[TenTec] My take on how to "fix" the Orion.

[Geoffrey S. Mendelson]

Posted to my blog: http://geoffstechno.livejournal.com/

The following is my take on how to "fix" the Orion. Feel free to comment,
but unless it is unprintable, post it to the list. :-)

For those that are not ham radio operators, the only manufacturer of
amateur radio transmitters from "the old days" left in the U.S. is a
company called Ten-Tec. They produced wonderful radios almost 30 years
ago. I have two of them in my "shack". An Argonaut 509 and Triton IV
digital. Both are over 20 years old and except for a loose dial cord on
the Argo, they both work well. They lack lots of modern features, but
they communicate quite well.

Ten-Tec's current offering is a radio called the Orion. It is now in
it's second generation called the Orion II. The Orion is a software
defined radio. It looks like a radio, complete with lots of knobs to
tweak, it acts like a radio but inside all of the control functions and
low level signal processing is done by a computer.

There's the rub. Ten-Tec's software genius left the company, and they
have not been able to replace him. I'm not surprised, they are in the
middle of the U.S. Southeast (in Tennessee) which IMHO is even more
remote to what's happening than Jerusalem. The cost of living is low and
so are the salaries. If a genius embedded software designer and
programmer were to make $100,000 a year, which is an enormous salary in
Tennessee, now they would be paid $300,000-$40,000 in silly valley, the
demand is so high.

The other issue is of course, working for a well established company
with little or no growth potential. Ten-Tec is a solid business, it will
be around in a year, or five, or ten. Most start-ups won't. BUT... If
you are good enough to run a design/programing team, you can write your
own ticket. Any job for a startup will either leave you on the street or
a millionaire (or both) in a year or two. So what if you take a job for
a company that goes bust. Save your money, work for another startup and
if it hits, it will hit big and you can buy all the radios you want and
retire. If the second startup fails, go to a third and so on. Eventually
you will hit it big or burn out and take a job for Intel. :-)

A development team must be put together that is both productive and easy
to support.

So how does one make the Orion work? Here's my take on it:

1. Replace the CPU. Get rid of the DragonBall, a rehashed 1980's
processor from Motorola and replace it with something more powerful. In
a $4300 100 watt output radio the difference between a $10 processor
that uses less than a watt, or $40 processor that's hundreds of times
more powerful, uses six watts and has a more common instruction set
(larger pool of programing talent) is trivial. If you want to separate
the processing into three separate processors along with their own
memory and control chips, it would raise the retail price to about
$5,000. Not a very big jump. It probably would be less because digital
signal processing chips would no longer be necessary and they run very
expensive commercial (paid for) code.

The processor I have in mind is made by AMD, and Intel has similar
chips. VIA claims to, but their claims often exceed the actual hardware
by miles. Transmeta chips would have done well too, but they never were
able to sell enough to keep in business and went under.

The processors use the X86 instruction set, the same as in any PC. PC
programers are easy to find. Really good ones are hard to find, but
nowhere as hard as finding ones that program DragonBalls.

2. Split the code.

Use one processor to control the radio functions. This code, by it's
very nature has to be kept proprietary.

Use the second processor to run the digital signal processing. There is
lots of public domain DSP code out there, so development cost is lower.
Not only that, since it really does not control the actual radio, it can
be released under the GPL or a BSD artistic license and the world
becomes your extended development and testing group.

Use a third processor to run the user interface and display. The Orion
has a nice full color display, but even the latest version has trouble
keeping up. With a 1gHz x86 processor, it will nicely run the display,
talk to a users computer and poll the various knobs and switches and
transfer their settings when they change to the processor that runs the
radio.

It could even have an ethernet interface for both remote control of the
radio itself, and a digital data in/out.

Since the radio would have modular (in pieces) code, each piece can
easily be developed on it's own and then integrated after it's tested.
Since it runs on CompUSA type cheap PCs, it can be tested by a small
army of quality assurance (QA) testers long before it ends up inside a
radio. The testers can be signal processing and programing technicians,
not obscure DSP chip programmers or ham radio operators. You could hire
them out of the local technical college not have to woo them away from
the big companies.

In conclusion, what the Orion needs is not only a redesign and
reprogramming, but a rethink on how to develop it.

Geoff. 


-- 
Geoffrey S. Mendelson, Jerusalem, Israel gsm at mendelson.com  N3OWJ/4X1GM
IL Voice: (07)-7424-1667  IL Fax: 972-2-648-1443 U.S. Voice: 1-215-821-1838 
Visit my 'blog at http://geoffstechno.livejournal.com/