Microwage RF detector

Ouch, not a common band. This is usually done with a spectrum analyser.

Unless you have a lot of TX power, you have a problem.

Search for w1ghz antenna book, and look at the 24 ghz ham stuff. You need a waveguide to sma adapter and then a diode detector module. Ones that go that high are not cheap. The waveguide will probably have to be bigger then what is used at 24 so you are not "waveguide beyond cutoff"

The open waveguide will be a 3dB antenna, not much gain at all. The transition from SMA to the detector will be about 2 dB of loss, So your already behind the curve. Something like this.

They do show up on ebay, but usually not tested.

Mind you, microwaves are a RF safety issue, don't get your eyes forward of the dish. Be aware that high SWR from you sticking a probe in the wrong place could blow out your receiver front end in some cases. And even if you do have RF, it doesn't tell you how much or if your modulating correctly. This is why a RF power meter or a spec an with a calibrated antenna is used.

In other words, time to find a RF technician or engineer with the correct gear. Steve

Why not a simple dipole with a diode at the middle? You could make that on a slice of fr-4 in a few minutes, for less than lunch money.

Use something like an SMS7621-079 maybe.

John

John, at 18 on FR4?

You sure?

Steve

Why not? Things scale. People use dipoles+schottkies at way higher frequencies than this. The only reason to use FR4 is because it would be an easy way to solder that _really_tiny_ part down. It would probably be easier to use real wires for the antenna itself, because they'd be easy to trim to length to tune the thing. With enough transmit power, it wouldn't need tuning at all.

A patch antenna would be interesting. Googling that would give lots of hits. You could hack one on copperclad with an x-acto knife.

That Skyworks diode has Xc around 50 ohms at 20 GHz, a little high but not lethal. Aeroflex has an 0402 schottky that's only 0.08 pF, which would be even better.

The problem isn't quantified, so we don't know how much signal we'd have. Close to a transmitting dish, I'd assume a bunch.

It's surprising what fast stuff you can do with tiny surface-mount parts on pc boards, things that used to need plumbing.

John

Hello,

With a microwave diode you should be able to detect levels of -30 dBm (or even significantly lower with good match).

With a 3 dBi antenna this converts to a plane wave power of about 30mW/ m^2 (for example transmitter with 1.5W EIRP at 2m distance).

For this purpose you should use real schottky diodes intended for SHF. That are the ones with reverse voltages of about 4V or less and capacitances of 0.25pF or less. An example is BAT15 (smd, not beam lead). If you can get special mmW diodes (capacitances around 0.1 pF or less), and tune them, you will get more output voltage with the same incident Power Flux Density.

To get all EMF generated by the junction, the load should be larger then the so-called "video resistance" (factor 10 gives you 90% of EMF). Diodes with relative high forward voltage have high video resistance and need a buffer (for example op-amp with CMOS input stage). By doing this, you can use the diode without bias avoiding thermal issues.

Best regards,

Wim PA3DJS

Thanks for all the FB. I am starting to think it's more trouble than it's worth. I was thinking of a small unit (0.6" aerial, diode and meter

- maybe even a cap) which when in the vacinity ~2m would detect the rf from the dish (0.3-0.6m dish) - certianly not in front.

Really tiny? 0.060" length is a mid-size boulder :-)

That is true. Except the stuff inside the sampler heads you showed us, that probably won't fly with SMT packages.

It's really a pain to solder. But I'm thankful it's not a beam lead.

Agoston Agoston showed me a 20 GHz sampler he did, surface-mount parts on FR4. A lot of stuff that used to be hybrids can be done now surface mount, too.

Oh, here's that other detector diode:

Cute little devil.

John

Why not ask the intended receiver's RX power level? Or a bolometer?

Have they done it monolithic yet? Should be fairly simple to, say, lay down an aluminum dipole on top of SiO2 dielectric, on top of fairly conductive (n+?) silicon, for a not-too-awful antenna element, then put a rectifying schottky junction in the middle, plus tracks leaving the site for power out. Then repeat the array a brazillion times. Then, scale it down, so the antenna is around 800nm wavelength or so (should be possible with today's newest processes?). Does that work?

Tim

an aluminum dipole on top of SiO2 dielectric, on top

put a rectifying schottky junction in the middle,

times. Then, scale it down, so the antenna is

processes?). Does that work?

I believe that's been done actually, someone wrote about this here in this NG some years ago. It'd be very interesting to read what happened with that project.

True. He seems to be a guy who rolls up his sleeves and gets stuff done. It is similar with phenolic. Lots of people say it ain't any good for anything past audio. Wrong. I have a few VHF/UHF splitters here that are done in two-layer phenolic.

The first 144MHz transistor power amp I ever built was also on phenolic because back in those university days buying FR4 would have required tapping the beer kitty and that was off limits.

But that really needs IR reflow soldering or at least a hot air station.

Hello,

I don't know the power of your uwave unit, but when it is a data link, you will not have sufficient EIRP in the side lobes to detect with a simple detector with low directivity antenna.

So best is to check the equipment's data first and do some math before spending time and money.

Best regards,

Wim PA3DJS

I've used these printed Vivaldis before at 10 ghz.

You need .085 line to the diode, the Vivaldi is designed to use the coax, minus some outer jacket, as its matching structure.

If you ask Jeff Kruth, microwave surplus vendor extraordinaire , owner of the above auction, for a detector module and two inches of .085 hardline with the connector to the vivaldi, I'd bet he puts a package together for you. Should be very inexpensive, and odds are he has a preamp laying around as well. The detector gets a load resistor and then goes direct into your voltmeter.

Doesn't the 18 ghz link have have diagnostics? It had to be aligned somehow......

Steve

Oh quitcherbitchin. It's ZERO POINT ZERO EIGHT PICOFARADS.

John

I don't know about visible, but it's been none in the THz range.

John

But only if you connected it with one strand each plucked from a super-fine litz wire :-)

Uh! IR reflow or hot air for something that big? I routinely do 0201 without anything special. Just a stereo microscope and a metcal but you sure have those, don't you? I (almost) just can't imagine anything easier to solder...

Tim Williams a écrit :

And package that in a mint black epoxy package? :-)

I think Phil Hobbs wrote something about this when he still was at IBM...