Unstable broadband connection

[DeltaVee595]

I have had years of totally reliable connection with Virgin Media until about 3-4 weeks ago. Since then I have been experiencing 'reliably unreliable' connection, with reduced bandwidth and dropouts tending to occur morning and evening. 

A VM engineer visited (excellent service!), replaced my old hub with a nice new Hub5, but this has not made any difference. 

BQM shows latency spikes matching the periods of dropped connection (see attached files). Ookla indicates my bandwidth is going from 368Mbps/ 18ms ping /30ms jitter to 6Mbps / 160ms ping /1184ms jitter in the space of a few mins. 

I've tried a factory reset on the hub, tried the hub in modem mode and confirmed that the problem persists across ethernet and WiFi. I mostly use ethernet so this is not a WiFI problem. 

I would welcome any assistance or advice about resolving this. 

[jbrennand]

See if the PostRS errors and T3 errors have all reset to 0. Then check every hour or so to see if they start reappearing - they shouldn't. If they do (particularly the postRS and T3 ones) you have a problem (noise) that only a Tech visit will sort.

Your drop off (red spike) was at ~10.30 am - what do the logs report around that time?

[DeltaVee595]

Thanks for you help with this John. It is very much appreciated. 

In the two hours since my most recent factory reset I can see 3 Pre RS errors, 0 Post , 0 T3 Timeouts. I will continue to check as you suggest. 

The factory reset has wiped the log so I can't tell what was happening for sure but I believe I reset the power on the router around 10:30am. 

BTW: can you recommend a source of information about cable modems where I can learn how to interpret the information about noise and errors? 

[jbrennand]

---

@DeltaVee595 wrote:

BTW: can you recommend a source of information about cable modems where I can learn how to interpret the information about noise and errors? 

---

Most of the info you need is in two "pinned" posts at the top of this board...  from Sephiroth and Andrew G.

[Sephiroth]

The Pre-RS errors are corrected in flight (by algorithm) and are nothing to worry about unless they are running riot - in which case the Post-RS errors (uncorrectable data) will be rising.  If the Post-RS errors are not rising, then the downstream is likely to be performing OK and the upstream might be suspect (more difficult to diagnose).

John has identified your BQM red spike as a 'drop-off'.  Or was that when you rebooted the Hub?   Is your 'drop-off' an event when you've lost the Internet?  Your PC will tell you if you are connected to the Internet.  

And do keep posting the stats(incl. the Network Log) and the BQM.   We'd want to see whether or not the T3 events are still occurring because the BQM does show minor red lines at the top of the chart which may well coincide with the T3 events.  In which case the upstream is likely to be playing up - either through oversubscription or fault.

[DeltaVee595]

3.0 Downstream channels

Channel Frequency (Hz) Power (dBmV) SNR (dB) Modulation Channel ID

1	331000000	-1.9	41	QAM 256	25
2	147000000	-2.6	40	QAM 256	2
3	155000000	-2.5	40	QAM 256	3
4	163000000	-2.1	41	QAM 256	4
5	171000000	-1.7	41	QAM 256	5
6	179000000	-1.6	41	QAM 256	6
7	187000000	-1.6	41	QAM 256	7
8	195000000	-1.3	41	QAM 256	8
9	203000000	-1.4	41	QAM 256	9
10	211000000	-1.1	41	QAM 256	10
11	219000000	-1.3	41	QAM 256	11
12	227000000	-1.7	41	QAM 256	12
13	235000000	-2	41	QAM 256	13
14	243000000	-2.1	41	QAM 256	14
15	251000000	-2.2	41	QAM 256	15
16	259000000	-1.9	41	QAM 256	16
17	267000000	-1.8	41	QAM 256	17
18	275000000	-1.9	41	QAM 256	18
19	283000000	-1.9	41	QAM 256	19
20	291000000	-2.1	41	QAM 256	20
21	299000000	-2.1	41	QAM 256	21
22	307000000	-2.1	40	QAM 256	22
23	315000000	-2.1	40	QAM 256	23
24	323000000	-2	41	QAM 256	24
25	339000000	-1.8	41	QAM 256	26
26	347000000	-1.8	41	QAM 256	27
27	355000000	-1.8	41	QAM 256	28
28	363000000	-1.5	41	QAM 256	29
29	371000000	-1.6	41	QAM 256	30
30	379000000	-1.6	41	QAM 256	31
31	387000000	-1.7	41	QAM 256	32

3.0 Downstream channels

Channel Locked Status RxMER (dB) Pre RS Errors Post RS Errors

1	Locked	41	210372	13661
2	Locked	40	430783	572958
3	Locked	40	496103	353987
4	Locked	41	283712	265888
5	Locked	41	527238	949717
6	Locked	41	390757	624739
7	Locked	41	166387	168700
8	Locked	41	274595	86781
9	Locked	41	350244	106998
10	Locked	41	210269	74800
11	Locked	41	249718	67671
12	Locked	41	381749	89169
13	Locked	41	336646	69467
14	Locked	41	205685	60952
15	Locked	41	379615	104746
16	Locked	41	257558	44448
17	Locked	41	342590	65395
18	Locked	41	330447	95760
19	Locked	41	182043	65767
20	Locked	41	175313	53455
21	Locked	41	166655	65227
22	Locked	40	219578	76428
23	Locked	40	253085	47039
24	Locked	41	176225	13382
25	Locked	41	203245	16433
26	Locked	41	190176	21235
27	Locked	41	164346	11411
28	Locked	41	138619	4606
29	Locked	41	132048	4236
30	Locked	41	156324	5683
31	Locked	41	177009	15316

3.1 Downstream channels

Channel Channel Width (MHz) FFT Type Number of Active Subcarriers Modulation (Active Profile) First Active Subcarrier (Hz)

33

94

4K

1840

QAM 4096

1108

3.1 Downstream channels

Channel ID Locked Status RxMER Data (dB) PLC Power (dBmV) Corrected errors (Active Profile) Uncorrectable errors (Active Profile)

33

Locked

41

-2.1

592666390

22574

[DeltaVee595]

[Sephiroth]

Doesn't look good - the downstream post-RS errors.  The Forum Team will get to this thread in a day or so and I hope they'll agree that this needs sorting.  Do please make sure that your coax cable is tightly screwed in at both ends (long shot because the tech has been but you never know).

Also it would really help if you could see whether or not a neighbour has the same issue,

Keep us posted.

[DeltaVee595]

@Sephiroth wrote:

John has identified your BQM red spike as a 'drop-off'.  Or was that when you rebooted the Hub?   Is your 'drop-off' an event when you've lost the Internet?  Your PC will tell you if you are connected to the Internet.  

---

Thanks for you reply Sephiroth

The yellow spikes coincide with progressive performance degradation - bandwidth reduces, ping time shoots up - and sometimes this results in total loss of connection for a few minutes. 

Additional data from hub posted below. 

Is there any reason I should further delay trying to book an engineer visit? 

[Sephiroth]

The purpose of booking a tech (they are not engineers) would be for them to pre-assess what's going on and then referring the matter to Networks.

The tech has the necessary gizmo to detect noise.  Your challenge will be that the tech catches the right moment(s).

[DeltaVee595]

A very helpful Virgin Media Engineer visited today and found issues with excessive upstream levels and some issues with cable connectors between our house and the cab. 

Hopefully, this is fully resolved now.

Thanks @sephiroth @jbrennand @fizz for your help.