Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2 3 ... 6

Author Topic: Serious interference around 700Khz  (Read 26721 times)

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Serious interference around 700Khz
« on: February 18, 2013, 04:54:20 PM »

I thought I'd post this for interest to others, and because I'm open to suggestions.

Whilst trying to work out where the intermittently high error seconds and occasional loss of synch were coming from, I wandered around the house with an AM radio.

At around 700Khz I found a VERY loud buzzing sound.  Now and again it would rhythmically drop out with an almost morse-like pulsing. dit dit    dit dit.

Yesterday I blamed the bargain NAS and it's PSU, which I got off ebay and has caused me varying degrees of woe ever since, so I disconnected it.

But today I found the noise louder than ever.

Unplugging my gigabit switch made the sound decrease drastically in intensity, and the pulsing disappeared.

Aha!  Switch or PSU faulty / bad design.

Except, wandering around the house, the noise came back.  I could detect it everywhere in the house.

I went outside, and it was a lot quieter, but holding the radio to the house wall, I could hear it very clearly.

Up to the shed, it was very loud there too.

Near the phone line, I could hear the usual hiss of ADSL, but no buzz.

It seems that the noise is being radiated throughout the house ring main, and up to the shed via the mains cables, because isolating the shed from the mains made the radio fall silent.  (not because the radio went off - it's battery  :lol:  But the loud buzz was replaced with a faint hiss when the mains was disconnected from the house).

Definitely mains bourne.

And network related, since I could quieten it by disconnecting a network switch.

Near the downstairs PC, the noise was very loud, disconnecting the gigabit switch there made it al but fall silent.  Aha! So it's that!  :graduate:

Replaced switch with a 100mbit one I had spare, only faint noise.  But gets louder.  So not the switch then  >:(

Unplug PC (which was "soft" off anyway).  Silence.

Plug PC in again - noise.

Power up PC, noise back but reduced.

Switching it on and off causes the noise to vary on the radio in a different room, and as I'm typing it's just started buzzing very loudly again, even though the PC in question is "soft" off. 

I've now unplugged the suspect PC, and 700Khz is pretty much silent now.

So, I'm assuming here that the PC PSU is at fault.  Radiating noise into the mains and into the network wiring via the NIC.

Fair assumption, or did I miss anything?

Ian
« Last Edit: February 18, 2013, 05:07:00 PM by sheddyian »
Logged

JGO

  • Reg Member
  • ***
  • Posts: 729
Re: Serious interference around 700Khz
« Reply #1 on: February 18, 2013, 05:43:18 PM »

Something to consider:=   If it is mains borne then a power line filter as close to it as possible should see that route off. Possibly an internal one is dying ?  I think it is possible to buy external filters (from Maplin?) without breaking the bank, which could reduce if not eliminate the problem. ( Some people swear by ferrite cores but which grade of ferrite ?)


The problem with a transistor radio is that it is quite difficult to tell relative strengths as it's AGC levels them out, you could supress the conducted component but the receiver still pick up direct radiation from the PSU at close range.
Logged

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #2 on: February 18, 2013, 07:21:41 PM »

Something to consider:=   If it is mains borne then a power line filter as close to it as possible should see that route off. Possibly an internal one is dying ?  I think it is possible to buy external filters (from Maplin?) without breaking the bank, which could reduce if not eliminate the problem. ( Some people swear by ferrite cores but which grade of ferrite ?)


The problem with a transistor radio is that it is quite difficult to tell relative strengths as it's AGC levels them out, you could supress the conducted component but the receiver still pick up direct radiation from the PSU at close range.

If it's the PC's internal ATX PSU (which it seems to be) then I'll get a replacement one.   The noise could be heard on the radio all over the house, on both floors, and up the shed!  That's not good.

I'm currently listening to the radio, still tuned to the frequency the noise was heard on.  Earlier it was a gentle hiss, now the sun has gone down there are distant foreign stations breaking through now and again.  But no interference.  So does seem that it was the PC.

Just wait and see if the modem loses synch this evening.

Ian
Logged

Black Sheep

  • Helpful
  • Addicted Kitizen
  • *
  • Posts: 5717
Re: Serious interference around 700Khz
« Reply #3 on: February 18, 2013, 07:34:28 PM »

"I could detect it everywhere in the house".

Firstly, I personally would set the frequency to 612Khz as recommended by our Chief Engineers Office. Not sure what else you may be picking up at 700Khz ?? It would appear 'everything' ??  :)

'Noise' is obviously all around us, it's whether it's 'service affecting' noise !! You may well just be picking up normal noise' ?? I've been on a few cases whereby our 444B REIN tester screams at you 'It's here, it's here', when put next to a PSU or a baby monitor etc etc, when in reality it isn't actually a real-life 'Noise' issue.

The only true way to find out if you have REIN, is to test the spectral density across your circuit, to indeed see if the 'noise' is drowning the signal out. Also, you could have a poor MPF (pair of wires from Exchange to premises) showing a low AC Balance ?? This would mean that 'normal noise' would have an effect on your circuits stability.

I'm not trying to play down your efforts here, just making sure you don't throw out every PSU you have in the house for no reason.  ;) :)
Logged

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #4 on: February 18, 2013, 07:51:03 PM »

Firstly, I personally would set the frequency to 612Khz as recommended by our Chief Engineers Office. Not sure what else you may be picking up at 700Khz ?? It would appear 'everything' ??  :)

Thanks for your reply, and I bow to your knowledge, sir  :baby:

"around 700Khz" was because I stood next to the modem and slowly moved the dial from the low end upwards.  And at around 700 got a very strong, loud buzzing noise, which I was suspicious of.

Quote
'Noise' is obviously all around us, it's whether it's 'service affecting' noise !! You may well just be picking up normal noise' ?? I've been on a few cases whereby our 444B REIN tester screams at you 'It's here, it's here', when put next to a PSU or a baby monitor etc etc, when in reality it isn't actually a real-life 'Noise' issue.

Ah now, that's the question.  The symptom is high errored seconds and losing synch in the evening, sometimes repeatedly.  Resynch after this is 2 - 3 K lower than usual.

Quote
The only true way to find out if you have REIN, is to test the spectral density across your circuit, to indeed see if the 'noise' is drowning the signal out. Also, you could have a poor MPF (pair of wires from Exchange to premises) showing a low AC Balance ?? This would mean that 'normal noise' would have an effect on your circuits stability.

I have absolutely no idea there - except that, up until this point, I've had a fast synching (20 - 21k) very reliable connection for over a year.  So line condition (& AC Balance?) is/was OK ?

Quote
I'm not trying to play down your efforts here, just making sure you don't throw out every PSU you have in the house for no reason.  ;) :)

No, not at all - that's why I posted here, to see if I was doing the right thing and leave it open to suggestions!

Maybe the meaning was lost a bit in my rambling post - I'd suspected several PSU's along the way, then ruled them out when the noise came back.  But for the last 3 hours with the rogue PC disconnected, the noise has not returned.  That one, then, is probably the source of the noise.

So, with the rogue PC switched off, I'm waiting to see what happens to the modem stability :)


Ian
Logged

Black Sheep

  • Helpful
  • Addicted Kitizen
  • *
  • Posts: 5717
Re: Serious interference around 700Khz
« Reply #5 on: February 18, 2013, 08:02:08 PM »

The trick is to leave it set to 612Khz, and then go looking for a 'Buzzing noise'. Bearing in mind 'false positives'. You are doing absolutely fabulously from your comments, and it would seem like you might have found a particular noise source ?

It really is a dark art, tracking REIN.

As an aside, and this may have been covered already ?? We have had issues with 21CN circuits dropping below 3dB SNR especially as night sets in. This causes problems with errors and L.O.S. events (Loss of service). I don't know which ISP you are with, what package you are on, or whether you are subject to DLM ?? Just thought it worth mentioning. :)
Logged

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #6 on: February 18, 2013, 08:23:15 PM »

As an aside, and this may have been covered already ?? We have had issues with 21CN circuits dropping below 3dB SNR especially as night sets in. This causes problems with errors and L.O.S. events (Loss of service). I don't know which ISP you are with, what package you are on, or whether you are subject to DLM ?? Just thought it worth mentioning. :)

I'm with TalkTalk, and it's a 21CN LLU circuit.  Prior to this problem, I've not had any problems with LOS or high errors - I get a good synch around 20 - 21K.  The reported noise margin rarely moves away from 6db, and if it does, not by much and not for long.

I believe there is DLM, as yet I've not seen it permanently increase the noise margin. 

Out of interest, what's a good/bad figure for total error seconds?  (this modem only reports an overall total).  Currently it's showing 115 errors since I rebooted 3 1/4 hours ago.  Good, bad, ok, mediocre?

Ian

Logged

Black Sheep

  • Helpful
  • Addicted Kitizen
  • *
  • Posts: 5717
Re: Serious interference around 700Khz
« Reply #7 on: February 18, 2013, 08:39:24 PM »

Hmmm ?? The £64 million dollar question !!!

Various ISP's used to allow between 10-20 Errored Seconds (or, CRC's) per minute. From an engineering perspective, I consider any kind of steady incrementation as a potential fault. Of course, it comes down to the almighty dollar, and we are to a degree, responsible to the ISP (not the EU), as they are paying our wages and it is their circuit. If the deem 15 ES's a minute to be ok, then who are 'we' to argue.

Long and short of it, I'm, no expert on what is good, bad or indifferent when it comes down to historical data. I tend to work on the 'Here and now' when faulting, I'm afraid. I'm sure some of the others will have a better line of input on this ??
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: Serious interference around 700Khz
« Reply #8 on: February 18, 2013, 08:41:20 PM »

Quote
I'm with TalkTalk, and it's a 21CN LLU circuit.

All of us TT-ites in 'the know' will request the setting of a 6 dB target SNRM with the (TT) DLM disabled. Then, if required, adjustments of the DS target SNRM is made manually by tweaking with a Broadcom chip-based modem . . .
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #9 on: February 18, 2013, 10:11:34 PM »

Hmmm ?? The £64 million dollar question !!!

Various ISP's used to allow between 10-20 Errored Seconds (or, CRC's) per minute.

Well, that's a starting figure to consider.

Currently I still have synch, and have 128 total error seconds after 5 hours.  Which I work out to be 0.42 per minute, if I'm doing the maths and understanding it correctly.  So doesn't seem too bad?

Quote
From an engineering perspective, I consider any kind of steady incrementation as a potential fault.

Completely agree.  I've got a record of sorts of different readings of error seconds and also noted how long it had been running for, so maybe I'll see if it's got better or worse. 

But so far it's looking like the PC with the noisy PSU has been the cause of the problem.

Ian
Logged

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #10 on: February 18, 2013, 10:14:43 PM »

All of us TT-ites in 'the know' will request the setting of a 6 dB target SNRM with the (TT) DLM disabled. Then, if required, adjustments of the DS target SNRM is made manually by tweaking with a Broadcom chip-based modem . . .

I keep meaning to get TT to switch DLM off, and never get around to it. 

That said, so far it's not caused me any problems, and if I was having reliability issues I think I might want DLM to try and improve things at a cost of lower synch speed, at least in the short term.

Also, nether of my current modems have the ability to have their target SNRM tweaked, as far as I can work out.   :(

The TalkTalk one definitely doesn't; the room-heating TP-Link one does have some commands, that didn't seem to do anything when I tinkered, but haven't had a proper look yet.

Ian
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: Serious interference around 700Khz
« Reply #11 on: February 19, 2013, 12:12:03 AM »

. . . nether of my current modems have the ability to have their target SNRM tweaked, as far as I can work out.   :(

The TalkTalk one definitely doesn't; the room-heating TP-Link one does have some commands, that didn't seem to do anything when I tinkered, but haven't had a proper look yet.

Eric has very conveniently complied a list of modem/routers which have Broadcom xDSL chipsets.
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

sheddyian

  • Kitizen
  • ****
  • Posts: 1159
    • My Shed Blog
Re: Serious interference around 700Khz
« Reply #12 on: February 19, 2013, 12:27:31 AM »

Eric has very conveniently complied a list of modem/routers which have Broadcom xDSL chipsets.

Ah yes, I'd had a look at that list the other day, a worthwhile resource.  Sadly, my modem isn't on it.

The main chip on the modem has a weedy little heatsink stuck firmly to it, so I can't see the manufacturer.  However, several other chips nearby have "Trend" printed on them, so that's probably a giveaway.

If I do go back to using this TP-Link modem, at the very least it's getting a much bigger heatsink, though I'm seriously considering a little fan.  It gets worryingly hot.  In my view, if you can't hold your finger on it for more than a few seconds, it's too hot.

Ian
Logged

JGO

  • Reg Member
  • ***
  • Posts: 729
Re: Serious interference around 700Khz
« Reply #13 on: February 19, 2013, 08:04:21 AM »

" The trick is to leave it set to 612Khz, and then go looking for a 'Buzzing noise'. Bearing in mind 'false positives'.
It really is a dark art, tracking REIN."

What is so special about 612 kHz ?

If it is IMPULSE noise, then it will have a wide spectrum ( Car ignition interference used to stretch all the way to Band I TV at  40 + MHz)  and the more important point is to avoid confusion with a strong broadcast station. Yes it will have peaks but spark transmitters have been illegal since 1928  ! or is someone still commemeorating the Titanic ?!!!!
Logged

Black Sheep

  • Helpful
  • Addicted Kitizen
  • *
  • Posts: 5717
Re: Serious interference around 700Khz
« Reply #14 on: February 19, 2013, 09:24:41 AM »

It's not 'special', it's just the frequency our Chief Engineers Office recommend. I believe it's due to there not being much else around that particular frequency to cause disruption. However, one can hear a (I think it is) French radio station until approx. 0830 in a morning.

I'm no expert on the science, but I'm quite good in practice at finding REIN at 612Khz .............. not 700Khz, or 40Mhz, just 612Khz. ;) :)
Logged
Pages: [1] 2 3 ... 6