cancel
Showing results for 
Search instead for 
Did you mean: 

Does 'load' make a difference to connection stability ?

GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Does 'load' make a difference to connection stability ?

Hi,

i have recently moved to PN fibre from a bonded TT and SKy ADSL setup as BT have recently brought fibre to the local cabinet.  We went live on Thursday and started off with a disappointing 3Mb down with 10Mb uploads.  The line also kept resetting itself.  I have reported this and am waiting for someone in the PN FTTC? Team to pick up the ticket.

 

As I still have a live Sky BB link into the house I swapped the Ethernet connection back to Sky to give everyone in the house reliable access. However if I logon to the PN router via Wi-Fi I notice that I have a 40Mb download connection. 

Over the last few days I have tried swapping the main network connection between SKy and the new PN router.  If the router is only being us by one tablet or phone via Wi-it gives a speed of 40Mb or more - as soon as I connect the home network to it this drops to 2 or 3 Mb and becomes unreliable.

There are about 20 devices on my home LAN most connected by Ethernet including two separate Wi-Fi access points as it is a largish house.

Any ideas of anything I can try ahead of more formal help from the PN support team ?

 

 

 

10 REPLIES 10
MrSilver
Pro
Posts: 550
Thanks: 82
Fixes: 9
Registered: ‎05-10-2016

Re: Does 'load' make a difference to connection stability ?

Strange one, not heard of anything like that.

Could you introduce some of the network at a time, e.g. Just some of the wifi devices?

There was a problem with amazon echos or chromecasts or something like that which was killing the hubs until they got patched.

Others check is how much data is the hub using when it's slow? Could it being saturated with downloads from updates now it's on a faster line that wasn't as easy to see before?

Do you have a switch you can mirror a port on to check with wireshark what's going on the network?

Finally, do you have access or a friend you can borrow from another router to test with?
Anonymous
Not applicable

Re: Does 'load' make a difference to connection stability ?

@GrahamMackie - A couple of other questions to add to those of @MrSilver.

Out of curiosity what Fibre package are you on?

Also when you say it drops to 2 / 3Mb how do you know this?

GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Re: Does 'load' make a difference to connection stability ?

Hi

 

Thanks for the responses.  Having spent most of the holiday break trying to improve things I have come to the conclusion that the issues are so random that my 'load' theory was just a coincidence as I have managed  to get 45Mbps or so at various times over the long weekend.  However the line does spend more time around 6 or 7Mbps than it does at 40Mbps and often drops to 2 or 3 Mbps.

 

I am using the BT Wholesale speed checker to check speeds but also Speedtest.net.  What is clear is that the line is very unstable and resets every few hours or so.  the longest it has stayed up is 13 hours however it is more usual for it to reset every two or three hours.  The reset is always preceded by the following message in the event log:

PPP LCP Send Termination Request [User request]

I realise that this constant resetting doesn't help with line speed.  Hopefully the FTTC team at Plusnet will be able to do something as my decision to move ISP is not playing down well at home at the moment ! 

 

MrSilver
Pro
Posts: 550
Thanks: 82
Fixes: 9
Registered: ‎05-10-2016

Re: Does 'load' make a difference to connection stability ?

I am guessing you are on the 40 download 2mb upload product?
GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Re: Does 'load' make a difference to connection stability ?

No - I'm on the Superfast product so should be getting up to 70. However out neighbours moved to FTTC before us and they are getting 60 down and about 20 up.
Anonymous
Not applicable

Re: Does 'load' make a difference to connection stability ?

@GrahamMackie, It may be that other devices on the LAN are using the bandwidth, so unless you can do a speed test with everything apart from a single client disconnected the results you see for the speed test may be correct at the time the test was carried out.

Are you able to do this, then connect the devices one at a time to if any of them have an adverse effect on your speed. If you have any Windows machines I'd be inclined to start them first. Also, is there any pattern to this behaviour?

Also, your mention of a bonded TT and Sky BB links how can this work with two different vendors? What router do you have BTW?

GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Re: Does 'load' make a difference to connection stability ?

Hi,

 

I spent the weekend turning off everything and adding devices one by one - at one point I had all devices connected and was getting about 40 Mb download but after about 20 minutes it went back to 3 Mb.  PN team have now acknowledged a fault and have reported to Open reach.

I have a standard PlusNet Hub One which was supplied when I ordered the line.

Prior to this I have been using Sharedband which bonds two DSL circuits together.  We were 4+ miles from the Exchange on copper so the best we could get was 3Mb so I bonded a TT and a Sky circuit to get 6Mb down.  However this is a very expensive way of doing it as the Sharedband service with unlimited data was £50 per month plus the cost of the two DSL circuits.  

 

regards

 

 

GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Re: Does 'load' make a difference to connection stability ?

Looks like the FTTC team are on the case now and have posted some info against my fault ticket.

 

BD
BRAS adsl35000 Profile Info N/A
ND509 - Potential performance issue has been identified as performance test has failed. Please pass to BTW for further diagnostics.
PTTR
Type Date Direction Profile Speed
TAP-1 1-1-2017 (12:54) DownStream 42215 6580
TAP-1 1-1-2017 (12:54) UpStream 20000 15490
GEA Service Test
Test Outcome Pass
Test Outcome Code GTC_FTTC_SERVICE_0001
Description GEA service test completed and no fault found but unable to check for customer equipment connected to modem.
Main Fault Location OK
Appointment Required N
NTE Power Status PowerOn
Sync Status In Sync
Downstream Speed 35.8
Upstream Speed 18.6
Profile Name 0.128M-80M Downstream, Interleaving Low - 0.128M-20M Upstream, Error Protection Off
Copper Line Test Pass
Bridge Tap Not Detected
Radio Frequency Ingress Not Detected
REIN Not Detected
Cross Talk

Not Detected

 

Looks like it is going to be passed the BT SFI team for invetsigation

GrahamMackie
Hooked
Posts: 9
Registered: ‎31-12-2016

Re: Does 'load' make a difference to connection stability ?

The BT engineer was here for a few hours today and has checked the circuit back to the DP and then back to the cabinet. He says that there is nothing wrong with the circuit itself. So the issue must be within my home.

He removed the existing dual faceplate and we tested the circuit using the test port and achieved 30Mb or so. He replaced the faceplate and it dropped back to 6Mb so he has replaced the faceplate with a new NTE5C 'single port' faceplate and the circuit is now using a microfilter which it didn't need before. The engineer didn't have any dual phone/DSL faceplates on his van.

Before he left we tested the circuit and we were getting 9Mb down. He said that it should 'train itself' back up again but if it didn't that I was to contact PlusNet and ask for a new router. Within minutes of him leaving my wife complained that the internet was down. I have since tested it a few times and it is currently running at about 2Mb down with no upload.

The BT engineer didn't have any explanation for the frequent dropouts. However he did have the 30Mb 'cap' removed from the line.

Since he left I have been using the BT Wholeslae speed test site and speedtest.net to test the line speed. Looking at the historic results from Speedtest over the last week or so it seems that the only time I get 'good' line speeds is when I am connected to an IP address starting with 51.6. On every occasion where I have logged poor line speeds the IP address doesn't start with 51.6.

Examples of 'slower' IP addresses include:

143.159.223.227 - 4Mbps
37.152.194.81 - 2Mbps
195.99.43.121 - 3.6Mbps

Again this could be a red herring but at the moment the router is connected to 51.6.107.121 and is giving me 31Mbps with just one client PC connected to it via a single ethernet cable.

MrSilver
Pro
Posts: 550
Thanks: 82
Fixes: 9
Registered: ‎05-10-2016

Re: Does 'load' make a difference to connection stability ?

This is a slim chance that when you get an IP speedtest doesnt know its picking a test server that is miles away, unlikely as it usually works. To test manually select "new server" and pick a London based one like voda is pretty good usually.

In addition, try speedof.me and fast.com as other test sites to make sure its not a funnt of speedtest and your IP address.

Main thing to keep an eye on is the sync speed on the router, that shouldnt be dropping down. 

Best one to really try is http://speedtest.btwholesale.com/ do all 3 tests, the third changes your login details but bypasses plusnet so you can check your speed when slow, then test.