Kitz Forum
Broadband Related => ADSL Issues => Topic started by: g3uiss on October 01, 2018, 05:03:49 PM
-
Strange problem arisen over last few weeks.
Most mornings at 0919 my DS SNRM fall to nearly zero on my ADSL line, there is a corresponding drop on my VDSL line but only to about 4db. Until today, the ADSL held sync, but today it fell so much it lost sync. I guess its a big bust of noise at a frequency in the ASDL range.
Strangely it doesn't happen every day but more than others, and more strangely it happens less days on the VDSL .
I have put some current stats at :-
www.west-lodge.com/adsl/index.htm - ADSL line
www.west-lodge.com/vdsl/index.htm VDSL line (Regretfully no snapshots uploaded recently )
There is nothing in the house with a timer or device that triggers at 0919. I suspect it must be external. It was happening when I was away and the house unoccupied and nothing running ?
-
I also have a noise burst which happens each morning that I've never been able to get to the bottom of and has been ongoing for a few years now.
It will cause my line to generate quite a few errors - as per the attached you can see throughout most of the day I only get 1 CRC, but the whatever it is will cause a sudden burst to around 2800-3000 which will last for about a minute.
That spike is the reason why I have to monitor my line, because sometimes the errors will stick until I reboot my modem. If I don't then DLM will penalise me.. which is a PITA because by the time DLM catches up, the errors will have already stopped about 18hrs ago, but will take weeks to recover. :/
-
Interesting. Very similar, I get about 3000 CRC on my ADSL2 when it happens at 0919, I did at first think it might be a new electric train. The cable from my UG DP goes over the railway bridge which you will know was recently electrified, but firstly there is nothing going under the bridge at 0919 and I guess it couldn't just be 1 train. (However Northern cancel more than run ;D )
I'm going to have to live with it, like you unless it gets so bad that the line syncs every day. This morning the CRC was less and the SNRM dipped to just above 0 so it stayed in sync.
Do you get a similar SNRM drop to match and is the time constant ? Interesting if it is as its just a bit later than me ???
Have sent you a personal PM also
-
No corresponding dip in SNRm which is strange considering the amount of errors. The only thing I can think of for that is there is a micro-second dip that dsl stats doesnt catch as it only records every 1 min. In that instance CRCs would still show as they are based on accumulated errors rather than real time.
However that would not account for the fact that the CRC/ErrSec spike will usually last over 2 recordings. I suppose there could be a couple of spikes, but you would have thought that because it happens daily and has been happening for years, that DSLstats would have caught it at least once.
>>> firstly there is nothing going under the bridge at 0919
Mine isn't always at exactly the same time. It will usually be between 9.15 to 10:15 but can also happen any time between 8 - 11:30. Today's was at 10:24 & 10:25.
I have seen similar on a few other lines just not as noticeable and I have in the past wondered if it is something at the DSLAM side.
-
The question is, what happens at that time. Something getting switched off, or switched on? Heating turned off, or an oven turned on? (By neighbour) if so, something needs interference suppression.
Is there any chance that a mains filter might help? (No, because it is RF, going straight into the DSL line, not coming in by mains.)
-
Help !
The issue continued similar to as outlined by @Kitz. Mine was always at the same time, but apart from the spike in CRC and drop in SNRM I didn't until today, get any more re syncs.
Today, exactly 1 hr earlier I got the same, but this time a re sync. I assume whatever causes this is linked to a timer not changed back to GMT perhaps. However at 14:14 another busrt's of noise and a re sync. This time (I assume) the DLM banded me as the SNRM is now 12 (from 6), with consequent speed drop.
I'm really surprised the DLM would do that on a second re sync ?
Do I report to ISP, as they aren't going to see / believe my observation on CRC burst I doubt, but at some time I would need the SNRM resetting ?
(My VDSL line shows a similar bust of CRC but less and a drop of 2db in the SNRM, but to date no re syncs).
Stats at www.west-lodge.com/adsl/index.htm (no snapshots)
Suggestions more than welcome
Added (1626) I also note I’m missing a big batch of tones that were there before, something that has been debated before, but only odd tones, now a great big gap ! Also QLN looks flat, could this be the balance of the noise burst at that caused the re sync ?
Tony
-
Oh dear you have my sympathy,
I've just had a similar problem fixed (although the engineer has no idea what fixed it). The guess was substantial interference from another line/outside source which went away on the disconnection of a line in the area (I got the line as my second as there were no other pairs going to the DSLAM and the company had just closed).
If it's the same then your QLN will be noisier on the resync, but if you give it a few minutes and force as resync you will go back to your original speed and QLN - or will do until the resyncs cause the DLM to intervene. So you may have to turn off your modem for 30 minutes to be safe. The only certain thing for me was that rain caused the noise for my problem, your problematic source seems to be on a timer. I think of it as crosstalk++ (or a real PITA).
The bad bit is that if you get an engineer then nothing will show up, as your pair is fine... Ask your neighbours if they are suffering the same, as you may need numbers to get someone to do something about it. At least you've got a time when it happens, but I suspect the duration will be too short to be useful.
-
Thanks. I think your right. I forced a re sync and the speed and SNR returned to normal, so did the bit loading and QLN.
I guess when the re sync happened yesterday it synced while the bust of noise was on.
Thanks
Tony
-
Well this issue eventually went away in mid January still got daily CRC bursts at 0919, but it was rare that it re synced.
It’s returned this last week with vengeance, same time but CRC peaks are higher but at the same time.
Knowing what’s been said, it’s unlikely anything can be done to find the source (@kitz found the same) and contacting neighbours is unlikely to find anything unless they monitor their connection ( unlikely) or happen to notice a re sync at that time !
So I think there are 2 options
1) ignore it as the ADSL2 DLM is unlikely to do much with a resync ever 24 hrs
2) ask for interleave to hope the FEC might just reduce the CRC,s enough to hold sync.
I think @bcat is right OR would never find it.
Tony
-
Hi g3uiss
I would try double filtering it
Sounds a strange thing to do I know, but you would be surprised how many times that has resolved an issue
Many thanks
John
-
John
That’s an easy thing to try, I assume you mean plug a standard filter into the phone part of the faceplate and then the HG612 into that ?
Tony
-
Hi g3uiss
Ahh sorry I forgot that there may not be any phone devices plugged in
If they are, double filter the phone side
Do you have any extensions of master nte or alarm system connected to phone side - if you have, filter alarm also. Alarms do put a lot of noise on a line
You could try double filter on dsl side, but you would need to change to correct plug to do so.
Many thanks
John
-
John
Ok there is just a single phone so I can double that.
No extension or alarms etc. Also no timers that kick in at 0919 every day. The CRC spike lasts 1 DSL Stats minute. SNRM DIPS TO -1.
On my VDSL circuit noticeable but very slight CRC rise and about 1db snrm fall
I feel it’s an external timer somewhere on the 800m run and affecting the ADSL2 frequencies.
Tony
-
Hi g3uiss
Ahh sorry I forgot that there may not be any phone devices plugged in
If they are, double filter the phone side
Do you have any extensions of master nte or alarm system connected to phone side - if you have, filter alarm also. Alarms do put a lot of noise on a line
You could try double filter on dsl side, but you would need to change to correct plug to do so.
Many thanks
John
Doesn't the DSL side typically have ZERO filtering except in the modem itself? So double-filtering that part should do nothing at all.
-
Hi
Yes it is meant too but I have seen where there is dsl on the phone side as well as dsl side...
One client had moved and plugged router into phone line but still syncd with very poor figures
Many thanks
John
-
Given its most likely an external issue of a spurt of noise, would interleave or getting the SNRM increased to 9db be a possible solution, if so which is likely to produce the better result?
-
Considering the SNRM variations which I was experiencing in each day, I increased my upstream SNRM target to 9dB. I was getting upstream SNRM sharp drops/rises of 4dB or so each day and that way it would make sure that my worst-case minimum daily SNRM would always be high enough after a drop so that the value would then be so low as to risk data corruption
-
Thanks @weaver. I will give that a try, in my case it drops to -1 if I added a headroom of 3db it should still stay in sync.
-
Interesting the last two mornings ( 10 minutes earlier) the modem didn’t re sync, but I got a PPP drop on the router. It looks like a shorter but bigger CRC spike and SES also but there wasn’t a SNRM drop. So I even more confused as what if anything to do. The PPP drop is only as long as it takes to re connect. Perhaps I should just hope the interference just goes away ?
On the VDSL. Circuit I see the massive CRC spike 1000 for and a 2db SNRM drop but stays in sync. The total ES daily is around 750 on DSL Stats.
Tony
-
Do you have any industrial units or similar close to you as this sounds like the type of interference you'd get from a badly suppressed large electric motor or machine being started. Another possibility is a poorly suppressed vehicle that is sending wide-band interference for a short period as it starts-up and drives off.
:)
-
I have a similar issue. Since January my line experiences a huge spike in errors twice daily at roughly a 12 hour interval causing a retrain. It is not exactly 12 hours, but usually some time just after 06.00 and 18.00 hours so I think it may be some human activity around these times not a timer event. OR investigated and gave the line a clean bill of health despite seeing the step change in their line history stats. Neither they nor my ISP were prepared to investigate further. I live in a small village with a few shops, hotels and a garage, but I suppose any one of these might have some equipment that could produce this issue.
My line syncs at 67Mbps, but I have capped it at 50Mbps which appears to reduce the effect on the line and stops the re-syncs. I am on the 3db profile, capping increases the SNRM to 8.8. Interestingly I have not noticed any reduction in performance at 50Mbps, probably not that surprising given our usage pattern.
-
Thanks. Sounds like @vic has the same. I’ve no industry around and nothing like a vehicle as it’s to precise on the timing. I wasn’t going to even try ISP or OR they won’t find any fault.
Mine is ADSL2 so I can’t cap via the HG612 but I’m going to ask ISP to increase SNRM to 9 ( from 6). The line is secondary so the speed drop won’t be too significant. I’m just going to wait a few days and see if it goes away again as it has before. I guess being ADSL2 one retrain a day won’t cause any long term issues anyway, unless I’ve missed something?
-
@vic0239 - there’s a lot of this kind of thing going around it seems :)
I had this visible in the form of SNR spiking for several weeks but it was many times per day; also it was downstream-only.
In addition I have one line with variable upstream high-level/low-level SNRM on a 24 hour cycle with ‘high’ so many hours in the day, ‘low’ so many hours in the day. This is a square wave with duty cycle that is very very roughly 50%, well, not near 0% or 100% anyway. Also a second line shows a smaller less regular squarewave.
-
Well this is interesting. I’m away from home at the moment and checking my stats (I know I’m sad, but it has been raining a lot) I note that there have been no DS errors since we left. Zero.
I switched a few things off before leaving, one of which was the central heating boiler. I had discounted this despite it being newly installed just before the problem started as it cycles many times during the day and my errors occur at just these two time frames.
However, I’ve been developing a theory that the culprit could be the thermostat on the hot water cylinder. My reasoning is that the new boiler is much more efficient and the temperature of the water is higher resulting in the thermostat cutting off whereby it may never have in the past and the hot water timer is on at these times.
Does this theory hold water (sorry)? Could such a device cause such a spike of errors? I’ll obviously experiment when I get home, but would be interested if this is plausible.
-
@vic. I’ve exhausted anything in my home that occurred at 0919 presisly everyday. It must be external, and the hope of finding that zero. So it’s a case of living with it I’m sure. The CRC spike is over 1000, so it’s a big chunk of interference.
Frustrating to say the least.
-
. . . I’ve been developing a theory that the culprit could be the thermostat on the hot water cylinder. My reasoning is that the new boiler is much more efficient and the temperature of the water is higher resulting in the thermostat cutting off whereby it may never have in the past and the hot water timer is on at these times.
Does this theory hold water (sorry)? Could such a device cause such a spike of errors? I’ll obviously experiment when I get home, but would be interested if this is plausible.
Yes, that is feasible. Any load switched on or off could be the cause of an RF spike and/or an EMI pulse via the mains wiring.
Some years ago NewtronStar went all round his home, fitting RC suppressors to all switches, etc, in an attempt to reduce (or eliminate) the errors on his VDSL2 service that he was seeing.
-
As b*cat has said, any load switching device such as a heating/zone thermostat can generate noise either directly or via mains wiring, had an occurance on my daughters system where the oil fired boiler thermostat was generating a considerable amount of noise clearly visable on the router stat's and also causing a "stutter" on all the house terrestrial TV's running on an amplified distribution system from outside aerials.
In this case I believe the interference was transmitted through the mains certainly in the case of the house TV's, not so sure about the interaction with the modem/router.
Anyway, replacement of the boiler stat. completely cured the issue on TV's and modem/router.
Likely stating the obvious, but it's all down to a process of elimination, being able to identify/eliminate if possible any local events which correspond to the modem/router recorded disturbances as logged by DSLStats.
Initial quick checks can be carried out on the heating/zone stats. by adjusting their respective set points loading and unloading, however, this is not completely fool proof as set point adjustment can be quick and "clean" whereas normal temperature change which is usually a much slower action can produce contact "noise" when approaching the set point from either direction, the classic symptoms of a dodgy thermostat.
Then, as g3uiss has said, after all local sources of interference have been eliminated it's down to possible external sources which is a completely different ball game.
-
Thanks to all for your replies. I’m home tomorrow and the hot water will be switched back on pronto. I will keep a watchful eye on the line stats in that first hour. If my theory is correct I should see a spike. It looks a fairly simple device to replace if it is the culprit.
-
That's it! :yay:
Sitting at the computer I can hear the pump running and was able to detect when the thermostat switched it off. Nice little blip at each occurrence. Unfortunately the cylinder thermostat is now obsolete, so I will need to source an equivalent replacement.
@g3uiss I was feeling your pain until I came up with this theory, attempting to find the source of such interference is futile when it's outside your premises.
-
Now that you have identified the source of interference I would next try manually changing the tank stat set point up and down around the current temperature point so activating/de-activating your CH pump (assuming that's what the stat controls) and checking if you still get the modem/router blip.
Compare this with normal tank temperature changes instigated stat operation with respect to noise generation.
If the former does not produce interference but the latter does then I would believe that the stat is indeed "noisy" when approaching it's switching point from either or both directions and needs to be replaced.
If interference is produced by both permutations then I would consider first fitting some form of Capacitive or Res/Cap suppression across the contacts before considering replacement.
-
Where can we get a suitable big fat mains voltage rated capacitor from ? That would cure it I would think ?
I assume that if one’s interleaved death is greater then it might help ? Unless the duration of the glitch is too great anyway, because of arcing, not just simple one-off rise dI/dt from one transition. Crappy contacts where voltage bounces (did they used to call it being in need of ‘de-thump’ ?) would be much much worse altogether, taking us back into the 19C.
-
Where can we get a suitable big fat mains voltage rated capacitor from ? That would cure it I would think ?
Devices, more often than not, called a "suppressor" are an encapsulated, series connected, RC pair and can be purchased from any of the big electrical/electronic retailers . . . RS Components, CPC Farnell, etc.
-
Thanks for your input tiffy. I performed your test and the spikes were produced when manually turning the thermostat to a lower temperature setting i.e. valve closing and pump stopping. I did this a couple of times with the same result. I had already ordered a replacement having sourced a Siemens unit which claimed to be an equivalent which I have now fitted. It is working satisfactorily and no spikes when manually adjusted, so fingers crossed for no more interference during normal operation and I can remove my speed cap.
Apologies to the OP for hijacking the thread.
-
@vic0239
OK on your further testing as per suggestions.
As previously advised, with these results I would have first gone for a "suppressor" as b*cat correctly describes, however, as you have now sourced and fitted a replacement tank temperature stat which has cured the problem, a good result achieved.
It's quite possible that your replacement stat has a suppressor fitted as standard across it's contacts, many now do.
Also apologies to g3uiss for thread hijacking and good luck with your external noise source identification.
-
As previously advised, with these results I would have first gone for a "suppressor" as b*cat correctly describes, however, as you have now sourced and fitted a replacement tank temperature stat which has cured the problem, a good result achieved.
I did investigate sourcing a suppressor, but those I though might be suitable cost more than the new thermostat. ??? I’m no expert though so was probably looking at the wrong type.
-
Also apologies to g3uiss for thread hijacking and good luck with your external noise source identification.
No problems I’m glad you have identified your issue. I’m bewildered as how to proceed, or just live with it as I can’t filter it out, as it’s happing somewhere down the line.
I think it’s central heating. It started 1st October last year and stopped about late May. This year it’s started a bit earlier but it’s been quite cold here. We haven’t yet turned ours on but will shortly.
[Moderator edited to fix the above quotation.]
-
I think it’s central heating. It started 1st October last year and stopped about late May. This year it’s started a bit earlier but it’s been quite cold here. We haven’t yet turned ours on but will shortly.
If you think it's an external central heating related source then I would imagine it would be quite near at hand, if mains bourne then likely to be someone on the same distribution phase, ie. probably not next door.
The only other easily achievable suggestion would be to try some form of mains filtering on your modem/router, these are available in either wired or encapsulated format, encapsulated much easier to fit but more costly, check some of the sources suggested earlier by b*cat such as RS or CPC/Farnell, worth a try.
-
Devices, more often than not, called a "suppressor" are an encapsulated, series connected, RC pair and can be purchased from any of the big electrical/electronic retailers . . . RS Components, CPC Farnell, etc.
What am I actually looking for - looked at RS and not clear
-
@g3uiss
Have had a quick browse ref. mains filters, somewhat shocked regarding cost especially with RS Components.
Cheapest option I can find is here:
e-Bay Plug In Filter (https://www.ebay.co.uk/itm/Maintrab-6A-Filter-Transient-Mains-Spikes-RFI-Protection-Audio-Computer-OMQ4-01/253386753287?hash=item3aff070507:m:m4LTM7_yMZG6pRPcdaCjJqw&var=552550257092)
Listed as old stock but I don't believe the related technology has changed radically.
Another option but considerably more costly:
Amazon 6 Way Filter (https://www.amazon.co.uk/Tacima-6-Way-Mains-Conditioner-black/dp/B00UB0G4DQ/ref=dp_ob_title_ce)
The earlier model of this item listed on Amazon appears to refer more to RF Interference protection as well as surge protection and mains conditioning:
Earlier Model Amazon 6 Way Filter (https://www.amazon.co.uk/Tacima-Conditioner-Frequency-Interference-Filter/dp/B000PS5700)
If you are a "hands on" person and have the resources then making your own filter would likely be the cheapest option with something like this:
e-Bay Bare Mains Filter (https://www.ebay.co.uk/itm/Mains-filter-Schaffner-FN2020-16-06-16A-Data-sheet-supplied-110-250-VAC/183978632996?hash=item2ad5fb2724:g:vLwAAOSwUfZdbpVq)
Would need to be mounted in preferably a metal box with a short mains lead input and a 13 A. socket for output.
Of course absolutly no guarantee that any form of mains filtering on your modem/router will solve your issue, it's just part of the process of elimination.
-
Mine still happens practically every day at around 10am give or take an hour either side.
Never got to the bottom of it although don't know if I dare say this, but it has been causing less Err/Secs for the past couple of months.
-----
g3uiss - do you have a smart meter?
-
Yeah these type of issues are the worst.
The issues that are there all the time are better as then a visiting engineer witnesses it and will then have an easier time of fixing it.
On my ADSL service I had noise bursts at random times during office hours and always would be worst in the morning with the initial burst been akin to a shotgun blast at the line. I later found out it was known by the engineers and was told off the record where the problem lies. It was exchange side, and sure enough on VDSL the problem is long gone. So I was lucky in that respect.
-
Mine still happens practically every day at around 10am give or take an hour either side.
Never got to the bottom of it although don't know if I dare say this, but it has been causing less Err/Secs for the past couple of months.
-----
g3uiss - do you have a smart meter?
No @kitz no smart meter.
-
Yeah these type of issues are the worst.
The issues that are there all the time are better as then a visiting engineer witnesses it and will then have an easier time of fixing it.
On my ADSL service I had noise bursts at random times during office hours and always would be worst in the morning with the initial burst been akin to a shotgun blast at the line. I later found out it was known by the engineers and was told off the record where the problem lies. It was exchange side, and sure enough on VDSL the problem is long gone. So I was lucky in that respect.
That’s really interesting as my other line is VDSL and doesn’t suffer. I put that down to the affected frequency being at the low end.
-
I've been looking around for an example of the circuit diagram for a mains EMI filter. Attached, below, is typical example. (It's not a recommendation for that actual Chinese made device!)
If that circuit diagram needs to be "translated" into English sentences, please just ask. I'm sure quite a few of us will be capable of doing so . . .
-
Very close format and practically identical component wise to this unit e-Bay Bare Mains Filter (https://www.ebay.co.uk/itm/Mains-filter-Schaffner-FN2020-16-06-16A-Data-sheet-supplied-110-250-VAC/183978632996?hash=item2ad5fb2724:g:vLwAAOSwUfZdbpVq) I listed earlier in the thread, also made in China, majority of things are these days.
The "cable tails" varient as illustrated by b*cat could make termination and fabrication into a suitable enclosure if desired a bit easier, save a bob or two on crimp terminals etc..
There are a lot of package termination variations readily available with various current ratings which shoud be chosen with regard to what is going to be powered through the filter.
-
I've been looking around for an example of the circuit diagram for a mains EMI filter. Attached, below, is typical example. (It's not a recommendation for that actual Chinese made device!)
If that circuit diagram needs to be "translated" into English sentences, please just ask. I'm sure quite a few of us will be capable of doing so . . .
Great, no difficulty with circuit diagram being well used to those. Just thinking of a neat way to house and terminate in an AC plug and socket.
-
Very close format and practically identical component wise to this unit e-Bay Bare Mains Filter (https://www.ebay.co.uk/itm/Mains-filter-Schaffner-FN2020-16-06-16A-Data-sheet-supplied-110-250-VAC/183978632996?hash=item2ad5fb2724:g:vLwAAOSwUfZdbpVq) I listed earlier in the thread, also made in China, majority of things are these days.
In this particular case, I would not consider purchasing a used item.
-
The ultimate test, as the approximate time of the interference is known, would be to temporarily power the modem/router from a sealed lead-acid battery of sufficient capacity. If the problem is still seen, then it cannot be due to transmission via the mains supply.
-
As all my kit is UPS protected, I removed the AC to the UPS. It ran on battery over the time period and the errors remained unchanged.
-
So mains-borne transmission can now be ruled out.
-
I know I'm a bit late on parade as it's now November. I'm just wondering how you are getting on with you line problems?
One way I cured mine was to fit a common mode choke on the incoming line like this:
(https://i.postimg.cc/jdQQkgKb/FT-240-43.jpg)
Previously to fitting this when I used my transmitter on 1.9mHz it used to kill my modem stone dead, after fitting it I just lose some of the tones around that frequency, a modem re-boot brings them back when I've finished, it has generally cleaned up my line no end and just get a few CRC errors during the day. The torriod I used was 43 mix, I think 31 mix would be better as that is intended to cover lower frequencies that ADSL2+ uses.
You can get the torriods from here:
https://radio-store.co.uk/toroids-rings/12-fair-rite-ft240-31.html
You mileage might vary of course, like any EMC problems it is a case of experimenting to see what works. Incidentally, it will help with line balance as well.
-
Interesting I’ve tried a ferrite bead but not a choke. I will give your suggestion a try
Thanks
-
Having tried all the helpful recommendations nothing has solved the problem. I know @kitz has a similar problem, but mine is EXACTLY the same time daily. A daily resync on ADSL2 is not in its self a major problem, and if I didn’t know I likely might not know.
However similar to @kitz occasionally the resync happens during the burst giving a high SNRM and ES needing a manual resync to sort. Also it drops my VPN, again this re syncs but potential at a low speed.
Given OR wouldn’t be able to find this, are there any mitigation’s I could try to try and hang on to sync. Perhaps getting the SNRM altered to 9db so a drop might not go to minus and hold sync ? Or having the line interleaved? I would "Cap" the speed, but I don't think you can do that with ADSL2 on a HG612 ?
My VDSL circuit suffers the same but a small dip in SNRM for the same CRC burst, and ES burst, and always stays in sync. I’m assuming the interference is either at the lower tones, or nearer the exchange than the cab. Hence the reduced effect on my VDSL.
Suggestions welcome.
Tony