MOSFET Follies question...

[Gaz]

Want to try the IRF820 source follower in place of a typical 12AX7 cathode follower to free up a tube, but had a question about the gate-to-source maximum voltage.

Right now I measure 18VAC RMS going into the cathode follower measuring from grid to cathode with 100K cathode resistor. The IRF820 has a maximum input of 20V peak to peak. Keeping the 100K the same for the source on the IRF820, won't that be way too high, or am I not understanding this correctly?

Isn't 18VAC RMS about 51V P-P?

Thanks, I'm new to using these devices.

[nickb]

Something is wrong here. It's a source follower i.e. source follows the gate and there should be a very small AC voltage between them. There will be a DC offset of around 3 to 5V, just enough to satisfy the DC operating conditions.

Try this:
Kill any input signal and measure the DCV from S to G.
Apply signal and with an 10nf (or so) 630V cap in series with your meter lead, measure the ACV on G, S to G and S to ground,

[Gaz]

Nick, thanks, that helps. Sorry if that was confusing, but I haven't tried the MOSFET circuit yet, was just measuring the tube cathode follower as is. It was naive of me to think the voltages would be exactly the same when replacing the tube with a MOSFET.

The reason this came up was because I was noticing some source followers examples with 12V zener clamps from gate to source. Are those just precautionary for some fault condition? Based on your explanation it seems like it would be very unlikely then for the gate and source to ever see that voltage difference when the FET is setup to be a source follower.

[nickb]

Even so, the numbers you gave are off. The only significant difference is that the DC offset (grid to cathode) will be about -1V. The grid to cathode AC signal should still be small.

As far as protection is concerned, in real life things get complicated so fit the zener. It's cheap insurance.

[R.G.]

Nick has it correct. The gate can stand something like +/-20V from gate to source. In a MOSFET follower, the source follows the gate around, just like a cathode follower. A 12AX7 has an input range of about 2V, maybe less, from cutoff to what passes for saturation, yet it follows voltages of hundreds of volts as a follower.

A MOSFET will not conduct at all until the gate-source voltage gets up to a threshold voltage, usually about 3-10V depending on the MOSFET. After it's "enhanced" enough to start conducting, more voltage from gate to source lets more current through, generally about one amp per additional volt from gate to source. So in a tube CF setup, the gate-source voltage will always be sitting just a hair above the threshold voltage, as so little current is needed.

There are probably better MOSFETs today than back when I first wrote that up. Back then, the IRF800 series was about the best.The MOSFET world has come a long way in nearly two decades.

[J M Fahey]

Right now I measure 18VAC RMS going into the cathode follower measuring from grid to cathode with 100K cathode resistor. The IRF820 has a maximum input of 20V peak to peak. Keeping the 100K the same for the source on the IRF820, won't that be way too high, or am I not understanding this correctly?

Isn't 18VAC RMS about 51V P-P?

You are measuring wrong .
No way you have 18V RMS between grid and cathode, either your meter is cheating you by displaying DC as AC, a very common mistake in a lot of meters, or by holding those unshielded wires you are introducing hum.

As said above, GK AC voltage, or GS if you use a MosFet is small, 1V RMS or less.

And as suggested above, add a protecting Zener anyway

[Gingertube]

As Jaun says,
- add a 12V protection zener diode, cathode to the gate and anode to the source (15V would be OK instead)
AND
add a 1K Gate Stop resistor in series with the gate connection, with the resistor body as close to the gate connection as you can get.
MOSFETs are VERY high gm devices and so are susceptible to parasitic oscillation. The gate stop resistor will fix that.
The protection zener's gate connection can be to either side of that gate stop resistor, whichever side is easiest.
Cheers,
Ian

[Gaz]

Even so, the numbers you gave are off. The only significant difference is that the DC offset (grid to cathode) will be about -1V. The grid to cathode AC signal should still be small.

As far as protection is concerned, in real life things get complicated so fit the zener. It's cheap insurance.

Yes you and Juan were correct. I remeasured from grid to cathode on the cathode follower, and with the preamp full up only about 4.5VAC RMS. One of my probes was not making good contact when I initially measured. Sorry about that, but still learning from this thread!

[Gaz]

As Jaun says,
- add a 12V protection zener diode, cathode to the gate and anode to the source (15V would be OK instead)
AND
add a 1K Gate Stop resistor in series with the gate connection, with the resistor body as close to the gate connection as you can get.
MOSFETs are VERY high gm devices and so are susceptible to parasitic oscillation. The gate stop resistor will fix that.
The protection zener's gate connection can be to either side of that gate stop resistor, whichever side is easiest.
Cheers,
Ian

Thanks for the advice!

[nickb]

Yes you and Juan were correct. I remeasured from grid to cathode on the cathode follower, and with the preamp full up only about 4.5VAC RMS. One of my probes was not making good contact when I initially measured. Sorry about that, but still learning from this thread!

Still doesn't seem right. I expect perhaps 1Vrms or so. Also a expect a bad contact to lead to a low reading rather than high. Your meter can be confused when trying the measure AC in the presence of DC so use the capacitor is series with your meter lead and restest as instructed earlier to be sure. Also, if it's not a true RMS meter the non-sinusoidal waveshape will lead to reading errors. The only way I can think of for this to happen is if you were driving hard and had a significant capacitive load of several nF. I'm guessing this drives a tone stack which is not that capacitive.

[Gaz]

Still doesn't seem right. I expect perhaps 1Vrms or so. Also a expect a bad contact to lead to a low reading rather than high. Your meter can be confused when trying the measure AC in the presence of DC so use the capacitor is series with your meter lead and restest as instructed earlier to be sure. Also, if it's not a true RMS meter the non-sinusoidal waveshape will lead to reading errors. The only way I can think of for this to happen is if you were driving hard and had a significant capacitive load of several nF. I'm guessing this drives a tone stack which is not that capacitive.

I remeasured with the cap in series with the probe and measured 2.8VAC RMS with maximum amplitude (my meter is true RMS). It doesn’t go above 1 volt before 4 o’clock on the gain control. That’s seems about right with what you estimated, no? It is driving a tonesrack like you presumed. Thanks again for the help.

[nickb]

That's more believable. Right now I don't have access to my pc to confirm but it would help greatly if could see the waveforms on the grid and cathode.