Merlins 10k 100p grid input.

[RockinRocket]

Where would the best place to ground the100p?

Can I tie the 100p from V1 Grid to cathode pin (before the bypass cap resistor)?

[Phil_S]

Do that if you want interaction with the cathode.

I would ground it on the ground side of the cathode R/C.

[RockinRocket]

That what I thought thanks.

What about ground the 100p to the shielded input cable braid?
Im probaby wrong on that

[Colossal]

What about ground the 100p to the shielded input cable braid?
Im probaby wrong on that

I thought about that when I read your post. The grid resistor and added cap could go neatly on a terminal strip or just wired directly to the socket with the cap heatshrinked (heatshrunk?) in parallel (one leg tied to the shield braid) and isolated from the grid resistor.

[Ten Over]

You can tie the 100pF cap. from the grid to the cathode and there will be no interaction between the grid and cathode as long as the cathode is bypassed. The high frequency roll-off will be the same whether the 100pF cap. is connected to ground or to a bypassed cathode because the bypass capacitor gives the 100pF cap. a very low impedance path to ground at those frequencies.

You can get a similar high frequency roll-off by using a 10pF cap. tied from the plate to the grid and a 4.7K grid stopper.

[pdf64]

You can tie the 100pF cap. from the grid to the cathode and there will be no interaction between the grid and cathode as long as the cathode is bypassed. The high frequency roll-off will be the same whether the 100pF cap. is connected to ground or to a bypassed cathode because the bypass capacitor gives the 100pF cap. a very low impedance path to ground at those frequencies...

I think that the point of the 100pF cap is to achieve a similar degree of 'RFI rejection at the input' with a 10k grid stopper as with the regular 33k of input grid stoppering.
A cathode is usually bypassed with an ecap; they may have a fair bit of self inductance, creating a significant impedance at RF.
Hence it may be best to connect the cap to 0V, rather than the bypassed cathode.

...You can get a similar high frequency roll-off by using a 10pF cap. tied from the plate to the grid and a 4.7K grid stopper.

I'm not sure that would prevent the g1-k diode acting as a demodulation detector, in the presence of a significant degree of RF at the input.

[Ten Over]

I think that the point of the 100pF cap is to achieve a similar degree of 'RFI rejection at the input' with a 10k grid stopper as with the regular 33k of input grid stoppering.
A cathode is usually bypassed with an ecap; they may have a fair bit of self inductance, creating a significant impedance at RF.
Hence it may be best to connect the cap to 0V, rather than the bypassed cathode.

That's what I thought the point was.

Cap manufacturers are a little stingy on the inductance values, but I found a 22uF @ 25V e-cap with 30nH. So we have a 100pF cap in series with a 30nH inductor when the cap is connected to the bypassed cathode. The impedance of the inductor doesn't figure into the total impedance much until the frequencies get above the resonant frequency, so I will start at 300MHz. At 300MHz the total impedance of the series combination is 51 Ohms, at 600MHz it is 110 Ohms, and at 800MHz it is 149 Ohms. That makes a pretty impressive voltage divider with a 10K resistor.

[Ten Over]

...You can get a similar high frequency roll-off by using a 10pF cap. tied from the plate to the grid and a 4.7K grid stopper.

I'm not sure that would prevent the g1-k diode acting as a demodulation detector, in the presence of a significant degree of RF at the input.

Yeah, well, I've been using this technique for decades and it has kept every local radio station out of my high gain amps. Seriously, though, I am having difficulty seeing what the g1-K diode is going to demodulate when the RF has been killed by the local NFB.

[pdf64]

...You can get a similar high frequency roll-off by using a 10pF cap. tied from the plate to the grid and a 4.7K grid stopper.

I'm not sure that would prevent the g1-k diode acting as a demodulation detector, in the presence of a significant degree of RF at the input.

Yeah, well, I've been using this technique for decades and it has kept every local radio station out of my high gain amps. Seriously, though, I am having difficulty seeing what the g1-K diode is going to demodulate when the RF has been killed by the local NFB.

If RF gets demodulated by the g1-k diode, then suppressing RF after that point (eg by rolling off the high frequency gain of the stage) isn't beneficial; as it's not RF anymore.

[Ten Over]

If RF gets demodulated by the g1-k diode, then suppressing RF after that point (eg by rolling off the high frequency gain of the stage) isn't beneficial; as it's not RF anymore.

Are you suggesting there is a time lag during which RF is demodulated and when the local NFB catches up it suppresses the RF but not the demodulated signal? This doesn't seem to happen if you insert a 68K grid stopper and use the tube's input capacitance (primarily plate-to-grid with the Miller effect) to severely reduce RF at the grid. I don't see how an external plate-to-grid capacitor, which also severely reduces RF at the grid, is any different than internal plate-to-grid capacitance. All three strategies appear to me as though they reduce RF at the grid in order to eliminate RFI.

[pdf64]

Dunno about a time lag, but yes, I was thinking that the proposed low pass filter was somehow later in the signal chain than the g-k diode demodulator.
And I concede your point, ie that the regular internal Miller capacitance seems to work just fine at suppressing RF, so no reason that adding to it with your external 'plate-to-grid' cap shouldn't do the same.
Another 'learning experience' for me