What is stop criteria for TTT?

You will need an internal UE log for this when TTT is cancelled after starting.
Most likely only inside qualcomm dlf logs.

TTT cancellation is quite often if we consider rf environment.

Well, depends on the test case you are running, right?
If you have hyst configured, the value, etcā€¦

If you drive 1500 meters it is impossible not to catch one with regular values.
Like range 3dB and hyst 2dB.

As the logs are not available, can we just check theory on this?

I think theory is pretty clear too: leaving condition means TTT stopped.

I also see room for this. (And my understanding is, once you start TTT it will run until it ends, assuming here no reportOnLeave.)

Iā€™d say this helps a little bit more:

As per my understanding and seen UE logs earlier, As shown example of a3offset 3db and hyst 2dB
Entering condition is 5dB, leave is 1dB.

Report onleave just indicates whether or not the UE shall initiate the measurement reporting procedure when the leaving condition is met

So hyst is still needed to leave whether or not UE reports it. Ideally UE should report a3 if it meets citeria but while leaving, we r not allowing ue to report MR to avoid signalling.

Entering condition (TTT start) it is also not reported.
Only when TTT expires Measurement report is sent.

Yes thats the purpose of ttt to avoid pingpong due to fluctuations, to hold for 256ms as example then to report MR.
Does anyone use adaptive ttt here? Which adapts ttt based o UE speed.
We have fix 256ms for data and 100ms for voice in A3.

More important than adaptive TTT is a good filter coefficient.
Like fc3 in 3G and fc6 in 4G.

True, we have fc8 in our network.
We changed to fc4 but not seen much gain.
Can you please suggest how cells should be selected to try fc6?
Cells with more mro_pingpong or more mro_ too early handovers or any other.

Timetotrigger is not for MRā€¦

After measurement report the eNodeB has to trigger eventsā€¦
Letā€™s say your event A1 threshold is -100 and we aware that UE will send measurement report every ms to eNodeB.
Now assume that based on UE measurement report the RSRP of current serving cell is -100 so eNodeB has to trigger event A1 but behind the scene - when current cell masurement is -100 eNodeB will not blindly trigger the eventA1.
Instead what it does is, itā€™ll wait untill the TTT value, letā€™s say TTT is 312msā€¦ so the current cell RSRP which is -100 has to hold till 312ms then only eNodeB will trigger event A1.

So TTT job is to tell the eNodeB that when to trigger events.

The Measurement Report which sends UE is inclusive of Hyst n offset value.

So hyst is not deciding the entering or leaving condition but it is also one of the way to show serving cell is better!

Hyst is directly involved in entering or leaving criteria formula.

Yesā€¦
Itā€™s used to pretend your serving cell is better than neighbor.
But what Iā€™m trying to say is UE sends MR after applying everything to eNodeB them only TTT comes.

Yes like offset, hyst has direct role in entry n exit.

No, 1st TTT comes after TTT expires, UE sends MR.
Otherwise why would eNodeB pass TTT information to UE in rrcreconfig?

The question asked was, when TTT play the role?

At itā€™s expiry then TTT plays its role.
Role of TTT is to send MR at its expiry.

Hello All, let me clear question againā€¦
What event can stop TTT before expiry?

@marcengo and me are of opinion that ā€œentering condition should hold trueā€
@RFSpecialist says its leaving condition.

If anyone else have full proof justification please help.

This is from ShareTechNote explaning benefit of hysteresis:

image

It clearly says, if in our example if nbr is within range of 1dB to 5DB (+/-2dB hys) till TTT then UE sends MR.
So exit before TTT is only when leaving crieteria is met.