Page 1 of 1

Sine wave transmission

Posted: Thu Aug 06, 2015 8:27 am
by Nrekcah
Dear all,

I sent a sine wave using bladeRF board with the following parameters:

Tx board:
Frequency = 2.55 GHz
Samplerate = 5 MHz
Bandwidth = 28 MHz
TxVGA1 = -16 dB
TxVGA2 = 9 dB

Rx board:
Frequency = 2.55 GHz
Samplerate = 5 MHz
Bandwidth = 28 MHz
RxVGA1 = 12 dB
RxVGA2 = 0 dB

For transmission, I use a CSV files and tx and rx commands.

Sine waves at the input of DAC and at the output of the ADC are presented in this figure: http://www.hostingpics.net/viewer.php?i ... ission.png

The problem is that the received sine wave (at the output of the ADC) has a period that increases with time.
Does any one know the origin of this problem ?

Thank you in advance.
Nrekcah.

Re: Sine wave transmission

Posted: Fri Aug 07, 2015 3:53 pm
by bpadalino
Are you sure it isn't just a slight frequency offset from each other?

If you transmit just a tone (vector of 1+0j) what is the resulting frequency on the other board?

Re: Sine wave transmission

Posted: Mon Aug 10, 2015 1:45 am
by Nrekcah
Thank you Brian for the reply,

But how can I measure the resulting frequency if I transmit a vector of (1+0j) ?

Thank you in advance.

Re: Sine wave transmission

Posted: Mon Aug 10, 2015 6:11 am
by bpadalino
There will inherently be a frequency offset between the boards. Since one board is sending just a CW, and the other board is receiving CW+(f_off), the resulting baseband signal will be f_off.

Does that make sense?

Re: Sine wave transmission

Posted: Mon Aug 10, 2015 7:39 am
by Nrekcah
Yes Brian it appears that it is a frequency offset and especially a sampling frequency offset.
The offset in this case appears to be large.
Is it a method to avoid this problem ?

Re: Sine wave transmission

Posted: Mon Aug 10, 2015 8:44 am
by bpadalino
How large is large, and what is your center frequency?

Setting the trimdac to trim the VCTCXO to be very close to each other can make the difference much smaller.

Re: Sine wave transmission

Posted: Mon Aug 10, 2015 9:55 am
by Nrekcah
After 300 samples, the offset reaches more than 5 samples (the received sine wave is delayed by 5 samples from the transmitted one after 300 samples).
Yes, by setting the trimdac to be very close to each other, the difference becomes much smaller, but the offset persists.
I think that a digital algorithm must be implemented to correct the residual offset.
Did you know a digital algorithm for this purpose ?

Re: Sine wave transmission

Posted: Tue Aug 11, 2015 6:18 am
by bpadalino
Yes - you're looking for a PLL for synchronization. The basic technique you're looking for is Carrier Recovery. There are a whole slew of techniques that you can use depending on your system and what performance characteristics you need. I think entire books have been written about synchronizing wireless systems - it's a very common problem with many solutions.

Re: Sine wave transmission

Posted: Wed Aug 12, 2015 2:39 am
by Nrekcah
Ok thank you Brian.