From: George Nychis <address@hidden>
To: Bill Stevenson <address@hidden>
Cc: address@hidden
Sent: Thursday, November 20, 2008 8:17:17 PM
Subject: Re: [Discuss-gnuradio] does USRP need some warm up before it gets into its normal state?
Bill Stevenson wrote:
> Hello, all!
>
> I am using demo file
benchmark_tx.py and
benchmark_rx.py to test the average PER at a certain distance. Whenever I initiated a test, the first two or three PER varied violently. For example, the
expected PER is around 4%, in the first three tests, PER might be 30% and after the first three tests, each PER in each test would change within a constant range, say 4%. Does USRP need some warm up before it goes to its stable state?
>
It's a known issue with the USRP that there is a power spike when initializing. I forget the exact cause of this, Matt would probably know. I would throw away the first set of samples.
- George
Thank you George! Do you know how many samples I should throw away because sometimes it appeares to be 2 samples, while other times it appears to be 10 samples? Could Matt tell me what the size of the first set would be? Thank you!