How to find all interfered 2G cells on a network level

3G / 2G
,
1

Hi 2G experts,
For 2g interference, I remember it was not a straight forward judgment to decide if there is high interference or not, as far as I recall the band4 and 5 proportion has to be more than 30% of the total bands 1-5 (or something like this).

Can someone please provide guidance on how to find all interfered cells on a network level?

And after that, I want to have two columns in my final sheet.
“Column A:2G cell name”, “Column B:interference level in dBm” if possible rather than only boolean i.e. “Column B: interferred/non interferred”.

2

Hi, which vendor is this?

3

Huawei, but I assume it should be similar across all vendors isn’t it?
If there is another approach with another vendor kindly share it.

4

Yea, it should be.
In ericsson we used to set these parameters par cell, they are called LIMx, to define the ranges of ICMBANDs, and as u said they r from 1 to 5, 5 is ver bad, 4 bad, and so on untill 1 is good, but this depends on the ranges you define per cell.

5

I see. So when you say you define it per cell, are you talking about a threshold for alarm trigger?

6

Yes, the parameters called limit1 limit2, … limit4.

7

What about the logic of proportionality?
I mean there can be small spikes in band 4 and 5 but is too small compared to what’s reported under band 1,2,3.
Do you consider this factor as well to trigger the alarm?

8

You can, in general we see the proportions over the scan period.
And then you can set your alarming thresholds.
But one thing to consider here.
This can show an issue that can be related to External interference or something coming from your radio installation “feeder, triplexer, connector, antenna”.
So it is tricky.

9

Yes and great point, I got it but I actually just want to know if there is interference or not regardless of the cause i.e it’s due to HW or external source.

So at the end, can I assign a value for the interference in dBm? Just like rtwp and ul intereference in 4g/5g? For each cell as I mentioned I need to put the value of interference.

I know it’s a strange request maybe, but I want an interference database for each cell for all technologies but 2G is different, that’s why I asked.

10

For this, we used to run measurments, it will give you the exact level of interference.

11

Can you please elaborate? Do you mean an OSS counter?

12

From counters we can get ICMBAND only.
Lets say 20% of ICMBAND5, 30% of ICMBAND4, 10% ICMBAND3, 20% ICMBAND2 and 20% of ICMBAND1.

But from measurments, you can measure the level of interference on cell level on some frequencies you give to UE to measure.

13

Ok, can you remind me what is the unit of the counters ICMBANDx?
Is it time or traffic or what exactly?

14

This is ICMBAND distribution

image
image945×529 76.2 KB

This means the cells is recording most of the time high ICMBANDS.
ICMBAND delimited by LIMIT4, in our case 22, menas the level of interference is greater than -110 + 22 which is -88 dBm.

Now there is another step of check, from OSS measurements.

image
image1482×594 111 KB

To the left is the cell cofigured frequencies.
We can see high inteference on channel 1017 EPGSM.

To the right the measured frequencies which are not configured on the cell.
I could not show them all.
But the level of interference is high on all EPGSM bands.
Channels above 1000.

image
image1472×599 119 KB

15

Thanks for this informative plots.
Can I ask how to export it?

16

Cell – freq – number of samples – interference mediane – interference percentile.

And the percentile level you can asjust before you start the measurement.
I put in general 95%.
There are other fields as well.

17

So if you want to export this on the complete network on cell level as csv?
Is it doable I mean and easy to be done?

18

Yes, just a click.