RF Power Detector Circuit, Discrete Components. How does it work?

[Yansi]

Hi,

one more issue to solve in one of my little project and that is a power level detector. What would be a good circuit approach to measure RF level in a signal path?

Current frequency range of interest: 100kHz to 30MHz,  power levels of say up to 13dBm, measured range 20-30dB.

I think that using a single diode power detector, does not give very good and precise results. Therefore I am looking into some more interesting circuits.

I have tried to do a bit of research and come across this quite interesting circuit, from a Marconi 2018A/2019A RF generator.

I have attached the page from the service manual, where there is the output PA, some electronic attenuation stages, the power detector circuit, ALC loop and some signal conditioning circuitry for the ALC.

It seems that power detection is being done with D11 and D12, while IC1-b/4 is the amplifier for the detector. Wondering... what's the purpose of D13 in its feedback, and how does this detector topology work? I could not find any application note with anything similar to this.

I think that IC1-c/4 is the ALC servo, controlling the PIN diodes, based on the signal from the detector and the setpoint from the DAC chains below (interestingly enough, the reference signal also goes through an amp with a diode in its feedback - why?)

Thanks
Y.

[retrolefty]

I would consider using a Analog Devices AD8307 log amp/detector module. You might have to put a simple resister attenuator before the device to cover the highest Dbm level you might desire. Chip works great, very good dynamic range for the money and simple to interface. Here is one example module below. I recall paying over $15 just for the chip ten years ago or so.

https://www.ebay.com/itm/AD8307-RF-Power-Detector-Module-DC-to-500MHz-Transmitter-Power-Test-92dbm/263868813801?hash=item3d6fce91e9:g:E94AAOSwafhbbG6E:rk:1:pf:1&frcectupt=true

[Yansi]

Sorry, I am not going to use a $13  500MHz log detector IC with 90dB of range, for a 30MHz job with 20 or 30dB range required. (Even though I know well where to get AD8307 for cheap and already have them available)

I am asking for a discrete component solution, not an integrated one, which will probably end up costing more than 10 times what would be fully adequate to do the job.

[KJDS]

The diode in the feedback path provides temperature stability, otherwise you have an RF activated temperature sensor.

Use a single BAT63 on the +vev op amp input for the detector, and another in the feedback path. Use a TLC2632 chopper stabilized op amp and with careful layout you can get more than 40dB dynamic range up to a couple of GHz.

Make sure that the diode cathodes are facing away from the op-amp

[Yansi]

Thank you for commenting, however I am still a bit confused how does it work.

For example, what for is the diode with the negative bias, D11?   Is the L10-R42 important for the detector to function or what's its purpose?

The amplifier seems connected as a logarithmic one. I understand it may compensate for temperature of course, but what is the transfer function of such detector + log amp?

As a simple diode detector is (obviously) quite non-linear, does this circuit with the log-amp  correct for the linearity, or is there any other purpose of that?

Thanks

//EDIT: What do you mean by "Make sure that the diode cathodes are facing away from the op-amp"?

[Yansi]

Found an interesting web page covering this topic, with a few interesting citations at the end.

https://www.qsl.net/in3otd/electronics/power_detectors/power_detectors.html#ref1

I will probably cobble up something to test it.  Also good to know, that MCP6001 has very low input bias current, so it is suitable for this application. (I have these available, so will put them to the test).

However I am not sure, if I have any suitable small signal schottky diodes.  Plenty of BAV99 and others alike, but not much schottky :-/

[KJDS]

bias current flows out from the input of most op-amps, hence if you have the anode towards the input then current flows through the diode and gives a little forwards voltage, just enough to help the diode work as a detector.

...and I'll be amazed if you find anything that works as well as BAT63 diodes and a TLC2652. Input offset voltage and current are key to low level performance.

[Yansi]

I will currently be more amazed if I can find a small signal RF schottky diode within my junkboxes. Sigh! 

So it has to wait, I need to get some 

So this is what I came up with. I think it may be good enough to 20 or 30dB range, at some decent power levels (around 0dBm or so).

Pointing the diode outward would be nice, however I could not deal with the negative voltage generated, in a single supply scenario. (the small "bias" also explains those 4M7 resistors(R43, R45, R46) in the Marconi 2018A schematic).

(ignore the BAT48 descriptor, not suitable)


//EDIT: Typo.