Smeasure is for all intra-inter.
In Nokia threshold1 for a3 start measure and threshold2a, hysthreshold2a are used.
So when serving RSRP - hys threshhodl2a > threshold2a then EVENT A1 is to be sent, stopping measurements.
Thanks, good to know the parameters name for Nokia, which don’t have relation with the name of the events (threshold1 not direct relation with Event 1).
This is the general description for Ericsson 4G in Connected Mode:
Yes, that creates a lot of confusion as well.
Thanks all of you for your help 
It is observed in Nokia there is no cellindoffset and hysteresis used.
And there is a offset of 3 dB for a3 event and 2 dB for a5 event defined.
Is this Ok?
2 dB is so short. Check LNCELL and LNHOIF mos for target freq
For intra freq, a3offset + hysa3offset gives you the final a3 threshold, 4 dB is better.
Cell ind offset you can find in neighbour objects for speeding up or slowing down handover to target cell.
Actually for intrafrequency handover it was configured as 3 dB for a3 event and 2 dB for A5 event.
I see.
As i said, if you use 2 dB only for interfreq, there will be a lot of ping pongs.
Better to stay at higher bw layer till its spectral efficiency becomes low and pass to lower layer.
This is for Data. For VoLTE there are qci1 level thresholds.
I think that the hysteresis for event A5 will not cause ping-pong since it’s not a single event Handover.
It will depend on threshold1 and threshold2 setting.
2 dB for event A5 for me is more than enough…
Obviously it is dependent thresh 1 and 2 for A5 but as far as I understoof they use A3 for interfreq also, and 2 dB short for that.
If you use A5 it is all dependent on what thresh1 and thresh 2 set to target freq.
Oh… ok, yes… agree with you.
A3 for interfreq with 2 dB hyst may have ping pong issues.
According to this comment 
I understood that 2 dB hyst was configured for A5 event…
Hmm… If so, those 2 dB between which layers?
You went to low layer because it’s 2 dB better and come back at which RSRP?
More info is needed to comment out.
Yeah set 2 dB between each layer and you ping pong UE between 2 
Also not only thresh1 and 2 are important but also start measure should be 2 dB lower than thresh1 of serving cell for UE to have time to measure target RSRP.
Normally, I prefer to set A2 threshold at the same level than A5 Th1.
I think that 2 dB lower, gives a windows where the UEs only is measuring but can`t do anything.
Is there some recommendation to reinforce those suggested 2 dBs?
(besides this: “UE to have time to measure target Rsrp”, which I think could be a issue to think into account, but not a rule)
Yes it is not a rule, but if UE starts earlier than would have more time to determine candidates so you may avoid early handovers.
But also of course starting a2 start early means more measurement gap and lower throughput.
Trade off!
Between a2 and a1 stop meas how many db do you set?
4 dBs effective level, and 40ms (quick) for timetotrigger for A2 and 640ms (slow) for timetotrigger for A1 events → this is, ensuring good levels before to stop measurements.
For Ericsson case, there is the option to configure A1 and A2 independently, but generally it’s used the feature: Mobility Control at Poor Coverage to manage this kind of situations, and this feature only gives the possibility to set one threshold for A1 and A2, which is adding or subtracting to the current signal level according to the signal level fluctuation and the A1 or A2 events occurrence (see image, which show example levels, not the same I use)
In general terms, I think this option is better (and less risky) than setting independently A1 and A2.
I agree 4 dB or more, but timetotrigger I generally set the same for both a1 and a2.
Nokia setting:
- threshold1 = -50 dbm
- threshold2interfrequency = -100 dBm
- threshold2a = -96 dbm
- a3offset = 3 dB
- threshold3 = -116 dBm
- threshokd3a = -114 dBm
In this case my concern is intra frequency setting 3 dBm for a3 event and 2 dBm for a5 event is OK or not?
Same setting for interfrequency also.
There is no offset parameter for A5 event I think.
If 2dB is hysthreshold3, then total difference will be around 6 dB between serving and target cell, which seems to be OK.
3 dB offset is ok for A3 event used in case of better cell HO.
Don’t think both A3/A5 is actually required in a site… A3 can cover the cases of A5 as well if offset is increased a little.