Can FET transistor be used to replace RF isolator?

So the question is about things like RF circulator, but with two ports. What if replace isolator with FET, or maybe ever cascaded? How succesufull it would be? Thanks!

A Fet set up as an amplifier with an accompanying attenuator pad will work ok (maybe 15 dB of isolation), although it might add additional noise or intermodulation products. But since a fet and resistor pad will cost < 1S, and an isolator will cost >S25...it is an experiment worth running for a production job. Of course, you realize that you will have an "isolator" when you are through, no longer a circulator.

The whole point about a circulator is that it has 3 ports!
What goes in Port1 exits Port2. What goes in Port2 exits Port 3. What goes in Port3 exits Port1.
A whole lot of ferrite magnetic trick is usually needed to make it happen.

A 2-port device clearly cannot do anything like this.

If you perhaps mean a 2-Port device where energy into Port1 exits Port2, but not vice versa, then a FET amplifier does this. S21 (the gain) is normally much higher than the reverse transmission S12.

Isolation between two inputs can be done with various types of combiner, for example the "Rat Race", or the "Quarter Wave Hybrid". Those two are resonant things somewhat restrictive on bandwidth. Wideband directional couplers also provide isolation. This would be 4 ports with maybe one port terminated.

These arrangements also involve some loss from the network.

It's not clear just what you are seeking but here is an interesting design;

http://www.techlib.com/electronics/circtran.htm

The "active isolator" was invented and commercialized by ARA company ~30 years ago. It combined a wideband amplifier and an attenuator, covering frequency bands ~30 to ~300 MHz in sections.

In `2006 I needed a good wideband isolator covering 500-1800 MHz IF band. Ferrite isolators do exist but are expensive and need a heavy steel screen ue to low-intensity magnetic field.
So I re-invented the "active" isolator using RF 2743 MMIC amplifier with an adjustable ~15 dB attenuator. The results were better than the ferrite isolator, the device was cheaper and only needed ~ +5 VDC, 40 mA each.
I published my design in "www.highfrequencyelectronics.com" for anyone who needs it. At this moment I do not remember when exactly it was published. Try to look "Polivka,J.: Active RF isolators"

Simple answer is no, Fet cannot handle as much power as an RF Isolator can. Some may know but others may not. A circulator has 3 ports which normally consist Port 1 is the input, Port 2 as an output and Port 3 is loaded with a 50 ohm load. The load is there in case there are feedback from Port 2 will act as a fuse and will prevent any power going back to Port 1. The output of the amplifier is what normally connected to the input of the circulator and the antenna is connected to port 2. A circulator is design to protect the amplifier and case there any feedback from the antenna.

Read my paper as indicated above. If you need,you can use a FET instead of the MMIC I used. And yes, signal power is limited due to the active elements used. I used my "active" isolator for lo-level signals, like -10 dBm. Ferrite isolators can handle more power but the limit is ~ 1 W approx. Only special devices can handle kilowatts.

I agree if your power requirement is low. I'm a passive ferrite eng, my question for you is, are FET frequency limited? As far as passive device, especially ferrite product. There are 3 types of design. They are above resonance, resonance, and below resonance. For above resonance unit, you can easily design a unit that can handle of Kw of power. Special cooling and internal compensation are require to maintain it normal operation temperature. There are many factors in the design to make it work at the level of power. Some design require to work at a certain latitude, so beside power, we have to worry about corona arcing.

For below resonance design, I agree with you. The power handling for that is ~ 1W if you using an absorber as a load. There are chip resistors now that can work up to 18 GHz and we can push to power handling to ~5W.

Thank you for your comments!
I had to design my "active" isolator due to the fact that I failed to find a good low-frequency wideband ferrite isolator. THe bandwidth was 400 - 1800 MHz. Available ferrite isolators were huge, heavy, and expensive; for such frequency the magnetic field intensity is low and requires heavy steel case for shielding. So I tested my idea and it worked. In my paper above, I compared the parameters of m device with one expensive ferrite model. I achieved a full bandwidth with a very good match (RL>20 dB) and isolation >30 dB.
Later I found (added in References) that other designers invented and sold even full circulators with transistor amplifiers, for RF bands from 10 to 250 MHz, where even today ferrite devices are not available.

- - - Updated - - -

To add a word about FETs: the MMICs I used mostly utilize bipolar transistors. FETs can be used if you can match the gae input impedance to ~50 Ohms over a wide band. Otherwise a FET amplifier is better as a narrow-band amplifier.

Some MMIC do utilize FET stages even in wideband amplifiers. MESFETs have been designed for > 200 GHz, so if you find such device, there is almost no frequency limit.

Thanks for the clarification, base on the bandwith required in your design, you are absolutely correct. The frequency, size and cost would be a big factor if you choose to go with a passive design. I know that in a passive design, it will be hard to achieve the electrical characteristic that you can.

I don't know what operating temperature your design need to work at. Say it's for a commercial product an operating temperature is 0 - +60 degree. I would say the best we can do is 15dB for return loss and Isolation and about 1 dB for insertion loss.

If you require a higher frequency than we can achieve close to the specs you provide. We would do an octive design, an example of that would be 2-4 GHz, 4-8 GHz, 8-16 GHz. The power handling as you stated earlier would be ~1 W.

I'm very impress with your approach and congrats!

FET `s and isolator are good pare and no way to replace with *** marriage.