Signal Capture Differences and Modulation

[postman]

Hi there,

Attached are captures of an identical signal taken by two different devices. Device 1 is a bladeRF with a large monopole antenna that I do not know the exact specs of, and Device 2 is a RTL-SDR with the stock antenna.

Device 1 Spectrum Image: http://imgur.com/a/b7HkG#0
Device 2 Spectrum Image: http://imgur.com/a/b7HkG#1

One can see how they are related, but their characteristics are quite different. Can you think of why the spectrum looks so different, especially as it concerns the multiple frequency artifacts for the RTL-SDR and none for the bladeRF?

Additionally, do you have any thoughts as to what modulation may be in use here? The RTL-SDR suggests FSK, but I don't know what to make of the bladeRF capture...

Thanks!

[jynik]

Any chance you could share the raw captures? Playing around with the raw data a bit in Octave might help uncover some helpful details. (Perhaps take a gander at the real time-domain signal?)

You probably have far more expertise in this area than myself, but one thing I noted is that your FFT sizes are different between those two images...not sure if that would result in any major differences though.

For what it's worth, while screwing around one weekend, I happened to catch a signal similar looking to one in the "Device 1" picture. I believe my similar signal was ASK, and given the data I've decoded so far, I'm suspecting it's from an ISO 18000-6c UHF RFID reader (especially considering my proximity to one)...but I could be totaly wrong.

[postman]

Hey, thanks for the response. Sure, here is a raw capture: https://docs.google.com/file/d/0B9UP0pU ... VYazQ/edit. I don't have access to my BladeRF right now so this is with an RTL-SDR dongle and its generic antenna. The file parameters are:

samplerate 2.4e6
frequency 345e6
format: 8-bit unsigned quadrature/2 channels

I can get some better captures with the BladeRF tomorrow morning. Let me know if there are any details you want me to focus on and I can get you what you need. My end goal is to be able to replay these signals, and I figure knowing the modulation and the data in question is a good first step.

[sigblips]

See the spectrogram image below. The signal in your rtl_close_2.u8quand file is FSK with a 7300 baud rate. It's a carrier detect tone, followed by a preamble, and then the mark/space data bits.

rtl_close_2.u8quand_FSK.png

I noticed that the gain is too high and your signal is clipped. This causes distortion. The I and Q channels are the green and purple in the Waveform image below:

rtl_close_2.u8quand_waveform.png

It is a amazing that the FSK spectrogram image is so clear considering the clipping distortion.

Attached are captures of an identical signal taken by two different devices. Device 1 is a bladeRF with a large monopole antenna that I do not know the exact specs of, and Device 2 is a RTL-SDR with the stock antenna.

Device 1 Spectrum Image: http://imgur.com/a/b7HkG#0
Device 2 Spectrum Image: http://imgur.com/a/b7HkG#1

One can see how they are related, but their characteristics are quite different. Can you think of why the spectrum looks so different, especially as it concerns the multiple frequency artifacts for the RTL-SDR and none for the bladeRF?

You are comparing a 24 MSPS rate capture (bladeRF) with a 1 MSPS rate capture (RTL-SDR). That's a huge difference. Things will look very different at those scales.

[jynik]

You are comparing a 24 MSPS rate capture (bladeRF) with a 1 MSPS rate capture (RTL-SDR). That's a huge difference. Things will look very different at those scales.

That makes sense; sorry for the potentially irrelevant comment regarding the bladeRF capture.

Any chance we could get a bladeRF capture too? (In particular, one that's not saturated, and using a smaller bandwidth, if possible.) Even considering that sample rate, I'm still a little surprised the bladeRF capture didn't catch the change in frequency.

On IRC (#bladeRF), postman noted that the FCC test report for the device he's playing with specified the use of ASK modulation. However, it's very clear from the excellent description above that the RTL-SDR capture shows an FSK signal.

Is it possible that we're not even looking at two captures of the same thing? (Or that I'm just not looking at the right signal in the bladeRF capture ) In my comments above about the bladeRF capture, I realized was looking at the stronger signal (center freq +~1Mhz?). Any chance the FSK signal here is buried by the DC offset, or just a little tough to see due to the relatively large bandwidth of the bladeRF capture?

I see there's a weaker signal a few 10's of kHz above the center frequency in the bladeRF capture... I assume that's the FSK signal shown in the RTL-SDR capture? It seems like there's a constant tone for much of the capture...which would correspond with the carrier tone sigblips mentioned?

Perhaps try to calibrate out that DC offset a bit before getting a capture just to get rid of that distraction. If you suspect the bladeRF's frequency tuning is a bit off, it might be worth to double check the VCTCXO calibration and adjust it if needed...

Again... new to the RF world myself...take my comments/questions with a grain of salt. sigblips, please feel free to call me out on anything here that's partially/totally wrong.


As a side note, I know that some automatic DC offset correction is in the works. (I think bpadalino was or is planning to work on that...)