Gain of FET circuit

I thought it needed to be nonlinear to work. But I'm probably out of my depth on this one. If correct, your 2 requirements are inherently contradictory.

NT

The existing circuit is non-linear, I think it is because they used the thermocouple to measure the voltage. But I don't really know. Also the meter scale is non-linear. So, it may be they used a special hand picked tube to match the non-linear scale on their meter. I plan on measuring the drive voltage directly and changing the meter so it reads linearly. I may be all wrong here, but I think the Fet circuit will be linear from 1 volt to 5 volts input. Is that reasonable?

Mikek

It sounds like it would be easier to scrap both the tube and the meter (or adjust the meter to linear response if that's possible) and re-design a low capacitance linear meter driver from the ground up, frankly.

Once you just need a linear response and are OK with going solid-state you can do what you like, transistors are cheap, you can make a bootstrapped buffer with an op-amp meter driver, whatever.

Isn't That what I'm trying to do? Mikek

I'm only find precision rectifiers that go up to 50kHz, is there some frequency limit on precision rectifiers? Mikek

Yes. Low level broadband RF voltage measurement has many solutions. One very practical way over an enormous dynamic range and from LF to VHF is to use the AD8307 log amp and its sucessors. There is a famous QST ham radio magazine project and lots of others on the web. These are for the

If you want a linear response then you might inspiration browsing the website of the deceased LA8AK Norwegian ham experimentor - there is one ex HP circuit using two diodes in the collector of an amplifier that I remember looked fun to try.

piglet

If you are saying the follower has gain of 1, but the drain resistance reduces that, I don't think that is correct. The gain of the FET will cause a current to flow in the source to keep the source "following" the gate with only a small voltage difference (as determined by the gain). That's a given.

The resistance in the drain path would provide inverting gain, but that point is bypassed to ground via the caps. So the drain sees no AC signal and has no impact on the gain of the follower. Even without the caps the impact would be very small.

You can put an MMBFJ309 on a little breakout board for much cheaper than that.

Cheers

Phil Hobbs

After all this hassle, you are worried about $6?

I've earned at the bottom end of the middle class most of my life. However, I have lived frugally, saved and invested for 35 years, I can now live better, from the return on my nestegg than when I worked. But, that doesn't mean I'm going to change may habits. There are plenty of Fets, for $0.50, if that will deliver the performance I want, why waste $5,50? But if it is needed I'll buy it! With a bootstrapped circuit, I don't know that I gain anything with a lower Ciss. I also have at least 50 fets in my stock of parts. Mostly j102,

I only see precision rectifiers up to 50kHz, is there a reason they can't be used to 10MHz? Mikek

Using the tube characteristics for a 2A6 which this tube is supposed to be selected or optimized from, I get a cathode voltage of about 3.2 volts and a current of about 0.2 mA (based on a 15.68 kohm cathode resistance). I see that is not the same as what you have.

I don't get what C302 is for. If this circuit rectifies, is that a smoothing cap? Since it is the same value as the plate cap and the load resistance is a lot higher than the plate resistance, it would seem to be doing a lot of smoothing. I don't see how it can be rectifying. 4 Vpp will be fairly linear with a 3.2 volt bias between the grid and plate.

I can't say how your FET circuit is biased since the part number is not identified.

I like that, even a little better specs. I prefer a through hole, but I can work with the MMBFJ309. I only see precision rectifiers up to 50kHz, is there a reason the won't work to 10 MHz? Thanks Mikek

Ok, I think the follower has a gain of approx 1X. I'll work with that number. Thanks, Mikek

How is this circuit different from the one you have already built? The one James Arthur was helping you with will do this as well if not better.

Digikey has lots of jfets in the 50 cent range, and they can be sorted by capacitance.

The inductor is already being resonated with capacitance. A few pF one way or the other shouldn't matter. How much capacitance did that tube have?

Cg-s isn't bootstrapped, since the source has a cap to ground. It's a rectifier.

With radical enough opamps and diodes, they can. Seems like overkill. The old Q-meter was designed when people couldn't afford things like oscilloscopes or RF-detector ICs.

"Perfect" rectifiers are generally limited by slew artifacts in one way or another. The higher the frequency, the less time they have to settle out, and so the worse the error. There are two common circuits, one of which has to come off the negative supply rail and slew all the way to ground on alternate half cycles, and the other only has to slew by a diode drop.

Either way, small signals won't generate enough input error voltage to reach maximum slew, and so the details of the transient are strongly signal-dependent, which makes the circuit nonlinear unless the signals are slowly-varying.

You can get a full wave "infinite impedance" detector by using a differential pair and taking the output from the source leads. That may or may not be a win.

Cheers

Phil Hobbs

There are also multiple definitions of Q that are closely similar at Q values above 5 or so, but vary fairly widely below that. (The error goes like 1/sqrt(1+1/4Q**2) iirc.)

f_0 / FWHM != X/R at resonance != L network voltage transformation ratio.

Cheers

Phil Hobbs

I had simplified the schematic to get rid of some of the other functions.

Here's the full schematic of the tube circuit,

In the end gain is not going to be a problem, if I need any it probably won't be more than 2X or 3X.

Yes, a lot of smoothing I think you can see that in the needle movement as you make adjustments, hardly enough to be objectionable though.

I don't see how it can be rectifying. 4 Vpp will be fairly linear with a 3.2 volt bias between the grid and plate.

Are you getting 3.2 volt bias from the 100Mohm to ground on the grid? A while back, I was tld it rectified because any negative voltage on the grid put it in pinchoff and there was no output from the tube.

So the circuit on the top of page 2 here,

I can use the 2n4416, but at this moment I plan on the MMBF-J309.

Thanks, Mikek