Ever Tried A Reverse Bright Cap?

[jordan86]

My final "issue" on my single ended marshall is the Master Volume. It has a 1 Meg volume pot between the CF tone stack and the power tube. I've noticed that past 3 o clock it introduces a much harsher high end. I'm wondering if anyone has tried a reverse bright cap on a pot? I know the idea of bleeding off highs has been implemented on plate resistors (dull cap). Thinking a high pf cap tied from the wiper to ground lug. Or maybe something like 47pf in series with a 220K resistor.

I'd like it to be less noticeable at lower settings. Or maybe not noticeable until the highest settings. At lower settings on the pot there would be less resistance to jump, making it less noticeable yes? Being single ended the MV usually lives between Noon and 3 o clock. No lower. I'm real happy with the tone up to 3 o'clock. Just wondering how to isolate and address that without darkening the entire amp.

Schematic
https://el34world.com/Forum/index.php?action=dlattach;topic=27019.0;attach=89468;image

[thetragichero]

are the harsh highs coming from the master volume or from driving the tube harder?
if you've got a negative feedback loop, try bypassing the resistor with a small cap (47pf-470pf)

[jordan86]

are the harsh highs coming from the master volume or from driving the tube harder?

I had that thought. I’m not sure. It would seem the latter could certainly be true. I presume a higher value grid stopper going into the power tube could help with that? But would it make all settings on the MV darker?

There is no negative feedback.

[HotBluePlates]

... the Master Volume. It has a 1 Meg volume pot between the CF tone stack and the power tube. I've noticed that past 3 o clock it introduces a much harsher high end. I'm wondering if anyone has tried a reverse bright cap on a pot? ...

I don't recall the full explanation, but this is a known issue that goes something like this:

MV full-up is the actual "voice" of the amp.  But a lot of times people turn the MV down.  The resistance between the wiper and the input of the MV pot is a large-ish series resistance that rolls off treble.  So the amp sounds brighter with the MV turned up, darker with the MV turned down.

The fix usually proposed is to insert a ~100kΩ resistor from the wiper of the MV to the wire running to the output tube grid.  The has the effect of somewhat isolating the output tube grid from seeing a changing series-resistance at the MV pot.  The amp should tend to retain the "darker" sound of the MV being turned down all the time.

I don't really mess with master volume circuits, and can't tell you if that's a very effective fix.  But it's been in print for at least 20 years as a recommended change to JCM800 amps, so it will probably work here.

[pdf64]

My take on that mod is that it provides a more constant source impedance feeding the input Miller capacitance of the next stage, eg the LTP and power amp.
For a Marshall tone stack feeding a 1M audio pot, typical mid way tone control settings might result in a source impedance of about 100k. So with the master vol up full CW, its 1M track resistance and about 2M power amp input impedance with both be in parallel with that, so a total source impedance of about 80k.
As the master is turned down, the source impedance will increase to a peak of about 250k at the electrical mid point, maybe about 3 o’clock. After that, turning down further will reduce the source impedance, so about 90k at noon, 10k at 9 o’clock.
So assuming a Miller capacitance of 120pF, the resulting low pass filter's (LPF) corner frequency will be 16kHz, 5kHz, 15kHz and 133kHz respectively.
Adding 100k to the above source impedances and the LPF corner freqs come down to 7kHz, 3.8kHz, 7kHz and 12kHz respectively.
Hence adding 100k to the master vol pot's output will act to even out the change in source impedance (ie that will interact with the following stage’s input capacitance), over the range of the pot’s rotation.

[d95err]

I’n guessing the power tube is pushed very hard above 3 o’clock. AFIK, single-ended powertube distortion is often rather harsh (at least, that is what I hear in most soundclips of cranked single ended anps).

One idea is to simply use a voltage divider before the MV to limit the max signal level into the power tube.

This approach is used by e.g. Friedman amps.

[HotBluePlates]

... AFIK, single-ended powertube distortion is often rather harsh ...

By itself, distortion being in a single-ended tube is immaterial.  If anything it may be darker than in push-pull power tubes, because the even harmonics aren't being canceled in a push-pull stage.

Remember, the preamp tubes are all single-ended so if preamp distortion is okay we're probably not objecting to single-ended vs push-pull.

I’n guessing the power tube is pushed very hard above 3 o’clock. ...

True, it remains to be seen whether the harshness is due to distortion in the output tube or some sonic effect of the MV, as thetragichero pointed out.

But this is worthwhile to explore, as folks building master-volume amps over time have found obtaining a whole bunch of distortion in a single tube stage sounds worse/harsher than a little distortion per stage, spread out over multiple stages.

My take on that mod is that it provides a more constant source impedance feeding the input Miller capacitance of the next stage, eg the LTP and power amp. ...

This could be!  It makes more sense in light of this amp's topology.

I finally remembered what it was about in the JCM800:  more consistent bass roll-off at the coupling cap feeding the fixed-biased output stage.  Though here we lack that coupling cap due to the cathode-biased KT88.

[2deaf]

The control grid-to-plate capacitance of a KT88 is 1.2pF and the control grid-to-cathode capacitance is 16pF.  Transconductance is 11.1mA/V and the OT primary impedance is 2K5.  I get a total input capacitance of 50.5pF.

The input impedance of a KT88 is high enough for me to ignore.  With a worst case source impedance of 267K, the corner frequency would be 11.8KHz.  Under these circumstances, it seems unlikely that the low pass filter phenomenon is the cause of the original complaint.