Awful variation in 4g download speeds

[Rob195]

I live on an island with four masts.  Speeds from one are excellent (65-1105 d 25-45 u)  but the other three are very poor  usually 5-8d and 0u.  More often than not one of the three poor masts fails on speedtest completely.  I keep reporting and keep getting told that there is not a problem and it may be a one off........I have three weeks of speed test results now all consistently showing the same.  I moved back to EE after a breif time away to take advantage of a good value 40gb deal but I cant use it.  Very frustrated.  Incidentally the signal from each mast is usually excellent and making calls is not a problem - all masts give me a 3/4 bar 4g but I can park under two of them and get 8mbps at best.

[bristolian]

These figures are all giving measurements for the radio link between your phone and the serving network site, which appears to be good in all cases.

If there is an issue elsewhere, for example heavier data load causing slow speeds on individual sites or perhaps congestion elsewhere in the network.. these figures won't highlight them.

[Rob195]

Thanks I think I understand.  The signal indicated by the bars on the phone is the same at all four sites so that would fit with what you've observed from the numbers.  What would constitute a heavy data load do you think.  I spend a lot of time wandering around Castlebay and I dont see hundreds of folk streaming video on their phones.....in fact I cant remember the last time I saw or heard anybody on a moblie in Castlebay and most locals would have broadband in their home. If let's say there were 100 folk sat at home streaming video on their mobile not using their broadband would that slow things down...or would 50 or 25.  My daughter live in the middle of Telford and her mobile speedtest results are as good as my best.....

[mikeliuk]

Hi @Rob195 ,

In the above, you've reported you measured poor performance from a certain cell tower with poor signal to noise ratio although the other parameters seems ok. The ping of 429 ms was far above an acceptable unloaded value of around 60 ms or a good value of around 30 ms. The jitter of 54 ms is also pretty poor. https://usatcorp.com/faqs/understanding-lte-signal-strength-values/

eNb 17470

RSSI -63dBm

RSRP --85dBm

RSRQ -5db (red triangle down)

SNR 3.4dB (Green triangle up)

Speedtest result was 7.47 down 0.00 up Ping 429 Jitter 54.78

You showed details of poor measured performance, again where the signal to noise ratio is poor but the other parameters are ok. The ping of 480 ms is again bad and the jitter of 108 ms is bad.

eNb 17470

RSSI -63 dBm

RSRP -85dBm

RSRQ -5dB (red triangle down)

SNR 3.4dB (green triangle up)

Speedtest 7.73 D 3.61 U Ping 480 Jitter 108 3

The cell tower for which you measured good performance shows excellent signal to noise ratio where 13 dB would already be regarded as excellent. The ping of 29 ms is good and 38 ms is ok. Jitter under 5 ms is good and under 10 ms is ok in this context.

eNb17469

RSSI -60dBm

RSRP -86 dBm

RSRQ -5dB

SNR 23 dB

102.64 down 46.07 up Ping 38 Jitter 4.86

eNb 17469

RSSI -60dBM

RSRP -86 dBm

RSRQ -5dBM

SNR 23dB

down 71.94 up 28.62 ping 29 jitter 8.89

From the above data, it seems it is the signal to noise ratio which determines whether you get good performance or poor performance. If you move closer to the bad masts, it's possible the signal to noise ratio may improve and you may measure better performance. It's like a person shouting but from a great distance, or with some obstruction between the speaker and the listener, you can hear better if you move closer to test.

The ping of 480 ms may purely be a result of the bad signal to noise ratio. It's possible there is a problem elsewhere and there may be a little congestion somewhere on the data path, but you won't be able to judge this with any accuracy with a SNR of 3.4 dB.

It is possible that standing under the bad masts might then give you an excellent SNR but you then find other issues once the poor signal to noise ratio is resolved.

You may wish to look up the meaning of the "green triangle up" as it may not mean what you think it means given it does not show when the SNR is excellent. From memory, even an SNR as bad as 3.4 dB would not fully explain the bad performance as I think I've achieved reasonably responsive connections with such values and achieving around 15 Mbps down.

Typically when I see a bad SNR of under 6 dB, I force a reconnection to a better cell tower to get a SNR above 10 dB.

[bristolian]

Ultimately, regardless of endless detailed analysis which adds limited value, the recommendations to the OP remain unchanged. The signal quality is good in these cases and the issue is not poor coverage.

There are multiple possible causes for these slow speeds, but they are all network side and are only likely to be resolved by proper investigations being done.

[Rob195]

Hi

Thanks for all of that much of which I dont understand but your observations may be useful when I pluck up the courage to call the tekkies.

Cheers

Rob

[Rob195]

Thanks again.  As I commented above I dont understand much  (maybe all 🙂 of the numbers.  From what you know of the way things work and the numbers you've seen do you think it is something that can be fixed?

Cheers

Rob

[Rob195]

Hey out therein the community.  It would be great if I got that call from the technical team that I've been promised twice now.  Maybe whoever calls could read through the above so I'm not repeating myself.  I di spend an hour on chat last week too.....somebody was going to call me after that as well.

[Leanne_T]

Hi @Rob195

Thanks for coming back to us. 

Could you reply to my message and we certainly see what we can do to help 🙂

Thanks. 

Leanne.