Hard Clipping

[ViperDoc]

I'm very green when it comes to oscilloscopes (amongst all else!), but I put my 5F8A on the scope today and found mostly asymmetrical hard clipping throughout the circuit with the exception of V1 (12AY7) and more symmetric hard clipping on the second stage of the PI (V3) when turning up the knobs. I'm reading up on the basics of this through Aiken and Blencoe, but what would you recommend as most effective in softening clipping in a preamp? My thoughts were to change tubes, but I'm also very interested in how to adjust to taste.

Preamp cathode bias is where I'm looking at the moment. The other question I have is about power tube screen resistor values. I've seen discussion that suggests that the screen resistor value is relevant to both 1) lowering the screen voltage relative to that on the plates, and 2) the internal operation of the tube. I'm considering adding a series resistor to either the tail end of the choke into B+2 (thereby lowering all the preamp voltages a bit) or off the B+2 feeding the 470R screen resistors I have there at the moment.

The sound is incredibly strident when turned way up. It sounds great for moderate gain rock rhythm tones in the middle of the gain range, but cranked, this thing sounds like a burning porcupine. (If you catch my drift).

I appreciate your advice!

Brian

[pdf64]

Perhaps try a similar investigation on an amp whose overdrive you like; then compare and contrast the circuits and their overdrive characteristics.

[xtian]

Your ugly distortion might be a lead dress (parasitic oscillation) issue. Have you posted any photos of your build?

[ViperDoc]

Your ugly distortion might be a lead dress (parasitic oscillation) issue. Have you posted any photos of your build?

Good point. See photos just posted here: https://ampgarage.com/forum/viewtopic.p ... 06#p440006

[xtian]

Nice looking build, Brian. I would want to know how the overdrive sounds if you disconnect gNFB entirely. Also, with MV maxed (I've had trouble with PPIMV MVs making badness).

[Reeltarded]

It is pretty! Umm needs small grid stops.. less is better, input stage like 33k and then add smaller on later stages 10-15k and even put 10k on PI grids...

The input stage wants and needs it. It looks good.. small values on driven stages stops blocking distortion..

[ViperDoc]

Nice looking build, Brian. I would want to know how the overdrive sounds if you disconnect gNFB entirely. Also, with MV maxed (I've had trouble with PPIMV MVs making badness).

Indeed, I’ve been testing this with the PPIMV full up. Good advice.

Will play with NFB.

[ViperDoc]

It is pretty! Umm needs small grid stops.. less is better, input stage like 33k and then add smaller on later stages 10-15k and even put 10k on PI grids...

The input stage wants and needs it. It looks good.. small values on driven stages stops blocking distortion..

Thank you, sir. Do you mean 33K to replace each 68K on the input? I always run these amps with jumped channels, so the channel that uses both inputs would run at 34K in parallel, right?

I’ll post a schematic so I can make sure I understand which other grid stoppers you mean.

EDIT: HERE YOU GO:

VIPERDOC 5F8A SCH.pdf

[Reeltarded]

What does your scope say before and after V2 when the clipping takes over?

Is it possible V2 socket or tube has something going on.. I wonder. I had a trashy socket no visible sign of anything weird that did a very similar thing.

I would actually remove the PPIMV until the problem is solved. I do that first on any amp with it.

[teemuk]

Could you perhaps post a plc of waveforms on scope screen? Also what amplitude input signal are you using?

Those generic triode gain stages will not clip all that softly. And yes, asymmetric clipping is typical. Maybe your expectations do not meet reality. Something like a Tube Screamer or Big Muff actually clips way more softly than a generic 12AX7 common cathode amp.

The common cathode and direct coupled cathode follower stage should, in total, clip fairly symmetrically (but not softly, the cathode follower in particular hard clips). Differential amplifier - the phase splitter - has softer clipping characteristics, again fairly symmetric unless overdriven a lot. First gain stage won't clip with typical input signal levels (actually it's pretty hard to get any preamp stage to clip before power amp in these types of Fenders).

[Reeltarded]

Left message about shielding in other thread. Make sure it's covered before any other determination.

[ViperDoc]

Could you perhaps post a plc of waveforms on scope screen? Also what amplitude input signal are you using?

Those generic triode gain stages will not clip all that softly. And yes, asymmetric clipping is typical. Maybe your expectations do not meet reality. Something like a Tube Screamer or Big Muff actually clips way more softly than a generic 12AX7 common cathode amp.

The common cathode and direct coupled cathode follower stage should, in total, clip fairly symmetrically (but not softly, the cathode follower in particular hard clips). Differential amplifier - the phase splitter - has softer clipping characteristics, again fairly symmetric unless overdriven a lot. First gain stage won't clip with typical input signal levels (actually it's pretty hard to get any preamp stage to clip before power amp in these types of Fenders).

Understood. I'm not knowledgeable enough to have expectations, only interested to understand how it all works. When I tidied up the bias board and OT secondaries from yesterday, the blunt clipping waveform has become elusive! I'll see if I can find it.

Very interested to understand the common cathode design. I know Dumble separated the V1 cathodes on the Tweedle D and added LNFB on the PI, not applicable here in the same way, likely. I might try the cathode separation out a bit. When separating cathode components, you halve one and double the other if I recall. Need to look that up.

Here's a shot of the PI plates in various stages of gain. I need to further examine my scope, the scale controls turn into position controls when two signals are being watched at the same time. No idea why.

Here's the PI plates with a 110 Hz 500mV input signal (pink first triode, yellow second triode):

Both Normal and Bright Volumes 50%:

5F8A PI B50 N50.png

One Volume 100%, the other 0%:

5F8A PI B100 N0.png

Both Bright and Normal Volumes 100%:

5F8A PI N100 B100.png

[ViperDoc]

Nice looking build, Brian. I would want to know how the overdrive sounds if you disconnect gNFB entirely. Also, with MV maxed (I've had trouble with PPIMV MVs making badness).

I disconnected the gNFB and, as you would expect, the overdrive comes much earlier and is incredibly thick, definitely too hairy. I'm going to raise the NFB resistor value, though, to grab a bit of it back.

[teemuk]

I have to say I don't see anything abnormal in those waveforms. Looks like moderately symmetric clipping of tubes to me.

Here's a youtube clip that shows clipping of a single-ended amp towards both waveform lobes:
https://youtu.be/_6ulkga5ekg
Push-pull would be more symmetrical in how the waveform lobes clip but neither is overly "soft" at moderately large region like, say, this Tube Screamer circuit with solid-state diode clipping:
https://youtu.be/aA9-t3C9m7o

I have a hunch of what you have read and heard about tube distortion and I have a hunch that having that oscilloscope will also challenge most of that. ; )

[ViperDoc]

I have to say I don't see anything abnormal in those waveforms. Looks like moderately symmetric clipping of tubes to me.

Here's a youtube clip that shows clipping of a single-ended amp towards both waveform lobes:
https://youtu.be/_6ulkga5ekg
Push-pull would be more symmetrical in how the waveform lobes clip but neither is overly "soft" at moderately large region like, say, this Tube Screamer circuit with solid-state diode clipping:
https://youtu.be/aA9-t3C9m7o

I have a hunch of what you have read and heard about tube distortion and I have a hunch that having that oscilloscope will also challenge most of that. ; )

Indeed, there's nothing better than looking at the real thing. I also looked at the DC cathode follower in Blencoe's book, touting a way to smooth CF clipping into a "more creamy" distortion. I misunderstood that the CF stage has no appreciable net gain, but might clip, although I'm not sure how this would apply in this 5F8A circuit.

See here: http://www.valvewizard.co.uk/dccf.html

I see R1 values anywhere from 100R to 1M. I'm curious to see what this would do, so I might just throw this in and see what happens. I can always remove it.