FX series Loop design

[marshallguy]

I know some of you here are familiar with this schematic I posted recently and here’s my question relating to the FX Loop I’m using. I followed a design by Randall Smith Mesa Boogie Rectroverb. He uses two 2.2M resistors as a voltage divider on the grid as a stand alone cathode follower. I’ve seen it also done on other manufacturers designs with a 1M or 474k from cathode to grid tied between a 1k on cathode and 100k to ground.

The more I shift through Mesa’s schematics, the more I wonder if I should have following anything he did. I took a look at Peavey’s 5150 too. Jesus, did they go a little overboard?

After some research, maybe 2.2M is a bit high for a grid leak resistor and could be contributing to more noise?

The other way using a 474k grid to cathode would be a better option and lower the noise floor? Or, just change the current design from two 2.2M resistors to a lower value like 1M or less?

Thanks…

[sluckey]

It's a fixed bias CF. Input impedance is supposed to be high. Don't think in terms of grid leak resistor. You can find that exact same circuit in the 6G15. Merlin used two 10M in his example.

[marshallguy]

It's a fixed bias CF. Input impedance is supposed to be high. Don't think in terms of grid leak resistor. You can find that exact same circuit in the 6G15. Merlin used two 10M in his example.

Thanks sluckey for that feedback info.

[EL34]

Posted in wrong board..... moving

[PRR]

.....maybe 2.2M is a bit high for a grid leak resistor and could be contributing to more noise?…

They are in shunt with a lower impedance signal source. Like that, higher is less hiss.

[pdf64]

… After some research, maybe 2.2M is a bit high for a grid leak resistor and could be contributing to more noise?

If this was a sensitive circuit, eg with lots of gain after it, optimising the noise would be a consideration. But all the heavy gain is before it.

Nonetheless, there’s still the loop recovery gain stage, so you may want to take it a step further than Mesa.
To do that, yet still retain the benefits of fixed bias, the noise generated by the DC passing through the grid bias potential divider can be addressed in the same way as any fx pedal. See https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Electronics/Operational_Amplifiers_and_Linear_Integrated_Circuits_-_Theory_and_Application_(Fiore)/04%3A_Basic_Op_Amp_Circuits/4.03%3A_Single_Supply_Biasing
in particular the change from figs 4.3.1 to 4.3.2.
ie decouple the output of the potential divider and add a grid leak resistor between that node and the grid node.
By that, the DC in the grid leak resistor will be minimal, same as with a regular common cathode stage where the grid would be referenced to circuit common. So minimal noise, no matter how high a value, providing the total grid to cathode resistance limit for the valve type is observed.

[marshallguy]

… After some research, maybe 2.2M is a bit high for a grid leak resistor and could be contributing to more noise?

If this was a sensitive circuit, eg with lots of gain after it, optimising the noise would be a consideration. But all the heavy gain is before it.

Nonetheless, there’s still the loop recovery gain stage, so you may want to take it a step further than Mesa.
To do that, yet still retain the benefits of fixed bias, the noise generated by the DC passing through the grid bias potential divider can be addressed in the same way as any fx pedal. See https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Electronics/Operational_Amplifiers_and_Linear_Integrated_Circuits_-_Theory_and_Application_(Fiore)/04%3A_Basic_Op_Amp_Circuits/4.03%3A_Single_Supply_Biasing
in particular the change from figs 4.3.1 to 4.3.2.
ie decouple the output of the potential divider and add a grid leak resistor between that node and the grid node.
By that, the DC in the grid leak resistor will be minimal, same as with a regular common cathode stage where the grid would be referenced to circuit common. So minimal noise, no matter how high a value, providing the total grid to cathode resistance limit for the valve type is observed.

Ok, just to be clear did you say something to the effect to place a like 10k grid resistor between the the 2.2M resistors and the grid to V1a per schematic?

[pdf64]

… Ok, just to be clear did you say something to the effect to place a like 10k grid resistor between the the 2.2M resistors and the grid to V1a per schematic?

No, nothing like that, sorry.
Compare a and b of Fig 12.11 http://www.valvewizard.co.uk/cathodyne.pdf
Your R1 and 2 are currently 2M2.
If you want to optimise the noise performance (despite it being unlikely to yield much actual benefit / improvement), I’m suggesting to add Rg and C2 to your circuit. eg 1M and 10uF. Possibly reduce R1 and 2 a bit, eg 470k.

[marshallguy]

… Ok, just to be clear did you say something to the effect to place a like 10k grid resistor between the the 2.2M resistors and the grid to V1a per schematic?

No, nothing like that, sorry.
Compare a and b of Fig 12.11 http://www.valvewizard.co.uk/cathodyne.pdf
Your R1 and 2 are currently 2M2.
If you want to optimise the noise performance (despite it being unlikely to yield much actual benefit / improvement), I’m suggesting to add Rg and C2 to your circuit. eg 1M and 10uF. Possibly reduce R1 and 2 a bit, eg 470k.

Ok, I see now, yeah, probably not going to gain much from changing it. I’m trying to solve another problem that’s trending at the moment and trying to eliminate every possibility.

Amp in ringing over any note I play. Injecting 150mv @500hz what I’m seeing is the pure sign wave has this jagged rippling look to it. I removed the 180p Clean Ch. MV bypass treble peaking cap and the the volume at less then 2 there is improvement but not until I back the treble pot down to below 2 does it disappear. I also removed the 4.7uf cap @ V4a temporarily to reduce gain for now but it seems be near the input where the problem starts. I thought I cured that problem with the 390p to ground but injecting 1000hz instead of 500hz made it harder to detect.

[marshallguy]

.....maybe 2.2M is a bit high for a grid leak resistor and could be contributing to more noise?…

They are in shunt with a lower impedance signal source. Like that, higher is less hiss.

Thank you.

[marshallguy]

Posted in wrong board..... moving

I would like to if possible close this discussion.

Thank you everyone.

[sluckey]

Just stop posting comments and the thread will slowly drift away. There may not be any info in this thread of any value to you but someone may come along and find the answer to a question they have.