No samples from Generic RTL2832U

[airikka]

Hi,

I have two different RTL2832U USB sticks, but I can't get any output from any of them. I have tried on two different machines, one with debian jessie and one with debian stretch.
I get
/////////////////
$rtl_entropy
No configuration file found, continuing with defaults.
Options parsed, continuing.
Found 1 device(s):
0: Generic RTL2832U
Using device 0: Generic RTL2832U
Detached kernel driver
Found Elonics E4000 tuner
Setting Frequency to 70000000
Your device is capable of gains at...
: -1.00
: 1.50
: 4.00
: 6.50
: 9.00
: 11.50
: 14.00
: 16.50
: 19.00
: 21.50
: 24.00
: 29.00
: 34.00
: 42.00
Setting gain to 42.00
Doing FIPS init
Reading samples in sync mode...
.....and here it seems to hang forever.....
///////////////////
With the other stick
/////////////
$rtl_entropy
No configuration file found, continuing with defaults.
Options parsed, continuing.
Found 1 device(s):
0: Generic RTL2832U OEM
Using device 0: Generic RTL2832U OEM
Detached kernel driver
Found Rafael Micro R820T tuner
[R82XX] PLL not locked!
Setting Frequency to 70000000
Your device is capable of gains at...
: 0.00
: 0.90
: 1.40
: 2.70
: 3.70
: 7.70
: 8.70
: 12.50
: 14.40
: 15.70
: 16.60
: 19.70
: 20.70
: 22.90
: 25.40
: 28.00
: 29.70
: 32.80
: 33.80
: 36.40
: 37.20
: 38.60
: 40.20
: 42.10
: 43.40
: 43.90
: 44.50
: 48.00
: 49.60
Setting gain to 49.60
Doing FIPS init
Reading samples in sync mode...
.....and here it seems to hang forever.....
///////////////////

But when I run rtl_fm on a frequency, I get a lot of output.

Do you have any idea?

Thanks!

[airikka]

I have begun to debug this problem.
What I have seen so far is that the buffcounter, which should reach BUFFER_SIZE (2500), is increasing very slowly. It took about 13-14 minutes to reach 400.

[airikka]

I changed the 75ohm load to an antenna and this time it takes about 1 minute before i have reached the 2500 BUFFER_SIZE. But now i get

Failed: FIPS 140-2(2001-10-10) Monobit
Failed: FIPS 140-2(2001-10-10) Poker
Failed: FIPS 140-2(2001-10-10) Runs
Failed: FIPS 140-2(2001-10-10) Long run

all the time.

[pwarren]

Hi there!

The idea is that rtl-entropy only outputs useful entropy, so it filters things by various methods and tests it with the FIPS suite of entropy tests before writing out. If the dongle isn't receiving much noise, it doesn't output much!

After adding an antenna, the next thing to try would be mucking about with the gain settings. I have some dongles that only work well on their maximum gain setting, and others that only work well with 0 gain, so try that and see how you go!

[airikka]

I'm again using a 75 ohm load on the antenna connector and put the dongle in a shielded box. When using a lower sample rate, 2.8M, it fills the buffer much much faster. But I never get passed the FIPS tests. I have tried all valid gain settings on differents frequencies, but FIPS fails all the time.
Any suggestions?

Thanks!

[pwarren]

Next step would be to stick an antenna on and see if it gets anything. If it does it may be that your dongle's amplifiers are too quiet for the software to pick up much noise!

As an aside, I'd recommend a ChaosKey or a OneRNG if you need useful entropy!

[webernetz]

Hello guys,
I am currently testing with the same setup: DVB-T stick with 75 ohm load. (See a picture here: https://twitter.com/webernetz/status/809489705247133696 )

--> But I have the same issue as airikka: Nothing passes the FIPS tests. Do you have an idea why this happens? To my mind, the amplifier from the stick should generate some noise. (When I am plugging in the antenna again, much more random data comes in...)