PDA

View Full Version : HSUPA Low Throughput



MaikonX
2014-08-26, 10:33 PM
Hi experts,

What could be checked in our RNC to improve HSUPA Throughput? Any specific feature? Parameters? We use RAN 14.0. On DTs the rate is very low < 100kbps on most of the route. All Nodebs are IP with 4 carriers (mostly)!! RTWP is high! Lots of users per Cell. Lots of custumers complains!!

Pls any help I give reputation ;)

agenov
2014-08-29, 01:56 AM
You already answered i.e RTWP is high . that's one of the main problems nowadays.

br
alex


Hi experts,

What could be checked in our RNC to improve HSUPA Throughput? Any specific feature? Parameters? We use RAN 14.0. On DTs the rate is very low < 100kbps on most of the route. All Nodebs are IP with 4 carriers (mostly)!! RTWP is high! Lots of users per Cell. Lots of custumers complains!!

Pls any help I give reputation ;)

MaikonX
2014-08-29, 02:39 AM
You already answered i.e RTWP is high . that's one of the main problems nowadays.

br
alex

We all know this.

And what about changing the BACKGROUNDNOISE on UCELLCAC to something like 220 (-90dbm)! Field tests has shown improves on UL thr during busy hour. But what will be the "side effects" on this? Anyone can explain?

agenov
2014-08-29, 03:13 PM
never played with that parameter. Try and share :)




We all know this.

And what about changing the BACKGROUNDNOISE on UCELLCAC to something like 220 (-90dbm)! Field tests has shown improves on UL thr during busy hour. But what will be the "side effects" on this? Anyone can explain?

gudubettin
2014-08-29, 03:52 PM
Hello,
First you need to check transmission capacity and way for low HSUPA rates,
If RTWP is high, you need to adjust attenuation parameter,
Regards

MaikonX
2014-08-29, 11:33 PM
Hello,
First you need to check transmission capacity and way for low HSUPA rates,
If RTWP is high, you need to adjust attenuation parameter,
Regards

IuB Capacity is ok, all nobes are IP, this is not the case. What is the name of this attenuation parameter?

The changes in BNG Noise has improves a lot the UL throughput! Also you must turn OFF BNG Switch, if it is ON.

Shakeel_15
2014-08-30, 01:05 AM
Hello,

Could you mention the exact parameter names for background noise and bng switch?

BR


IuB Capacity is ok, all nobes are IP, this is not the case. What is the name of this attenuation parameter?

The changes in BNG Noise has improves a lot the UL throughput! Also you must turn OFF BNG Switch, if it is ON.

MaikonX
2014-08-30, 01:19 AM
Hello,

Could you mention the exact parameter names for background noise and bng switch?

BR

Yes, sure!

MOD UCELLCAC: CELLID=XXX, BGNSWITCH=OFF, BACKGROUNDNOISE = 220 ( 220 => -90dBm, by default is 61 => -106dBm). The formula is -112dbm + BGN/10.

Philippe
2014-09-01, 05:45 PM
If BGNSWITCH is set to OFF:
-The background noise used by the HSUPA scheduling is set to a fixed value determined by BACKGROUNDNOISE, by default to -106dBm (typical value of background noise if the spectrum is clean),
- The HSUPA scheduling algorithm will compare the value of BACKGROUNDNOISE + MaxTargetUlLoadFactor(e.g. 4dB) with the measured RTWP to decide if HSUPA resources can be allocated.

So if the background noise is bad estimated, you have risk to be sooner limited and have bad HSUPA performances.

If you put BGNSWITCH to ON, the Background noise value will be updated based on RTWP measurements and so you will have a value closer to real background noise in your network and you will not face the noise rise limitation as soon as in the previous case.

However, having RTWP of -90dBm is not normal and you should look for the cause of the interference (other system interfering such as 900MHz or 2100MHz not licensed equipment, repeaters, radar,...). This investigation can take time.

MaikonX
2014-09-01, 10:43 PM
If BGNSWITCH is set to OFF:
-The background noise used by the HSUPA scheduling is set to a fixed value determined by BACKGROUNDNOISE, by default to -106dBm (typical value of background noise if the spectrum is clean),
- The HSUPA scheduling algorithm will compare the value of BACKGROUNDNOISE + MaxTargetUlLoadFactor(e.g. 4dB) with the measured RTWP to decide if HSUPA resources can be allocated.

So if the background noise is bad estimated, you have risk to be sooner limited and have bad HSUPA performances.

If you put BGNSWITCH to ON, the Background noise value will be updated based on RTWP measurements and so you will have a value closer to real background noise in your network and you will not face the noise rise limitation as soon as in the previous case.

However, having RTWP of -90dBm is not normal and you should look for the cause of the interference (other system interfering such as 900MHz or 2100MHz not licensed equipment, repeaters, radar,...). This investigation can take time.

1o. OK that is the default parameter. But we can not always work with defaults.
2o. Yes that is how HSUPA scheduling will work. Only that the value of Noise Rise (dB) will depend on the MaxTargetUlLoadFactor (Noise rise dB = 10*Log(1/(1-MaxTargetUlLoadFactor)). Só this will not always be 4 dB. The default value of 75% leads to 6 dB in noise rise. In our network this value is 85% (8.2 dB) and some sites is even more 90% (10dB)!
3o. We agree the the best will be ON. But RTWP around -90dBm or even more can be easily reached by load on busy hour on our network. And this has tremendous impact on HSUPA.

With all that, I am just saying the we had a great improvement on HSUPA logs from drive test after changing those parameters.

Philippe
2014-09-01, 11:26 PM
I agree with your comments. Figures are example. The MaxTargetULLoadFactor will also depend on your design and the max noise rise you can afford whitout affecting uplink accessibility KPIs.