dabhand.sourceforge.net software
The only software that's here currently, is stuff which I've written to get the device working for study purposes. This is for experienced developers to use, mainly because it's currently no use to end users.
Important point to note: All of the analysis utilities here contain sampling parameters that I've made up given my very limited understanding of what the hardware is producing.
All of this is in CVS, check out the dabhand module if you want to have a play. Let me show you what I've got...
dabmate-detect.C
This scans the USB bus for the DABmate module, and dumps out a few parameters. e.g.
DABmate found. Manufacturer: dabMate Product: USB Digital Receiver Serial no.: 0001-80000000 Device has 3 settings. Green LED should flash for 5 seconds. Done.
dabmate-bulkread.C
This configures the DABmate to a particular frequency, and reads the captured signal using USB BULK READ requests. e.g.
./dabmate-bulkread 5d 02 > DUMP.bbc
It runs forever, so you'll want to interrupt it after 10 seconds or so, otherwise your disk will fill. At the moment, the captured data isn't likely to be any use to you. It can be fed to the analysis programs below.
Note: I suspect this method of reading data to be lossy. The dabmate-iso approach uses isochronous transfers, and seems to read more data.
Here's a table of frequncies which I've derived from a USB analyser.
| Band | Frequency / kHz | Value |
|---|---|---|
| 5a | 174928 | d4 d0 |
| 5b | 176640 | 80 d2 |
| 5c | 178352 | 2c d4 |
| 5d | 180064 | d8 d5 |
| 6a | 181936 | ac d7 |
| 6b | 183648 | 58 d9 |
| 6c | 185360 | 0d db |
| 6d | 187072 | ?? ?? |
| 7a | 188928 | ac d7 |
| 7b | 190640 | 58 d9 |
| 7c | 192352 | 0d db |
| 7d | 194064 | 80 de |
| Band | Frequency / kHz | Value |
|---|---|---|
| 8a | 195936 | 2c e0 |
| 8b | 197648 | 84 e3 |
| 8c | 199360 | 58 e5 |
| 8d | 201072 | 04 e7 |
| 9a | 202928 | b0 e8 |
| 9b | 204640 | 5c ea |
| 9c | 206352 | 2c ec |
| 9d | 208064 | d8 ed |
| 10a | 209936 | 04 f3 |
| 10b | 211648 | b0 f4 |
| 10c | 213360 | 5c f6 |
| 10d | 215072 | 08 f8 |
| 10n | 210096 | 2c f3 |
| Band | Frequency / kHz | Value |
|---|---|---|
| 11a | 216928 | d8 f9 |
| 11b | 218640 | 84 fb |
| 11c | 220352 | 30 fd |
| 11d | 222064 | dc fe |
| 11n | 217088 | 00 fa |
| 12a | 223936 | b1 00? |
| 12b | 225648 | 5d 02? |
| 12c | 227630 | 09 04? |
| 12d | 229072 | ?? ?? |
| 12n | 224096 | d9 00 |
| 13a | 230784 | 61 07 |
| 13b | 242496 | 0d 09 |
| 13c | 234208 | b9 0a |
| 13d | 235776 | 41 0c |
| 13e | 237488 | ed 0d |
| 13f | 239200 | 99 0f |
dabmate-iso.C
Very similar to dabmate-bulkread, this configures the DABmate to a particular frequency, and reads the captured signal using USB isochronous transfers. e.g.
./dabmate-iso 5d 02 > DUMP.bbc
It runs forever, so you'll want to interrupt it after 10 seconds or so, otherwise your disk will fill. At the moment, the captured data isn't likely to be any use to you. It can be fed to the analysis programs below.
This utility reads more data than dabmate-bulkread, so I suspect this to be a more reliable way to read the data. The iso transfer method lends itself to better support for realtime capture devices such as this, so it's not too much of a surprise.
dabmate-iso.C uses some very bodged code to support the ISO transfers, so you may have more trouble getting it to compile on your system. The problem here is that I'm using libusb, which doesn't support ISO transfers, yet.
freq-plot.py
This is a Python script, and you run it with one argument, a dump file, as captured by dabmate-bulkread above. e.g.
./dabmate-bulkread DUMP.bbc
This script runs a Fourier transform on a snapshot of data from the capture file, and draw a plot on the screen like this:
The plot above is made from a real capture of a good DAB signal. The line which stands out at x=1169 is of interest. Closer analysis shows that this is likely to be the DAB central frequency. The DAB OFDM scheme does not transmit data on the central carrier, which causes it to stand out.
phase-plot.py
This is a Python script, and you run it with one argument, a dump file, as captured by dabmate-bulkread above.
This script runs Fourier analysis over a small period of time on the data, and plots a particular value of the frequency domain (x=real, y=complex). In other words, the distance from the centre (0,0) represents the amplitude of the signal, the angle around the centre represents the phase. e.g.
The plot above is made from a real capture of a good DAB signal. Since the signal is COFDM, each carrier is moduled using QAM. So, I'd expect to see four clumps of data in the scattergram. Unfortunately, that's not what I'm seeing at the moment.
single-frequency.py
This is a Python script, and you run it with one argument, a dump file, as captured by dabmate-bulkread above.
This script runs Fourier analysis over a small period of time on the data, and plots two values on the Y-axis against time on the X-axis. The two values plotted are the amplitude and phase of a particular value of the frequency domain.
The plot above is made from a real capture of a good DAB signal. The blue line is amplitude, and the green line is phase. I'd expect the blue line to be much squarer. The phase line should be closer to one of four phases, so I'd expect it to have much clearer transitions between four discrete values. The above plot has some structure but seems to have a strong random flavour.
ofdm-analyse.C
This needs the FFTW3 library, available at www.fftw.org.
This compiles to an executable which reads the standard input for data, a dump file, as captured by dabmate-bulkread above. e.g.
./ofdm-analyse < DUMP.bbc
This script runs Fourier analysis on regular time intervals of the dump data, and dumps out the value of the symbols at a range of frequencies (1165-1179) represented in ASCII on the standard output. Each value is represented as a string of up to 5 characters. The number of characters represents the amplitude, and the characters will be A, B, C or D depending on the phase quadrant each symbol appears in. Here's a snapshot of output:
1058:DDDD CCC AA AAAAA AAAAA BBBBB AAAA BBBB CCCC CCCCC AAAA BBB AAA AAAA AAAA
1059:CCC CCC AA AAAAA AAAAA AAAAA AAAA BBBB CCCC CCCCC CCCC BBB AAAA AAAA AAAA
1060:CCC CCC AAA BBBBB AAAAA AAAAA AAA BBBB AAAAA CCCCC DDDDD BBBB AAAAA AAA AAAAA
1061:CCC DDD AAA BBBBB AAAAA CCCCC DD BBB AAAAA CCCCC DDDDD BBBB AAAAA DDD BBBBB
1062:CCC DDD AAA DDDDD AAAAA CCCCC BBB BBB AAAAA CCCCC DDDDD DDDD AAAAA DDDD BBBBB
1063:AAA DDD AAA DDDDD AAAAA DDDDD AAA BBB BBB CCCCC BBBBB DDDD AAAAA DDDD BBBB
1064:AAA BBB AAA CCCCC AAAAA DDDDD AAA BBB DDD CCCCC BBBBB DDDD AAAAA DDD BBBB
1065:AAAA BBB BBBB CCCCC AAAAA BBBBB AA BBB DDDD CCCCC AAAAA CCCC AAAAA BBB BBBB
1066:AAAA BBBB BBBB AAAAA BBBBB BBBBB BBB BBB CCCC CCCCC AAAA CCCC AAAAA AAAA BBBB
1067:AAA BBBB BBB BBBBB BBBBB BBBBB BBB BBB CCCC CCCCC CCCCC AAA AAAAA AAAA BBBB
1068:CCC BBBB AAA BBBBB BBBBB AAAAA CC BBB AAAAA CCCCC CCCC AAA AAAAA CCCC DDDD
1069:CCC BBBB AAA DDDDD BBBBB AAAAA CCC BBBB AAAA CCCCC DDDD BBB AAAAA CCCCC DDDD
1070:DDDDD BBBB CCC DDDDD BBBBB CCCCC CCC BBBBB AAA CCCCC DDDD DDD AAAAA CCCCC DDDD
1071:DDDDD BBBB CCC CCCCC BBBBB CCCCC CCCC BBBBB DDDD CCCCC BBBB DDD AAAAA CCCC DDD
1072:DDDDD AAA CCC CCCCC BBBBB DDDDD AAAAA BBBBB DDDDD CCCCC BBBBB DDD AAAA DDDD D
1073:DDD AAA AA AAAAA BBBBB DDDDD AAAAA BBBBB DDDDD CCCCC AAAAA CCC AAA DDD A
1074:BBB AAA AAA AAAAA BBBBB BBBBB AAAA BBBBB DDD CCCCC AAAAA CCC AA DDD DD
1075:BBB AAAA AAAA BBBBB BBBBB BBBBB AAAAA BBBBB AAAA AAAAA CCCCC CCCC CC DDD DDD
1076:BBB AAAA AAAAA DDDDD BBBBB BBBBB AAAAA AAAA AAAAA AAAAA CCCCC CCCC DDD DD DDD
1077:BBB AAAA AAAA DDDDD BBBBB AAAAA BBBBB AAA BBBBB AAAAA CCCCC CCCC DDD C DDD
1078:BBB BBBB CCCC CCCCC BBBBB AAAAA BBBB CCC BBBBB AAAAA CCCC CCC DDD CCC DDD
1079:AAA BBBB CCCC CCCCC BBBBB CCCCC BBBBB CCCC BBBBB AAAAA BBB DDD DDD CCC DDD
1080:AAAA BBB CCC AAAAA BBBBB CCCCC BBBB CCCC DDDD CCCCC BBBB BBB DDD CCC DD
1081:AAAA BBB DDD AAAAA BBBBB DDDDD BBB DDD CCC CCCCC AAAAA BBB DDDD CCC AAA
1082:CCCC DDD BBBBB BBBBB DDDDD DDD DDD CCCC AAAAA AAAAA BBBB DDDD BB AAAA
1083:CCC CCC CCC BBBBB BBBBB BBBBB DDD DDD AAAA AAAAA CCCCC DDDD DDDDD BBBB AAAA
1084:AAA CCCC DDD DDDDD BBBBB BBBBB BBB DDD BBBBB AAAAA CCCCC DDDDD BBBBB BBBB AAAA
1085:AAA CCCC DDD CCCCC DDDDD BBBBB B CCC BBBBB AAAAA CCCCC CCCCC BBBBB AAAA BBBB
1086:AAA CCCC DDD CCCCC DDDDD AAAAA DDD CCCC BBBBB AAAAA DDDDD CCCCC BBBBB AAAAA BBBB
1087:AAA CCC BBBB AAAAA DDDDD AAAAA BBBBB CCCCC BBBBB AAAAA DDDDD CCCCC BBBBB AAAA BBBBB
1088:AA AAA BBBB AAAAA DDDDD CCCCC BBBBB CCCCC DDDD AAAAA BBBBB AAAAA BBBBB AAA DDDDD
1089:CC AAA DDDD BBBBB DDDDD CCCCC BBBBB CCCCC CCCCC AAAAA BBBBB AAAAA BBBBB BBBB DDDDD
1090:C AAA DDD BBBBB DDDDD DDDDD BBBB DDDD CCCCC AAAAA AAAA AAAA BBBBB BBBBB DDDDD
1091:BBB AAA DDD DDDDD DDDDD DDDDD DDD DDDD AAAA AAAAA AAAA BBBB BBBBB BBBBB DDDD
1092:BBBB CCC DDD DDDDD DDDDD BBBBB AAA DDDDD BBBBB AAAAA CCCC BBBB BBBBB BBBBB DDD
1093:BBBB CCCC DDDD CCCCC DDDDD BBBBB AAAAA DDDDD BBBBB AAAAA CCCCC BBB BBBB BBBBB CCC
1094:BBBB CCCC DDDD AAAAA DDDDD BBBBB AAAAA DDDDD BBBBB AAAAA DDDDD AAAA DDDD BBBB CCCC
1095:BBB CCCC DDDD AAAAA DDDDD AAAAA AAA DDDDD DDDDD AAAAA DDDDD AAAAA DDDD BBB CCCCC
1096:DDD CCCC DDD BBBBB DDDDD AAAAA BBB DDDDD DDDDD AAAAA DDDDD AAAAA BBBB DDDD CCCCC
1097:DDD CCC DDD BBBBB DDDDD CCCCC BBBB DDDDD CCCCC AAAAA BBB AAAA BBBBB DDDD CCCCC
1098:BB CCC CCC DDDDD DDDDD CCCCC DDDD DDDDD CCCCC AAAAA AAAA AAA BBBBB DDDDD CCCC
The columns represent the values of frequencies, and the rows represent time-incremental samples. The frequency range was chosen as it contains the frequency shown in the FFT plot from freq-plot.py. There's a strong pattern here, although I was hoping to see the frequency carrying a static pattern, rather than one which rotates.
If you live in a place where there is VERY strong DAB reception, I'd be interested in the results you get.