actually not at all concerned w/ that amp. I'm simply curious due to another design of my own that is cathode biased and IS having crossover distortion issues. The clean channel of my design is basically a 5E3 w/ a baxandall tone stack and a 12dw7 (ecc832) for a gain/concertina splitter. I also have all my plate voltages quite close, within 10v of the 5E3 stock voltages. The other difference in my design is that I'm not using a tube rectifier but instead have a 200ohm resistor for simulation. Oh, and I ditched the 6V6 bypass cap altogether to rid of the crossover distortion. Just wondering why my new design has that awful thin fuzzy sound when cranked and the 5E3 sounds fantastic.

[lowell]

I just dug out an old schem of my 5E3 that I'd scoped and drawn waveforms of each stage on. Looks as though the plates of the 6V6 with the volume all the way up indeed DO NOT go into crossover distortion. Important to note is the absence of a power tube bypass cap. In fact it looks like moderate clipping w/ very rounded corners. Also one side of the OT plates have very asymmetrical waveforms. Hmmm.... I'm wondering if the absence of tube rectifier and use of 12dw7 splitter are causing more pronounced bias shifts in the power tubes.

[Enzo]

SO what does the waveform look like at the power tube grids? SImple clipping doesn;t need fancy phase shifting explanations, by the way. Have you ruled out clipping in the phase inverter?

[MWJB]

Hi Lowell,

I saw the thread about your one off build, why do you refer to it as "crossover distortion" specifically? Your other build seems to have a lot more stages than a 5E3, what is the clean rated output of your custom build (w/6V6 in cathode bias).

If you hit the power tubes with a big enough signal, beyond clean power, you can induce crossover distortion (it comes in earlier on the sweep of the volume pot). What would be handy with regard to your custom build would be some AC voltage figures at each stage (as Enzo points out, distortion earlier in the circuit could be what you are seeing transferred to the power tube plates/OT secondary)...this really belongs in the other thread.

A 200ohm resistor won't sound like a tube rectifier, too large a "sag" resistor will make the amp sound "wheezy", try bypassing it.

I'm not sure whether you are saying the 5E3 has a bypass cap, or not?

[lowell]

Hey,
I was thinking the same thing when this issue came about so I scoped all grids and plates. The grids of the power tubes only begin clipping around 75% volume. I'm not sure how I'd differentiate b/t the PI clipping and the power tubes grids clipping, unless I insert a pot at the power tube grids.

Also I'm only speaking of the clean channel here which has the same amount of stages as the 5E3. Also I have a switch that cuts the signal at the 6L6 grids so I'm only speaking of the 2 6V6.

Neither of these amps (5E3, my build) have cathode bypass caps.

Ok I'll bypass that sag resistor and see if I like it.

[lowell]

SO what does the waveform look like at the power tube grids? SImple clipping doesn;t need fancy phase shifting explanations, by the way. Have you ruled out clipping in the phase inverter?

It looks quite normal until the volume reaches 75% or so, even then it looks like compression/distortion, not really full grid clipping as the corners are still rounded.

[MWJB]

"The grids of the power tubes only begin clipping around 75% volume." I wouldn't worry too much about where the volume pot is set, it doesn't necessarily relate to output.

How many W are you getting at the speaker before clipping?

[tubeswell]

It looks quite normal until the volume reaches 75% or so, even then it looks like compression/distortion, not really full grid clipping as the corners are still rounded.

Interesting discussion. From what I understand from Merlin b's website, hard grid clipping doesn't occur anyway do to the 'island effect'. Does that sound right??

[lowell]

"The grids of the power tubes only begin clipping around 75% volume." I wouldn't worry too much about where the volume pot is set, it doesn't necessarily relate to output.

Depending on where the volume is though the grids of the proceeding stages may or may not clip. I thought that was relevant.

How many W are you getting at the speaker before clipping?

I'll get back to you on this. Also, where am I looking for clipping when measuring this? grids, plates, speaker out?

[lowell]

well didn't have time to do that stuff yet. however I had a thought. Does a tube rectifier, when sagging, counteract grid conduction and bias shifts thus taming crossover distortion? I can't see how this would be since it seems to me the tube's bias actually gets colder as the voltage sags. Unless I'm missing something.

[tbryanh]

"Does a tube rectifier, when sagging, counteract grid conduction and bias shifts thus taming crossover distortion? ... it seems to me the tube's bias actually gets colder as the voltage sags."

During sag, the lower plate voltage causes cathode current to decrease. The lower cathode current causes bias voltage to decrease. The lower bias voltage causes cathode current to increase to offset somewhat the decrease in cathode current from the lower plate voltage. (This action occurs immediately with un-bypassed cathode resistors. It is not so immediate with bypassed cathode resistors.)

A new balance occurs, and the bias, relative to the new plate voltage, is close to the same cold-ness or hot-ness as it was before.

Since bias voltage is lower, it is easier for grid conduction to occur. As a result, it is easier for crossover distortion to occur.

How much easier is it for crossover distortion to occur? Not sure. The voltage might sag 10% at most. This would probably put a 10% cap on the amount of increase in easiness, but the amount of increase in easiness might be much less than 10%.

[MWJB]

"Quote:
How many W are you getting at the speaker before clipping?

I'll get back to you on this. Also, where am I looking for clipping when measuring this? grids, plates, speaker out?"

W RMS are measured at the speaker out, what we are trying to establish here is whether your amp is making a realistic output before distortion sets in. This doesn't necessarily relate to whether or not you are going to get a tone you like (I notice for instance you are using a 410ohm cathode resistor at V1 - 5E3 uses 820ohms). An amp might make maximum clean W RMS at a quarter of a turn on the volume or at 90%, that's why I'm saying it's not relevant. Once we are happy that you are getting appropriate clean power (& crossover distortion specifically is not the issue), then we can work back & tweak the tone.