A 10Gb/s Wireless Link Using a W-Band Phased Array Chipset

[Hugoneus]

Dear EEVBloggers!

I am the founder and host of TheSignalPath.com. Some of you have asked me about the type of work and research that I do. Although most of the things that I work on are confidential, once I publish something, it becomes public domain. I'd like to show you a picture of one of my experiments which appears in my latest Journal of Solid States publication. The title of the paper is "A 70–100 GHz Direct-Conversion Transmitter and Receiver Phased Array Chipset Demonstrating 10 Gb/s Wireless Link". Unfortunately I can't link to the paper because it is IEEE copyright material.

The ICs that I have designed are a pair of transmit and receive phased arrays which operate between 70-100GHz and are intended for data communication. Using these ICs I was able to establish a 10Gb/s wireless link (16-QAM) with beam-forming. Can you identify all the equipment in the picture?  O0

If anyone is interested in more information, I'll answer any questions you may have.

[marshallh]

Real nice work. I can only understand the highest level of the RF stuff, what sort of data were you feeding the AWG?
Makes this 5gbps USB 3.0 I'm working with seem simple

OK with you if I screenshot the block diagram and put it up here?

[Hugoneus]

Real nice work. I can only understand the highest level of the RF stuff, what sort of data were you feeding the AWG?
Makes this 5gbps USB 3.0 I'm working with seem simple

OK with you if I screenshot the block diagram and put it up here?

Thanks!

The data from the AWG was 16- and 32-QAM constellation signals to achieve 10Gb/s data transfer.

You can share any data from the paper here if you like. Which portion of the USB 3.0 circuits are you working on?

[EEVblog]

And the floor at the base of the tripods is covered in drool! 
 

[Marco]

Unfortunately I can't link to the paper because it is IEEE copyright material.

The subject matter of the paper is entirely beyond me ... but I am interested in why you don't put a post-print on your site? I can understand why a lot of academics don't want to bother, IEEE makes it a bloody pain in the ass ... but you are allowed and you're putting in effort to communicate about it any way.

[HackedFridgeMagnet]

Wow nice stuff.
I suppose that radio range is considered far field at 100GHz.

[Hugoneus]

And the floor at the base of the tripods is covered in drool! 
 

Hehe... Yeah, you should see our RF equipment!

[Hugoneus]

Wow nice stuff.
I suppose that radio range is considered far field at 100GHz.

Yes, definitely!

[Hugoneus]

Unfortunately I can't link to the paper because it is IEEE copyright material.

The subject matter of the paper is entirely beyond me ... but I am interested in why you don't put a post-print on your site? I can understand why a lot of academics don't want to bother, IEEE makes it a bloody pain in the ass ... but you are allowed and you're putting in effort to communicate about it any way.

I can link to the IEEE website, but you would have to have IEEE subscription to view the paper itself.

[Aeon]

please link to the site

I can read it at my university they have a subscription for their students

[Marco]

I can link to the IEEE website, but you would have to have IEEE subscription to view the paper itself.

That's not what I meant, you're allowed (with the right disclaimers and citations etc etc.) to publish the accepted version  (ie. before final editing/formatting for publication) of your paper on personal, institutional and non commercial archival sites. As far as I can surmise from the IEEE website most of their journals even prepare a special version of the accepted paper with DOI references expressly for that purpose which you can download from the author portal (dunno if this is true for the journal in question).

[marshallh]

IEEE make the authors pay $1750 for the privilege of letting the world see their work on the IEEE site. There are methods to get around the paywall, just think a bit

Block Diagram:

Fig. 2. Simplified block diagram of the implemented 4-element phased array receiver (left) and transmitter (right) chipset. Each integrated circuit is equipped
with a global SPI interface for settings and calibration.

The antenna array shown in Fig. 15 is designed to achieve the wide bandwidth while maintaining good gain and efficiency.
While the proposed antenna and array can be realized in conventional microwave substrate such as quartz, we have chosen a high-speed PCB Megtron-6 [23] substrate for array fabrication
to further advance the knowledge and techniques of millimeter-wave antenna integration. The usage of such substrate material in this frequency band is a novel attempt to the authors’
best knowledge. Such PCB fabrication process would enable the integration of antennas into the rest of the control circuit and achieve simplification of overall module fabrication.

Fig. 26. Measured constellation and spectrum of the complete wireless link operating using an 88 GHz carrier. At a distance of 1-meter, a 2.5 GBaud/s 16-QAM (left) or 1.75 GBaud/s 32-QAM (right) link can be established with a measured EVM of 8.1% and 6.6% respectively.

You can share any data from the paper here if you like. Which portion of the USB 3.0 circuits are you working on?

I'm writing a FPGA MAC/Link layer controller, talking to a bare metal PHY that does serdes,8b/10b. All digital though. It would be crazy to see it move to wireless..

[Pippy]

Wonderful work!

Is the adjustable antenna phasing apart of what gives you the 10Gb/s or is the 10Gb/s just down to fact that you have GHz of RF bandwidth/spectrum up at 100GHz to play with?

[Hugoneus]

Wonderful work!

Is the adjustable antenna phasing apart of what gives you the 10Gb/s or is the 10Gb/s just down to fact that you have GHz of RF bandwidth/spectrum up at 100GHz to play with?

The independent phase control of each antenna element enables beam forming in space so the radiation pattern can be controlled in one dimension.

At 100GHz, there is a lot of bandwidth to go around. The difficulty is designing circuits which have sufficiently low noise figure and high linearity, RF/IF bandwidth to support 10Gb/s operation. Since this is double side band transmission with complex modulation, I/Q accuracy and calibration is also critical.

[mark03]

I can link to the IEEE website, but you would have to have IEEE subscription to view the paper itself.

That's not what I meant, you're allowed (with the right disclaimers and citations etc etc.) to publish the accepted version  (ie. before final editing/formatting for publication) of your paper on personal, institutional and non commercial archival sites. As far as I can surmise from the IEEE website most of their journals even prepare a special version of the accepted paper with DOI references expressly for that purpose which you can download from the author portal (dunno if this is true for the journal in question).

THIS!!!   

Not to hijack the thread, but I would like to take this opportunity to plead with all IEEE authors to please place a copy of your accepted paper(s) on your personal web site, or somewhere else where google can find it and anyone can read it.  All of my publications are freely available this way, it's perfectly legal, and the world will be a better place if everyone follows suit.  Think of it as a stop-gap measure until the govt gets around to doing their job, i.e. mandating open access for all federally funded work.

Full disclosure:  although I am published in IEEE journals, I am no longer a member, even though my company would pay the dues.  I will re-join as soon as members automatically gain access to the IEEE Digital Library.  Gotta take a stand somehow.