Forum: GNU Radio does USRP need some warm up before it gets into its normal state?

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
Bill S. (Guest)
on 2008-11-21 02:48
(Received via mailing list)
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?

Thanks a lot for all!!!

Bill
George N. (Guest)
on 2008-11-21 03:18
(Received via mailing list)
Bill S. 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
Bill S. (Guest)
on 2008-11-21 04:04
(Received via mailing list)
________________________________

From: George N. <removed_email_address@domain.invalid>
To: Bill S. <removed_email_address@domain.invalid>
Cc: removed_email_address@domain.invalid
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?


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!

Bill



Bill S. 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
George N. (Guest)
on 2008-11-21 04:05
(Received via mailing list)
Bill S. wrote:
>
> 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!
>

Just throw away your first 100 samples to be safe or something... it
doesn't really matter in the long run does it?  Same as if your USRP
took an extra microsecond or two to warm up and you missed them.

- George
Bill S. (Guest)
on 2008-11-21 04:32
(Received via mailing list)
________________________________

From: George N. <removed_email_address@domain.invalid>
To: Bill S. <removed_email_address@domain.invalid>
Cc: removed_email_address@domain.invalid
Sent: Thursday, November 20, 2008 9:04:34 PM
Subject: Re: [Discuss-gnuradio] does USRP need some warm up before it
gets into its normal state?


Thank you again, George! Did you mean that I just need to forget about
the packets transmitted in the first 10 or more microsecond by the
USRPs? Maybe I have a misunderstanding of initiation. Does initiation
mean every time I plug in the power cable of USRP to transmit or every
time I press the "enter" button to start the transmission once again
with power cable already on it?

Thank you!!!

Bill



Bill S. wrote:
>  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!
> 

Just throw away your first 100 samples to be safe or something... it
doesn't really matter in the long run does it?  Same as if your USRP
took an extra microsecond or two to warm up and you missed them.

- George
George N. (Guest)
on 2008-11-21 04:34
(Received via mailing list)
> Thank you again, George! Did you mean that I just need to forget about
> the packets transmitted in the first 10 or more microsecond by the
> USRPs? Maybe I have a misunderstanding of initiation. Does initiation
> mean every time I plug in the power cable of USRP to transmit or every
> time I press the "enter" button to start the transmission once again
> with power cable already on it?
>

Why don't you just try it and find out?  Start by thowing away 20
samples at start, and see if you see the spike.  If you do, try throwing
away 30 samples.  If you still see it, continue until you don't.

- George
This topic is locked and can not be replied to.