The last few days I have been using Scilab
to simulate various types of CPM. It is much
more fun playing with stuff rather than following
the maths. Well it is for me.

I have now finished the first book I bought on
CPM and have just started another book. The new
book is on estimating symbol timing and carrier
recovery. It covers PSK as well as CPM.

I enjoyed watching the AMSAT-UK Colloquium
on the BATC site this weekend. It has been many
years since I was last on the satellites. There is only
some much I can do at once.

I am also trying to figure out how best to mount
the QS1R I have into a 19inch rack without
damaging it. I have a number of similar units that
all have the same problem in that there is no easy
way of mounting them.

Thinking about writing a small
sound-card application to implement CPM.
In theory it should significantly outperform
PSK31 if I manage to implement it correctly!
Thinking of doing something based on
LRC H=2, M=2, L=3. Will make a good
training exercise either way.

Still working on the PCB.....
Reading about the performance of CPM
in a fading channel. So far it has only been
analysed in a Gaussian channel in the book.

Reminisced a bit about my short time at the BBC
after finding http://www.bbceng.info/

Not radio related but I was saddened to hear
of the death of Henry Allingham. I never met
the man but I feel I have lost yet another link to
my own past. My late maternal grandfather
served on HMS Barham as a Lieutenant
Instructor (Navigation) during the battle of
Jutland. A few months ago the last of my grandfathers
maths scholarship students, Prof Jack Good,
who worked for Alan Turing at Bletchley Park
also passed away.

RIP Mr Allingham.

I have now got to the section in my book that
describes various transmitter layouts.
I was going to use direct generation of the CPM
signal using a DDS. The book suggests an alternative
technique of applying the CPM to the PLL.
This would allow the use of a cheaper DDS.
However there are the effects of the PLL filter to
take into account. The books says the phase error
will be acceptable. We shall see. I can try both ideas.
I feel a bit stupid not thinking of this myself.

What I am a bit worried about are the modulation
sidebands of the CPM. The technique I am using will
not allow any form of pulse shaping or filtering.

I had a look at the local DSTAR repeater spectrum
and the sidebands are not as well suppressed as
I thought they might be.

SMD parts arrived from Digikey this morning.
I have done some test soldering and it appears
that I can hand solder 0402 parts. It seems
essential to use solder paste to hold the part in
place. A quick search of the internet seems to
show that the best way to do it is to use a cheap
hot plate. I will try hand soldering though.

Now I know why I gave up hardware and moved
to software!

I found an interesting section in my book about CPM.
"Digital Phase Modulation" ISBN 0-306-42195-X

quote

"It has been verified that it is possible to have
constant envelope digital modulation systems
which are more power efficient than QPSK and MSK,
for example. In some cases the power gain is
considerable; complex block or convolutional would
be needed to achieve it with traditional memoryless
modulation methods, and their check symbols would
increase bandwidth. In the schemes presented here,
the code is built into the modulator itself. The
redundancy of the transmitted signal lies in its tree
structure, and not in the insertion of extra check
symbols, and thus all the schemes have an information
rate that is larger or equal to one. That is of great
importance to the bandwidth occupancy of the
transmitted signal."

Still I have not got to the part that tells you how difficult
it is to obtain carrier, symbol and modulation index (h)
synchronisation yet!

MIL-STD 188-181C cheats by transmitting an MSK preamble
which is much easier to obtain the information from.

Still if something is easy to do then it is not worth doing
I suppose.

I have ordered the SMD parts I require to do the
IF board from Digikey in the U.S. Provided I
spend more than £50 the carriage is free.
I just have the VAT to pay on receipt with
no handling charge.

I have delayed submitting the PCB because I found
a clearance problem with the board. I had not left
enough clearance between it and the Saxo FPGA board
that I am using to interface between the P.C and the
ADC chip http://tiny.cc/XFYgu

I am currently working my way through the maths
associated with CPM using the book
"Digital Phase Modulation" ISBN 0-306-42195-X

The book was bought 2nd hand and appears to have had
two previous owners RACAL-MILGO and later DeVry
University. Judging by the library insert in the book
it has only ever been read by two people.
Considering it is a classic on this type of
modulation that seems weird.

The problem with maths of course is that you just can't
start in the middle of things. All maths depends on
other maths! So I am stretching back to stuff I last did
30 years ago. Learning Calculus the 2nd time round
is not as difficult as it was the first time. I never did any
sampling theory at University so some of this stuff is new.

The reason BSON didn't work too well was that
the PLL noise was too great for constellations
beyond 16 points which limited the video to about
1 frame per second. Also the webcam capture
used VFW (video for windows) which puts a heavy
load on the P.C and does not want to work on Vista.

Hopefully using CPM will allow me to increase the
data rate in a 25 KHz channel significantly.

Also I hope to find something better than VFW which
can be used on Linux as well as Windows.

I am also really interested in using QT4 to do the GUI.

I am really an ALGO person rather than a GUI one.
My idea of fashion is beige so my GUIs stink!

I worked on a commercial project a while back and after
the graphics artists waved their wands on the program
it looked oh so much better!!

I have decided to restart my blogging as Tweets
are too small. I tend to never finish a project so
if I tell people what I am doing I will look stupid
if I don't finish it!

The current major project is the development
of an experimental digital radio that will allow
moving video (and other things) using a 70cms
radio on a standard narrowband FM channel.

I have chosen to use full response multi-h CPM
generated by a DDS chip. The waveforms I have
chosen to experiment with are the
MIL-STD 188-181C ones. These are used
on UHF ~250 MHz military satellites,
so are proven technology.

For the prototype radio I have decided to use
a modified T7F packet radio. I will replace the
synthesizer with a DDS so I can have accurate
control of the transmitted phase.
On receive rather than using the hardware
discriminator I will use an A/D converter and
will do the demodulation of the 450 KHz IF
in software.

The demodulator will use the Viterbi algorithm
to estimate the received symbols. This is standard
practice when demodulating CPM.

CPM is a constant envelope modulation so can be
used with class C amplifiers and with receivers
that have limiters in their IF strips.

The approach I have chosen will be scalable to
other frequencies and data rates.


I am going to develop an application I wrote
last year called BSON (Be Seen On Air).
This application reads video from a webcam
compresses it using the BBC's Dirac codec
http://diracvideo.org/ and then transmits it.
Currently BSON uses QAM to transmit the data
but has only been partially successful.

Audio is transmitted using the Speex open
source codec. Speex is not really suitable for
this type of non internet application.

I will log my progress (or lack of it) in this blog
over the next year or so. When I have a basic
system functional I will invite others to help if they
want to.

I don't want to waste peoples time on something
that might never fly!