[Ham-80211] Activity - comments and idea

[Bob Keyes]

On Wed, 30 Jan 2008, Naber, Benjamin L. SPC wrote:

>  With the bandwidth requirements we have for 2m and 70cm, what
> modulation method would get the most effective throughput? Does having
> FEC lower that actual data throughput because more bits are being sent?

> ~Benjamin, KB9LFZ

Benjamin,
I believe that the specific bandwidth usage patterns of your area will
greatly affect the appropriateness of an individual protocol. Those hams
in rural ares or areas with little restriction due to military uses will
have much greater freedom than others.

> > I've been working on a project, just a draft right now, but will be
> > manifesting into the real deal in a little while. I would like to know
> > everyone experiences with different multi-level modulations like QPSK,
> > 8-PSK and the others that are being used. I like the QPSK or 8-PSK as
> > that is what we use for our satellite internet.

Even the particular satellite you are using will effect what is the the
most effective modulation scheme. Geosynchronous satellites have a higher
RTT, which affects the applicability of short packets with
acknowledgements. LEO satellites will have less delay, but may be harder
to track, and when they approach horizons of course will suffer distortion
due to more atmosphere to penetrate.

> Experiences are one thing, actual parts availability and
> bang-for-the-buck might be different. QPSK and 8PSK work great, and
> gives a pretty good tradeoff between power and bandwidth requirements.

That, in the end, is what really counts - how much can we get as
'commodity parts' and how much can we bend them to our needs or will?

> You will probably end up using off-the-shelf 802.11 chipsets, just
> because they work so well and are ubiquitous.

But the abilities of the chipsets, and our ability to 'bend' them, vary
greatly.

> The collection of
> modulation modes in 802.11 a and g (BSPK, QPSK, 16QAM and 64QAM) offer a
> good set of options to choose from. The challenge is getting deep enough
> into the chipset to use the modem directly without having legal access
> to the chip maker's proprietary code.

There's much happening with regards to the AR5K driver in Linux. There's a
move from the 'closed' HAL, provided by Atheros, towards the
reverse-engineered OpenHal and its relatives in the BSD and Linux worlds.
There is also much legal uncertainty here, as the US Gov't. requires that
spreading codes are standardized for DSSS, though I am not sure what
restrictions are in place for OFDM.

> Historically, 64QAM was what the telephone company settled on as the
> best compromise for long distance microwave hops. Your MILSATCOM stuff
> is mostly 32QAM although some systems go higher. Only if you're really
> short on power, and have excess bandwidth (or very low data rate
> requirements) would you want to choose PSK.
> >
> > What I'm working on is a TCP/IP setup that will have the functionality
> > of a wireless router, but will have the functions amateurs like to
> > see, and also be open source Linux so we can attract more computer
> > people in here. Seeing how this a group with people who are computer
> > oriented, there shouldn't be anyone griping "this is ham radio!" If
> > you are one of them, then pay no mind as I pass you by.

I think that anything that is not open-source (be it BSD or GPL) is not
ham radio. We have no use for proprietary schemes and they should not be
tolerated or even allowed. Considering this, we should try to move as much
Amateur Radio activity towards FOSS as possible (too much of it is
Microsoft centric, and this is a real problem for me).

> > If you are familiar with DD-WRT, it may have that setup as I like
> > using that on my linksys and Motorola WRTs.

I prefer OpenWRT which is why I develop under it. I believe it has a
greater future than DD-WRT because of its OS redesign from the ground up.
I don't fault you on a philsophical level for using DD-WRT though.

> That (and its cousins in the open and mostly-open source community) is a
> good start. The snag has always been the hardware abstraction layer that
> the chipsets need.
> > the main thing I like about this idea is that it will NOT require
> > drivers on a computer to work so an old Toshiba Satellite 75MHz Laptop
> > with an PCMCIA Ethernet port and TCP/IP support with Win3.1 will work.

Eh, I've got a big stack of such old laptops (I actually just chucked
anything that was less than 90 Mhz!) and I'd like to find uses for them -
but these can be on HF with less demanding protocols.

> That's a good start. But it has to be cheap, too.

Cheap is crucial. But how cheap? There should be a sliding scale of
expense that translates to performance. I'd like to think that with a lot
of elbow-grease and $50, a station can be assembled which will share some
compatibility with thousand dollar stations. The reason for this is both
from idea that some people may not have as great an interest or need to
have powerful stations, to those people who do not have the economic means
to afford more expensive hardware.

What bothers me right now, is the separation of the open-source Community
Wifi movement and that of Amateur Radio.  There is a an amazing amount of
enthusiasm and energy in the former, but experience and resources in the
latter.

At my club (Harvard Wireless Club - world's oldest station!), the emphasis
has been for some years on HF and contest operation. Tese are good folks,
but the number of undergraduates involved in the club in reaching
dangerously low levels.  I am told that other clubs are in serious danger,
and we may lose them entirely unless their decline can be stopped. I
believe this can only happen by amateurs seeking out and adopting the new
generation of digital hams. I don't want this to be interpreted as saying
I want to ban AM & SSB on HF, but I believe that digital radio shoudl be
acknowledged as the future of amateur radio and awarded spectrum as such.
But digital hams must make sure that our DSPs can detect analog modes and
operate with the minimal amount of interferences as possible. I think we
can do this. CPU power and DSP methods, along with speech recognition, are
growing by leaps and bounds. But such CPU power does not come cheap, which
is why we continue to need bands which are digital-only, for those
stations which are less sophisticated (because they need to be cheap).

I know I am going on at some length here, please forgive me, as I have had
too much coffee.

I believe that Amateur Radio must continue to justify its existance and
its allocations in light of the vast demand and monetary value of
bandwidth (even though, per radio act of 1934, bandwidth is not to be
sold!).  We must do this in a non-competition basis with commercial
providers, especially as those providers are paying more and more for
spectrum. Greedy governments will be increasingly casting eyes on
underused amateur spectrum in search of money. We must prove that the
spectrum is being used, and is being used wisely. This will involve some
rather introspection by the amateur radio community. These philsopical
arguements can go hand-in-hand with those in the FOSS (Free Open SOurce
Software) community. For instance, is it appropriate for a commercial
entity to utilize amateur radio spectrum when the information they
transmit is freely available to others without charge? How useful to
the public does this information have to be in order to to be a legitimate
use of the spectrum?

I have thought of this issue in relation to sensor networks. I believe
that meteorological information and climatological sensor information is a
legitimately public use of the amateur spectrum. I have done some
preliminary design of some possibly radical changes to sensor networks, as
a result of my involvement with sensor networks projects at Harvard
(CitySense.Net - though I am no longer employed by them). But what about
other information? If I were to say, put a sensor on amateur spectrum
which provided information on the water level of a certain reservoir,
would this not be more appropriate for a commercial or municipal license?
Or what about atmospheric carbon dioxide levels downstream from a
coal-powered utility plant, which would be primarily for the use of the
owners of that plant in proving their compliance with various regulatory
reuirements? But as it could also be used by their critics, and
atmospheric researchers in general, is it still legitimate, even though it
is "boring" to the rest of the amateur radio community? Perhaps these
issues can be addressed by considering how much of the valuable spectrum
they consume, and what benefit they produce - but do we want to make every
sensor transponder open to the bureaucracy and competitive sniping that
such policies might entail?

If fashioned correctly, a new policy might not only protect amateur
spectrum, but to increase it. For instance, I would think that an
automated digital NVIS system would be of great pragmatic value, while
also allowing amateur use and maintenance.

Lastly, I would like to find a way of FUNDING such research an
implementation of this infrastructure which may be useful to both amateur
and scientific use. Consider the advantage of volume-produced digital NVIS
gear, for instance. Of course, as a researcher in digital radio
communications, I;d like to think of some of that funding paying my rent
and an occasional conference, be it formal or a beer-gathering with
friends.

73,
Bob Keyes
N1YRK