On 4/17/2010 1:33 PM, Martin Ewing wrote:
On Sat, Apr 17, 2010 at 1:15 PM, Dr. Gerald N. Johnson
<geraldj@weather.net <mailto:geraldj@weather.net>> wrote:
Yet one of the problems with the Orion II is limited memory and
processor capability, compounded for the Orion with less of each. Where
the desired PC for a Flex radio has a gigabyte or two of ram with a CPU
running with a 2.5 or 3+ GHz clock speed, with a dedicated processor for
the video (think gaming computer) and a 32 or 64 bit memory interface,
preferably 64 bit. It might be that one option for Orions could be an
analog interface to the external PC with I and Q from the detectors for
the sweep section or after all the filters at 14 kHz audio so the Flex
radio (as well as other codes) could be applied in the station PC. Thus
enabling far greater computer power than the embedded hardware. Its
likely the limits of the embedded hardware that make things interact in
new versions of the firmware and that make correcting and adding so very
difficult. The fast PC could then control the Orion through the
conventional serial interface.
73, Jerry, K0CQ
OK, Jerry, you convinced me. Go for it!
Seriously, I wouldn't expect help from TT, but it could be done by a
very motivated ham. The thing is, someone with those skills would have
lots easier ways of getting a powerhouse SDR rig. Buying a Flex XXX is
a lot cheaper, if you value your time!
I figure at least a man year. I have many much simpler projects for
other things in mind anyway. It might take a year to become competent
enough in the embedded processors to get a rudimentary radio running.
I would say the Orion (O1 especially) manages to do an awful lot of good
stuff with its limited resources. The DSP chips take most of the
computational load. The rest is user interface, command & control, etc.
There's half a megabyte RAM limit (19 address lines) for the control
CPU, I didn't look at Orion 1 schematics deeply enough this morning to
find external RAM for the DSP chips.
What benefits would you possibly hope for by bolting on a super PC
backend (that you couldn't buy somewhere else)? It might be easier to
design a "proper" hardware front panel to put in front of a Flex 5000. :-)
Might be just another computer interface to the computer control software.
All that said, I do wish the Orion hardware was a bit more open, giving
the user access to the IFs, I/Q, etc. for experimenters. Also, a much
more solid computer control port. And a better frequency standard,
external freq. locking, and on and on... But we go to war with the
rigs we have.
73 Martin AA6E
Its just often a lot easier to do very fancy performing software when
there are GB and GHz to play with which allows using higher level
languages and procedures that won't fit in the embedded processors
without doing all the programming in assembler which is very slow to
program while it can be the very fastest of executing, what I did for
the Z80 30 years ago. My productivity (though the programs too much more
memory and time) went way up when I converted to C and went up more when
I got a 32 bit C compiler for OS/2 instead of MS-DOS so I didn't have to
do the work of dumb Intel extended addressing. Most of the megabytes of
my software has no direct user interfaces, it all runs in batch modes
with command files for the varying weather tasks.
73, Jerry, K0CQ
|