Friday, February 20, 2009

Ticking the last box: Glonass L1

Hello!
We had to wait long enough for some RF filters. Half a hour after receiving them, the Maxim MAX2769 which is fitted on our NSL USB dongle tuned correctly on Glonass frequency. I had to quickly make up a bit of Matlab code, and off I went acquiring Russian birds over Nottingham skies.
Result: 8 satellites. Pictures attached.

Fig. 1: The signal characteristics show some interference peak, and the gain should to be increased a bit.


Fig. 2: The typical histogram of acquired channels. This time is Glonass.


Some details of the channels:


Fig. 3: Details about the individual channels.


I hope the MAX2769 has no secrets any more :)


The next post will be something new, about a new exciting Software Defined Radio GNSS receiver!

Michele

8 comments:

Anonymous said...

Hi,

did you use the MAX2769 in wideband
mode ( LPF ) or let it track each
FDMA channel sequentially ?

Andre

Michele said...

Andre,

I i used the MAX2769 in wideband (LPF) mode.

Cheers,
Michele

Unknown said...

Just curious, what crystal reference and synthesizer frequencies were you using to detect glonass with the MAX2769?

Michele said...

Hi Kyph,

don't remember very much actually! I am pretty sure that I used a 16.368 MHz TCXO reference and that channel 0 was centered at 6.12MHz.

Cheers,
Michele

Unknown said...

Hi, thanks for your quick reply. Also thanks for the blog, I kept reading more recent posts and found it very interesting.

Did you find out where that interference in glonass came from? I was surprised to see a strong peak in this band (after all it doesn't take much power to blind a gnss receiver)

Michele said...

Hello Kyph,

sorry I totally confused the data set! The one I posted here on the blog is certainly acquired using I&Q configuration of Primo, so the Glonass CH0 is likely to be in baseband or so. You can see that by the shape of the filter. That peak looks really like some armonic of clock, but I really cannot guarantee as it has been some time ago I wrote this.

Cheers,
Michele

Unknown said...

Hello again.

Yeah, I see now that probably explains the spurs. If f0 is 1602MHz there would be one harmonic at +2MHz and another at -14.3MHz so you're probably right.

I have another question. When measuring the front end with wideband lowpass configuration do you remember having any problems? I've been playing with a MAX2769 these days and I've measured 3dB bandwidths well above 15MHz, while according to the datasheet its nominal BW should be 9MHz (double side 18MHz = 2*9MHz), but when I asked Maxim I got no answer.

Michele said...

Hi Kyph,

Well done.
As you can see from the spectrum, the bandwidth does indeed look wider than 9MHz SSB. If I'd have a little spare time I could connect the MAX2769 to the spectrum analyser and measure accurately the 3dB bandwidths, but I cannot guarantee now as it is not a high-priority issue for us now. But I promise I'll keep you posted if I get round doing the test.

Cheers,
Michele