Packet Loss when router is remotely busy

[#1]

Thank you, I have updated the ticket with my availability.

[#2]

In other news, I noticed that some hand-sprayed white arrows have appeared around the BT manholes that have water running into/out of them.
Hopefully this is a precursor to some road-mending/ re-surfacing type work to stop them flooding.

[#3]

Hiya @Highside, thanks for getting in touch and for updating the ticket. I have now booked in the new visit and have updated the fault ticket for you here: https://www.plus.net/wizard/?p=view_question&id=223659602

[#4]

The engineer came today. He was the same chap as came last time!

This time he said he did see some problems and went and fiddled with the roadside cabinet.
He says it is definitely fixed now and that the internet speed should increase over the next 10 days.

This is just as well because our downstream sync speed has dropped to 23 Mbps (from 27) which is the slowest I have ever seen it!

So far today there has been zero packet loss, which is hopeful, but it is in the evenings that the internet gets thrashed to death and the problems usually show up, so I'll let you know...

[#5]

Cheers for the update, that’s promising!

Looking at your ThinkBroadband Quality Monitor, things don’t look too bad now, but how’s it gone tonight?

[#6]

The saga continues!

Plusnet have been very good at chasing this up, calling me up to ask if things are OK, and arranging for engineer visits.
The last visit was on Friday 29th April.

A small team arrived and used a cherry picker to get at the wires at the top of the pole across the road. They said they re-made a connection over there, although they didn't seem to think there was anything actually wrong in the first place.

The chap said that the first engineer had found some "battery voltage" on the line and done some work in a manhole that cured that problem but that since then nobody can see anything wrong.
He showed me his test box that was measuring electrical resistances and capacitances of the copper between me and the roadside cabinet and they were all well within tolerance and balanced etc.

But sadly, the sync speed has gone down yet again. I am now sat at a steady 23Mps (it has reduced 29... 27... 25... with each intervention).

Also, the thinkbroadband monitor here: (not sure if the link below will work):

Broadband Quality Monitor | thinkbroadband

shows approximately 10% packet loss around 8PM on Sunday 1st May, which is really bad!

The problem (apart from the speed now being about 23% lower that it should be) is that we only get packet loss when the internet is busy, and the engineers that come here don't seem to do anything that makes the internet busy.

As I said, the last chap just seemed to measure the electrical characteristics of the copper between me and the cabinet. This does not seem to be addressing the main issue.

[#7]

I'm really sorry this is still ongoing and I can appreciate this must be frustrating.  I can see our faults team has arranged for an escalated engineer visit for tomorrow and hopefully they can get to the bottom of it. 

Please let me know how it goes

[#8]

Oh dear, not good news.
Although it does not appear to be for want of trying, things are not improving.

Another engineer came this afternoon and spent hours here. I was at work, but my son let him in.

The engineer said he swapped us to another pair of wires.

He also disconnected all the extension sockets in the house, so the only phone socket that works is the master socket.

While I appreciate that if there is a problem with the extension wiring it can affect the broadband speed I think it is pretty unsatisfactory for the guy to just disconnect all the extensions and then leave without making sure that the extensions had anything to do with the problem.

Nothing he has done has improved anything. Sync speed is sitting at 23.1mbps (which might actually be marginally slower again!). The Thinkbroadband monitor says packet loss is the same as ever and this is borne out by our experience when playing Overwatch, it reports the same packet loss as before.

I'm pretty miffed that I'll now have to get all the extensions wired back in again, but will wait a few days so that we can see if the Hub Two decides to speed up a bit after it has "settled down". I am extremely sceptical that sync speed is going to improve since this is the third (or is it fourth?) time that something has been fiddled with and we have been told to wait for a few days because it might speed up after a while, but it never has.

[#9]

Well, I was wrong, the Downstream sync speed HAS changed.
It has got SLOWER. Now sitting at 22.304 Mbps.

[#10]

Going back to basics...

1. There should have been no need to disconnect the extension wiring as that should have been connected to the front face-plate of the NTE socket - do you have the latest NTE5c master socket installed?
2. Is the phone line quite - especially at the time you notice the packet loss?
3. The packet loss is most discernible at 8pm - what else is happening in the household at that time?
4. Can you please post the router stats?
5. Is this a Plusnet Hub One?

If the answer to the last point is "Yes", I would turn to router stats and run the BTHH5a version 24x7 and start looking for evidence of RFI (REIN).

[#11]

1. The socket says Master Socket 5C on it.
I think the engineer wanted to leave the extensions disconnected but leave me with a working phone and internet when he left. With the front off the master socket there is only one test socket so he would have had to choose between plugging in my phone and my modem. I think that is right anyway.
But yes, of course he should have done his tests and put things back when he found no improvement!

2. The phone line is quiet at the moment. In years gone by we have had crackles that caused problems but they were fixed and have not returned. I mention this simply to show I know what you mean by a noisy line. The packet loss happens when the internet is busy. I have not tried making/receiving a phone call at this time. Is this worth doing as a test?

3. 8 PM is when the internet gets busiest. Four adults are in the house and by that time one or two people are streaming videos and it is in the evening that Jonathan plays his computer games like Overwatch.

4. I can't see how to get routerstats out of a Hub Two. Attached is as screenshot from www.broadbandchecker.btwholesale.com which may or may not help. No mention of RFI.

5. It is a Hub Two.

[#12]

I found an information page in the Hub Two which might help:

Product code:Plusnet Hub Two

Serial number:<redacted>

Firmware version:v0.05.03.07154-PN

Firmware updated:Unknown

Board version:R01

GUI version:1.3 10_03_2021

DSL uptime:0 Days 20 Hrs 23 Mins

Data rate:5.924 Mbps / 22.304 Mbps

Maximum data rate:5.870 Mbps / 25.428 Mbps

Noise margin:6.0 / 5.7

Line attenuation:8.0 / 24.2

Signal attenuation:6.3 / 28.0

VLAN id:101

Upstream error control:Off

Downstream error control:Off

Data sent / received:2.8 GB Uploaded / 19.3 GB Downloaded

[#13]

NTE5c - the extension wiring should have been left attached to the face plate - you can do that yourself as that is within the user's domain / responsibility.

It is not possible to connect a router to the test socket other than via a dangling filter, which will have been supplied with the router.

To me those line stats look odd - compare them to mine...

DSL uptime:21 Days 17 Hrs 7 Mins

Data rate:5.958 Mbps / 63.027 Mbps

Maximum data rate:5.958 Mbps / 62.690 Mbps

Noise margin:6.0 / 2.7

Line attenuation:7.4 / 19.9

Signal attenuation:7.4 / 25.7

US figures are not too dissimilar.

DS attenuation figures are not vastly different, but the DS sync is substantially different.

Your DS SNRM is on the nail.

So big question - why are two circuits having a similar attenuation performing so significantly different?

[#14]

Well, at least Plusnet are not giving up!
I got a call just now to say they will be chasing this up further and another engineer (I'll have to allocated them a parking space if this gets much more regular) will likely be around on Thursday morning.
I'll keep the forum posted.

[#15]

Townman, you say the downstream figures are not vastly different but:
Noise margin me: 5.7 you: 2.7
Line attentuation me: 24.2  you: 19.9
Signal attentuation me: 28 you: 25.7.

I don't know what the ranges or scales of those numbers are.
For all I know the difference between 2.7 and 5.7 is the difference between "brilliant" and "awful".
Same for the others.

Is it relevant that upstream figures look much more similar than downstream? How does the physics of that even work? There is only one pair of wires!
The copper wires to my roadside cabinet are about 950m long I am told. Do you know how long yours are?

Much shorter than that I am guessing, judging by the 63Mbps you get.