[hpsdr] HERMES schedule question

Fri Jun 18 15:08:56 PDT 2010

Wow --  That sounds like a very ambitious project, Ray.  I'm interested in what you come up with.  Keep in mind that software-wise, things are going to be moving towards a Gigabit Ethernet network interface (See the Ozy II Project) because USB can't handle enough of what we're sampling at the antenna.  So, given that in this case your data is going to get packetized at some point in time, it might be just as easy to get a fiber->copper Gig-E converter.  I'm not trying to discourage you in any way, but just trying to let you know that the USB on Hermes is about to be eclipsed by Ozy II to a certain extent.

On Jun 18, 2010, at 10:07 AM, Ray Page wrote:

> ***** High Performance Software Defined Radio Discussion List *****
> 
> Thank you, Phil.
> 
> I'll be on TeamSpeak tonight, but in listen mode. I have a lot of catching up to do. I realize it's a little early to be thinking about derivatives of HERMES, but an idea struck me as I was reviewing the HERMES design. Unless there is a major flaw in the concept, I am willing to peruse it and fund the project myself with the work donated to the HPSDR community. 
> 
> The core of the idea is to have the option to locate the RX and TX functions at the antenna using fiber-optic cable. The HERMES design is the perfect starting point. It would be partitioned at the FPGA/ADC and FPGA/DAC connections. A full-duplex 3.125Gb fiber-optic link  will carry raw ADC and DAC (or P/E) data. There will also be dedicated control/monitoring data paths as well. The fiber can loop through multiple RX and TX modules as desired with only one-each selected at any given time. Otherwise, one could populate more than one fiber-optic transceiver on the main processing board to realize multiple concurrent RX/TX modes as needed. Both RX and TX functions can also be integrated onto a single FO board to reduce cost. 
> 
> I am very comfortable with designing custom FO data transmission systems, as I have quite a bit of experience in this area. I already have (and own) IP that  uses a pseudo-random, self-synchronous polynomial to guarantee DC balance and a Gaussian noise profile. I believe that the likelihood of interference will be greatly reduced due to the fact that the sensitive circuits are isolated from each other by links that will have their uncorrelated noise contribution spread over a very large bandwidth. The link exhibits around 200ns pipeline delay plus propagation delay of the FO cable itself. In other words, the data paths are real-time---not packet based. It will be nearly trivial to transmit contiguous 16-bit x 130MHz sampling data over the fiber. In effect, the fiber becomes a virtual full-duplex parallel bus. That is exactly how the interface is presented inside the FPGAs.
> 
> The FPGA on the main board will need to be changed to a Cyclone IV -GX. Smaller Cyclone GX's are required on RX and TX modules.
> 
> PROS
> ====
> - Shack can be located up to 200m away from the antennas using low-cost multi-mode fiber exhibiting the equivalent of 0dB loss. Note that the 200m limitation is per FO hop, so a single FO pair can loop from one tower to another if someone is so fortunate to have this problem.
> - Single FO pair can select any arbitrary combination of an RX and TX module in the chain. ie: RX on one band and TX on another.
> - For those wanting to operate more than one TX or RX module concurrently, then more FO pairs can be added back to the base unit.
> - Lightning protection for the shack. Only the RX, TX, and power supplies at the antenna are at danger.
> - The FO cable itself is immune to RFI/EMI, so routing of the cable is simplified.
> - FO cable is far cheaper than large copper RF coax.
> - RX and TX modules can be plug-n-play, since the control/monitoring channels can identify their capabilities automatically.
> - Monitoring includes status of each FO connection in addition to RF module status for easy troubleshooting and performance monitoring.
> - This topology is extremely flexible. Start simple, and then expand to your hearts content (never). 
> - Each RX and TX module contains everything needed to connect to the antenna. RX modules will have any required filters, attenuators, etc. TX modules will include power amps and output filters as needed. 
> 
> CONS
> ====
> - More expensive than HERMES for a simple setup.
> - More complex than HERMES for a simple setup.
> - RX/TX modules exposed to weather/temperature extremes if mounted on the tower.
> - It doesn't appear that SERDES-enabled FPGAs (at least not Altera) are available in leaded packages---only BGA.
> - My bias and lack of experience as a HAM may prevent me from thinking of another CON.
> 
> Finally, as you may be aware from an earlier post, I am testing for the first time tomorrow for my Technician license along with my son, Carson. So, I do not have the real-world experience of operators like yourself. My idea is based on what I imagine my needs will be. Please feel free to temper my ideas.
> 
> Thank you,
> 
> Ray Page
> Georgetown, TX
> 
> 
> _______________________________________________
> HPSDR Discussion List
> To post msg: hpsdr at openhpsdr.org
> Subscription help: http://lists.openhpsdr.org/listinfo.cgi/hpsdr-openhpsdr.org
> HPSDR web page: http://openhpsdr.org
> Archives: http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/

--
Jeremy McDermond (NH6Z)
Xenotropic Systems
mcdermj at xenotropic.com

 1276898936.0