airspy_rx Data Type 4 does not seem to work and other suggestions

[blort]

airspy tool airspy_rx does not seem to work (data is all zeros) when type 4 (unsigned int 16-bit Real) samples are selected. Type 3 (signed int 16-bit Real) seems to work (data varies and is non-zero).

root@odroidXU:/dev/shm# airspy_rx -r myrawfile -f 104 -t 4 -a 1 -v 3 -m 5 -l 10 -d
airspy_rx v1.0.0 27 Nov 2014
frequency_MHz -f 104.000000MHz (104000000Hz)
sample_rate -a 1
sample_type -t 4
biast -b 0
vga_gain -v 3
mixer_gain -m 5
lna_gain -l 10
Device Serial Number: 0x16C463C8335CA9C7
Stop with Ctrl-C
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
^CCaught signal 2

User cancel, exiting…
Total time: 6.6347 s
Average speed 5.0000 MSPS Real
done
root@odroidXU:/dev/shm# od -x myrawfile
0000000 0000 0000 0000 0000 0000 0000 0000 0000
*
376000000
root@odroidXU:/dev/shm# ls -al myrawfile
-rw-r–r– 1 root root 66584576 Dec 3 08:43 myrawfile
root@odroidXU:/dev/shm# rm myrawfile
root@odroidXU:/dev/shm# airspy_rx -r myrawfile -f 104 -t 3 -a 1 -v 3 -m 5 -l 10 -d
airspy_rx v1.0.0 27 Nov 2014
frequency_MHz -f 104.000000MHz (104000000Hz)
sample_rate -a 1
sample_type -t 3
biast -b 0
vga_gain -v 3
mixer_gain -m 5
lna_gain -l 10
Device Serial Number: 0x16C463C8335CA9C7
Stop with Ctrl-C
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
Streaming at 5.00 MSPS
^CCaught signal 2

User cancel, exiting…
Total time: 3.6306 s
Average speed 4.9982 MSPS Real
done
root@odroidXU:/dev/shm# od -x myrawfile | more
0000000 ffb0 ff50 ff70 ffb0 ff60 ffd0 ff40 fef0
0000020 00c0 ff80 ff90 fee0 ff30 ffe0 ffa0 ffb0
0000040 ff80 ffb0 ffb0 ff30 ff90 ffa0 ff70 ffa0
0000060 ff90 ff80 ffb0 ff20 ffb0 ff80 ffa0 ff60
…

root@odroidXU:/dev/shm# airspy_info

Found AirSpy board 1
Board ID Number: 0 (AIRSPY)
Firmware Version: AirSpy NOS v0.9.9 Beta
Part ID Number: 0x6906002B 0x00000030
Serial Number: 0x16C463C8335CA9C7
Close board 1
root@odroidXU:/dev/shm#

The Odroid XU is a quad 1.6 Ghz Arm A15. Maximum processor load on one processor was 9%.

I got the same result on a radxa.

Using the USB 3.0 on the Odroid XU I recorded a perfectly good wav file using sample type 0 at 10 Mhz

root@odroidXU:/dev/shm# airspy_rx -w -f 104 -v 3 -m 5 -l 10 -d
Receive wav file: AirSpy_20141203_093020Z_104000kHz_IQ.wav
airspy_rx v1.0.0 27 Nov 2014
frequency_MHz -f 104.000000MHz (104000000Hz)
sample_rate -a 0
sample_type -t 2
biast -b 0
vga_gain -v 3
mixer_gain -m 5
lna_gain -l 10
Device Serial Number: 0x16C463C8335CA9C7
Stop with Ctrl-C
Streaming at 10.00 MSPS
Streaming at 10.02 MSPS
Streaming at 9.99 MSPS
Streaming at 10.01 MSPS
Streaming at 10.00 MSPS
Streaming at 9.90 MSPS
Streaming at 10.02 MSPS
Streaming at 10.01 MSPS
Streaming at 10.01 MSPS
Streaming at 10.00 MSPS
Streaming at 10.00 MSPS
^CCaught signal 2

User cancel, exiting…
Total time: 11.1642 s
Average speed 10.0032 MSPS IQ
done

I’m not sure I’m reading the MSPS “Streaming at 5.00 MSPS” correctly, but when the frequency is set to 10 Mhz, I thought the number of samples per second would remain at 10 Mhz – not decrease to 5 Mhz. Is this an error?

I think it would be useful to have airspy_rx report the number of bytes per second. For sample types 3 & 4 at 10 Mhz sampling rate, I would expect it to be 20 MBytes / second – anything less and I would suspect a performance issue. It might be nice to have airspy_rx tell the user when the actual sampling rate is less than the expected sampling rate. I’m sure you could figure that out without being told by the program, but newbies might benefit from that kind of helpful message. When the sample rate is messed up, what you hear sounds like the chipmonks! (https://www.youtube.com/watch?v=DbYYwGmi3_8)

Something to consider is enhancing airspy_rx or adding airspy_tcp to run over a gigabit ethernet interface to SDRsharp.

If you would like some help modifying airspy_rx.c I would be glad to volunteer. blort.blort AT yahoo.com

If you are taking votes for additional sampling rates, I vote for 8 Mhz and 4 Mhz. I have some Linux computers that top out at 8.2 Mhz and others at 5.5 Mhz when using sampling type 0 at 10 Mhz. So 8 and 4 Mhz would be a nice (Christmas??? (Dec 25) / new years??? (Jan 1) / ground hog day??? (Feb 2)) present.

I’m sure that you are as excited about this device as we are. This device has an amazing potential that we are all eager to make full use of.

Blort

Benjamin,

Would you like some help with airspy_rx or airspy_tcp?

Please contact me at blort.blort AT yahoo.com if you are interested.

Blort

[bvernoux]

Thanks blort for that amazing log
I will check it again with same test to check if it's not an issue with odroidXU
Also about airspy_tcp we are not planning to develop such tool (in short term as we are ultra busy on lot of things) but you are welcome to do it and we will add it in trunk

Best Regards
Benjamin

[touil]

Fixed

[blort]

Benjamin,
I've started on airspy_tcp.c.  FYI, I also tested the Airspy Type 4 data on a radxa which is slower than the XU but still is different.
Blort

 On Friday, December 5, 2014 8:00 PM, Youssef <notifications@github.com> wrote:

Fixed—
Reply to this email directly or view it on GitHub.

[bvernoux]

Blort it's fixed now check it again with latest trunk
Any feedback is welcome about airspy_tcp as lot of people will love such tool

Best regards
Benjamin

[blort]

Benjamin,
I tested with Type 4 data and it works great!
Blort

 On Sunday, December 7, 2014 3:27 AM, Benjamin Vernoux <notifications@github.com> wrote:

Blort it's fixed now check it again with latest trunk
Any feedback is welcome about airspy_tcp as lot of people will love such toolBest regards
Benjamin—
Reply to this email directly or view it on GitHub.