QUESTION: Homebrew Spread spectrum
 

Can anyone out there tell me a good source of practical information
(schematics and the like) for building SS transmitters and receivers.
I'm particularly interested in how I'd design the part of the receiver
that acquires then tracks the synchronisation signal from the
transmitter (correlation etc)

I did an initial scout around on the net but the only stuff I could
come up with was TAPR but there's seemed to have been cancelled.

I'm into experimenting, and not afraid of math (an electronic engineer
from a long time ago...)


Can anyone help me here...


Tim

Timothy C Holtom wrote:
Can anyone out there tell me a good source of practical information
(schematics and the like) for building SS transmitters and receivers.
I'm particularly interested in how I'd design the part of the receiver
that acquires then tracks the synchronisation signal from the
transmitter (correlation etc)

I did an initial scout around on the net but the only stuff I could
come up with was TAPR but there's seemed to have been cancelled.

I'm into experimenting, and not afraid of math (an electronic engineer
from a long time ago...)

Can anyone help me here...

I have Dixon's text on spread spectrum techniques, but that's perhaps
Too Much Information. The TAPR has a book out on spread spectrum in
ham radio, which may be more to the point. Have a quick look at
http://www.tapr.org/tapr/html/Fpub.ss.html.

And of course a google on `"Spread spectrum" book` will give lots and
lots of hits on textbooks.

But as for synchronization, you have to know roughly where in the
FH or PN sequence the transmitter is, and then adjust your receiver
clock to sync up, and then do early-late tracking to keep the RX clock
synced up with the TX clock.

If I were doing it, I'd always start at the beginning of the FH or
PN sequence for each transmission, and build my receiver to expect
the TX to do that. Otherwise it starts looking too much like crypto
for me to want to deal with it.

--
Mike Andrews

Tired old sysadmin

Timothy C Holtom wrote:
Can anyone out there tell me a good source of practical information
(schematics and the like) for building SS transmitters and receivers.
I'm particularly interested in how I'd design the part of the receiver
that acquires then tracks the synchronisation signal from the
transmitter (correlation etc)

I did an initial scout around on the net but the only stuff I could
come up with was TAPR but there's seemed to have been cancelled.

I'm into experimenting, and not afraid of math (an electronic engineer
from a long time ago...)

Can anyone help me here...

I have Dixon's text on spread spectrum techniques, but that's perhaps
Too Much Information. The TAPR has a book out on spread spectrum in
ham radio, which may be more to the point. Have a quick look at
http://www.tapr.org/tapr/html/Fpub.ss.html.

And of course a google on `"Spread spectrum" book` will give lots and
lots of hits on textbooks.

But as for synchronization, you have to know roughly where in the
FH or PN sequence the transmitter is, and then adjust your receiver
clock to sync up, and then do early-late tracking to keep the RX clock
synced up with the TX clock.

If I were doing it, I'd always start at the beginning of the FH or
PN sequence for each transmission, and build my receiver to expect
the TX to do that. Otherwise it starts looking too much like crypto
for me to want to deal with it.

--
Mike Andrews

Tired old sysadmin

Timothy C Holtom wrote:

Can anyone out there tell me a good source of practical information
(schematics and the like) for building SS transmitters and receivers.
I'm particularly interested in how I'd design the part of the receiver
that acquires then tracks the synchronisation signal from the
transmitter (correlation etc)

I did an initial scout around on the net but the only stuff I could
come up with was TAPR but there's seemed to have been cancelled.

I'm into experimenting, and not afraid of math (an electronic engineer
from a long time ago...)


Can anyone help me here...


Tim

The '88 ARRL handbook had a spread-spectrum project. You'll need to
find a buddy to implement one, too.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Timothy C Holtom wrote:

Can anyone out there tell me a good source of practical information
(schematics and the like) for building SS transmitters and receivers.
I'm particularly interested in how I'd design the part of the receiver
that acquires then tracks the synchronisation signal from the
transmitter (correlation etc)

I did an initial scout around on the net but the only stuff I could
come up with was TAPR but there's seemed to have been cancelled.

I'm into experimenting, and not afraid of math (an electronic engineer
from a long time ago...)


Can anyone help me here...


Tim

The '88 ARRL handbook had a spread-spectrum project. You'll need to
find a buddy to implement one, too.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Tom Bruhns wrote:

W7WKR was doing some experimenting in that area a few years ago (w/FCC
permission), and likely would share some info. I don't have his email
address handy, but perhaps you can find it with a search.

Since very accurate time is available, it should be possible to get
close in the sequence. A typical way to lock onto pseudorandom things
is through correlation. But how do you optimize things when there's
fading and very weak signals to begin with?

Well, the nice thing about spread spectrum is that it effectively
integrates all that under-the-noise signal and produces something with
a lot more amplitude. And yes, the autocorrelation properties of the
PN sequences are very important in synchronization: the output of the
correlator is something close to 1/n for out-of-sync and 1 for in-
sync.

The difficulty is knowing roughly where in a long sequence to start
the search, and that's why the military GPS receivers use the C/A
(Coarse Acquisition) signals to get the Precise Positioning Service
parts of the receivers in the ball park. Once they've done that, it's
the same old game: slew the clock and look for a jump in the output of
the correlator.

--
Mike Andrews

Tired old sysadmin, math jock very much interested in SS

Tom Bruhns wrote:

W7WKR was doing some experimenting in that area a few years ago (w/FCC
permission), and likely would share some info. I don't have his email
address handy, but perhaps you can find it with a search.

Since very accurate time is available, it should be possible to get
close in the sequence. A typical way to lock onto pseudorandom things
is through correlation. But how do you optimize things when there's
fading and very weak signals to begin with?

Well, the nice thing about spread spectrum is that it effectively
integrates all that under-the-noise signal and produces something with
a lot more amplitude. And yes, the autocorrelation properties of the
PN sequences are very important in synchronization: the output of the
correlator is something close to 1/n for out-of-sync and 1 for in-
sync.

The difficulty is knowing roughly where in a long sequence to start
the search, and that's why the military GPS receivers use the C/A
(Coarse Acquisition) signals to get the Precise Positioning Service
parts of the receivers in the ball park. Once they've done that, it's
the same old game: slew the clock and look for a jump in the output of
the correlator.

--
Mike Andrews

Tired old sysadmin, math jock very much interested in SS

A company I used to work for manufactured a frequency hopper spread
spectrum unit.

It was a 'slow' hopper, with Ack/noack for packets so it would retransmit
a missed data packet which covered the case when a particular channel was
occupied.

The acquisition method would also work for a pretty fast frequency hopper
too. It used a master unit that sent a sync signal burst every time it
jumped to a new channel. The remote that was looking for the master would
jump in a duplicate pattern, but at a slower jump rate until it heard the
sync signal from the master. the remote would then start hopping at the
correct rate, and follow the master as it hopped.

Acquisition might be a bit slow, but it did work nicely, and it allowed
other remotes to be turned on at any time and acquire the network.

==============

I have never tinkered with an amateur radio version, but it seems to me
that the ARRL has a book on spread spectrum and there are some FCC mandated
requirements for the PN code, or the hopping pattern for freq hoppers, to
allow the FCC to be able to listen in.

The FCC also limited which frequency ranges Spread Spectrum can be used
on, so I'd do a bit of digging before trying to build hardware......



Jim Pennell
N6BIU

A company I used to work for manufactured a frequency hopper spread
spectrum unit.

It was a 'slow' hopper, with Ack/noack for packets so it would retransmit
a missed data packet which covered the case when a particular channel was
occupied.

The acquisition method would also work for a pretty fast frequency hopper
too. It used a master unit that sent a sync signal burst every time it
jumped to a new channel. The remote that was looking for the master would
jump in a duplicate pattern, but at a slower jump rate until it heard the
sync signal from the master. the remote would then start hopping at the
correct rate, and follow the master as it hopped.

Acquisition might be a bit slow, but it did work nicely, and it allowed
other remotes to be turned on at any time and acquire the network.

==============

I have never tinkered with an amateur radio version, but it seems to me
that the ARRL has a book on spread spectrum and there are some FCC mandated
requirements for the PN code, or the hopping pattern for freq hoppers, to
allow the FCC to be able to listen in.

The FCC also limited which frequency ranges Spread Spectrum can be used
on, so I'd do a bit of digging before trying to build hardware......



Jim Pennell
N6BIU

----- Original Message -----
From: "Laura Halliday"
Newsgroups: rec.radio.amateur.homebrew
Sent: Tuesday, March 30, 2004 8:32 PM
Subject: QUESTION: Homebrew Spread spectrum


How to synchronize to direct-sequence spread
spectrum should be obvious (think Costas Loop).

A sliding-window correlator is a common approach,
otherwise. Three correlators in parallel. One runs
a little ahead, one is on time, the third runs a
little behind.

The LO free-runs (deliberately) a little slow or
fast. When the locally-generated PN sequence lines
up with the received sequence, the outputs from the
early/late correlators track the LO, while the
on-time correlator provides the output.



Thanks for the info, Laura. I never did any work with a Direct sequence
system, but this makes a lot of sense.


Jim Pennell
N6BIU

----- Original Message -----
From: "Laura Halliday"
Newsgroups: rec.radio.amateur.homebrew
Sent: Tuesday, March 30, 2004 8:32 PM
Subject: QUESTION: Homebrew Spread spectrum


How to synchronize to direct-sequence spread
spectrum should be obvious (think Costas Loop).

A sliding-window correlator is a common approach,
otherwise. Three correlators in parallel. One runs
a little ahead, one is on time, the third runs a
little behind.

The LO free-runs (deliberately) a little slow or
fast. When the locally-generated PN sequence lines
up with the received sequence, the outputs from the
early/late correlators track the LO, while the
on-time correlator provides the output.



Thanks for the info, Laura. I never did any work with a Direct sequence
system, but this makes a lot of sense.


Jim Pennell
N6BIU