The linear Diode Detector

what article? what circuit?

NT

Sorry about that!

Mikek

It's a simple RC time constant at the highest modulation frequency.

F = 1 / (2*Pi*R*C) where: F= highest modulation frequency R= load resistance C= capacitor across the load resistance For your AGC, the RC time constant will be however fast you want your AGC to respond to changed in signal level, without also demodulating any AM modulation on the carrier.

Can this circuit be normalized from 50 ohms to a 470 ohm input impedance? Without affecting its linearity. Looking at figure 9, the 68 ohm, can be raised to 640 ohms, but what about the 4.7k ohm resistors, what do I do with those values?

Mikek

What you could do is replace R1 R2 (those two 4.7k) and replace with a

piglet

Yes, that's the figure 4 modification. I do plan on using that. Good plan on the pot.

I probably should add the load of the meter to the output before adjusting for best linearity.

As far as equipment, I'll just need to go with what I have for testing the linearity. I have an HP 8640B, A Tek 2465B and a Boonton 92BD RY millivolt meter.

Thanks, Mikek

Seems silly. Why not add a diode drop worth of offset before the detector diode? That would do the temp comp and also fix most of the low-end deadband.

Would you draw that for me, I'm not sure where the diode drop should be. I'm driving the detector from a collector, with 470 ohms to B+.

Thanks, Mikek

Who needs self-biasing for an active detector?

A fast peak detector might be done something like this:

Vcc Vcc -+- -+- | | |/Q1 |/ Q2

| |>. | | |>. | | | C1 --- | '--------- | -' --- | | | | | === .-. | + - | | .------+------(M)----->| | VR1 22K | | '-' C302 --- [22k] R301 |

Wide dynamic range, and temperature-stable, too.

Cheers, James Arthur

Thanks Jeff, So the cap is just a filter, not involved in the linearization process. In this case I have no modulation, just want to convert the RF to DC.

Mikek

I must remember to plant some of those geranium diodes this spring.

:)

There are a zillion possible circuits. You could do something like this...

The PNP does the offset and temperature correction, and adds some current gain for your 470 ohm source.

If you use a schottky diode, it will have a few tenths of a volt of positive offset at the output, easily fixed. A silicon diode would be slower, but would have less offset.

It might be interesting to have the output resistor go to V- instead of ground.

It could also be made into an averaging detector, instead of a peak detector. Too many possibilities.

Hey, this is goofy:

If you're driving a meter, the diff output ain't bad.

Are you building a receiver with an AGC (automagic gain control) that is driven by a linearized detector? If so, you might need some detail on AGC design (threshold, maximum SNR, dynamic range, attack time, release time, hold time, modulation immunity, noise limiter interaction, etc)? For example:

On the other foot, if all you want is an indication of received RF level, may I recommend one of Analog Devices wide dynamic range RF detectors: These are commonly used for QRP RF power meters, VSWR meters, and could probably be convinced to drive an S-Meter.

There are also various chips made for cell phones, where a linear RSSI (receive signal strength indicator) is required. For example, AD8367. Notice the built in detector.

If there's no modulation, what are you building? (I'm curious).

DSO scopes that can easily handle 10MHz and do all sorts of measurements are on the market -- do you really need the "RF-ish" sort of stuff?

Fun. If you ran the op amp off a single supply, moved the parallel RC to the negative side, and returned the input resistor there, you could maybe make the second stage diff amp's job easier.

Cheers

Phil Hobbs

John, you're breaking the rules: A crystal radio set does not contain any power supplies or active devices. It seems that the OP (amdx) is attempting to invent a better crystal set.

No, I'm attempting to replace the tube amp/detector in a Boonton 260A.

Mikek

I may be getting the details wrong, but this circuit works well, down to well under a diode drop. D1 and the cap form a "normal" RF detector, with some diode drop. The op-amp and associated parts raise the output voltage up by about the same amount as D1 drops it. The trick is that the resistor needs to be different (larger, IIRC) than the line impedance to account for the peak factor of the current through D1.

D1 |\ o----->|---o------------|+\ | | >----o---------o | .---|-/ | --- | |/ | --- | | | | | | o----|

With a lot of help, I want to replace V301 the in the Q circuit. I don't expect 50kHz to 50MHz, but If I can make it work to 10MHz, that would be great.

V301 functions as a high input impedance amp and detector. I have the high input impedance transistor amp, now I want to rectify the output to get DC to drive a meter. Here's the original circuit around V301.

Thanks, Mikek