Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Black Sheep]

By way of a short story ..... I did a repair on a local VDSL circuit on Friday morning. They were running at 27Mbps with a crackly phone line.

Fix completed and I performed a full DLM reset ..... their router came back into synch at 47Mbps. I had a slight niggle that they were closer to the Cab for that speed, and so decided to check again via a remote test, yesterday. They were now in synch at 76Mbps !!!

Have you ever seen those cartoons where their eyes go like saucers and nearly pop out of their head !! 

So, maybe things will improve for your circuit in time ??

[TheVoid]

Just did another speed test and it's shot up to 60mbps! I guess it's going to take time to get to the highest level.

Looking at those stats above, it looks like my line can go up to 80mbps though right?

[TheVoid]

Here's a photo of my master socket. It was like this when we moved in 3 years ago. Looks a bit bulky to me! Ignore the black box below it, it's my mobile phone signal booster.

[Black Sheep]

Yes, they're bulky-looking ... but it's exactly the kind of socket I'd want to see on a DSL installation.

We have moved to a MK4 SSFP now, but believe me, the MK3 you have is more than adequate.

Others may be able to answer your question around your stats, and further speed increases ??

[re0]

Just did another speed test and it's shot up to 60mbps!

I would say that a one-off speedtest is probably not the best way to gauge any improvement since there are so many factors that can impact the reliability of a speedtest (such as computer specifications, other users on the network and even the web browser in use). In this case, we should stick to monitoring the sync speed and statistics from the router directly.

I guess it's going to take time to get to the highest level.

There is no real time frame that can be applied to how long it will take for your connection to reach its potential. It depends on the DLM and whether your line is within the defined thresholds for improvement. Since SRA (Seamless Rate Adaption) is not used on xDSL products that are on Openreach's network, it's not exactly a smooth transition upwards towards 80 Mbps but rather the link (connection between your router and the hardware in the cabinet) will be dropped (usually early hours in the morning) and the connection speed will be renegotiated with the defined parameters from the DLM if any changes are due.

Looking at those stats above, it looks like my line can go up to 80mbps though right?

Please bear in mind that the quoted 80 Mbps on the DSL Checker refers to SYNCHRONISATION SPEED as opposed to ACTUAL THROUGHPUT on the network. Actual throughput while synchronised at the full 80 Mbps under good coniditions will be around 75 Mbps as there will be a small overhead.

The key words used here are "up to". It depends on so many different conditions. So you may be able to synchronise at 80 Mbps one day, and after a rescync maybe 76 or 77 Mbps. Seeing as you have stated you have received 75 Mbps in the past, if you were referring to actual throughput (through downloads and speedtests) then there was a good chance that you were synchronised at or near 80 Mbps.

Edited to clarify some information.

[TheVoid]

Well my line stats seem to have improved which would explain the jump to 60mbps, however frequent speed tests are still saying an average of 49mbps with the occasional result at 60mbps...

Data rate:
20.00 Mbps / 65.36 Mbps

Maximum data rate:
23937 / 66324

Noise margin:
15.3 dB / 6.1 dB

Line attenuation:
14.7 dB

[re0]

I can see your data rate has improved by around 7 Mbps. It looks like your SNRM target is still 6 dB, but it is quite difficult to see whether the DLM has made any positive changes to your line thus far since you do not have access to in-depth statistics on the Smarthub. It may be the case that the DLM has adjusted some parameters outside of SNRM (which is the only one we can really obervse on your device); perhaps G.INP has been enabled since that approx. 10% jump in sync and attainable (maximum data rate) seems rather suspect to me unless it is simply caused by variations in conditions throughout the day.

I once read that there was a way to tell whether G.INP had been enabled through the IP profile, but perhaps someone here can confirm or debunk this myth?

60 Mbps throughput would be about right for the synchronisation speed. If you're seeing an average of 49 Mbps then just make sure that you are doing your testing over a wired connection (wireless can support such a speed quite reliably over short distances, but there is always risk of interference and other devices on the network using bandwitdth). Also, you should bear in mind that other factors outside of your control can play a part in the throughput you see such as congested backhaul links that your ISP uses which may - at peak times - see lower speeds.

Out of interest, which speedtest facility are you using?

[TheVoid]

Out of interest, which speedtest facility are you using?

I used Speedtest.net

[Moderator edited to remove the unnecessary bulk of the quotation, just leaving the question to which the answer has now been given.]

[re0]

Might be worth trying https://www.thinkbroadband.com/speedtest as it shows both single- and multi-threaded test results. http://www.dslreports.com/speedtest has become another favourite of mine since it seems to be fairly reliable and shows BufferBloat statistics along with ping times to different locations. But if you notice the same variance as with speedtest.net, then perhaps it's because one or more of the issues I have previously highlighted.

[TheVoid]

Still only getting 49mbps - 59mbps. Not sure paying £4 a month extra warrents the very slight increase in speed?

[re0]

Still only getting 49mbps - 59mbps. Not sure paying £4 a month extra warrents the very slight increase in speed?

In a lot of cases, just the increase in upstream is worth it alone for the extra cost. So, while the increase in downstream may be very slight, if you do a lot of uploading the 20 Mbps will be a big help compared to 10 Mbps (or rather, under good circumstances, ~18 Mbps compared to ~9 Mbps in terms of actual throughput).

As @gt94sss2 said previously in relation to a "boost visit":

If you want to see if BT will agree to a boost visit it has to be requested within 28 days I believe.

As I stated previously, there is no defined period of time before your connection reaches its full potential (since we are essentially waiting for G.INP to be enabled here, which has been known to take many weeks). If you think that you should be receiving more, then perhaps you should follow the advice of @gt94sss2. Otherwise, the best advice is to just sit tight but keep an eye on the stats for improvements.

[Alex Atkin UK]

Still only getting 49mbps - 59mbps. Not sure paying £4 a month extra warrents the very slight increase in speed?

Speed tests are notoriously unreliable, I have seen consistently low results before on my line while still hitting 76Mbit in real-world usage.

Have you tried http://www.dslreports.com/speedtest as it tends to make the most effort to push the test as hard as possible.

Plus as others have said, you haven't even give the line a chance to stabalise yet.  You could yet end up at full sync.

[re0]

I think calling a 25-50% increase on download and a 100% one on upload a "very slight" increase is a bit harsh.

In the OP's case, the increase on the downstream is about 13% after upgrading. Thought I can totally agree that 100% uplift on the upstream is not very slight at all.

Speed tests are notoriously unreliable, I have seen consistently low results before on my line while still hitting 76Mbit in real-world usage.

I can agree to a degree. They are only reliable and will show the max throughput if there is sufficient bandwidth from the host and in transit (through backhaul), and/or if the computer is able to reliably run the speedtesting facilities. Speedtests are only good and accurate at measuring throughput if the conditions are right.

Have you tried http://www.dslreports.com/speedtest as it tends to make the most effort to push the test as hard as possible.

I have previously suggested this as I have found it useful for measuring other metrics such as latency and BufferBloat.

[Alex Atkin UK]

I just find that sites like speedtest.net and thinkbroadband speed test have had issues with my connection, while practically every other real-world website has obtained faster speeds.  The speedtest.net app especially seems to be bottlenecked by my phone somehow, as running iperf3 on the same phone can get 300Mbit to a local computer over the WiFi, so it really shouldn't be bottlenecking to the Internet.

[j0hn]

Speedtest.net consistently gives me speed test results that are higher than my actual sync. Seeing as though that is impossible, I don't believe their results any more.

ThinkBroadband's speed test has good routing to most ISP's. Their test gives single thread and multi thread results.
It's never once given me a result higher than what my IP profile is.
It's by far the most accurate/reliable test for myself.