Power failure knocks 10mb/s off the line speed.

I really had no choice but to knock the power off to the modem yesterday. On reconnecting, thought I'd try a speedtest, 10Mbps loss. Ok, it's gone back up by another 3Mbps overnight but I'd have thought OR systems would be a bit more intelligent. After all, having been up for well over 3 months at one rate (lightning strikes caused the last failure), should one single cut immediately bring down the speed? In fact, thinking about it why does it get involved with playing with the line speeds on a stable line anyway. Haven't they ever heard of if it ain't broke, don't try fixing it.

Why did you have to turn off the modem? It could be relevant.

Probably not the BT systems but your modem deciding on the sync rate that can be achieved on connection.

As modems don't reconnect at higher speeds of their own accord then if the modem has not resynced overnight then the increase might be due to IP profile catch-up game at PlusNet.

The key is what are the actual connection speeds and stats from the modem, plus the Openreach DLM looks at error rates too.

While I might be an electrician, working live on 230volts isn't really recommended. I was swapping out some 13amp sockets for ones with USB sockets built in. So off went the MCB for about 30 mins or so. Wasn't a case of on and off as I changed each one but a simple off, then back on after finishing the whole lot.

One thing I noticed today was that on Plusnet, my line speed has also dropped from 78Mbps down to 67.4Mbps. I do wonder whether it's worth my while having words or wait until it goes back up.

Ummmm.

The immediate thought has to be what those inbuilt USB mains sockets are all about. I don't think you can assume they aren't the cause. What are they for, and is something now plugged into a couple?

Can you see the error stats on the Draytek?

What RobertoS is hinting at is that the switched mode PSU producing the USB might be a noise source and could be interfering with the VDSL2 signal.

One step to mitigating this if the sockets are to remain is changing RJ11 lead from a flat to twisted pair wire and at £2.99 its a cheap change to make https://www.amazon.co.uk/gp/product/B008IKVAP8/ref=a...

Agree with both answers so-far. One of the biggest impacts to VDSL2+ signals is noisy PSUs, which is why Christmas tree lights etc can often have such a big impact on broadband speeds.

It is very likely the PSU is generating electrical noise which is being picked up by the telephone wiring in your home, and more likely the RJ11 cable to the router, which is probably close to one of the USB power sockets.

10Mbps does not just fall off with a power failure. It has to be something changed and introduced more noise.

In reply to a post by RobertoS:

The immediate thought has to be what those inbuilt USB mains sockets are all about. I don't think you can assume they aren't the cause. What are they for, and is something now plugged into a couple?

Those USB sockets are used for charging only. Something like these USB sockets The wiring that they were connected to was a spur off the ring at the back of the house and the modem sockets are at the front. And no, nothing is yet plugged into them.

Edited by 69bertie (Thu 28-Jul-16 19:18:45)

In reply to a post by 69bertie:

In reply to a post by RobertoS:

The immediate thought has to be what those inbuilt USB mains sockets are all about. I don't think you can assume they aren't the cause. What are they for, and is something now plugged into a couple?

Those USB sockets are used for charging only. Something like these USB sockets The wiring that they were connected to was a spur off the ring at the back of the house and the modem sockets are at the front. And no, nothing is yet plugged into them.

Even so, they're switch mode power supplies with mains attached. These supplies can be less stable (and therefore more noisy) with no load on them. It may not be the model you have, but see https://www.youtube.com/watch?v=zoZ1_aEDPos , especially the section from 3:37 to 4:49.

It got scarier after 8:21.

Makes no difference if something is plugged in or not. They are still a PSU, and often with this kind of device, especially if bought for cheap like the ones you linked the PSUs are erratic.

Even if you removed the offending sockets now, DLM will have likely already kicked in, so it would probably take a few weeks before speeds went back to normal.

Appreciate you do not want to believe the sockets to blame, but it most likely is.

In reply to a post by RobertoS:

It got scarier after 8:21.

Indeed - that defect is potentially lethal.

Too true. But that socket looks like it was made in a back street factory! As for the earthing arrangement, don't think I've ever seen one like that, ever. I've seen many with two earth terminals but always with a bar going across between the two. That screw hole damage to the socket would make me wonder if the person installing it had trouble (they usually need deep boxes) and ripped the bar off trying to make it fit. Anyway I'd like to think MK know what they are doing...at least I hope so. As to being safe, there is so much electronics hard wired nowadays etc (just take any RCCBo apart that is probably already fitted to your mains fuseboard), that as long as you don't buy cheap, I don't there is much of an issue.

Anyway, been in touch with Plusnet, seems I have to wait 72 hours. I will do that.

There's an interesting thread on the IET forums that suggests that currently BS5733 and BS1363 don't cover outlets which have built-in USB charging ports although they are included in the upcoming BS1363-2.

Another interesting comment is the MK sockets cost around £2.20 a year to run under no-load conditions.

Link

In reply to a post by 69bertie:

After all, having been up for well over 3 months at one rate (lightning strikes caused the last failure), should one single cut immediately bring down the speed? In fact, thinking about it why does it get involved with playing with the line speeds on a stable line anyway. Haven't they ever heard of if it ain't broke, don't try fixing it.

I've been considering starting a thread 'VDSL not fit for purpose?', any disconnection knocks my speed down to around 2.5 from 4, more often than not this requires my to contact TT to get it back up again.

So if 4 meg is all it provides when 'working well' then you have a stupidly long (for VDSL) D side. So the thread should be more along the lines of, 'what other products might I use' as VDSL isn't for me.

In reply to a post by Zarjaz:

So if 4 meg is all it provides when 'working well' then you have a stupidly long (for VDSL) D side. So the thread should be more along the lines of, 'what other products might I use' as VDSL isn't for me.

Or maybe 'the office bods at OR say my line should be 7 to 11 meg what should I do?',
or maybe 'I can see where the fibre runs from my house, but BT rationalization of cabinets many years ago left our village with long lines'
or maybe 'roll on 2020 and the USO'
or best of all 'according to latest BDUK2 map field at back of my house will have superfast in 2018, can I put my modem there'

Have a walk around your house with a 'Medium wave radio tuned to about 612 kHz and listen for noise or REIN or Repetitive Electrical Impulse Noise ,
see
http://www.kitz.co.uk/adsl/rein.htm

For some makes I could well believe it would cost some money but MK say otherwise for theirs. Zero cost on standby with automatic disconnection as soon as the USB plug is removed MK USB Tech specs
Even so I look at all the power adapters scattered around the house on nigh on everything and they all have wasted heat coming off them. Case of what is convenient I suppose.

One thing I have recently noticed is my computer's motherboard fed USB sockets are still usable (I use them for charging the phone) even though the computer is shutdown. Which means my gold rated PSU is still producing power even though it looks off.

Edited by 69bertie (Sat 30-Jul-16 10:03:07)

I turn the surge-protected socket off when I close down my desktop PC. I've noticed for ages that when I turn the socket on the Ethernet port light on the router comes on even before powering up the PC. So PSU isn't completely controlled by the on/off power button at the front.

I've never known of any computer to 100% shutdown, must admit to have never tested the USB, but plug in a ethernet cable and the socket will light up, and a few weeks back I borrowed an old keyboard and mouse as my wireless adaptor packed up, and the backlights stay on (back in the day the mouse and keyboard where off a 'state of the art' gaming m/c with green backlight), also if you check MB for LED's you will probably find these stay lit.

Isn't that on purpose - for wake on LAN functionality?

In reply to a post by 69bertie:

One thing I have recently noticed is my computer's motherboard fed USB sockets are still usable (I use them for charging the phone) even though the computer is shutdown. Which means my gold rated PSU is still producing power even though it looks off.

In many cases this is deliberate, so the ports can be used for charging devices. I am writing this on a 6 week old Lenovo ThinkPad X1 Yoga, which has one of the USB 3.0 ports on the device and a couple of the USB 3.0 ports on the OneLink+ Dock permanently powered. These ports have a 'battery' symbol under the USB logo to denote they are permanently powered.

In reply to a post by 69bertie:

For some makes I could well believe it would cost some money but MK say otherwise for theirs. Zero cost on standby with automatic disconnection as soon as the USB plug is removed MK USB Tech specs

Are you sure? On which page of the Tech Spec does it say the USB Integrated Sockets power down when te USB plug is removed?

In reply to a post by MHC:

In reply to a post by 69bertie:

For some makes I could well believe it would cost some money but MK say otherwise for theirs. Zero cost on standby with automatic disconnection as soon as the USB plug is removed MK USB Tech specs

Are you sure? On which page of the Tech Spec does it say the USB Integrated Sockets power down when te USB plug is removed?

Page 5:

Dust proof USB port shutter
Zero standby power consumption, auto disconnect when plug
removed

Its certainly interesting as I haven't installed any USB sockets myself because unless they physically disconnect the power you cannot run a insulation resistance test with them fitted.

That said, the MK sockets are so expensive at retail that its kinda silly to even bother. I assume the op got them at bulk due to being an electrician.

Edited by alexatkin (Sun 31-Jul-16 04:45:58)

I'd have expected some power used, as you say for LAN wake up purposes but it still came as a surprise to find the USB sockets still live at switch off. Not that I'm complaining, very handy not to have to bother with USB plug in adapters.

In reply to a post by alexatkin:

In reply to a post by MHC:

In reply to a post by 69bertie:

For some makes I could well believe it would cost some money but MK say otherwise for theirs. Zero cost on standby with automatic disconnection as soon as the USB plug is removed MK USB Tech specs

Are you sure? On which page of the Tech Spec does it say the USB Integrated Sockets power down when te USB plug is removed?

Page 5:

Dust proof USB port shutter
Zero standby power consumption, auto disconnect when plug
removed

Its certainly interesting as I haven't installed any USB sockets myself because unless they physically disconnect the power you cannot run a insulation resistance test with them fitted.

That said, the MK sockets are so expensive at retail that its kinda silly to even bother. I assume the op got them at bulk due to being an electrician.

Go back and read the document again ... No where does it say the Sockets with Integrated USB have zero power consumption when not in use.

Many years ago the power button was a hard button that actually killed all power to the system. Then they moved to soft power buttons - the only way they can work is if the PSU is supplying power to the motherboard as that is where the power switching occurs.

Plus they bought in services such as wake on lan - the only way to do that is to keep the lan circuitry live otherwise it can't see the incoming power on instruction.

And as others have mentioned the USB power can be powered so they can charge devices even when the system is "off".

In the early days you could get away with changing the memory and the like just with the power switch off but these days you either need a power switch on the PSU to isolate the motherboard power or you have to unplug or switch off at mains.

Wake on LAN isn't the only service ... wake on mouse & keyboard nowadays requires USB power.

In reply to a post by WWWombat:

Wake on LAN isn't the only service ... wake on mouse & keyboard nowadays requires USB power.

I did say:

services such as wake on lan

I didn't fancy listing every reason the m/b would still need power.

In reply to a post by MHC:

Are you sure? On which page of the Tech Spec does it say the USB Integrated Sockets power down when te USB plug is removed?

Page 5 in the gloss underneath Dust Proof USB shutter ..... 'Zero cost on standby'

In reply to a post by 69bertie:

In reply to a post by MHC:

Are you sure? On which page of the Tech Spec does it say the USB Integrated Sockets power down when te USB plug is removed?

Page 5 in the gloss underneath Dust Proof USB shutter ..... 'Zero cost on standby'

Read the document again, VERY carefully. No where does it say that the USB socket has "Zero cost on Standby" , "Zero standby power consumption", or anything similar. And does it have a dust proof shutter?

Your making an easy mistake Bertie. It took me a while to spot where, as MHC is not being particularly helpful.

Page 5 is not about the USB-enabled 13amp sockets. It is about devices called USB charging modules. It was the dust shutter text that led me to it, as the ones you are using clearly don't have USB shutters.

This is a video of a test that shows a socket with inbuilt USB drawing no current when no device is plugged in. It demonstrates there is zero cost on standby as stated in the MK document.
https://www.youtube.com/watch?v=NRiTuIdbU8g

Edited by deleted (Wed 03-Aug-16 08:41:49)

The last few seconds of the commentary says "it most likely does [draw some current] but it's so negligible...".

Totally different brand

Didn't its booklet also say it took a little?

In reply to a post by BatBoy:

https://www.youtube.com/watch?v=NRiTuIdbU8g

Off topic, but amusing and perhaps relevant to current politics, in the panel of clips provided at the end of that one there is this one.

In reply to a post by cheshire_man:

The last few seconds of the commentary says "it most likely does [draw some current] but it's so negligible...".

Typically they can provide 2A at nominally 5v so 10W which equates to around 45mA at 230v. Allowing for losses there is probably around 50 to 60 mA needed at full power. In a quiescent state that will drop to anywhere from 2 to 10 mA depending on design and efficiency. Say 5mA - that will equate to around 10000Whr per year. It may be small but it is NOT zero.


edit: to tweak a few figures

Edited by MHC (Wed 03-Aug-16 11:02:38)

In reply to a post by MHC:

It may be small but it is NOT zero.

The video shows the device draws zero current.

I do not care what that video shows. It does not show the device under discussion. And where is the proof that the meter is calibrated? Purchased from Lakeland ... says a lot.

There's no reason why those results wouldn't apply to the MK socket, and the MK socket documentation states the device draws zero current on standby.
Case closed.

I have just got some figures for an MK Logic Plus range Socket with USB outlets. The quiescent current is actually just under 7mA - which seems higher that I might have expected.

NO IT DOES NOT.

They devices are of different design, from different manufacturers and the MK documentation does not state it draws zero current.

Yes it does, as has been shown earlier in the thread.

WHERE?

Page 5.

Page 5 does not refer to the socket version. READ the document completely and accurately..

In 2013 MK Electric was one of the first UK manufacturers to introduce a range of award winning USB charging modules, and now this technology is being integrated into our market leading twin socket.

Do you understand what "integrated" means?

I can also read a specification - something you cannot.

You appear to be labouring under the misconception that the dust cover turns the device off. I'm happy to have pointed out you are wrong.

1. You cannot read the document

2. You believe that tests on one manufacturers device apply to all other - WRONG.

3. Actual results that I have now posted show that when the USB ports are not in use there is still power used.

I wonder if the dust covers do actually operate micro-switches? I've never seen a usb port covered except on mobile phones.

I have no need to read technical documents, further I have no need to carry out experiments - I just ring a man who can. MK Technical department confirms the device draws 12mA on standby.

So, total proof that I am correct and YOU were wrong.

It also demonstrates that you definitely did not read the document correctly.

You mean the things 69bertie has bought?

That will be per unit as well of course - he said he'd done a few.

But more importantly now, I wonder if one or more is putting out noise it shouldn't?

You didn't say 12mA at any point.

I said just under 7 mA. The email from MK that I received at 12:21 states 150mW which equates to 6.5 at 230v.

I see where you're going wrong - there are 2 sockets.

One power module and the 150mW is the total.

And by the way - there are two mcroswitches, one on each USB port that will turn off the power when the cover closes.

Even more good news

But more importantly now, I wonder if one or more is putting out noise it shouldn't?

Seems extremely likely.

Update. We had a power cut today. speed test tonight And even Plusnet have my line speed back up at 77.2 Mb Whether DLM would have relented anyway as the speeds had started to shift upwards over the last couple of days. But it's taken nearly 3½ months for the speed to finally return back to what I call normal. Really is time DLM was cremated!

Edited by 69bertie (Sat 12-Nov-16 19:56:49)