[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Symbol rates. USRP2 vs USRP1

From: Valerio, Danilo
Subject: Re: [Discuss-gnuradio] Symbol rates. USRP2 vs USRP1
Date: Thu, 23 Apr 2009 10:25:34 +0200
User-agent: KMail/1.11.2 (Linux/2.6.27-14-generic; KDE/4.2.2; x86_64; ; )

Hi Colby!

In the RX path the difference in the adc rate should be compensated by the fact that you multiplied the decim by a factor of 1.5.
[In your bbn_80211b_rx.py (line 98)]

In the TX path Andrea figured out that:

Rate = 2 * [FPGA_processing_rate / (spb*interp_rate)]

The BBN guys suggested to use spb=4 and interp_rate=32 for communicating with a wifichipset.
In fact, with USRP1:
2 * [64 MS/s / (4 S/bit * 32)] = 1 Mbps

On the contrary, in USRP2 the FPGA_processing_rate is 100 MS/s.
Therefore, in order to transmit at 1 Mbps, we need to set the (spb*interp_rate) accordingly:

2 * [100 MS/s / (4 S/bit * 48)] = 1Mbps.

We were wondering how it is possible that spb=4 can represent a barker sequence.
We have tried some different combinations of spb and interp_rate (e.g. spb=8 and interp=24), so that the result is always 1 Mbps.
But we still have no success.
By "no success", I mean that the wifi chipset monitoring the air receives frames that seem to be randomly generated.

Any suggestions?
Anyway, thanks for uploading your branch. ;-)


On Wednesday 22 April 2009 22:24:19 Colby Boyer wrote:
> Hi all,
> As some of you know, I am working on porting the BBN 802.11b code to the
> USRP2. I understand that the ADC/DAC rates are higher than the USRP1. What
> are the effects, if any, on the rate at which symbol are sent through the
> air? I am suspect that this could be the reason I cannot decode sent
> packets.
> Colby

reply via email to

[Prev in Thread] Current Thread [Next in Thread]