Favorite Ways to Control Compression

[jabguit]

Interesting. What does LDR stands for please?

Light dependent resistor

[Tobyk]

He wrote that above: 10nF and 470K

He also wrote 6M8... I was just wondering if there were any legitimate target values that he had already played with.

6.8M is the grid leak bias value Valco used a lot. Mr pdf64’s idea was that a low value (470K) here instead maybe could cause compression, but he hasn’t tried it.

[dorrisant]

LDR = Light Dependent Resistor. A Vactrol is an example.

I see jack beat me to it... more coffee please.

kitten coffee.gif

[jabguit]

LDR = Light Dependent Resistor. A Vactrol is an example.

I see jack beat me to it... more coffee please.

kitten coffee.gif

I concur........

[pdf64]

I rigged up a grid leak biased CC ECC83 stage, and tried to test whether the gain was affected by signal level, ie compression, using a 400Hz sine wave, 20mV to 1.4Vrms (the output clipped above that).
Whether the grid leak was 470k or 10M, it didn’t.
I was surprised, even though the effective bias changed with signal level (VDC on grid), any effect on gain was negligible.

Sorry for the diversion

EDIT: on further investigation, using my bench PSU connected between 0V common and the 470k grid leak, -3,9V bias was required to halve the stage gain, ie gain reduction of -6dB.
-4.5V quartered the gain, -12dB gain reduction, but the operating point was by then so altered that the max 'cleanish' output signal (ie 'headroom') was greatly reduced. Only 10VDC across the 100k anode load, 0.1mA idle current. However, the overdriven output waveform was very rounded, asymmetric, no harsh clipping, as per a typical cold clipper.

So grid leak biasing wouldn't be likely to achieve any useful gain compression as an input stage.

To check whether it may be worth trying it as a 2nd or 3rd stage, according on the gain structure of the preamp, I tested a combination of fixed bias at the cathode with a series pair of LEDs, and grid leak bias.
Again, unfortunately that achieved practically no gain reduction of the stage, as the signal input was raised from 50mVp-p to 5Vp-p.

320V HT, 100k anode load, 1M load, 0.1uF input coupling cap.

[dorrisant]

Sorry for the diversion

Not a diversion... Good information.

[Tobyk]

I rigged up a grid leak biased CC ECC83 stage, and tried to test whether the gain was affected by signal level, ie compression, using a 400Hz sine wave, 20mV to 1.4Vrms (the output clipped above that).
Whether the grid leak was 470k or 10M, it didn’t.
I was surprised, even though the effective bias changed with signal level (VDC on grid), any effect on gain was negligible.

Sorry for the diversion

Thanks for testing it for us!
To answer the OP, I will go out on a limb and say: pentode screen bypass pot might be the best way to control compression.

[pdf64]

To try and investifate what was going on, and exhaust all possibilities for grid leak biasing and signal compression, I made a couple of further tests, see edit to my post below.

[Tobyk]

To try and investifate what was going on, and exhaust all possibilities for grid leak biasing and signal compression, I made a couple of further tests, see edit to my post below.

So you by accident found out that 470K grid leak bias acts as a cold clipper, but without the loss of gain – that's cool!

[pdf64]

To try and investifate what was going on, and exhaust all possibilities for grid leak biasing and signal compression, I made a couple of further tests, see edit to my post below.

So you by accident found out that 470K grid leak bias acts as a cold clipper, but without the loss of gain – that's cool!

I think it was cold clipping due to the fixed bias to the cathode, courtesy of the series pair of LEDs, about 3.2V I think (I’ll check).
A cold clipper can have a fully bypassed cathode (as with the LEDs) and so have high gain, yet still cold clip. It will also have grid current clipping too, though the grid leak bias tends to push against it (that being the bias shift that I was thinking would cause compression)..

[bepone]

EF83 is small vari-mu pentode and can be used for compression

[Tobyk]

EF83 is small vari-mu pentode and can be used for compression

So you’re saying remote cutoff pentodes generates compression, just by installing them instead of a sharp cutoff one? Or are there other factors involved? I’ve tried using a 6sk7 instead of a 6sj7 in v1, no difference in gain, tone or compression. Maybe needs a later gainstage?

[brewdude]

I always wonder if I chase the same kind of “compression” as others.

Is the goal to limit output peaks? And/or, raise the volume of quiet passages?

Is the goal to have output volume stay the similar as you control of amount of “gain” with the guitar volume?

Is it to make the amp feel less strident? Less ice picky?

Does a clean tone need different set of variables then a dirty tone?

I have a whole bunch of other questions…
However, I can’t really offer suggestions or solutions to any of these mostly rhetorical questions.

Though, I am curious if anyone is amused that the dilemmas we seek as guitar players are often in pursuit of diametrically opposed goals?
Bright and fat?
Warm and crunchy?
Compressed and dynamic?
To me, it seems the magic happens surfing the lines between these attributes searching for conditions that lead simultaneous goals. There are always trade offs.

[bepone]

So you’re saying remote cutoff pentodes generates compression, just by installing them instead of a sharp cutoff one? Or are there other factors involved? I’ve tried using a 6sk7 instead of a 6sj7 in v1, no difference in gain, tone or compression. Maybe needs a later gainstage?

according to the curves from datasheet, to acchieve compression effect + 2H because effect is applied only to half sinusoide, you need 15Vpp.
guitar cannot provide this voltage swing so one valve stage is needed in front of this stage, or clean boost pedal but with limited effect (7-8Vpp cca).

[Tobyk]

So you’re saying remote cutoff pentodes generates compression, just by installing them instead of a sharp cutoff one? Or are there other factors involved? I’ve tried using a 6sk7 instead of a 6sj7 in v1, no difference in gain, tone or compression. Maybe needs a later gainstage?

according to the curves from datasheet, to acchieve compression effect + 2H because effect is applied only to half sinusoide, you need 15Vpp.
guitar cannot provide this voltage swing so one valve stage is needed in front of this stage, or clean boost pedal but with limited effect (7-8Vpp cca).

Thanks, have you tried this tube yourself?
Also, what does ’2H’ mean please – 2nd harmonic?