Slight FTTC speed drop

[#1]

My VDSL speed has dropped a bit from a measured 37Mbps to 35Mbps

FAST

Yout Internet speed is

35 Mbps

I recently decided to retest the link from my PC to Hub 2 router using my little TP-Link USB WiFi device instead of Ethernet connection, with the router placed near where it might be if ever I went the FTTP route. I moved the router from its normal LJ style extension socket (also a master socket!) to the 'proper' Openreach NTE5C master socket (Oh dear! - shouldn't have done that?).

Since then, I've had a few disconnections and now I find the measured download speed has dropped to 35Mbps from the normal, long term measured 37Mbps (I am on a 40/10 Mbps service).

I knew I shouldn't have touched that NTE5C, normally I leave it strictly alone. Best to not even look at it, in case it takes offence...

Could this 35Mbps speed correspond to any banding of the line? To date I have been impressed with FTTC's long term reliability and stability and I don't want things to go the same way as ADSL 2+.

[#2]

@pvmb A drop from 37 to 35 has nothing to do with 'banding' - if you were subject to that the speed reduction would be much more severe.

Have the disconnections involved a loss  of DSL? Has your SNR  changed?

Is the TP-link device still in play?

[#3]

@pvmb 

Small differences like that could be explained by the fact that many speed testers will dynamically choose what they think is your nearest and least loaded speed test server, before performing the test.

Therefore from test to test, you might be comparing different servers, that will be different distances from you, or be under different loading at different times of day, etc.  So not surprising to see a 2Mbps variation without knowing more information.

I rarely use speed testers, but instead check my modem's line stats, to see how it's performing.

Not having used the "Fast" speed test before, I just ran it to see if it showed which server it used, here it shows "Slough" -

[#4]

@pvmb 

Hmmm,  I'm not impressed with that Fast speed tester.

I just logged in to my router so that I could monitor my WAN and LAN speeds, and ensure I had no other network traffic running

I then ran that Fast tester, TEN times in succession, and got varying results from 63Mbps to 69Mbps.

I know from years of tuning my setup, that variation IS NOT my line or equipment.

Running the SamKnows tester TEN times in succession, gave me consistent results every time within 0.1Mbps.

.

[#5]

You tracking me, @outcast ? Every time I answer a post, you pop up within seconds.

[#6]

If you use the FAST test, using non-default settings (which can be set to operate over minutes, rather than seconds and a bit more).

As far as I can see, SAMknows doesn't have those types of settings.

Using the FAST tester with parameters set at their maximum might indicate that even FTTP connections are not as such. I've a couple of 5 minute tests that fall outside the MGL.

[#7]

I have used the TBB speed test facility ever since I moved to Zen (and a static IP) on both FTTC and FTTP and found it consistent.  I also use speedtest.net  (Ookla), which also gives consistent, and comparable, results.

[#8]

---

@grumble wrote:

If you use the FAST test, using non-default settings (which can be set to operate over minutes, rather than seconds and a bit more).

---

Initially I thought that was interesting, and something new I've learned today which might be useful.

So I went in to my router and setup rate limiters on my WAN up and down traffic, so a restricted bandwidth less than the VDSL IP profile, would give consistent throughput during tests.

As advised, I changed the FAST settings to run for 300 seconds,

Then ran multiple extended duration FAST tests, with results varying from the limiter rate -3Mbps to limiter rate +4Mbps.

Conclusion : USELESS !  🙄

I'll stick with my favourite Waveform test. 😎

.

[#9]

---

@jab1 wrote:

@pvmb A drop from 37 to 35 has nothing to do with 'banding' - if you were subject to that the speed reduction would be much more severe.

Have the disconnections involved a loss  of DSL? Has your SNR  changed?

Is the TP-link device still in play?

---

Yes, the DSL has subsequently dropped a few times (no drop in the last 5 hours) - I blame me going anywhere near the NTE5C. I am still using the TP-Link, until I finish testing. But, when originally set up, I immediately tested the speed and at that time it was the usual 37Mbps. I don't know about any change in the SNR. Currently 18.8 / 16.1 dB which seems about normal from memory.

Interesting! The Think Broadband Test gives 35 Mbps (or slightly over) while reporting "Postcode average for last three months" as 37.3 Mbps/8.8 Mbps. 🤔

[#10]

I doubt the error log will tell us much, but if you could export it and attach the resulting .csv file to a post along with your Technical Log, we might spot something.

[#11]

Strange those quoted TBB speed figures of "average for last three months" correspond pretty well exactly to my usually expected measurements from FAST even though I don't usually use the TBB speed test. Do they have something in common, or do speed testers share figures?

Still showing a FAST measured speed of ~35 Mbps after 15 hours with no drop. One other quirk I have noticed, which may be completely irrelevant, is the sync speeds shown on the Hub 2 home page:

Downstream sync speed: 39.999 Mbps

Upstream sync speed: 10 Mbps

Whereas, normally these would be:

Downstream sync speed: 40 Mbps

Upstream sync speed: 9.999 Mbps

Here are a couple of examples of DSL failure - I have little doubt they were due to the main NTE5C socket:

17:37:58 04/08/25 WAN connection WAN2_INTERNET_PTM connected
17:37:39 04/08/25 DSL Link Up: Down Rate=14361kbps Up Rate=9999kbps; SNR Margin Down=5.9dB Up=5.8dB
17:37:31 04/08/25 WAN Auto-sensing detected port DSL WAN
17:36:46 04/08/25 WAN Sensing Auto sensing Complete interface selected
17:36:46 04/08/25 WAN Sensing Auto sensing Running
17:36:46 04/08/25 WAN connection WAN2_INTERNET_PTM disconnected [ERROR_NO_CARRIER]
17:36:45 04/08/25 DSL Link Down: duration was 67 seconds
17:36:05 04/08/25 WAN Sensing Auto sensing Complete interface selected
17:36:05 04/08/25 WAN Sensing Auto sensing Running
17:36:05 04/08/25 Success - secondary DNS servers
17:36:05 04/08/25 Success - primary DNS servers
17:36:05 04/08/25 WAN connection WAN2_INTERNET_PTM connected
17:35:46 04/08/25 DSL Link Up: Down Rate=39999kbps Up Rate=10000kbps; SNR Margin Down=14.3dB Up=17.1dB
17:35:38 04/08/25 WAN Auto-sensing detected port DSL WAN
17:35:05 04/08/25 DSL Link Down: duration was 31 seconds
17:34:42 04/08/25 DSL Link Up: Down Rate=28134kbps Up Rate=10000kbps; SNR Margin Down=5.4dB Up=13.6dB
17:34:34 04/08/25 WAN Auto-sensing detected port DSL WAN
17:34:02 04/08/25 WAN Sensing Auto sensing Complete interface selected
17:34:01 04/08/25 WAN Sensing Auto sensing Running
17:34:01 04/08/25 WAN connection WAN2_INTERNET_PTM disconnected [ERROR_NO_CARRIER]
17:34:00 04/08/25 DSL Link Down: duration was 31793 seconds
08:44:31 04/08/25 WAN Sensing Auto sensing Complete interface selected
08:44:31 04/08/25 Success - secondary DNS servers
08:44:31 04/08/25 WAN Sensing Auto sensing Running
08:44:31 04/08/25 Success - primary DNS servers
08:44:31 04/08/25 WAN connection WAN2_INTERNET_PTM connected
08:44:15 04/08/25 DSL Link Up: Down Rate=39999kbps Up Rate=10000kbps; SNR Margin Down=15.4dB Up=16.9dB
08:44:07 04/08/25 WAN Auto-sensing detected port DSL WAN
08:43:16 04/08/25 WAN Auto-sensing last connection port: DSL WAN
10:03:07 03/08/25 WAN Sensing Auto sensing Complete interface selected
10:03:07 03/08/25 WAN Sensing Auto sensing Running
10:03:07 03/08/25 Success - secondary DNS servers
10:03:07 03/08/25 Success - primary DNS servers
10:03:07 03/08/25 WAN connection WAN2_INTERNET_PTM connected
10:02:49 03/08/25 DSL Link Up: Down Rate=39999kbps Up Rate=10000kbps; SNR Margin Down=15.4dB Up=16.9dB

Here is a typical past DSL link statistic

22:39:29 31/07/25 DSL Link Up: Down Rate=40000kbps Up Rate=9999kbps; SNR Margin Down=16.5dB Up=14.3dB

I will be ending my test of the remote router positioning soon and see what the results are after I return it to its normal LJ extension phone socket and linked to the PC by Ethernet cable.

I know the drop in speed is quite small, but I remember ADSL 2+ ratcheting down in speed over the years whenever 'something' happened - and never successfully recovering long term.

[#12]

---

@pvmb wrote:

I recently decided to retest the link from my PC to Hub 2 router using my little TP-Link USB WiFi device instead of Ethernet connection, 

---

So instead of testing over Ethernet you are using wireless? This makes any comparisons totally irrelevant. Both tests should be using the same Ethernet cable.

---

@pvmb wrote:

. I moved the router from its normal LJ style extension socket (also a master socket!) to the 'proper' Openreach NTE5C master socket

---

You can’t have two Master sockets, you are normally using an extension. If you want the best possible speed you should locate the hub at the real master socket location, fit a filtered faceplate and move all the extensions to the faceplate filtered output. If necessary run Ethernet cable to your computer location.

When you move to Full Fibre you can have the ONT installed by the master socket then use the Ethernet cable to connect the ONT WAN  to your router, which you can put back in your preferred location.

[#13]

Couple of points.

I do have two master sockets!

The LJ extension socket I normally use is also a master socket - i.e. it has all the componets of a standard master socket. Including ring capacitor and the old surge arrester.

The router is currently plugged in to the NTE5C 'Test' socket - in order to minimise NTE5C issues. The WiFi adaptor is using 5G, it is rated as "867 Mbps on 5 GHz". Unless I am mistaken - which is possible - originally the link tested at 37Mbps before the subsequent DSL drops, which I have no doubt were due to the NTE5C master socket.

But, as I said, I will be moving the router back to the extension socket in a day or so and will see then what happens when it is reconnected via Ethernet. I will report if it reverts to 'normal' 37 Mbps. I hope it does!

[#14]

---

@pvmb wrote:

31/07/25 DSL Link Up: Down Rate=40000kbps Up Rate=9999kbps; SNR Margin Down=16.5dB Up=14.3dB

@pvmb 

With line stats like that, plugging your router in to the master socket (where the external landline comes in) using a short DSL cable, and removing the extension phone wiring (which you won't be needing after the imminent PSTN phone switch-off), if you then upgraded from FTTC to SoGEA, I estimate that when your SNR margin drops to 6.0dB, that your 40/10Mbps line speeds would increase to a minimum of 60/15Mbps, and could reach 70/18Mbps.

Worrying about whether your speed test shows 37Mb or 35Mb would be irrelevant, as your speeds might double !.

As you've previously mentioned that FTTP is available to you, obviously that is the much better choice, with better reliability, better speeds, and at the cost shown below is probably cheaper than what you're paying for you 40/10 FTTC. -

[#15]

Testing completed:

Router with 5GHz WiFi moved from master socket test socket back to location by extension socket and plugged in. Still testing at 35 Mbps download on 5GHz WiFi..

Router WiFi turned off and plugged into PC via Ethernet cable. Still testing at 35 Mbps download (rather than previous, stable 37Mbps). Will we ever really know?

Ah! The vagaries of Broadband... 🤔