Re: Zero Bias Mode Triode Gain Equation
Steve Conner, Wed Feb 16 2011, 09:49AM

I think you need to think more carefully about what the tube is actually doing. Is it acting as a cathode follower? A grounded grid amplifier? +15v relative to what?

Your link is broken, so I can't see any schematic you might have posted.

Re: Zero Bias Mode Triode Gain Equation
sngecko, Wed Feb 16 2011, 08:45PM

Weird about the link. If you just copy/paste it into the address bar, it certainly works. (I thought I tried it).

Anyway, this tube is not being used in any standard way, and I'm not sure how to describe it so that we can all understand.

The explanation given by one implementer, John Doran, can be found here:

"This is a “cascode” connection of a power field-effect transistor (the actual constant-current source) and a large, high-voltage vacuum triode. The triode drops the excess voltage (that is, any difference between +HV and the voltage across the laser) while passing the laser current on to the drain terminal of the FET. The operational amplifier ensures that the voltage across the resistor R is equal to VC. (John Doran)"

See attached .gif for the relevant schematic snippet.

Is there a technical name for which we're using the 3-500z?

Re: Zero Bias Mode Triode Gain Equation
Steve Conner, Wed Feb 16 2011, 09:02PM

Ok, that's a standard configuration, it's a cascode.

Best way to analyse it is look at the extremes. When the FET is on as hard as it can, there's 0v on the cathode, 15v on the grid. Look at your tube curves, will it pass the amount of current you need with +15 on the grid, and HV supply minus laser tube drop on the plate?

Now look at the other end. How big a negative grid voltage does the triode need to "cut off", that is to reduce its cathode current to a negligible amount, with the full supply voltage on the plate? Your FET voltage rating has to be that plus 15, or it will blow.

It's hard to analyse the circuit. The constant current drive to the cathode makes the plate impedance infinite, and so the voltage gain of the circuit is determined only by the load. You're using a discharge tube, a non-linear load, so you can't use the usual methods. But if you check it at the two extremes like I suggest, it should be ok at all points in between.

Re: Zero Bias Mode Triode Gain Equation
sngecko, Thu Feb 17 2011, 12:52AM

Thanks Steve. I should have looked more closely at the cascode configuration. Will do as you suggest.

Re: Zero Bias Mode Triode Gain Equation
sngecko, Thu Feb 17 2011, 01:42AM

Given a classic 3-500Z Triode and IRF830 in a cascode configuration shown below:

With FET full-on and the grid at +15V, a plate-to-grid voltage of 5000V (HV:15000V, dropping 10000V) I estimate, via the 3-500Z datasheet, a max current of 150mA, well above my regulated 80mA maximum. (Incidentally, this implies that the minimum plate-to-grid voltage at 80mA is ~3500V).

With filament-to-grid voltage of 500V (max on IRF830), and plate-to-grid voltage of 20000, well, we're completely off the charts. However, the 1mA line is fairly linear and can be approximately expressed as y = 0.0075x, where x is the plate-to-grid voltage and y is the filament-to-grid voltage. So, if the plate-to-grid voltage was 20000V, the filament-to-grid voltage would be ~150V -- again, at 1mA and easily within our 500V FET limit. So, if we were to approach the operating limit of the IRF830 FET, I think we can assume that we would be in negligible current range.

Does this reasoning pass the insanity-check phase...?

Re: Zero Bias Mode Triode Gain Equation
sngecko, Thu Feb 17 2011, 02:15AM

Oops. That equation should read: y = -0.0075x, resulting in a filament to grid voltage of -150V.

Re: Zero Bias Mode Triode Gain Equation
Steve Conner, Thu Feb 17 2011, 09:51AM

Yep, seems OK. If it were me though, I'd still put a 400V TVS across the MOSFET just in case.

And some sort of clamp on the 15V supply too, to save any other electronics connected to it if the tube decides to have a plate-to-grid arc. Which is quite possible if you're running it at plate voltages so far above its rating.

Earlier you said the "deamplification", and I was having a fun time trying to figure out what that meant. I guess it would be how much the cathode voltage changes for a given change in plate voltage, and as far as I can see that is just 1/mu.

(Mu is -dVp/dVg for a common cathode stage with a hypothetical plate load of infinite resistance. But in this circuit, with the grid held fixed and the cathode driven, dVg = -dVk.)

I still recommend my original method though, because you can't assume mu constant over the full range you'll be using.

Re: Zero Bias Mode Triode Gain Equation
sngecko, Thu Feb 17 2011, 02:47PM

Let's see:
(1) The IRF830 has a built-in MOS diode clamp source-to-drain. Do you think a 400V TVS would improve the safety margin a lot over this?
(2) The grid also has a neon lamp voltage spike suppressor for just that eventuality (plate-to-grid arcing). Sorry, that's not on the schematic snippet.
Thanks especially for the insight into tube dynamics.

Re: Zero Bias Mode Triode Gain Equation
Xray, Mon Feb 21 2011, 02:42AM


While reading through this thread, I see no mention of plate dissipation. As the number implies, the 3-500Z can handle a maximum plate dissipation of 500 watts. Therefore, if you attempt to operate it at its max plate current (150mA) with a voltage drop across the tube of 10000 Volts, you will greatly exceed the max plate dissipation.

Back in my Ham radio days, I ran a 4-400 in grounded-grid mode with an input power of 1KW. With the pi-network output tank properly tuned for a low vswr, the plate of the tube barely glowed red. But when I changed bands, and the vswr increased, the plate glowed orange very quickly while transmitting, which indicated that I was exceeding its max plate dissipation of 400 watts. So, I suggest that you keep an eye on the plate during operation of your laser. If the plate starts to light up like a lightbulb, then you'd better shut it down real quick!

Re: Zero Bias Mode Triode Gain Equation
Steve Conner, Mon Feb 21 2011, 08:05AM

Well, that's the nice thing about tubes, they give you advance warning of the magic smoke getting loose.

From earlier posts I got the idea that the operating point was 80ma and 5000v, a plate dissipation of 400w. But a real good design wouldn't be complete until you had proved that the 500w limit wouldn't be exceeded for any current setting, tube drop and line voltage. Well, not exceeded by too much for too long, at least.

X-ray, how did that work out? I've got a 4-400 here and would like to build an amp with it some day. Did you just connect both grids to ground?

Re: Zero Bias Mode Triode Gain Equation
Xray, Mon Feb 21 2011, 05:29PM


No, you can't tie both grids together. The suppressor grid needs to be tied to B+ or to the plate. The control grid is grounded, and the input power from the exciter transmitter drives the cathode. In grounded-grid configuration, the input power adds to the output power, so it's not wasted energy. It's been many years (mid-1970's) since I built that 4-400 linear amp, so I don't recall many details, but you can find many examples in ham radio websites and forums.

Re: Zero Bias Mode Triode Gain Equation
Xray, Mon Feb 21 2011, 05:31PM



EDIT: Change "suppressor grid" to "screen grid". Sorry!

Re: Zero Bias Mode Triode Gain Equation
sngecko, Wed Feb 23 2011, 07:09PM

If you check under Current Regulator, you will notice that the current regulation control and instrumentation is limited to 80mA. Expected operating points are between 2-40mA. You will also notice in the current control circuitry that when the plate dissipation is above 500W, a warning LED lights on the instrumentation panel.