My newly acquired sweep generator (wobbulator) doesn't have an inbuilt RF detector circuit unlike some others (some Wavetek models). So, I decided to build one. I was quite surprised at how many variations of this relatively simple circuit are out there and now I'm not quite sure which one I should build.

The basic types break down into a) those that block DC and those that don't and b) those that filter/smooth and those that don't. Here are some examples:

enter image description here

The circuit above with slight variations is often suggested for my use-case. It seems to rely on the input impedance of the oscilloscope (in my case) which is typically 1MOhm, with the resistor in the diagram chosen to provide an RMS output. Am I right in thinking that this version will only rectify, but not smooth the RF signal?

Here are variants of the one above, but with smoothing (the 2nd one is w2aew's rf probe circuit):

enter image description here enter image description here

There are also the non-DC blocking types that are used in AM detector circuits, but I think I can ignore those ones.

So, my questions are: has anyone built anything like this for their sweep generator and which circuit did you use?

Why wouldn't I want to include the 2nd capacitor to filter the rectified RF (the sweep range is from about 100 kHz to 120 MHz)? (I'm thinking the second types have a much lower input impedance).


I realize some out there will have no clue of what I'm talking about. So, here's a great youtube video that explains everything: https://www.youtube.com/watch?v=szC2RJRmlgI

  • 1
    \$\begingroup\$ Most modern scopes that people use will handle in excess of the bandwidth you state so probably nobody builds them anymore. It might help if you state the application. \$\endgroup\$
    – Andy aka
    May 25, 2018 at 14:50
  • \$\begingroup\$ Hi Andy, it's for a sweep generator I bought to do some filter analysis and other general educational stuff. Ideally, you demodulate the DUT output before feeding it into the vertical input of the scope in XY mode (but then, I'm sure you know all this!) Yes, it's a little 1970's - but that's where I'm at!:-) \$\endgroup\$
    – Buck8pe
    May 25, 2018 at 14:54
  • 1
    \$\begingroup\$ um, you want to measure the frequency response of a measurement device with this utterly frequency selective, and uncalibrated circuit?! Will not happen. Your scope is way better than that. \$\endgroup\$ May 25, 2018 at 15:23
  • 1
    \$\begingroup\$ so, what is your use case here? You say you want to visualize the frequency response of your scope, but why do you want to do that? To which end? \$\endgroup\$ May 25, 2018 at 15:25
  • 2
    \$\begingroup\$ @MarcusMüller: The idea is to use a scope to view the frequency response of a circuit. The sweep generator puts out two signals. One is the RF sweep, the other is a sawtooth synchronized with the sweep. Use a scope in XY mode. The generator sawtooth drives the scope X (horizontal,) the output of the RF detector drives the Y (vertical.) The display resembles the display of an RF analyser. It actually IS an RF analyser. Just uncalibrated and built of standard lab equipment than being a purpose built, ready made device. \$\endgroup\$
    – JRE
    May 25, 2018 at 16:35

1 Answer 1


I'm going to be radical, and suggest a break with the 1970's.

RF detectors have become a much more mainstream kind of thing. Your smartphone (and probably the cellphone you used to have) has an RF level detector in it to monitor its own power output. There are absolute boatloads of integrated RF detectors to choose from.

For your use case, you'd probably want a detector with log response, and a fairly wide dynamic range (say maybe 40dB.) Too little would be frustrating. Too much would be expensive.

You also only want to cover the RF range your generator can produce.

As an example, Analog Devices has a bunch of detectors with various frequency ranges and dynamic ranges.

At a quick look, the LT5537 seems like a good place to start.

0Hz to 1GHz, 83 dB, and cheap.

The circuit itself is pretty simple:

enter image description here

You get linearity, and predictable response. Uncalibrated, it'll probably outperform anything you could build out of discrete parts.

You will have to make a small PCB, and solder SMD parts.

I don't work for Analog. Don't get kickbacks for recommending them, either. They were just the first example of the kind of part I meant that Google found.

Max, and Linear, and probably others make them, too.

  • \$\begingroup\$ A break with tradition is good, so +1 from me. That said, there are some disadvantages here too, right. Firstly, it's an active solution and needs a supply. OK for the most part, but can be a pain. The other thing I notice is that it has a 50 ohm input impedance which means you have to match your DUT output to 50 ohms. The diode detector isn't like that (as I understand it - I could be wrong), so it has simplicity on it's side (insertion losses and poor linearity aside). \$\endgroup\$
    – Buck8pe
    May 25, 2018 at 18:15
  • 2
    \$\begingroup\$ The diode detector actually should be matched. Its just so inaccurate that nobody bothers. Its own frequency response (and impedance) would be all over the place, so whatever bit of inaccuracy the mismatch causes just isn't worth the bother. Power is pretty much a no brainer. Small 5V wall wart, 1.8V (or whatever needed) linear regulator on the PCB with the detector, you're golden. Heck, even a couple alkaline cells (with regulator) would do fine for months. \$\endgroup\$
    – JRE
    May 25, 2018 at 18:24
  • \$\begingroup\$ In other words: just treat it like the old diode detectors and ignore the impedance matching. It will still work better than a diode detector. \$\endgroup\$
    – JRE
    May 25, 2018 at 18:27
  • \$\begingroup\$ Good info and one to try. \$\endgroup\$
    – Buck8pe
    May 25, 2018 at 18:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.