[srslte-users] AGC

[Mathias Rønholt Kielgast]

With regards to my last point, I found that I of course should not be looking at average power of the signals. With the PAPR with OFDM, average "digital power" should of course not be one. 
Instead, I have been looking at the maximum amplitude of the samples. This makes more sense, however, not entirely. The maximimum amplitude is usually around 2/3, and then once in a while around 1.1. Could you explain this behaveour?

Yours sincerely,
Mathias Rønholt Kielgast
Master's Student, Wireless Communication Systems
Aalborg University

________________________________________
From: Mathias Rønholt Kielgast
Sent: 31 March 2017 09:03
To: Ismael Gomez
Cc: srslte-users at lists.softwareradiosystems.com
Subject: RE: [srslte-users] AGC

Hi.

So far it looks like it to some extend works, albeit not for PRACH.
However, I have found that the "digital power" (found using srslte_vec_avg_power_cf()) is much lower than I'd expected. I had expected a digital power of one, but it is usually 0.087. How can that be?


Yours sincerely,
Mathias Rønholt Kielgast
Master's Student, Wireless Communication Systems
Aalborg University


From: Ismael Gomez [ismael.gomez at softwareradiosystems.com]

Sent: 28 March 2017 10:56

To: Mathias Rønholt Kielgast

Cc: srslte-users at lists.softwareradiosystems.com

Subject: Re: [srslte-users] AGC






The issue is low SNR and eventual problems with UHD drivers.



On Tue, 28 Mar 2017 at 10:51 Mathias Rønholt Kielgast <mkielg11 at student.aau.dk> wrote:





At least for the PRACH, it tries to estimate PL and set transmission power accordingly (as it should).
However, it seems it is consistenly set too low, my eNodeB (OAI) does not register the attempts. As I have not got any further I have not been able to see the performance of the rest. Perhaps I'll try to set PRACH gain to get further in the random access procedure.

Do you know what is the issue with the power control resulting in the bad performance?










Yours sincerely,
Mathias Rønholt Kielgast
Master's Student, Wireless Communication Systems
Aalborg University













From: Ismael Gomez [ismael.gomez at softwareradiosystems.com]
Sent: 28 March 2017 10:47
To: Mathias Rønholt Kielgast
Cc:
srslte-users at lists.softwareradiosystems.com
Subject: Re: [srslte-users] AGC










I'm not sure what happens now if tx_gain is not set. I think it tries to use UHD commands to set the gain and maybe use some digital gain control too, but in our experience the performance is not very good I think.


On Tue, 28 Mar 2017 at 10:44 Mathias Rønholt Kielgast <mkielg11 at student.aau.dk>
 wrote:


Hi,

Thanks for the response.
If the logic you talk about is not used, is it then not the power control applied when no tx gain is set in the config file? Is there additional logic not used at the moment, but developed?

Yours sincerely,
Mathias Rønholt Kielgast
Master's Student, Wireless Communication Systems
Aalborg University


From: Ismael Gomez [ismael.gomez at softwareradiosystems.com]
Sent: 27 March 2017 13:53
To: Mathias Rønholt Kielgast;
srslte-users at lists.softwareradiosystems.com
Subject: Re: [srslte-users] AGC

Hi Mathias,

The logic for PUSCH/PUCCH power control is there in the code but not used. With the general-purpose front-ends we typically use is a bit difficult to achieve 1 ms resolution power control. Let us know what is your experience if you try
 it.

Regards

On Mon, 27 Mar 2017 at 10:35 Mathias Rønholt Kielgast <mkielg11 at student.aau.dk>
 wrote:

Hi SRS,

You write in comments/output messages, that the TX AGC is not functioning properly.
Is it known in which element there is trouble and what is required to fix it? It would very much like to have AGC working in both RX/TX for some tests.

Yours sincerely,
Mathias Rønholt Kielgast
Master's Student, Wireless Communication Systems
Aalborg University

_______________________________________________
srslte-users mailing list
srslte-users at lists.softwareradiosystems.com
http://www.softwareradiosystems.com/mailman/listinfo/srslte-users