Inter System HO from 3G to 2G ; NSN versus Huawei

[kyokuman]

Hi Experts,

Regarding HO from 3G to 2G :

- Huawei uses 2D/2F

- NSN uses 1F/2F

As I am about to swap from NSN to Huawei, do you know how it can affect the behavior/performances?

Thanks for your help !

[wolverine]

There isn't much difference in my opinion (assuming you set the threshold the same). If anything event 2D implementation is slightly better as the UE only sends this once to trigger the HO while for event 1F the UE needs to send this for ever RL in the AS.

[kyokuman]

Thanks.

That's the thing! After swapping few sites to Huawei, ISHO attempts increased by 4 times. Thresholds were mapped with NSN (-112dBm)

Huawei advised to increase 2D threshold to -110dBm, in order to decrease ISHO attempts.

When using 2D, conditions must be fullfilled on AS basis.
Why with NSN, it is more restrictive, all cells must fullfilled criterias.

[kyokuman]

What I wrote below doesnt explain the difference in terms of ISHO attempts I guess.

But I just noticed that Huawei uses an hysteresis (Hystfor2d) :
Qused < 2D_threshold - Hystfor2d/2

NSN, as far as I know, doesnt use any hysteresis. Condition to be fullfilled is Ecno_Threshold (FMCS)


You guys have probably some comments!? Thanks a lot ! I will add Rep for any answer

[agenov]

What I wrote below doesnt explain the difference in terms of ISHO attempts I guess.

But I just noticed that Huawei uses an hysteresis (Hystfor2d) :
Qused < 2D_threshold - Hystfor2d/2

NSN, as far as I know, doesnt use any hysteresis. Condition to be fullfilled is Ecno_Threshold (FMCS)


You guys have probably some comments!? Thanks a lot ! I will add Rep for any answer


As far as i remember this is 3GPP so all the vendors must obey. Check 25.331 (if i am not wrong)

BR

[boring]

hi,

NSN triggers compressed mode when all cells in AS has triggered an event 1f which is completely equivelant with Huwaei (and E*******) behaviour of event 2d.

any difference in the amount of ISHO attempts is merely due to different RSCP thresholds. for a large UE population 1 or 2 dBs can make big difference!

you should remember also that NSN is using other triggers for CM apart from RSCP (e.g. EcNo, UE Tx power and some internal triggers). perhaps the others are causing this amount of ISHO attempts.

br

[kyokuman]

Hi guys,

I have only ISHO due to RSCP activated.

Huawei explained the difference of number of HO by saying that NSN uses 1F while Huawei uses 2D.

For me, the reason is more likely the hysteresis that is not used by NSN.
So the final threshold is simply different.

I just increased 2D_RSCP threshold by 2dB and number of HO attempts are same now, compare with before swap.
so now, I have :
-110dBm, while with NSN it was -112dBm.
Number of HO is same now.

More comments are of course welcome

[tilangminh]

anything else ??

[boring]

Hi guys,

I have only ISHO due to RSCP activated.

Huawei explained the difference of number of HO by saying that NSN uses 1F while Huawei uses 2D.

For me, the reason is more likely the hysteresis that is not used by NSN.
So the final threshold is simply different.

I just increased 2D_RSCP threshold by 2dB and number of HO attempts are same now, compare with before swap.
so now, I have :
-110dBm, while with NSN it was -112dBm.
Number of HO is same now.

More comments are of course welcome

I think you are spot on now. essentially both approaches behave similarly and it is good to know that there is this difference between the two vendors because of the hysteresis

br

[twenty8]

Hi,

In NSN system, my understanding is that there is a time hysteresis which is normally set to about 100ms. So e1f is only send when the HHORscpThreshold (eg. -115 dbm) is meet for a period of 100ms (the hysteresis).

In Huawei system, as you mentioned the e2d is send when TUsed2d-H2d/2 , my understanding is that H2d has a default of 6 which makes it 3db , so if we set the trigger condition as same for NSN and HW , ie HHORscpThreshold = -115 dbm as well as TUsed2d = -115 dbm , we would actually get an earlier trigger in Huawei system, as the e2d is send when -115 - 3 = -112 dbm.

I believe this is the reason for the difference.

[balachandrun]

huawei uses TimeToTrig2D along with hysterysis(Event 2D Trigger Delay) with 320 ms to wait for RSCP or EC/IO has met the thresholds to avoid fading..

[kyokuman]

Guys,

I have mapped all the parameters.

I have set the additional hysteresis of Huawei to 0.

I have mapped the timetotrigger etc...


Number of Inter RAT HO attempts is much less wfater SWAP!


Still I don't have explanation.


If I want to have the same amount of HO as before, I have to set 2D threshold to -10dB, while it was -12dB with NSN.

But is it the solution ?

[Junio]

Hi my freind I want to have the solution for resolve SHO.FailRLAddUESide.CfgUnsup
SHO.FailRLAddUESide.Isr
SHO.FailRLAddUESide.InvCfg
SHO.FailRLAddUESide.NoReply

The last is more important
the vendor is Huawei

[agenov]

Guys,

I have mapped all the parameters.

I have set the additional hysteresis of Huawei to 0.

I have mapped the timetotrigger etc...


Number of Inter RAT HO attempts is much less wfater SWAP!


Still I don't have explanation.


If I want to have the same amount of HO as before, I have to set 2D threshold to -10dB, while it was -12dB with NSN.

But is it the solution ?


i believe it is not necesseary to have absolutely the same number of attempts something more less InterRAT HO is not so bad. Anyway, if you want to be as close as possible then define threshold+hyst to be equal as NSN settings. Something else you should take into consideration is this kind of handover is based on coverage not quality so if you have before swap quality HO as well and now you don't have this may result in the difference you are observing.

BR

Alex

[agenov]

Hi my freind I want to have the solution for resolve SHO.FailRLAddUESide.CfgUnsup
SHO.FailRLAddUESide.Isr
SHO.FailRLAddUESide.InvCfg
SHO.FailRLAddUESide.NoReply

The last is more important
the vendor is Huawei


The last "No reply" usually is provoked by radio issue but could be and hardware problem as well in the NodeB. THe best you can do is drive test or/and LMT trace to track the problem in more details.

BR
Alex

[wolverine]

You should also check that the CPICH values for both systems is the same as all these events on based on that. NSN measures CPICH on top of the cabinet. Where does Huawei measure it? Have you introduced new MHAs? They could increase/decrease loss. Fedders kept the same?

[Junio]

Thank for your answer.

[Junio]

Hi all
i must to propose the solution for to resolve for my intership :
at moment my solution are :
P-CPICH, pilote pollution ,add TMA, ajust the parameter of antenna(azimut,tilt) and RTWP

i don't know if it's a best way ?

[kyokuman]

Guys !

I think this topic is about HO behavior comparison between NSN and Huawei.

Nothing else !!

Thanks

[Junio]

good my freind you ara right.
thank

[boring]

You should also check that the CPICH values for both systems is the same as all these events on based on that. NSN measures CPICH on top of the cabinet. Where does Huawei measure it? Have you introduced new MHAs? They could increase/decrease loss. Fedders kept the same?

actually this does not matter because in both systems the UE makes an absolute measurement, which means it does not matter what is the site configuration

this means that if all parameters are the same then the amount of ISHO should be identical.

a stupid question in the meantime. the levels of traffic have increased after the swap? also has something (in terms of parameters) pushing more traffic from 2g to 3g?

[kyokuman]

Hi,

2G parameters have been mapped as well.
So there is no reason traffic is now more pushed to 3G.

All the 3G parameters are mapped. But in some cases it is not really possible.

When mapping exactly the thresholds to previous NSN values, I have increase of HO, and decrease of traffic.
After adding 2dB on the 2D thresholds, I have now same amount of HO, and traffic is normal as before.


Huawei pretends the Handover behavior is different because NSN uses 1F, and Huawei 2D.

I am still doubting.

And even I added 2 dB in the 2D thresholds, I am not convinced that this is the solution.
BECAUSE :

NSN : CS threshold to start Compressmode = -12dB
Huawei : now is set to -10dB !

Don't you think -10dB is very high (For Ecno trigger Inter RAT HO) ??

I think it is !

There are probably another reason (penaltytimes, other timers....) I am not awaare of (for now at least).


:confused::confused::confused::confused:

[cococrunch]

Hi Kyokuman,

Did CPICH power increase? Both NSN and Huawei CPICH reference point are after the RF module. -

If CPICH increase/decrease then you increase/decrease the coverage, BUT then your RSCP threshold is the same which in turn may reduce/increase the ISHO attempts.

[monnnos]

look .. for ISHO in NSN is triggered by e1f which is periodic and AS cells should cancel it by e1e
but for HUA 2D is triggered by best server ( 1st cell in AS) which is better in my opinion
you can check INTERRAT CS THD for EcNo(2D EcNo) to be -12 dB or higher ( -14 will decrease ISHO attempts but may increase CS drops slightly)
also INTERRAT CS THD for RSCP( 2D RSCP) to be -115 dBm ( this also wll decrease the ISHO but may increase CS drops specially on border cells that no server except them for long distance )

i think you 'd better increase 2D/2F RSCP-115/-112 and make 2D/2F EcNo -12/-9

reps and thx are welcomed :)

[gprastomo]

Maybe the coverage is not the same between nsn n huawei, after swapping to huawei coverage are shrink by 2dB, regardless of the hyst.

last time i did swap to nsn, and suddenly the AC failure is so high and i noticed that NSN is more sensitive so coverage are larger than siemens.