[Skimmertalk] Using Skimmer with RTL-Dongle SDR?

[Douglas Theriault]

As Pete N4ZR mentioned Rick N1GP and I built up a couple skimmers using 
7 RTL SDR Dongles several years ago.  We wanted to build up a small node
for CW/RTTY skimming across 7 bands.  Rick blazed the path getting the
skimmers to work with the sticks and his initial write up n the rtl-sdr
blog.

We selected an oDroid C1 to run the rtlhpsdr software after
experimenting with several variants of  embedded boards.  There is a
picture of a board set I built for HF skimming using 7 sticks on my QRZ
page near the bottom.  We wanted to find an embedded board that had USB
2.0 speeds and 1Gb Ethernet.  Finding a board that would support 7
sticks was a challenge but the oDroid C1 in the end worked well for us.

At the time we could not find RTL sticks with 1ppm xtals or tcxo
options.  The sticks we had seemed very sensitive to temperature
swings.  When spotting I was drifting > 200Hz and struggled with this;
we even tried to compensate dynamically in software but this didn't work.
When skimming the sticks seems to run hot; Rick had chipset unsolder on
him in one instance. So keeping the sticks cool but stable was fun.

For my setup I ended up building a PLL fed by a 10Mhz tcxo to generate
the 28.8mhz clocks for each stick.  I modified each stick removing the
xtal and feeding a clock from the PLL.  The PLL also generated a stable
125mhz clock for an up converter mixer to move HF spectrum to the 120mhz
to 155mhz frequency the RTL sticks could receive at.  The PLL effort
stabilized my HF spots.  With the higher quality sticks now available
folks may have better luck without going down the route I took.

Next issue I struggled with were aliasing errors, especially on 10m
band.  I experimented with various FIR tables to see if I could filter
them out but in the end did not have good luck.  If the tables were too
long I could run only 2-4 sticks before dropping packets. 

These alias errors seemed to be worse when there were strong signals. 
This may have been due to my HF up converter mixer design having too
much gain causing overload on the sticks.  I struggled with my design to
keep it flat across the HF band. 

As I remember, there is a IF stage in the RTL sticks and am not sure if
the errors I experienced were due to something in the design of the
chipset or not.  I was/am new to SDRs not having detailed documentation
on the chipset led to more questions than answers.  Still was a very fun
project and I learned a heck of a lot about SDR's.

At the time the HL v1.2 project caught our interest and became the focus
of our skimming pursuits.  For HF the HL boards were fantastic option
for skimming.

I often thought it would be nice to try the RTL sticks for VHF/UHF
skimming especially since you do not need an up-converter mixer board. 
Weak signal beacon monitoring would be great.  Unfortunately at my QTH I
am limited on what I can throw up for antenna's and have not pursued at
this time.

I have a repo on github as well for the project Rick and I collaborated
on. 

   https://github.com/dtheriault/hydra.git

As John found out there is a config file for the software, rtl_hpsdr.conf 
which defines the IP address for the skimmer instance and several other 
configuration options.

The copy on my github repo has comments which describe the config
fields.  In addition there is a wiki that kept track of notes/pictures
we exchanged during the bring up of the 7 stick HF skimmer. 

Feel free to contact me directly if there's any questions on the
software.  Its been a while but I'm sure Rick or myself can help answer
questions you may have.

Apologize in advance if this email started a new thread, I just joined
the list this morning in order to  post my comment.

-- 

73, Doug - NO1D