Broadband Speed Miles Off Minimum Guaranteed

[TreborMint]

Hello There!

Does anyone have any advice for contacting ee in regards to my broadband speed being a long way off the minimum guaranteed? We commenced the service in our new rented home two weeks ago so have given plenty of time for everything to settle down but we are only getting between 18-20Mbps download against a minimum guarantee of 66.8Mpbs. Furthermore even now after the bedding in period the connection seems to drop / router restarts quite often.

Connection to the Smarthub is by ethernet, which is then in turn connected to the mastersocket. I think this is quite an old version, which has "ADSL v1.0" printed in the bottom right corner:

Next to this is  a smaller socket just for the phone. There is no connection from the front of the master socket to this. There are a lot of other sockets for phones around the house which appear to be fed from a splitter that plugs into the front of this additional phone only socket. No devices are plugged into any of the sockets around the property apart from the Smarthub into the master socket.

Our Smarthub stats are:

Our speedtest is on this link: https://www.thinkbroadband.com/speedtest/1623140660394579255

The BT availability for our property is:

Whilst the £5 off per month for not meeting the guaranteed speed would be welcome I would much rather be getting a quicker and more reliable service. Before I contact Customer Services any advice as to how best for me to encourage ee to improve the speeds rather than just offering me the money off would be much appreciated!

Many thanks for your help!

Rob

[mikeliuk]

Here is a good reference: https://www.speedguide.net/faq/what-is-considered-good-dsl-noise-margin-snr-355

Basically, higher is better but it will be targeted down over the course of several days to increase bandwidth available, subject to the line remaining stable.

This is also another good explanation although the page appears somewhat dated: http://www.robertos.me.uk/html/noise_margin-snr-snrm.html

Here is some random discussion perhaps including some non-distinguished network engineers: https://networkengineering.stackexchange.com/questions/43791/what-is-snr-margin-and-line-attenuation

Edit: I think the key to understanding the noise margin is to distinguish between the target and the instantaneous value. Instantaneously higher is better but might also indicate a wider issue. An analogy is that I leave a lot of space between my car and the car in front. At all instances, a larger value is better because it is safer but a large value may indicate I'm an erratic driver who needs the large space to avoid crashing. On the other hand, if the large space persists throughout the day, I may lower my target following distance overnight so that I drive closer to cars the next day. Again, at all instances the next day, a higher value will be better and shows I might be able to reduce the distance the next day. But if it is observed during the day the following distance frequently fluctuates from a good high value to near zero, I may need to keep my target value high merely to avoid crashing the next day.

Edit2: the connection speed of 20 Mbps to 10 Mbps can be expected between 1.6 km to 2.8 km from the cabinet. https://improvebroadbandspeed.co.uk/guides/speed-vs-distance-from-cabinet-for-fibre-broadband-fttc/ If the distance is known to be shorter, there may be a problem.

Edit3: here is a reference from Cisco which is a recognized authority in network engineering: https://www.cisco.com/c/en/us/support/docs/long-reach-ethernet-lre-digital-subscriber-line-xdsl/lre-...

Edit4:

Noise margin (aka SNR margin) = SNR - line attenuation

For fixed line attenuation, noise margin rises with SNR, and higher is better for both. I don't think anyone would argue that low SNR is ever good, or that high line attenuation is ever good. This is the why instantaneously higher noise margin is better at that point in time.

https://hellocomtec.com/collab/knowledge-base/article/adsl-speeds-snr-and-attenuation

Edit5: clarifying based on below input that in Edit4 line attenuation is non-zero and therefore noise margin is not the same as SNR. No intention was meant to claim equality and my apologies if this was unclear above.

[pip11]

@mikeliukYou obviously have no grasp of the difference between noise margin and SNR. They are not the same. You obviously have no idea how DLM works either.

Posting google search drivel as in your above post helps no-one.

[mikeliuk]

Hi @pip11 ,

Many thanks for your input here. I'm always open to learning and correcting my mistaken understanding of things.

Please may you help me to understand what I have posted above is wrong or unclear?

(For example, in the above I show an equation where the noise margin is derived as an expression involving the SNR so I don't see any claim these two things are the same? Quite the opposite of being the same when one is derived as a non-identity expression involving the other.)

Edit: perhaps you mistakenly took the minus sign to be a hyphen and thought I was equating noise margin with SNR when that was not at all my intention?

[IT-4BusyPeople]

Thanks @XRaySpeX - I should of paid more attention to the status information 🙂

also thanks to everyone for the discussion.

With all the discussion about S/N etc. I think we might be overlooking something: My understanding of xDSL is that TX & RX speeds are negotiated as a pair. For EE this ratio is 4:1, so when I got noise on the line both rates got negotiated down. However, @TreborMint isn't seeing this, their RX rate indicates not only that they should be getting a TX rate of  76Mbps, but that the equipment at the other end of the line believes this should also be the case and thus has negotiated TX/RX rates accordingly.

I hope @TreborMint has contacted EE CS and arranged an engineer to visit.

From a web search, I located (on TomsHardware) a query about poor downlink and good uplink speeds, the offered advice was to replace the consumer modem/router ie. faulty hardware. Unfortunately, the actual solution wasn't posted and I doubt @TreborMint has a spare modem to plug-in...

[mikeliuk]

Agreed that something seems off because the upstream of 20 Mbps is right at the top band and might be expected at 150 m to the cabinet equating to roughly 76 Mbps downstream as you say. On the other hand, the actual downstream below 20 Mbps puts a cabinet distance estimate of over 1 km away. (Again the weirdness that upstream performs better than downstream which is the reverse of most people's experience except for when the network is congested with everyone pulling data in one direction, i.e. down.)

The good thing is that it's usually relatively easy for engineers to determine a root cause for something which looks impossible (i.e. why should a wire or fibre perform so differently compared to the maximum theoretical rate in each direction absent congestion in the network).

https://www.thinkbroadband.com/guides/fibre-fttc-ftth-broadband-guide

https://kitz.co.uk/adsl/DLM_profiles.htm

[XRaySpeX]

@pip11 : Thanks for your support.

[TreborMint]

Hello all, thanks for your input and sorry for not responding until now. 

I haven’t contacted ee cs yet. Before I do I’m going to try to get the dsl-Ac68u I was using before the modem firmware went funny going so as to rule out a faulty smart hub, but if that doesn’t work I’ve an ee bright box here as well. Could that work?

Also is it worth me trying with the modem connected through a filter plugged into the first face plate in the property? This unit (pic earlier in the thread) doesn’t look like a master socket though, and I’m a bit confused as to why it isn’t because I thought the first faceplate should always be a master socket. 

Really appreciate your help so far 🙂

[XRaySpeX]

A Brightbox 2 will work on Fibre; a Brightbox 1 won't,  being for ADSL only.

It's got an internal test socket -> It's a master socket.

[Mustrum]

Just to add, the Asus AC68 has one of the worst modems you could use.

But I suspect your line will now be banded, and need Openreach intervention to restore - they should also sort out your internal wiring - you will need to ensure you know where you want the master socket to be, if the one you are using is not the first one in chain.

By all means try the BB2 and post the stats from a dangly filter, aven the AC68, but unless you see improvements then the call to EE needs to happen sooner than later.