[srslte-users] Please help with srsENB DL problems!

[Ismael Gomez]

Can you send enb.log too? (log level info is enough)

On Mon, 30 Oct 2017 at 07:04 Ilshat Shaidullin <ilshat.desu at gmail.com>
wrote:

> I forgot to attach console trace in the last e-mail. Maybe it will help to
> figure out what's the problem.
>
> CASE 1 -->  UE is about 1-2 meters from USRP.
>
> When I test DL speed, I get this:
> ------DL-------------------UL----------------
> rnti   mcs   brate   bler  snr  phr  turbo  mcs  brate   bler
> 46     16   7.2M   0.0%    20    40   1.0    12   211k   0.0%
>
> When I test UL speed, I get this:
> ------DL-------------------UL----------------
> rnti   mcs   brate   bler  snr  phr  turbo  mcs  brate   bler
> 46    6.0   542k   0.0%    22    39   1.0    16   7.6M   0.0%
>
> Both DL and UL speeds are good.
>
>
> CASE 2 --> UE is further than 2 meters from USRP.
>
> When I test DL speed, I get this:
> ------DL-------------------UL----------------
> rnti   mcs   brate   bler  snr  phr  turbo  mcs  brate   bler
> 46     13   733k    46%    15    34   1.0    11   128k   0.0%
>
> When I test UL speed, I get this:
> ------DL-------------------UL----------------
> rnti   mcs   brate   bler  snr  phr  turbo  mcs  brate   bler
> 46    4.3   274k   1.7%    16    33   1.0    13   6.0M   0.0%
>
> UL speed is good (only small decrease compared to the previous case).
> DL speed is very low (large decrease of speed compared to the previous
> case) and BLER is high.
>
>
> Best,
> Ilshat
>
> 2017-10-30 14:02 GMT+03:00 Ilshat Shaidullin <ilshat.desu at gmail.com>:
>
>> Hello, Ismael!
>>
>> Thank you for replying me.
>>
>> I downloaded srsENB configuration files from repository, but nothing
>> changed. As only DL speed suffers when distance is larger than 2 meters, it
>> looks like that problem is caused by RS (Reference Signal) power. Maybe I
>> am wrong.
>>
>> I looked through the code and couldn't find out how pdsch-ConfigCommon
>> from SIB2 (information about RS power) is used.
>> I see that phy object is initialized with phy_cfg_t structure. But I
>> don't see how pdsch_cnfg from phy_cfg_t structure field is used.
>> In srslte (enb_dl.c, refsignal_dl.c) I couldn't find how RS power can be
>> changed.
>>
>>
>> Best,
>> Ilshat
>>
>> 2017-10-27 19:40 GMT+03:00 Ismael Gomez <
>> ismael.gomez at softwareradiosystems.com>:
>>
>>> Hi Ilshat,
>>>
>>> Make sure your reference signal power and pusch/pucch power settings are
>>> correct in sib.conf
>>>
>>> Regards
>>>
>>> El vie., 27 oct. 2017 9:33, Ilshat Shaidullin <ilshat.desu at gmail.com>
>>> escribió:
>>>
>>>> Hello!
>>>>
>>>> I have a question about 2 DL problems in srsENB. I use the following
>>>> configuration:
>>>> - USRP B210
>>>> - 25 PRB bandwidth
>>>> - Max DL MCS Limit 16.
>>>>
>>>> When phone is within about 1-2 meters from USRP, everything works
>>>> great. UE reaches 6.7 Mbps in DL and 4-5 Mbps in UL. But, if UE is further
>>>> than 2 meters from USRP, DL speed is very low (0.1-0.3 Mbps) although
>>>> signal strength is good (-77 dBm) and UL speed is high (4 Mbps). And if UE
>>>> is about 6-7 meters from USRP it can't stay attached for a long period. It
>>>> disconnects and attaches to commercial eNB,
>>>>
>>>> Also there is a problem, when UE tries to attach and the distance
>>>> between UE and USRP is about 6-7 meters. Signal strength is about -80 dBm
>>>> and is the best among other commercial eNB.  Phone performs many RA
>>>> procedures, but they fail. But phone can easily attach to commercial eNB
>>>> with signal strength -86 dBM and which is situated quite far from our
>>>> office.
>>>>
>>>> I have a feeling that the same thing causes these two troubles.
>>>>
>>>> Please, help me to figure out what is wrong.
>>>>
>>>> Best,
>>>> Ilshat
>>>> _______________________________________________
>>>> srslte-users mailing list
>>>> srslte-users at lists.softwareradiosystems.com
>>>> http://www.softwareradiosystems.com/mailman/listinfo/srslte-users
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.srsran.com/pipermail/srsran-users/attachments/20171030/23345729/attachment.htm>