@marcengo can you be more clear?
It does not wait for TTT expires in case neighbor cell is worse than offset+hyst?
Or in case neighbor cell is worse than offset-hyst?
This is the key question.
If neighbor is not satisfying entry during TTT is running, TTT stops and then no question of MR or handover.
If serving it at -90 and a3-offset is 3 db. Hyst is 1 db. Then nbr satisfy entry at -86 dBm.
Now if entry is stoping criteria of TTT, it will stop if neighbor falls below -86.
If leaving is criteria, then TTT will stop when neighbor falls below -88.
So which one gives genuine neighbor?
Obviously TTT stop if entry condition is not satisfied will give it.
If it is like you say you would not need an hyst.
Role of hyst is to play with+hyst at entry and -hyst at leave condition (TTT stops).
-Hystv is used to tell eNodeB that triggered event is no more valid now.
Once MR is sent for any event, then leaving criteria start to check.
And this condition should hold true till TTT.
This sentence is same as specs:
We have a different understanding.
In both 3G and 4G MR is sent after TTT expires and condition is still fulfilled.
I am saying that condition to stop TTT for A3 is offset-hyst you are saying it is offset+hyst.
Yes.
At least specs (36.331) says same thing as i understand.
What would be the role of hyst in this case, if I may ask, if it can be included in offset?
To report eNodeB that event is no longer valid and cancel the action if eNodeB is planning anything (like HO preparation).
Event is not valid untill it is triggered that means at the expiry of ttt and condition is fulfilled.
Till that time there is no event.
Meaning eNodeB doesnât know anything until it gets the measurement report.
Leaving criteria is checked after MR is sent for entry condition.
Have you actually seen a network that uses leave condition measurement report?
I havenât seen any.
That is to avoid signalling load.
Secondly once eNodeB take action like HO command, which is done within very short time right after eNodeB receive MR.
There is no point in sending MR for leaving condition.
Hence if MR for leaving condition is not sent, then hyst has no meaning other than HO margin increase.
So you say there are 2 events for A3 in 3GPP: one for entering condition reported to eNodeB and another one for leaving condition reported to eNodeB?
Yes. But mosty leaving event is not used. We have disabled it in our network.
Can you please let us know for which vendor and by which parameter you disabled it?
Samsung.
REPORT_ON_LEAVE in EUTRA-A3CNF command
REPORT_ON_LEAVE= false
I donât know SamsungâŚ
And this is how hyst is explained on ShareTechNote:
Case 2 would be our case.
I have not seen A3 events âreport on leaveâ configured for other vendors.
Not actualy.
This case says neither entry condition nor leaving condition satisfied. (Confining in pink zone). Our case is when entry is satisfied.
I think itâs up to vendor whether they want to stop reporting on leaving. In order to avoid signalling.
Specs have given flexibility to use or not use.
@RFSpecialist, my initial assumption was: TTT always runs until the end.
When it expires, with all measurements done, the UE would decide if it should send the MR or not.
If all measurements met the entering criteria, then it would send, if not, donât send.
However, the decision to wait until it expires makes no sense because if all the L3 filtered measurements must meet, as long as the UE have at least one not meeting, why wait, right?
Itâs is implementation choice, and it doesnât wait until the end of TTT. 