The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> spectrum fluxerating
https://designers-guide.org/forum/YaBB.pl?num=1213330654

Message started by iewgnem on Jun 12^th, 2008, 9:17pm

Title: spectrum fluxerating
Post by iewgnem on Jun 12^th, 2008, 9:17pm

I'm new on this forum, not sure how much traffic it gets but hopefully theres enough :P

anyway, I'm trying to digitize a single tone 315mhz signal, currently i'm running a signal chain like this:

Antenna -> SAW -> LNA -> SAW -> VGA (315 MHz) -> ADC (1 to 50MSPS) -> FPGA [ FIFO -> SERIAL] -> MATLAB

as you can see I'm undersampling the signal directly with a high analog bandwidth ADC, since its single tone undersampling seem much more economical. the signal goes into a FIFO in the FPGA and gets batch serialized. The sampling clock are sent through the FPGA to the ADC chip, its generated by a NCO with output converted to a clock signal and distributed using a global clock line.

I'm using a pretty basic transmitter (Linx 315MHz), its OOK stuck at 1.

Anyway, after each sample, I do a FFT on it and see where the peak is. Since the signal frequency is 315MHz and I'm sampling at a known frequency I could calculate where the peak should be. The problem is, the peak are not at where they are supposed to be, for example, when I'm sampling 315MHz at 20MHz, with aliasing and all there should be a peak at 5MHz, but its not, there are several other peaks but none at 5. futhermore, the peaks varies very slightly between each session, drifting higher very slowly and seemingly following a real world time rather than session count. I tried many different sampling frequencies and the peak never showed up at predicted places, and also seem to drift around unpredictbly, sometime even disapear, or multiple peaks appear and none at predicted frequencies. if I shutdown the transmitter, all frequencies disapear to be replaced by white noise, so I'm pretty sure signal is being received.

So that's the problem, what do you guys think? signals are being received but just not at the right frequencies, there seem to be no reason for the slow drifiting.

Title: Re: spectrum fluxerating
Post by pancho_hideboo on Jun 13^th, 2008, 10:22pm

iewgnem wrote on Jun 12^th, 2008, 9:17pm:

I'm using a pretty basic transmitter (Linx 315MHz), its OOK stuck at 1.
Anyway, after each sample, I do a FFT on it and see where the peak is.
Since the signal frequency is 315MHz and I'm sampling at a known frequency I could calculate where the peak should be.
The problem is, the peak are not at where they are supposed to be, for example, when I'm sampling 315MHz at 20MHz,
with aliasing and all there should be a peak at 5MHz, but its not, there are several other peaks but none at 5.
futhermore, the peaks varies very slightly between each session, drifting higher very slowly and seemingly following a real world time rather than session count.
I tried many different sampling frequencies and the peak never showed up at predicted places,
and also seem to drift around unpredictbly, sometime even disapear, or multiple peaks appear and none at predicted frequencies.

Show me your spectrums with your sampling time and capture time.
The aboves are quite natural results since you observe modulated signals.
Do you care about a length of capture time in FFT ?
You have to set capture time as multiples of 1.0/5MHz, 1.0/2.5MHz or 1.0/0.5MHz.