Mircostrip directional coupler???

Note that the URL didn't wrap, trying the , it may need to cut and pasted

I'm looking for information on designing what I believe is a microstrip directional coupler. I don't want to do stripline, since I would like to keep this to a 2 layer PCB. Google is not turning up anything overly useful other then typical stripline / microstrip stuff and GHz stuff.

Basically it is a largish microstrip which carries the RF energy, and it linearly couples slightly to two smaller microstips, one on each side of the largish microstrip. The small microstrip outputs are rectified with a diode and filtered with a cap. The output is given as a DC voltage dependant on power. One of the small micostrips is used to measure forward power, and the other is arranged so that the diode is forward biased when the power flows in the other direction giving the reverse / return power.

Here is a website selling such a device as an expensive adaptor for an RF connector:

Note that they call it a stripline when it is actually a micrpstrip, and I'd love to see the stated 500w or even 1kW flowing through it continuously.

Application:

Forward and reverse (return / SWR) power indicator. Hobbyist use low power (fractional watt) ~100 MHz personal FM transmitter. The material will be standard FR4 with solder mask. The dielectric constant of the FR4 will not be held to tight tolerance like some more exotic RF materials.

The purpose is more of an educational one then the actual use.

There are lots of articles, in books and on the web, about microstrip couplers. But MiniCircuits makes some small, cheap surface-mount ferrite-transformer directional couplers that might be less hassle.

You can use a single coupler for fwd/rev measurement if you can keep both ends terminated properly.

John

minicircuits has an appnote about the subject

Rene

Have you looked in publications like the ARRL Handbook?

There is the issue of bandwidth, the fL, fM, and fH ranges for MCL, and then the OP says 1KW thru-power rating which is 1e6mW or 60dBm, seems a bit high for the ones you're talking 'bout.

His application was stated to be "fractional watt." See above.

Somebody else sells ceramic planar surface-mount directional couplers... I have some around here somewhere... but I seem to recall the frequency range being higher than he wants. Murata? Somebody like that.

I have a pretty good collection of RF texts, and directional couplers are mentioned sparsely or not at all. Do you know of any good refs on microstrip couplers?

I do have one book that shows how to do a dc from discrete Ls and C's.

John

useful

the

diode

the

flows

it

transmitter.

What's interesting, was on Wednesday, we did a short preproduction proto run of a customers board, and I programmed the vision and other pick and place parameters for what I thought was a little RF transformer from MiniCircuits - Looking now at the MiniCircuits website, it is actually a directional coupler!

I've just ordered the literature on their website.

That's actually one of the things I'm having the most problems with finding information on - proper termination! The power carrying microstrip calculations seem quite doable, and there are many online stripline/microstrip calculators. The side microstrips shouldn't be much harder to calculate then the middle one. I'm assuming the termination is what sets the energy transfer ratio, along with the design of the side microstrips. I am assuming it is energy transfer, something like a current transformer, and not a voltage ratio. A voltage ratio would have many problems introduced like the diode losses, mainly Vf drop. Diode storage times and capacitance would also have an effect.

The interesting thing is I've never taken an RF course other then what I tought myself, as I temporally left University as I was finishing up my BSc degree and working on the 3rd year of my EE degree, and have had a hard time finding time to go back to finish in the last few years. I've managed to take the time to go back and take a course last summer (A- in a difficult Probability and Statistics course!) and a very difficult Differential Equation course (Considered a filter course, main topics cover a lot in first and second order, Laplace transforms, and Fourier series all in one semester) this winter term, which after not taking a math course for about

useful

it

the

diode

on

the

flows

RF

=8

when it

transmitter.

Actually, the 1 kW was in reference to the specs for the cheap little unit in the website that uses about a 1/8" wide microstrip printed on standard PCB material, likely 1 or 2 ounce copper weight. Humm, lets calculate that out - 1 kw into 50 ohms is just under 4.5A assuming no major imaginary power currents are flowing. I was wrong, that amount of copper could be just acceptable.

My power level is fractional watt. Frequency range is about 88 to 106 MHz, main intension is a personal FM transmitter capable of transmitting a few hundred meters / yards, mainly for use at a friends old farm house where we sometimes go for a weekend getaway. The major reason to include it is for SWR measurements, and to help educate myself with RF.

Jeff

useful

it

the

diode

on

the

flows

RF

when it

Thanks,

I ordered their literature about an hour ago!

Jeff

useful

it

the

diode

on

the

flows

RF

when it

transmitter.

No,

That is one book I need to buy. I've yet to even see a copy of those handbooks. I've had the link to their website for that purpose on my desktop for months, (along with the 160 or so other icons - good thing my desktop is

Jeff

Normally, all the microstrips would be 50 ohms, both the main line and the coupled one off to the side. It's the spacing and coupled length that determines the amount of coupling.

Because it's symmetric, it samples forward and reflected power at the same time. For it to work as calculated, the two lower ports should each see a 50 ohm load. That can be tricky if you load them into diode detectors, which vary in impedance. Adding an attenuator pad at both ends would help.

Things work sort of backwards from the way they may look. The reflected power pickoff port is on the same side as the power *output* port. Lines couple backwards!

A "3=port" directional coupler just has one port terminated internally.

Several people - Analog Devices, LTC, Maxim - are now making nice integrated wideband RF detectors, with either log or linear output. Most are fairly hi-z input, so could be shunted to net 50 ohms.

John

connector:

RFSim99 (freeware); tools-->component-->coupler. What it doesn't tell you is that for the given coupling, the coupled section needs to be

Cheers, Tom

For quick and dirty, take a piece of coax and slide an insulated wire underneath the braid. The "forward" end is at the transmitter and the "reflected" is near the antenna. At 100 MHz., you ought to be able to do it with an 8" or so length of coax.

The older ARRL manuals used this technique from HF through VHF with surprisingly good results.

Jim

Dumb mistake. Forward is near the antenna and reflected is near the transmitter.

Jim

Most directional couplers work the opposite way, reflected power port on the same end as the load. Line-line coupling works backwards! That's why near-end crosstalk is worse than far-end.

John

Thanks for the link. It's sometimes surprising what wiki has.

Looking around some more on wiki, I found this, which is more typical of what I have seen before. It is as you describe, but with seperate forward and reverse coupled lines.

Here is a circuit found on some amature radio website:

Actually that makes complete sense, as that would be the direction of propagation of the returning wave. I would expect it to first couple with the side a wave starts entering.

I'll keep that in mind!

The MAX2016 looks promising, I'll have to look at the data sheet in more detail when I get a little time

Some interesting pictures of strange microstrip couplers:

Many of these new log-detector buggers are so small its very difficult to use them since they're made for GSM market (see ad series, nano smd). And the ones that are not are much more expensive. Damn.

SioL

use

Small is no problem for me - I can mount just about any part bigger then an

The MAX2016 comes in a 28 lead QFN, which we have mounted well over 10,000 similar parts within the last year. If the design work is done correctly, QFN's pose little problems, even for one offs.

useful

it

the

diode

on

the

flows

RF

connector:

when it

transmitter.

Thanks, I downloaded RFsim99 - looks interesting. I'll play with it a bit on the weekend.

At 100 MHz, there is not even close to enough room on a typical PCB to approach 1/4 wavelength, which when accounting for some propagation velocity reduction from the FR4, would be around 1 meter in length!

The couplers I have seen used for this purpose were in the order of 1/25 wavelength, although at higher frequencies.