Hi Michael, thanks for your input. Here's a point by point response.
# It seems like the target keeps changing.
Quoting from the Press Release at Hamvention 2012:
"The radio will support data rates from 4800-56K+ bps with selectable modulation methods including GMSK, FSK, and 4FSK. The UDR56K will operate in the 70cm band (420-450 mHz.) at up to 25 watts."
As of today we have reduced the band to 430-450, due to availability of SAW filters and increased the upper limit to in excess of 100k.
As far as dropping legacy 56k support. We would rather spend engineering effort on a modern FEC modem.
We are still on target.
# Even if a higher speed is possible, it may not be deployable in many locations due to path issues.
We are releasing 9600 packet and 4800 D-STAR for full compatibility with the installed base.
Past that, you are absolutely right about the difficulties at higher speeds, so let me clarify our position.
Although lab work is required for verification prior to deployment, the real world is all that matters. We are working with some distinguished modem designers who have decades of experience in this area.
As we announced at DCC in November:
The UDRX will ship with a "Channel Sounder" which will allow all of our users to explore the field performance of candidate modems after they are released (all modems are open-source software plug-ins). I expect many hams will provide the data from mobile stations reporting back to an Internet-connected station where the data will be available on our website.
This data collection and optimization period will last for the better part of the year and result in a short list of modems/rates for different conditions. A negotiation protocol can be added later to establish the fastest reliable connection for a given situation including fall-back as you suggest.
# Also, I hope whatever you do includes an analyzer that shows …
RSSI and BER are built-in as are forward/reflected power. In the process of testing our receiver, it became necessary to inspect the data at various points and display the results graphically. We currently have a crude file based tool that allows us to collect a sample and post-process it using open source tools for things like constellation
The next step will be to connect to the receiver over a socket and display the results using something like WebGL. We will provide the socket but the rest is a substantial software task and we expect some talented Ham will pick this up. I have spoken to a couple of candidates already, but this is not a deliverable item for NW Digital Radio at this time.
All of our software is open-source so anyone may take our internal tools and build on them. If you have a strong math background and the requisite lab equipment, you can design your own modem using our base 9600 FSK/4800 GMSK as a reference.
# I really, REALLY want this to work
Nobody wants this to work more than I do. Except maybe my wife and my increasingly nervous banker.