circulator vs isolator

[nnills]

Hi,
I have a small question. In this ebay advertisement you see a circulator/isolator with 2 connectors and one stub/termination/something. What is this third side and could I replace it with a connector to make a circulator instead of an isolator?

https://www.ebay.com/itm/California-Microwave-RF-Circulator-Isolator-7-12-GHz-10dB-IL-0-7dB-SMA-Tested/272880556384?hash=item3f88f2d960:g:GxQAAOSw1VJZ23Ts

[nnills]

Same thing, but what are these lenses???

https://www.ebay.com/itm/Harris-Microwave-RF-Isolator-Circulator-8-8-10-8-GHZ/223222462108?hash=item33f918769c:g:9kgAAOSwqS5b4~oK

[DaJMasta]

In theory, sure, that should just be a termination load, but I don't know if in practice there are other passives or a compensation network or what.

I think to make the manufacturing easy, I would design the third contact point to be configurable either as a termination or a connector, make things modular and straightforward.


Finding the right connector could be tricky, though, since you probably want a good match to the original internal dimensions to prevent creating additional resonant modes inside the cavity - but it may not be a huge deal.



I think the lenses are supposed to be for a different auction by the same seller and they uploaded the wrong picture 

[David Hess]

The third side is just a transmission line termination.

[TheMG]

It is a termination but you can not simply unscrew it and replace it with an SMA connector. You have to completely disassemble the circulator because the termination is soldered to the stripline inside the circulator. The process can be very delicate and often there is a lot of glue holding various parts in place which the manufacturer has carefully placed to achieve the desired characteristics.

In other words, the disassembly and re-assembly to replace the termination with a connector is very likely to disturb the characteristics of the device. When you're dealing with such high frequencies it really doesn't take much to totally screw it up.

Here's a couple pictures I've taken of the internals of a typical microwave coaxial isolator from my junk stash.

[coppercone2]

its often not even soldered, I am convinced some of these parts they get the connector or terminator with a little tab sticking out, put a dash of glue into place and stick it inside. Isolators seem to have a trapazoid/wedge geometry in them that aligns the little tab correctly.

I was able to repair an isolator that had a dented SMA connector on it by ripping it out, then screwing a new 'short tongue' wave-guide W40(?) sma connector with a tiny dash of silver epoxy on the tab.

I don't know another assembly method ( other then cutting it open, conventional solder and welding, totally heating the thing up with solder paste or some kind of interesting laser optics/pnumatic clamp. ). I think its just designed with glue.

The thermal time constants of all the parts to be soldered are so low, I was not even sure you can do it if you heat everything up with a laser and smash the thing together at the smallest gap which would allow a laser beam to heat the stuff inside. MAYBE if you got a bus-bar and heated up the whole thing to make a ultra thin soldering iron you can do it, but I have a feeling that if you assembled it this way, then tightened the screws, it would not be a reliable joint with out some kind of precision alignment to let the solder solidify in the correct space.. it would be under tension/torque.

I would recommend ultra high conductivity silver epoxy.


The circulator displayed above.. is pleasant. Mine was totally welded around the perimeter, and nicely painted. The amount of work that would go into restoring it would be ridiclous.

If I was sure that you can heat the whole thing and use solder paste without cracking anything, I would do that, but I have no idea what the reflow (cooling) profile would be... maybe low melting point solder. IMO don't bother unless you have a mechanically damaged one.

I do think the one above is built a bit bootleg, since they use a standard SMA connector with a solder barrel, rather then a waveguide insertion flat tab. That looks like its meant to bond to a wire not make a solder lap joint. I suspect that the inner pin is going to be less mechanically reliable with thermal variations then the (more) proper type..

[TheMG]

I do think the one above is built a bit bootleg, since they use a standard SMA connector with a solder barrel, rather then a waveguide insertion flat tab. That looks like its meant to bond to a wire not make a solder lap joint. I suspect that the inner pin is going to be less mechanically reliable with thermal variations then the (more) proper type..

Well, it came out of some decommissioned early-1980s telecommunications equipment. Frequency range is 1.7-1.9GHz so actually pretty low frequency for microwave stuff. The details are less critical at lower frequencies I assume.

[coppercone2]

I do think the one above is built a bit bootleg, since they use a standard SMA connector with a solder barrel, rather then a waveguide insertion flat tab. That looks like its meant to bond to a wire not make a solder lap joint. I suspect that the inner pin is going to be less mechanically reliable with thermal variations then the (more) proper type..

Well, it came out of some decommissioned early-1980s telecommunications equipment. Frequency range is 1.7-1.9GHz so actually pretty low frequency for microwave stuff. The details are less critical at lower frequencies I assume.

I think the solder joint is just less mechanically reliable.. though the barrel is strong, the little waveguide insertion thing is very flimsy. I don't know if its better to have it bend or to have strain in a barrel... I think with the big pin, when you solder it to the substrate board with a big glob of solder, it ends up pushing on the dielectric sleeve more? But almost EVERYTHING I have seen does this method because the part is 3 times cheaper IIRC. I never saw it crack either, but I never serviced too much of this equipment. I don't even think its a real established procedure to reinsert the RF cables either.. maybe someone dealing with 1000$ physics grade RF connectors professionally would know. I don't think there could be a free lunch though, with the solder pin cable connector being equal to one that solders properly according to solder theory, impedance aside, with only stress and mechanical movement over temperature change (which I guess translates to surface contact area in threads and pin varying).

[David Hess]

Solder is not mechanically strong so if at least one side of the joint does not provide enough compliance which will be the case here with such a short center pin, a solder joint will eventually crack.  So some type of spring connection is better.

[coppercone2]

by waveguide connection I mean the SMA/2.5 that have a little 'tongue' that sticks out. I guess it is a leaf spring?

[pigrew]

Another reason to not heat them too much is that they contain magnets. Magnets may demagnetize at higher temperatures....

(I don't have any direct experience with heating or repairing circulators, so this may not be a huge issue.)