Still no cigar

The SI tables are still not working. I have a set of SI tables provided by
Rob M0DTS and they work.

I have code that provides identical tables, same contents, checksums etc
and they don't work. Obviously I am missing something. I will figure it
out eventually.

On the theme of GUIs I have decided to use QT4 developed by Trolltech
(now owned by Nokia). http://qt.nokia.com/ I am setting up the
development environment as I type so more on that later.

I have also been looking at all the DVB applications available for Linux.
I just wish I knew what they all did. Hopefully down the line it will be possible
to add receive support to my software using a suitable Linux tool.
I will have to do the demodulator but once I have the program stream extracted
it should be possible to feed that to a 3rd party tool for display.

I am ramping up my enthusiasm for the new UHD (Universal Hardware Driver)
from Ettus Research http://www.ettus.com/

This will allow cross platform support for their products. It will also allow closer
integration between my code and the USRP2. Currently no support for the WBX
transceiver board I use exists but it is coming. So watch this space.

I am always keen to hear from anyone reading this Blog, if you have any
comments, criticisms, advice or questions please contact me, I don't byte.

I am getting there

Today I have a picture of the Tonna 24 cms 23Ele Yagi
that I am going to use for TV transmission. I know some
of you have seen these before but some haven't.

Minor success to report today, I finally managed to get
my code to read Elementary stream packets from the
Hauppauge PVR card, turn them into transport stream
packets, then do all the FEC etc and transmit the result
using the USRP2.

The waveform you can just about see on the Spectrum
analyser under the antenna is the resulting 4 Msps
DVB-S signal. It is a lot more stable than it was before.
The CPU load to do all of this was 50% on one of the
cores. The others are doing virtually nothing.

I still have to fix some problems with the SI packets I am
sending but I am pleased with how things are going.

I am starting to think about a GUI for it :-)

Me and my DATV system

This is me sitting in front of the camera!
This is being sent at 2Msps.

This is the DATV system located at the end of the workshop.

Visible is the Linux monitor, the TV monitor, camera and STB.
Below the STB is the DATV rack with the exciter and Linux machine.

Work progresses

I have ordered a load of aluminium angle
to build boxes to go around the GH Engineering
PA amplifiers. I will post some pictures
when the job is done.

USRP2 Topless photo!

I am sure you all want to see what my USRP2 looks like with it's top off!

Things are going a bit better now, having changed the way my
software talks to the USRP2 I am now getting a much more reliable
DATV transmission. I still have a load of code to write but it looks
as if the whole thing is going to work (a huge relief).

The last couple of days I have received some 18 GHz SMA relays
from the U.S and some bulkhead N to semi-rigid cable units from
the far east as well as a waterproof plastic box for the preamp.
So on the hardware side all that is missing is the PA module which
should be here soon.

G4DDK 24cms preamp

As promised I have uploaded a picture. This is of the G4DDK ultra
low noise preamp I built for use on DATV.

Work on sorting out the drop outs of the DATV signal progresses.
After some testing it seems the Ethernet link between my P.C
and the USRP2 is fast enough to sustain the DATV signal so
there is hope that I will be able to get the whole to work reliably.

More DVB

I spent far more time that I meant to on
DATV S/W development yesterday.
I have started to integrate some of the MPEG2
transport stream code into my DVB-S modulator.

One problem is keeping the modulator fed with the
correct number of bits. The original S/W I used
required calculation of the various stream bit rates
so that the correct number of null packets can be
added to pad it out to the wanted transmission bit rate.

To automate the process I have added the ability
for my code to monitor it's transmit queue length.
When the queue length falls below a certain threshold
it will now inject null transport packets into the
transmitted bit stream to pad the bit stream out to
the wanted transmit bit rate.

Gradually I will add more of the required packets
into the bit stream from my MPEG2 code and drop
the corresponding packet types from the original
set-up.

I still have a problem though, the link between
the USRP2 and the P.C keeps dropping out for
a fraction of a second which causes the
transmission to momentarily break up.

What I am trying to do is well within the capabilities
of the USRP2 so I suspect I have not set my Linux
box up correctly to deal with the Ethernet interface.
Hopefully I can get this sorted out otherwise I am
going to have a very very major problem.

I know this Blog is getting a bit dull without any
pictures but most of the hardware I am using is
standard off the shelf stuff that can be found using
Google but I will try and put up some pictures later
in the week. I have not figured out how to take sexy
pictures of software yet (suggestions on a post card).

DVB-S Encoder

Not much done since my last blog.
I still need to figure out the best way
to do the sample rate conversion.

Tested the code puncturing modes in
DVB-S all of them worked except the
7/8 rate code. The comments in the
code and the code matched, unfortunately
the puncturing mask in the spec didn't!
Now fixed (both code and comments).
The code puncturing is used in -S and -T.

Spent some time looking for a suitable
waterproof box for the mast head preamp.
All the ones I could find in my workshop
had the wrong type of holes in them.