What's the truth on Miller and pentodes? Many references vaguely say it isn't a problem in pentodes. Does that mean that there is zero Miller effect?
Also, which spec is the correct grid capacitance to use in calculations? The spec sheet lists:
cg1(a) = 3.8 pF
ca(g1) = 5.1 pF
cag1 = max 0.05 pF
cg1f = max 0.0025 pF
This is for an EF86 or EF806 used as at the input of a preamp. I'd like to push the input impedance to 3 Megohms or maybe a bit more as seen by the driving source. The pentode will be configured as a normal gain stage.
JJ EF806 spec sheet pdf
Also, which spec is the correct grid capacitance to use in calculations? The spec sheet lists:
cg1(a) = 3.8 pF
ca(g1) = 5.1 pF
cag1 = max 0.05 pF
cg1f = max 0.0025 pF
This is for an EF86 or EF806 used as at the input of a preamp. I'd like to push the input impedance to 3 Megohms or maybe a bit more as seen by the driving source. The pentode will be configured as a normal gain stage.
JJ EF806 spec sheet pdf
flysig said:
For audio frequencies, it's seldom a problem.
Only if you connect them as cathode followers or grounded grid amps. For all grounded cathode/source/emitter amps, there is always Miller Effect. The reverse transfer capacitance for that tube is 0.05pF. Even with a gain of 100, that comes to a Cmiller of 5.05pF. That's not a problem.
hey-Hey!!!,
The Miller effect is what the g2/screen addresses. Miller deals with the amplified voltage at the output( the plate) and the capacitance between g1( or just plain grid in a triode ) and the anode. Since the device is amplifying the voltage difference between the two( the amplification factor is the delivered one, and not mu; Vanode:Vgrid, yes? )the capacitance appears larger; more current flows into it...since the capacitance isn't changing, we get additional current through the amplified voltage. Q=CV; take time derivative d/dt( Q=CV ) and get i=CdV/dt the anode voltage changing by time( and it is amplified, remember?).
cheers,
Douglas
The Miller effect is what the g2/screen addresses. Miller deals with the amplified voltage at the output( the plate) and the capacitance between g1( or just plain grid in a triode ) and the anode. Since the device is amplifying the voltage difference between the two( the amplification factor is the delivered one, and not mu; Vanode:Vgrid, yes? )the capacitance appears larger; more current flows into it...since the capacitance isn't changing, we get additional current through the amplified voltage. Q=CV; take time derivative d/dt( Q=CV ) and get i=CdV/dt the anode voltage changing by time( and it is amplified, remember?).
cheers,
Douglas
Re: Re: Re: Pentode Miller Capacitance, and which spec value to use?
Grounded cathode means that you're inputting the signal between the grid and cathode, and taking output from the plate and cathode, where the cathode is common to both. Doesn't matter if you include a cathode resistor. If you don't bypass it, you simply reduce the g(m) and the gain by degeneration. The reduced gain will cut down on the Cmiller a bit.
Grounded grid operation inputs the signal between the cathode and grid, and takes the output from the plate and grid. Here, the grid becomes the common element. In this type of operation, you don't want any grid resistors and/or excessive impedance to ground, otherwise, you make an oscillator. That's why the UHF triodes have multiple grid connections: you're supposed to ground them all to put stray impedance in parallel to reduce it.
flysig said:
Grounded cathode means that you're inputting the signal between the grid and cathode, and taking output from the plate and cathode, where the cathode is common to both. Doesn't matter if you include a cathode resistor. If you don't bypass it, you simply reduce the g(m) and the gain by degeneration. The reduced gain will cut down on the Cmiller a bit.
Grounded grid operation inputs the signal between the cathode and grid, and takes the output from the plate and grid. Here, the grid becomes the common element. In this type of operation, you don't want any grid resistors and/or excessive impedance to ground, otherwise, you make an oscillator. That's why the UHF triodes have multiple grid connections: you're supposed to ground them all to put stray impedance in parallel to reduce it.
hey Miles,
You've got it right, but allow me to put a different spin on it. With grounded (or common, damn that term) cathode, signal is applied between grid and ground, if there is a bias resistor there it will degenerate the signal and reduce gain if not bypassed, if adequately bypassed, the cathode might as well be tied directly to ground. With grounded grid, we apply signal between cathode and ground and tie the grid to ground (as the name implies) with a low impedance and resistance connection. In this case we eliminate Miller because the grid screens the cathode from the plate.
cheers,
Douglas
You've got it right, but allow me to put a different spin on it. With grounded (or common, damn that term) cathode, signal is applied between grid and ground, if there is a bias resistor there it will degenerate the signal and reduce gain if not bypassed, if adequately bypassed, the cathode might as well be tied directly to ground. With grounded grid, we apply signal between cathode and ground and tie the grid to ground (as the name implies) with a low impedance and resistance connection. In this case we eliminate Miller because the grid screens the cathode from the plate.
cheers,
Douglas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.