Disconnections and slow speed

[#1]

Barry,
Sweet!!! Edits crossed over!  
What you are seeing is exactly what I expected!  2.5db noise is coming from the laptop PSU / screen.  The absence of this looks like it gave you an additional 1Mbps in synch speed (circa an additional 15%) but still a way short of what your line should be capable of.  Now you are running fully powered after an unpowered resynch, the SNRM is buzzing around 4dB.  But is bouncing - would be nice to identify the source of that so that you see a line as flat as mine.
Right, now you know where this source is coming from, we can leave the laptop plugged into the power supply with the screen left permanently on.  This will avoid two known causes dinging the graph and nay dings on the graph can be put down to other sources.
Make no mistake, you've done a brilliant job here.  Treat yourself, garb a coffee and two chocolate hobnobs!
...and buy a branded laptop PSU.  I was watching an investigative TV programme the other night, looking at imported goods.  I think it was trading standards who impounded a shipment of laptop power supplies.  They took them to the lab to run safety tests.  Not only did the insides have metal parts bridging the HV / LV sections, but the casing failed the flame test, burst into flames and set fire to the "simulated carpet" below the test rig.  Has made me think twice about buying equivalent components.
Kevin

[#2]

Thanks a lot Kevin,
I would treat myself to a cup of tea but I'm afraid to put the kettle on in case it sends my router stats all over the place haha!
What I'll do is just show the stats from the last test I did following your guide:
Re-start routerstats
Reboot router while routerstats is running
PSU mode, screen on: 30 mins
Battery mode (PSU still plugged into mains), screen on: 10 mins
Battery mode (PSU unplugged from the mains), screen on: 10 mins
PSU mode, screen off: 10 mins
Battery mode, screen off: 10 mins
Have a nice break if you're having one, maybe I'll have one too ;-).
Thanks again,
Barry.

[#3]

Barry,
The PSU is certainly an issue.  What I cannot be sure of is does the screen cause a problem on its own or is that a function of additional loading on the PSU.  Keep an eye open for other sources.
Good luck with concluding this.  Please drop me a PM when you have more progress to report.
Kevin

[#4]

Hello Kevin,
I'm just looking at supposedly genuine IBM / Lenovo power supplies on Amazon. It's hard to tell these days if they are genuine though I think, or even if it will perform any better. I'll have to take a chance I suppose. I have some lead in the garage, I was thinking of enclosing the PSU in that to see if that helps (is that a silly idea?).
Thanks again, will let you know if/when anything happens.
Barry.

[#5]

Possibly a silly question, but how close is the laptop and its PSU to the modem/router and telephone socket?
I think most if not all computer / TV screens and switch mode power supplies will output some RFI, detectable at close proximity, but not detectable and not causing any problems from a couple of metres away.

[#6]

Hello ejs,
Thank you, it's a few feet away and realistically as far away as they're able to get here. I'm going to get another power adaptor anyway so I'll see how that goes.
Wrapping it in lead made no difference, perhaps it has to be really thick. Worth a try though!
Thanks again,
Barry.

[#7]

Hi guys,
just a brief comment on a few things as I haven't read all the detail since my previous post.
Master socket/extension socket - you can just 'snip off' the capacitor by it's wires to "convert" a master to a secondary, but as you've already ordered another secondary you may just as well hang on the current one 'as is' until some future point when you think there may be a use for it.
Attenuation changes can of course be due to poor or dirty connections. Things to check are that the wires to terminals 2 & 5 are nice and tight in their ADC terminals at all points. Another one is the plug into socket being a bit dirty. Next time you have things disconnected, plug in and out the plugs into sockets several times to help remove any surface oxidation (on filters as well).
Also the Master socket faceplate into the back-plate - and when it's pushed in make sure it feels a reasonable tightish fit - have a look at the contact wires in the back plate, they should all be angled down by about 30° or so.
The other thought I had about wiring was the BT Openreach drop-wire. If this is an old "flat" type then it will be causing a lot of the extra noise pick-up between dusk and dawn (note that generally this extra noise is rarely related to street lighting or other appliances etc) it's usually predominantly the increased MW/AM radio propagation.
If the BT incoming cable is a round black cable about 5.4mm diameter it's probably OK - again next time you have things disconnected, remove the backplate from the box and have a look at the wires behind - you should find a white wire and an orange one connected the the A & B terminals with a green wire and a black one unused. If the drop-wire has recently been replace it may only have the one pair (a white wire and an orange one).
The other quick comment was about Switch Mode PSUs. As already mentioned they can go south or west (depending on where the expression came from!!). As ejs mentioned, the proximity the modem/router may have an effect, but when connected wireless you may not see any change if it's further away when plugging it in or out of the laptop. That however would likely be NOT the case when connected by ethernet. I have seen examples of significant drops in SNRM due to SMPSs going bad, noticable when connected by ethernet.
The best way to check this sort of issue is to monitor with Routerstats on one computer, whilst switching on/off & disconnecting the one you want to check.This equally applies to Desktop PCs, not just laptops.
Screening (in lead or otherwise) is unlikely to make much difference as the noise is also & usually transmitted & radiated by the wires.
I'll try and come back later and see if there's anything else I can usefully help on.

[#8]

Thanks for the helpful  info Anotherone.
When the new extension socket comes and I've fitted it, I'll go round and check all the screws and contacts. I'll give them a squirt with contact cleaner just to be sure as well.
The 'drop-wire' is round and about the diameter you mention. The orange and white wire are used, and there are green, black and yellow unused wires.
I have a supposedly genuine power adaptor coming for my laptop, I'll do some comparison test when it arrives.
Thanks again,
Barry.

[#9]

The new (well, used) IBM branded power adapter arrived and appears to have made a big difference. I've attached some router stats showing the old adapter for a while, then battery for a couple of minutes, then the new adapter. As you can see, there's a sort of sine-wave thing going on over time but I don't know whether that's a problem or not.
Please let me know what happens next as I'm still running on the master socket downstairs and it's getting to be a bit of a pain.
Many thanks,
Barry.

[#10]

Hi Barry,
Ditching that duff PSU has made a massive difference 4dB is worth around 1Mbps.  Where next.
In the morning, restart the router and post the stats.  For your router / line to be performing correctly, we need to see 9Mbps or better.  If it is less, raise a fault here http://faults.plus.net - you've done enough internal testing / kit swapping to move forward.  Chris will or only pick this up and double check before deciding what next.
Keep a watch on that sine wave something is putting out a small level of cyclic noise - it could get bigger!
Kevin

[#11]

Hello Kevin, surprised to hear from you.
I'll do what you suggest. The initial good figures drifted down and this evening are all over the place currently 6.5 but have been down to about 5.5. I plugged the old power supply back in for a while to see what would happen and although it went down, it wasn't by nearly so much, maybe just over 1dB. Maybe something else is interfering in the evening but I don't know what. Currently nothing is switched on that wasn't on during the day. Anyway, I'll try again tomorrow. Many thanks,
Barry.

[#12]

Barry,
For clarity, you are going to see night time interference from MW/AM radio, generating 1.5dB to 2.5dB variation.
I'm back from the weekend and away from Wednesday - just checking up on progress of a couple of threads.

Kevin

[#13]

Kevin, I'm not convinced we got to the bottom of everything yet. That cyclic variation in SNRM is not normal but if it's "stable" it may be something that can be "lived with" once the possible cause is known. Right now I'd be more concerned about how the line changes throughout 24hrs on continuous connection.
Barry, do you have the SNRM graphs for yesterday evening? And any through the night out of interest, and through dawn into daylight?
I'd want to look at exactly how the SNRM is changing from daytime through dusk into nightime.
I expect you've rebooted now which I would have suggested this morning as Kevin did, so a current SNRM graph would be good as well.

[#14]

Thanks Anotherone,
I woke up in the early hours and couldn't get back to sleep and so decided to re-start the modem using the procedure mentioned somewhere in this thread. I have the stats just before and after that which I have attached. I'll leave the computer on with routerstats running later this afternoon right through the night and post those tomorrow if you think it will help.

[#15]

It will be interesting to here the experts view on this because I have never seen that sort of variation before