Mysterious problem with broadband / internal wiring...

Hi all,

I wonder if anyone might be able to help out. I recently order fibre broadband with TalkTalk and connected it up yesterday (the go-live date). Currently speeds are almost identical to what they were before we got fibre, around 4.5-6.8 MBPs.

I have been on the line with technical support and have done my own digging and the confusing results are as follows:

- Plugged into the face plate of our master socket, the router is receiving at 7 MBPs, and a speed test on my laptop using the wireless connection is a shade below that, as explained above, something around 6.5 MBPs.

- When I plug the router into the test socket, suddenly the router is downloading at 25 MBPs, but my laptop speed tests still show a speed of about 6.5 MBPs.

This suggests separate, but confusingly interlinked problems.

1. Faulty internal wiring, because the router download speeds are more than 3x faster when connected to the test socket.

2. Faulty wireless, because even when plugged into the test socket and receiving proper fibre speeds, the wireless connection is barely 25% as fast.

The confusing thing is that when I plug the router into the faceplate again, the speed on my laptop stays about the same, even though the router's reported speed drops from 25 to 7 MBPs. Surely if there is an internal wiring problem shaving that much speed off the router's potential maximum, the wireless connection should also drop when the router speed drops?

Is it possible that there are two separate issues? That along with the wiring problem, the wireless is only semi-functional and is coincidentally capped at about 7 MBPs, which also happens to be just below the lower speed of the router (when plugged into the faceplate).

It just seems unlikely that the wireless would be faulty on a brand new router, and only slightly faulty, so that it works fine, but with steadily lower speeds...

I have looked at the master socket and there is no bellwire attached. There are two blue/white connectors at both 2 and 5.

Can anyone help?

Thanks very much in advance!

I think the wireless is a separate problem, as it is completely separate from the WAN connection. Though of course, as you point out, it can't exceed the WAN speed.

The two pairs to the master faceplate suggests two extensions. What do you have on those extension, and is whatever is there filtered?

Are there only two extensions? In the house, whether used or not.

Please could you use the Address option of this checker and copy/paste the table of estimates.

(Just a small request for your next post. You appear to be talking megabits per second throughout, usually written as Mbps. MB is megabytes, and a megabyte is 8 megabits, so it gets confusing . Particularly as connection speeds are always Mbps and speed test results usually MBps).

Thanks very much for getting back to me. Point taken about Mbps, and about the problems probably being separate. I will be able to check the wired connection soon anyway, which should help clarify that issue.

Regarding extensions, there are 4 additional extensions around the house. There is also a fifth, which looks like a master socket, which I think was installed 15 years ago right at the top of the house, for broadband on a separate line, but I think it is defunct now. Can't get a dial tone from it, doesn't work for anything.

There is one additional phone plugged into one of the extensions, and a new microfilter from the fibre starting pack is being used there.

Shall I go and check for bellwire in all the extensions? And remove 1 line each from 2 and 5 at the master socket to determine which extensions are connected where?

Thanks again.

Don't remove any wires! Putting them back requires a special tool if you want to avoid risking wrecking those connectors.

The rest needs thinking about, and I have to go out for a while.

I assume no burglar alarm?

Don't forget that link to a checker .

OK, I'm leaving the wires for the moment.

New information:

1. I found a fifth extension being used for the Sky box. At this extension, a BT splitter is being used for some reason, with two vacant slots on the front of it. As in, the cable from the Sky box goes into the bottom of the splitter, which is plugged in, and neither of the other two slots are being used.

I tried removing the splitter (leaving the socket empty and Sky unplugged), and something weird happens each time. First of all, the broadband connection drops for a minute. Then, when it comes back, I run a speed test and the ping and the download speed are the same, but the upload speed halves. from about 1.9 Mbps down to 1.0 Mbps. I have done this cycle twice and it was the same each time.

2. I checked another extension point and there are two wires in each of 2 and 5. I removed the bellwire.

3. We do have a burglar alarm, but it is on a separate line. I don't know if this could have anything to do with it.

4. I tried the checker you linked me to (tried it yesterday as found the same link elsewhere), and it gives the following response:

"There is no data available for this number. This could be either because it is not a BT line or it is a new BT number that has just been provided. Most new numbers will appear on the checker 24 hours after BT has installed the line."

However, I also tried the checker at this link: http://www.productsandservices.bt.com/products/broad...

And it tells me:

"Yes! You can get superfast BT Infinity
FIBRE BROADBAND
You could get
22Mb-29Mb
Estimated download speed range*
20Mb
Minimum guaranteed speed*"


Urgh. Thanks again... and no problem, I can wait for when you're back.

Tried the address checker linked at the bottom of the same page (on the BT adsl checker) and this time it worked and gave me essentially the same results as that other page.


" Downstream Line Rate(Mbps) Upstream Line Rate(Mbps)
High Low High Low
VDSL Range A (Clean) 29.2 22.4 3.8 2.2 -- Available --
VDSL Range B (Impacted) 20.7 10 3.4 1.1 -- Available --"

Latest update:

- I realised the Sky input lacked a microflter, so I tried that with a microfilter, and without, and in the test socket and in the faceplate etcetc. No significant differences, but I noticed it wasn't connecting to broadband anyway (and why would it, since it's only plugged into a normal extension socket?), so I've just disconnected the Sky completely.

- I went around the house and disconnected all bellwire.

- In doing so, I understood the configuration of the extensions. It is a bit illogical, due to locations as well as the order in which the extensions have been added.

Essentially:

One of the two sets of wire at the main socket connects to the extension where the Sky was connected, which is the only extension on that loop.

The other four extensions are connected in series (the path loops along the house and back again in crazy fashion, but I have identified the last extension on the circuit (as it has only one set of wire - one wire each at 2 and 5), which is also the most recent extension, at the top of the house.

- With all of this done, I looked at the WAN speed again when plugged into the faceplate, and it had risen slightly from ~7 Mbps to exactly 8015 Kbps. Download speed on laptop was still about 6.8 Mbps and upload had stabilised back to about 1.9 Mbps.

- Now I have reconnected the router to the test socket. The WAN speed (from router specs) is back up to 25,104 Kbps, while speed tests on my computer are showing exactly the same results as before - download of 6.8 Mbps, upload 1.9 Kbps.


I am leaving it like this for a while to see if anything changes, but I assume the fact that the router is showing 25 Mbps means that it can't be anything to do with my ISP down-regulating me for frequent disconnections?

In which case, I'm stumped again. Bellwire gone, nothing else connected to extensions, Sky box unplugged, and still the WAN speed when router is connected to the faceplate is 30% of that when connected to test socket. The faceplate slots straight into the test socket, everything's disconnected and the router's microfiltered... what could possibly make such a massive difference...? Even if there's old wiring elsewhere in the house, why would it make a difference? Surely the signal doesn't need to go through any of that?

I'm going to lug a desktop down to the router to try a wired connection and check the wireless, but apart from that, I'm kind of inclined to disconnect all the extensions one by one (since we don't desperately need them, at least not at the moment) and see what happens... If it does come to that, and there is no internal network to speak of, and it still doesn't solve the problem, could it be a fault in the master socket itself?

For now, I'll wait and just try the wired test.

Thanks.

Even with bell wire removed the effect of extension wiring can still be significant hence why fitting a VDSL faceplate to isolate the effect of the extensions is a common thing.

As for why speeds are not better, its not clear if you are referring to wireless or over Ethernet speeds.

Yes, I know I've written a lot and for someone who isn't here it could be confusing.

I am indeed concerned about how the wireless is functioning, but I think it might be a separate issue from the issue concerning the internal wiring. I haven't yet been able to check a wired connection, but will do this evening hopefully, and that should cast some light on the wireless situation.

However, the difference between connecting the router to the faceplate and to the test socket is currently so great (WAN speed: 8 Mbps vs 25 Mbps) that I am most interested in resolving that.

I think fitting such a faceplate might be a good idea, since I have tidied up the network in every other way (easily) possible, short of disconnecting extensions. Do you know if I can fit that kind of faceplate myself (technically, legally, etc....)?

Thanks for your help.

In reply to a post by canary107:

I think fitting such a faceplate might be a good idea, since I have tidied up the network in every other way (easily) possible, short of disconnecting extensions. Do you know if I can fit that kind of faceplate myself (technically, legally, etc....)?

Yes you can. It's designed so you can.

You want a Mk3 Openreach VDSL interstitial filter plate. (Ignore the crazy price on the first one, something badly wrong there). The second one at £12.97 is the one.

You can see in the first pic how it fits, though it doesn't show your wires. Dead simple. Just remove the faceplate as before, with its wires still attached, plug the new Mk3 filter into the test socket, and put the faceplate on the front of the result. There are gaps for your wires to go through.

The Mk3 should come with longer screws to go through the combination. If there are two types of screw, use the ones like you already have.

Then you don't need dangly filters anywhere. But you can't plug your router in to any extension socket.

Edited by RobertoS (Thu 29-Sep-16 18:31:42)

Is the benefit of this faceplate only that it eliminates the need for microfilters, so that it will solve the problem if and only if the microfilters were the problem? If so, I guess it's unlikely to make a difference, since all my microfilters can't be so badly and identically faulty. I will still try this of course.

I wonder if someone could also explain why there is a difference between the test socket and master faceplate socket at all? All the advice is that routers should be plugged in at the master socket, so that extension wiring doesn't interfere with the signal. But then why is there any difference between the test socket and the normal master socket...? Particularly if there are no other devices plugged in anywhere in the house. That advice makes it sound like interference from extension wiring is cut out by plugging in at the master.

You may have read some incorrect advice - there's plenty of it around . Or got confused by more complex scenarios where the writer has not been clear.

The faceplate has the extension wiring attached, as you know. By removing it, and plugging into the test socket behind, you eliminate (in most cases) the extension wiring.

I say "in most cases" because there is a setup, usually historic, where there are extensions connected to junction boxes between the master socket and the point the incoming wire reaches the property. In this case removing the faceplate doesn't eliminate extensions.

That's why I asked you about all the extensions earlier, to see if that was a possibility. It's called star wiring. It is bad for ADSLx and often disastrous for FTTC. One of the main reasons for the years when only engineer installs were allowed was so they could detect and eliminate it.

It doesn't even matter if the extensions have nothing connected - their mere existence can cause trouble.

This interstitial faceplate has an inbuilt filter which protects the broadband side from any problems caused by extension sockets. Though even then there can occasionally be problems, making the test socket the only clean one.

Re-reading your post, I think what may have confused you is that connection at the master is always best even if there are no "problems" as such. Simply because on most FTTC installations every extra metre of phone-type wiring matters unless you are very close to the cabinet. Particularly if the extension wiring itself is poor quality.

So connecting at the master is a question of overall length and quality, not interference. Interference is what all the filtering is dealing with.

You have probably read about CAT 5 and Cat 6 wiring. Lots of people have that installed, replacing normal phone wire, if they really need to use an extension for broadband. It is also the wiring used for ethernet cabling.

When used to reach a normal extension for broadband the extra distance still counts. When used to carry an ethernet signal from a modem or router then (generally) it doesn't.

I hope that helps.

Edited by RobertoS (Thu 29-Sep-16 19:43:08)

In reply to a post by canary107:

Is the benefit of this faceplate only that it eliminates the need for microfilters, so that it will solve the problem if and only if the microfilters were the problem? If so, I guess it's unlikely to make a difference, since all my microfilters can't be so badly and identically faulty. I will still try this of course.

It only takes one to be faulty...

Faceplate eliminates need from filters and stops interference picked up by the extension wiring from affecting the VDSL2, so its a two fold benefit

Wires that are connected even with nothing plugged in are great at picking up radio interference (hence why many FM radios come with a length of simple wire to pick up the radio signal)

Maybe I've misunderstood you, but my point was that I've tried different microfilters all over, so for it to be a microfilter problem, they would all have to be identically faulty.

It may be a misunderstanding.

IanBB was saying that if you had a number of extensions and each had a filter then only one of them needs to be faulty to screw up the line.

I think your saying you have tried replacing your filter with a number of other filters and therefore ruled out it being "the" filter that is the problem.

RobertoS, thanks again for such a comprehensive response.

I don't think we have such an historic setup with junction boxes, as a) I can't see any evidence of it, and b) the WAN speed when plugged into the test socket appears to be exactly as it should be.

I do understand that this faceplate has the potential to solve the problem. MrSaffron's comparison of all that wiring to an aerial makes sense to me, and particularly given the illogical order in which extensions have been added in this house, we really do have a lot of wiring all over the place. Some of it will also be old and probably poor quality.

Anyway, I have ordered the faceplate, and I noticed a lot of reviews talking about exactly the kind of speed uplift that I am looking for, so I'm optimistic.

Thanks again.

I see. Actually that's helpful, because I probably hadn't grasped that one faulty filter elsewhere could have that impact, but it is still unlikely to be the problem here, because I removed all other appliances and filters and tried different filters.

A very common cause of this kind of slowdown is a house alarm. They often have no filter. Just a thought.

Even if it's on a different line??

I even checked it the other day and I was able to call our normal home phone by plugging a phone into the other master socket by the alarm and checking the number calling. It's definitely a different line.

Plus, the fact that the WAN speed when plugged into the test socket is above 25 Mbps according to the router (which is a good fibre speed).

Surely this pretty much rules that out as the cause.

Ok that isn't the problem then. Do you need to have the extensions connected up?

Not necessarily, and I am still willing to experiment with disconnecting. But I have ordered this supplementary faceplate, so I will see how that goes first.

Ok can you connect to the test socket and run a speed test using Ethernet.

Ensure the wireless is turned off on your laptop.

Do you get 25Mbps?

Finally what laptop are you using for the test - it is possible the wireless card inside the laptop is not very good.

I always recommend naming the 5ghz and 2.4ghz networks differently, you should do this as well. If you need help doing this look here

http://community.talktalk.co.uk/t5/Community-Help-Ar...

For now just keep the channel on auto, but change the wireless name for 5ghz.

It might be useful to try the Quiet Line Test on that Alarm Master Socket, as it is likely that its outside wiring runs alongside your other (BB) line for some distance.

Update 1) I did the quiet line test on the alarm line, and it was silent, so that seems OK.

Update 2)

OK, so I have received and installed the interstitial faceplate, and it seems to have solved half the problem. Where previously downstream speeds for WAN were 25 Mbps in the test socket and only 7 Mbps in the faceplate, the router is now plugged into the new, normal faceplate socket and WAN speed is at 21-23 Mbps, which seems fine to me (close enough anyway, particularly accounting for some variability).

However, the other half of the problem still exists. I have done everything you have all suggested and speeds to devices have not improved. I brought down a desktop and connected it to the router by ethernet, I re-ran speed tests on my S5 phone and my laptop, I renamed the wireless 5G strand (and I did tests separately, so tested the ethernet with all wireless devices disconnected), and speeds are still the same. All between 6.5-6.8 Mbps.

So at this point, I'm thinking the router must be the issue. If the router is getting speeds of 21-25 Mbps from the exchange (so interference from internal wiring has clearly been eliminated), and even a computer on a wired ethernet connection is only getting 6.8 Mbps, there must be something wrong with the router's output, right?

For the record, my wireless card is the Intel Centrino N 6235, which actually does have some issues (probably driver related) and occasionally crashes bringing down the computer with it. But mostly it works OK, and it certainly should allow sufficiently high speeds. And in any case, three devices are getting exactly the same speeds, including a different computer on a wired connection.

Is it time to contact TalkTalk and demand a new router...?

Oh, and in case it might help, I've pasted most of the router/wireless settings here that seem potentially relevant:

DSL Connection Status

Internet Connection
DSL synchronization status: Up
Connection status: Showtime

Line Quality
Upstream line rate (kbit/s): 2000
Downstream line rate (kbit/s): 21462
Upstream noise safety coefficient (dB): 13.4
Downstream noise safety coefficient (dB): 6.1
Upstream interleave depth: 1
Downstream interleave depth: 2417
Line standard: VDSL2
Upstream line attenuation (dB): 11
Downstream line attenuation (dB): 25
Upstream output power (dBm): 6.9
Downstream output power (dBm): 10.3
Channel type: Interleaved
DSL up-time: 0 days 5 hours 59 minutes 53 seconds

System Information

System Information
Product type: HG633
Hardware version: H.1.01
Software version: v1.15t
System up time: 5 days 1 hour 18 minutes 13 seconds

2.4 GHz Frequency Band
ON/OFF: ON
Security mode: WPA-PSK/WPA2-PSK
Channel: Auto(9)
Transmit power (percentage): 100%
Region: Great Britain

5 GHz Frequency Band
ON/OFF: ON
Security mode: WPA-PSK/WPA2-PSK
Channel: Auto(36)
Transmit power (percentage): 100%
Region: Great Britain

Broadband Information
DSL synchronization status: Up
Connection status: Showtime
Upstream line rate (kbit/s): 2000
Downstream line rate (kbit/s): 21462
Maximum upstream rate (kbit/s): 4780
Maximum downstream rate (kbit/s): 31200
Upstream noise safety coefficient (dB): 13.4
Downstream noise safety coefficient (dB): 6.1
Upstream interleave depth: 1
Downstream interleave depth: 2417
Line standard: VDSL2
Upstream line attenuation (dB): 11
Downstream line attenuation (dB): 25
Upstream output power (dBm): 6.9
Downstream output power (dBm): 10.3
Channel type: Interleaved
DSL up-time: 0 days 6 hours 30 minutes 13 seconds

Beyond that, a few other settings that might be relevant:

- DDNS is not enabled
- All 3 dynamic routing options are 'Passive' and 'Off'
- Operating mode: Auto
Current connection type: VDSL

Then there are some statistics for LAN, Wireless and Internet, with a table in each section with columns for Interface, and Received/Sent Byte, Packet, Error, Discarded. I can paste these if useful.

Thanks everybody.

One last check that I can think of. I hope you still have the desktop machine down there.

It often happens that after connecting wired, while wireless is running on that machine, that it takes a while for the ethernet to take over the connection from the wireless. If it ever does. I've been caught out a few times that way.

When I want to swap, I now make sure I either disable the wireless on the computer or go into the Network page in Control Panel and make sure I disconnect it there and that the ethernet is in charge.

Ha, thanks for the suggestion, but that's definitely not the problem, as the machine doesn't have wireless...

I'm teetering on the edge of calling in, but might wait a bit for any other suggestions first.

Just can't see how it can be anything other than the router now, unless I fundamentally misunderstand this system (which is admittedly perfectly possible...).

If the router is getting the right speeds from the exchange, and 3 separate devices, connected both by wireless and cable are all getting similar and severely restricted speeds, that seems to point to a router problem to me...

Try turning off the broadband hardware for an hour.

Why this might work, is that some providers use a mirror of a 'connection speed' setting and limit traffic based on that. This normally affects PlusNet users, but believe you are with TalkTalk and so this should not be a factor, but it is possible that all the kit at your end is now OK and needs the providers end to notice the connection speed change.

An hour off, gives their systems time to expire your connected session and relearn you when you connect again.

I'm going to do this shortly. Only thing is, wouldn't the limit be noticeable already in the WAN speed? I'd have thought they'd limit the speed to the router, rather than using the router to 'filter' it out.

But I will try.

Does that mean you are normally using powerline connectors to get to the desktop computer? If so, unplug them all and see what happens with a wired connection then.

We actually just used to have the router plugged into an extension which was next to this desktop. More recently, the desktop hasn't been connected at all while I try to sort this out. The plan is to wire an ethernet extension to it once this is all resolved, or else use powerline if it can be done without disruption.

We do also have powerline extensions in a couple of places for other things, but those have been disconnected/switched off from the start of my experiments. At this point, they aren't even plugged in.

No, this is a limit placed at the IP level, for BT Wholesale based connections (which you are not I believe) this is called the IP Profile.

You amazing human being! It's worked. Download speed up to 20.2 Mbps. That's close enough for me (1.2 Mbps down on the WAN speed).

Upload is 1.84, which is also as it should be (although it was already there before. Incidentally I don't understand why it would have worked right in that direction before but not the other way, but it doesn't matter.

Thank you all so much, and particularly MrSaffron, RobertoS and Ukhardy07.

Very pleased for you

The 2.4Ghz WiFi on the HG633 isn't great so stick to 5Ghz when you can.

My experience with Talktalk has been overall really good. Customer service is dire but the connection I personally find performs better than my BT one, it weirdly just "feels" a lot faster to load pages. I maintain both a BT & talktalk connection.

Excellent!

Thanks, I will do.

I've had mixed feelings about TalkTalk as there has been a bit of sneaky behaviour with new contracts over the past couple of years, as well as dodgy customer service. However, I do appreciate that when I've had legitimate issues recently, they've been quick to offer decent upgrades and perks. Think they've also tried to improve the customer service actually.

And so far the fibre is a massive improvement on the broadband previously. Steady and fast.