Slow Fibre again

Hi

The problem isn't getting any better, don't want to call out an talktalk engineer as they will charge me because it is working ok now its evening time its the worse. for example last night the internet wasn't working at all, had a sync speed and everything, but no throughput speed so I couldn't ping google or access any sites, had a external ip from talktalk though. The wireless is still also terrible but that's a different subject, i have since had enough and decided to get the HG612 out and run a couple of tests, i noticed the SNR fluctuates a bit, i watched it go from 8.1 to 6.6 for 30 seconds then back to 8 again after it keeps doing that, 8.1 then 6.3 then back to 8.1 never used to do that when i last used it about year and half ago. The connection is still banded and my homeplugs still have issues where the speed is really slow.

I have included the stats from the hg612 please can someone have a look and tell me if there is something that might be wrong?

# xdslcmd info --stats
xdslcmd: ADSL driver and PHY status
Status: Showtime
Retrain Reason: 0
Last initialization procedure status: 0
Max: Upstream rate = 6064 Kbps, Downstream rate = 30160 Kbps
Bearer: 0, Upstream rate = 2000 Kbps, Downstream rate = 22399 Kbps

Link Power State: L0
Mode: VDSL2 Annex B
VDSL2 Profile: Profile 17a
TPS-TC: PTM Mode(0x0)
Trellis: U:ON /D:ON
Line Status: No Defect
Training Status: Showtime
Down Up
SNR (dB): 8.0 12.9
Attn(dB): 24.4 0.0
Pwr(dBm): 11.7 4.3
VDSL2 framing
Bearer 0
MSGc: 34 330
B: 47 48
M: 1 1
T: 51 6
R: 16 16
S: 0.0681 0.7651
L: 7520 711
D: 951 1
I: 64 68
N: 64 68
Counters
Bearer 0
OHF: 810115 219674
OHFErr: 139 0
RS: 206522024 1268200
RSCorr: 200938 0
RSUnCorr: 13984 0

Bearer 0
HEC: 2600 0
OCD: 132 0
LCD: 132 0
Total Cells: 151824115 0
Data Cells: 6430080 0
Drop Cells: 0
Bit Errors: 0 0

ES: 8 0
SES: 0 0
UAS: 260 260
AS: 3530

Bearer 0
INP: 8.00 0.00
INPRein: 0.00 0.00
delay: 16 0
PER: 4.35 16.13
OR: 73.43 166.64
AgR: 22472.21 2166.36

Bitswap: 2949/2949 0/0

Total time = 1 hours 3 min 10 sec
FEC: 200938 0
CRC: 139 0
ES: 8 0
SES: 0 0
UAS: 260 260
LOS: 0 0
LOF: 0 0
LOM: 0 0
Latest 15 minutes time = 3 min 10 sec
FEC: 2416 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Previous 15 minutes time = 15 min 0 sec
FEC: 54126 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Latest 1 day time = 1 hours 3 min 10 sec
FEC: 200938 0
CRC: 139 0
ES: 8 0
SES: 0 0
UAS: 260 260
LOS: 0 0
LOF: 0 0
LOM: 0 0
Previous 1 day time = 0 sec
FEC: 0 0
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
Since Link time = 58 min 49 sec
FEC: 200938 0
CRC: 139 0
ES: 8 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
LOM: 0 0
#


Thanks

The key is whether the stats are any different at night and also unplug all the home plugs, there is a small possibility they might be interfering with the VDSL, and then do speed tests directly into the router using an Ethernet cable.

If the 6.3 to 8 while a swing of note, should not mean you have almost zero throughput. The key is a speed test every 30 minutes from now until 10pm or so, and see if there is a gradual drop in speeds as more people go online or if it is a sharper drop that coincides with sunset.

Also make a note of the first IP address beyond the hub during the day and when you have problems try pinging that, i.e. if that works and others don't then points towards an issue in the TalkTalk network and nothing to do with your broadband connection itself. Also considered using alternate DNS servers?

I will try unplugging the homeplugs and see if that improves

The 6.3 to 8 was earlier in the day, in the evening its around 5.5- 6 when i was on the super router. would it be best to use the speedtest on here or would speedtest.net be best for tracking the gradual drop.

i will try that if it gets worse later, now you mention it i seems silly i havent tried that yet, do you know which could be the best?

Our test would be better as can compare your results to others at the same time, and we are a little more sensitive to congestion issues even on the multiple thread version, plus the throughput graphs give extra info that the basic figures can mask.

Surely a fluctuating SNR margin won't, in itself, cause any slowdown unless there's a very high bit error rate and a lot of retries (although maybe g.inp retries don't show in error stats). If it got bad I would expect to see resyncs happen frequently,

I assume the throughput tests have been done direct-wired. I do not trust homeplugs to be reliable for throughput tests as they certainly are vulnerable to extraneous noise. If there's sufficient noise around to drop the VDSL SNR margin by 2dB, then it might be even more noticeable on homeplug networks.

I'd also recommend doing some tracert testing when under load as that can give a clue as to where the delays are occuring. It's not exactly rigorous of course, but often you can see which of the legs (or combinations of legs where nodes don't respond) are being delayed. If the problem is in the "first" mile it ought to show up as a big increase in latency in the first hop (however, many nodes that covers) if done under load conditions.