Non-Magnetic RF Circulator

[Coe Schlicher]

I noticed on this blog that a technology my company is commercializing was being discussed a few months ago.
"RF Breakthrough: Non-magnetic circulators"
Namely an RF Circulator using no magnets. As I am the CEO of the company, I thought I might offer to answer questions about the technology as this group seemed to be asking the right questions.
Our web site is
www.siaudio.com
The circulator technology we licensed from UT is explored somewhat on the site.
I wondered if anyone had any questions about the technology I might be able to answer. Just post them and I will answer them. As a side note, the inventor of the technology Andrea Alu is the company's CTO, so it is likely I will answer the question correctly after running my responses by him.
http://www.utexas.edu/news/2014/11/10/radio-wave-device-alu/
http://www.nature.com/nphys/journal/vaop/ncurrent/full/nphys3134.html

[miguelvp]

Just to add the link of the thread the OP is referring to:

https://www.eevblog.com/forum/chat/rf-breakthrough-non-magnetic-circulators/

[rs20]

Cool! I'll start with: do you actually sell parts we can buy? I don't see any products per se on your website, just a lot of cool research.

[Coe Schlicher]

We do not have parts you can buy. We are currently in the process of deciding what performance attributes would make up the most interesting first applications for the technology.
We have designed a circulator to address the needs of RFID reader. This market had performance specifications we already met with our original designs so we agreed to build a few for a full up demonstration of the technology in the field.
If you have any feedback about what type of circulator you would want I can tell you when you might reasonably expect to be able to buy one.
Currently the technology uses all off the shelf parts and is simply scaled for different tradeoffs between, power, frequency, linearity and insertion loss.
The fact that the technology can be build on chip also opens up many possibilities previously not open to existing magnet based circulators. Namely most things handheld or where weight and size are a factor.
One other exciting piece is that we have demonstrated much higher isolation than existing magnetic circulators. Above 30dB with good instantaneous bandwidth.
The time frame for the technology is as follows.
Next 6 months. Circuit board versions with off the shelf discrete components are being developed for some early adopter companies where the circulators size and weight open up new possibilities. In August the technology is being tested in a full up trial which we are very excited about. Lab results are looking promising that we can meet or beat all of the performance necessary for the application with a 100 time reduction in size and weight.
Next 12 months. Increase the power handling and linearity to address larger markets that utilize circulators.
Next 24 months. Create on chip embodiments of the technology for select applications. One of which would be helping to enable full duplex.
 

[T3sl4co1l]

Personally, I'd love to have a wideband one for power applications, but I'm sure I'll be waiting quite a long time for that.  I expect this will be mostly a small signal, or at least low power, thing, so an emphasis on high isolation and linearity would seem prudent.

To be fair, I'm not in radio to enough depth that I'd be deciding to use a circulator (or not), so I may not be of much help here.

Tim

[daqq]

Thanks, looks great! Could this be applied for frequencies in the 9GHz regions? And what kind of power can you send through it?

[Coe Schlicher]

To address the question about frequency range. The core invention works all the way from the acoustic bandwidth to the optical. We demonstrated the acoustic version first which was featured on the cover of Science magazine last year. So yes the 9Ghz range is possible.
The maximum power that can be pushed through it is not theoretically limited. That is not to say it has been publicly demonstrated yet.  If there are discrete components that can handle the power then we can. The technology is readily scalable.
The interest to date for the technology centers on the fact that it is not dependent on rare earth materials and can be miniaturized. Magnets are big and relatively expensive. PCB and discrete components are not.
The criticisms to date center around insertion loss and linearity. Both of which we have improved significantly since the original publication making the technology a true competitor to existing ferrous circulator technology.

[T3sl4co1l]

As I recall, it depends on nonlinear capacitors, correct?  Varactors aren't usually offered for tuning transmitter antennas, and convincing the manufacturers to make new parts would be about as successful as pushing water uphill

There are plenty of nonlinear ceramic capacitors, but are they even stable enough to work?  What's the sensitivity of something like this?  Is it still modest bandwidth (as I recall, the original article was for an RF bandwidth sort of application), or "wideband" in the octave sense, or "wideband" in the decades sense?

Tim

[Coe Schlicher]

Do you mean instantaneous bandwidth or total tunable range. Ferrite based circulators are passive and have a wide bandwidth of possible frequencies they can cover. If the channel you want to operate at is in the middle of the circulators range then it will work well. Our circulator works differently and is tuned to the center frequency of the channel you with to use. Often an octave or more of range. Ours current embodiment works differently and can not currently handle the instantaneous dynamic range needs of say radar (although a next generation will have much higher instantaneous dynamic range). Our circulator can be tuned to have very high isolation at specific frequencies. The pulse rate of the varactors directly correlates to the frequency where maximum isolation occurs. So a single circulator can be tuned for different frequencies.
I will think more about your question and get back to you varactor alternatives as accurately pointed out they are non-linear by their very nature.
 Also your comment concerning the manufacturers about custom parts is well taken and on our previous technologies unless you start the conversation with I need 10 million of these they wont consider changing anything about a component.
As I look at the survey people seemed most interested in isolation levels so I will put together some information specifically about that specification.

[Coe Schlicher]

Isolation
The informal survey above indicates isolation is the key performance attribute of note, which is not surprising. This is encouraging for our technology as our isolation is much higher than a conventional magnetic circulator. There is no theoretical limit, if we had ideally loss less elements then the isolation would be infinite. We have demonstrated >70 dB isolation which is full 50dB more than a top performing magnetic circulator. That said, we don’t think there is a point to push it further, since the impedance matching at the antenna port will ultimately limit the isolation performance between TX and RX ports. In practice no antenna is matched better than 30-40 dB.

Varactors
Varactors or tunable capacitors are common in printed circuits and IC design. The key advantage of our technology is that we are replacing magnetic materials, which are much more difficult to get and difficultly integrated on chip or in a printed circuit. There are also applications that use circulators where weight and footprint are very important. Such as anything using a circulator that work as a handheld device. In those application we offer some significant weight and size improvements.