Kitz Forum
Broadband Related => ADSL Issues => Topic started by: hake on May 14, 2009, 10:18:40 AM
-
Much learned thought has been expended for my benefit in the past few days. I had a presumed problem with upstream performance. The advice offered and acted on has been of great benefit. The effect of the advice was to route the the ADSL cable away from mains electricity and to use passive microfilters on the low pass telephone cable plant.
However, the problem of reduced upstream speed and a puzzlingly low upstream SNRM remained.
It is with much embarrassment that I now present the reason for the upstream problem (and to a lesser degree the downstream also).
I had put an ADSL Surge Protector in line with the ADSL cable. :-[
I had become blind to the presence of the surge protector (obviously a cheap and exceedingly nasty device).
I mention this so that others might avoid my self-induced predicament.
I must thank to all those who gave their time to my 'problem'. I have, nonetheless, benefited from their expertise as the router's performance stats show:-
Rx: 4704kbps SNRM 16db
Tx: 448kbps SNRM 16db
Oh, I'm so ashamed. :-[
-
No need to be ashamed. You ought to be able to rely on a surge protector.
It really does look as though you've traced the problem, so that's great news. :thumbs:
-
Thanks for the feedback Hake, that's great news. :thumbs:
With any luck your downstream sync-speed will increase further over the next few weeks or months, as DLM allows the SNR margin to relax.
And following a few days behind each improvement in sync-speed, your IP Profile will be also be raised leading to faster actual download speed.
You now have a line that many of us would be envious of :)
-
I have you guys to thank. I found the offending item (now binned) on eBay. Where do I find a good ADSL surge protector?
-
I use this:-
http://www.argos.co.uk/static/Product/partNumber/9822980/Trail/searchtext%3EMASTERPLUG+SURGE.htm
Not the cheapest but doesn't affect the ADSL. I connected through it a while back and didn't seem to cause any problems at all.
-
Thanks jid. I will look into that.
-
Where do I find a good ADSL surge protector?
First, ADSL is a radio wave. Any high capacitance protector will degrade DSL 'radio' signals (for the same reason that conventional telephones eat DSL and need that blocking filter.) A DSL protector must be low capacitance to not 'eat' radio signals.
Second, the protector is not protection. Protectors do not stop or absorb surges. Effective protectors connect surge energy to the protection - a separate device. Protection is earth ground. To provide protection, a protector must connect short (ie 'less than 3 meter) from each phone wire to earth. This has been routine in every telco CO everywhere in the world for the past 100 years. To make a protector even better, a lower capacitance protector is connected even shorter to earth. Every meter shorter increases protection.
Third, distance between a protector and ADSL hardware increases protection. Wire has low resistance but higher impedance. Wire impedance is why wire length is relevant. An even shorter distance to earth means a better connection from telephone wire to earth ground - lower impedance (not to be confused with characteristic impedance). Increased separation between the DSL modem and protector means higher impedance. More surge harmlessly absorbed in earth when a protector is connected shorter from phone wire to earth. Less surge will find earth ground destructively via your ADSL hardware.
Important is how that wire connects to earth. For example, sharp wire bends, wire bundled with non grounding wires, inside metallic conduit, using a different earthing electrode from AC electric and cable, wire splices ... all adversely increase impedance to ground - decrease protection.
Better DSL protectors use semiconductor devices such as Sidactor or low capacitance avalanche diodes (such as http://www.semtech.com/pc/downloadDocument.do?id=532). At the consumer level, the protector must specifically state in numeric specs that it is rated for DSL operation. And, of course, it must have a dedicated connection to earth. Protectors do not stop or absorb surges. Protectors simply divert surge energy harmlessly in earth - and remain functional even after a direct lightning strike. You will be amazed little a protector need be when it diverts (does not absorb) energy even from direct lighting strikes.
Meanwhile it must also be low capacitance so that it appears non-existent to ADSL signals (radio waves).
-
Is there any particular reason that you need a surge protection device on your line - do you have an expensive modem? As has been mentioned proper protection involves bonding to earth, which can be easier said than done.
-
Is there any particular reason that you need a surge protection device on your line - do you have an expensive modem? As has been mentioned proper protection involves bonding to earth, which can be easier said than done.
You might also want to consider the fact that the master socket already contains a surge protector, and the router will almost certainly have one too. These won't save your router from a direct lightening strike on your house, but if that happens a charred router would probably be the least of your worries.
-
You might also want to consider the fact that the master socket already contains a surge protector, and the router will almost certainly have one too. These won't save your router from a direct lightening strike on your house, but if that happens a charred router would probably be the least of your worries.
Surge protectors are earthed so that direct lightning strikes do not even damage the protector. Protectors that don't provide lightning protection are only protecting profit margins.
BT has switching computers connected to wires all over town. With every thunderstorm, the typical CO suffers 100 surges. How often has phone service in your town been lost for four days while they replace that switching computer? Earth protectors are that effective and that routine.
So that lightning is harmless even to the protector, every incoming wire is earthed by a 'whole house' type protector. Energy diverted to and absorbed harmlessly by earth does not enter the building; does not cause damage.
All appliances contain surge protection. Even required by international standards. A protector cannot divert energy harmlessly into earth without that earthing connection. A protector without earthing protects only from a type of surge that is typically not destructive. Install and earth a 'whole house' protector so that the rare and so destructive surge does not overwhelm existing internal appliance protection.
To answer orainsear's question: these destructive surges occur typically once every seven years. Less frequently in the UK. But frequency varies significantly even within a town. Install and earth a 'whole protector so that a rare and destructive surge does not damage protection already inside every appliance.
-
Thank you for those thoughts. I have decided not to use additional inline surge protection. I have, however, dug up an inline ethernet surge protector which now sits between the router and an ethernet switch. It uses a mains ground. The ADSL signal should not be impaired by this arrangement.
-
To answer orainsear's question
Thanks for the reply but I was actually asking hake about his needs for a surge protector device in his setup :D
-
I have, however, dug up an inline ethernet surge protector which now sits between the router and an ethernet switch.
An inline protector means nothing sits in that wire. For example, are these your protector?
http://www.keison.co.uk/furse/furse19.htm
http://www.tripplite.com/en/products/model.cfm?txtModelID=151
See that ground wire? If not connected to earth, then no protection exists. The ethernet wire goes right through the magic box - nothing but wire. A protector connects from each wire to that ground wire. If that ground wire is not connected short to earth, the protector does nothing. What can it do? One side is connected to nothing. No earth ground means no place for electricity to flow.
Does a magic box provides protection? None do. The magic box does nothing if not connected to protection. A ‘magic box’ is only a connecting device - to connect a surge to protection. No earth ground means a protector does nothing.
Ethernet protectors typically are rated for voltages that are below phone line voltages. Phone line protectors are typically 300+ volts. Ethernet wire protectors typically are only tens of volts. The ethernet protector may regard a ringing phone as a surge. But since the protector circuit is not connected to earth, then it does nothing for surges AND hopefully does nothing to impeded telephone operation.
Just because it is a magic box called a surge protector, then you think it does something? Some protector too close to appliances can even make surge damage easier.
-
Yes, I see. Just like the safety valves on a steam locomotive's boiler, there has to be path for the unwanted energy to be channelled through.
That means that the cheap and nasty inline surge protector device I wasted my money on could not have been effective. How easily are the gullible parted from their money.
The ethernet surge protection I am now using can presumably dump the surge energy to mains earth (when plugged in). It's a shame that such devices fail a PAT test.
-
Hake,
I don't want to worry you (indeed, I've already confessed that I don't worry about it at all really), but whilst I'm sure putting protection on the ethernet cable does no harm, even if its is earthed I fear it is necessarily as it good as it might appear....
...I'm thinking that an incoming spike on the phone wire could find a path, via the router and its PSU, to the live or neutral mains conductors. From there, it would look for a path to ground which might be straight through your TV, computer, Hi Fi - or all of them. :'(
But I freely admit I'm no expert and hopefully somebody will put me right if they no better :)
getting back to Orainseer's question... do you feel that you're particularly at risk?
-
Actually it's not quite true to say that a surge protector without an earth does nothing. Some types of surge may exist between the two incoming wires, and an unearthed protector connected between the two wires can reduce the surge voltage. This is what the spark gap in the master socket does. But a comprehensive protector would indeed need a short, low-impedance connection to earth to be at its most effective.
But in truth, nothing will protect your equipment from a direct lightning strike to the wires nearby. This is true as much for the mains wires as for telephone wires. The amount of energy in a lightning strike is far too much for any practical domestic surge protector to be any use. The best you can hope for is to get some protection against lower level surges, such as distant lightning strikes. Whether it's worth bothering I have my doubts. I prefer to unplug everything when there's a thunderstorm nearby.
-
But in truth, nothing will protect your equipment from a direct lightning strike to the wires nearby. This is true as much for the mains wires as for telephone wires. The amount of energy in a lightning strike is far too much for any practical domestic surge protector to be any use.
Which is why no phone service for four days everywhere in town is normal? A typical CO suffers 100 surges during every thunderstorm. Phone service is lost everywhere for four days after every thunderstorm?
Direct lightning strikes are routinely earthed without damage even to the protector. If a protector works by stopping or absorbing surges, then no protector survives. That popular myth (blocking surges) promoted in retail stores is not what a protector does. Protector diverts (shunts, connects, bonds, conducts) energy harmlessly to earth. Massive energy must be absorbed somewhere - in earth. Direct strikes that are properly earthed means a homeowner does not even know a surge existed.
Proper earthing means a short connection - low impedance - ie 'less than 3 meters'. That protector plugged into the wall is not earthed. Low impedance means 'less than 3 meters', no sharp wire bends, no splices, wire not inside metallic conduit, wire separated from all other wires, etc. Interior AC mains wire violates almost every 'low impedance' requirement. AC receptacle neutral or ground wire cannot provide necessary earthing.
An ethernet surge protector cannot dump surge energy to earth (when plugged in). Worse, it may provide a surge with more paths, destructively, via nearby appliances.
To have no surge damage, BT puts protectors as close to earth as possible AND up to 50 meters distant from electronics.
An ethernet protector (tens of volts) may interfere with or be destroyed by higher phone line voltages. Obviously, an ethernet protector is unacceptable for phone lines.
Protector for phone lines:
http://www.keison.co.uk/furse/furse11.htm
also has the necessary stud to make that short connection (no sharp bends, no spliced, separated from other wires etc) to earth.
And finally orainsear's question - is this the rare location where surges are more frequent?
-
I prefer to unplug everything when there's a thunderstorm nearby.
Good advice and it's what I do myself.
If you did want to go down an earthing route I wouldn't use normal house earthing. I'd install some gas discharge protection devices which were earthed with at least 8mm cable, larger diameter if the run was too long, connected to at least 1.5 metre copper earth rods driven into suitable ground.
-
The think is, lightning is close close to an an ideal current source. It'll generate a voltage across any impedance according to ohm's law, based upon V = IR, where I is fairly constant. So, if you leave you're TV connected to the aerial, but unplug it from the mains, the impedance from arial to earth will be many, many megohms. So, the voltage of the spark that connects your TV to the wall socket will be many megavolts, and that's why your house may burn down, instead of just your TV being damaged. :)
But bear in mind also that most home insurance covers lightning damage, so I choose not to worry. Especially in the UK where rain and wind are bigger problems than electrical storms.
-
If you did want to go down an earthing route I wouldn't use normal house earthing. I'd install some gas discharge protection devices which were earthed with at least 8mm cable, larger diameter if the run was too long, connected to at least 1.5 metre copper earth rods driven into suitable ground.
As noted, house grounding must be part of the earthing system. Earthing that must be upgraded to also provide surge protection. For example, that 8mm copper wire (for AC electric earthing) must be routed without sharp bends, not inside metallic conduit, as short as possible (ie 'less than 3 meters'), separated from non-grounding wires, etc. All incoming utlities must connect to the same earthing electrode either via a 'whole house' protector (AC mains, telephone) or by wire (cable TV, satellite dish).
Again, a protector is nothing more than a connecting device. A connection that meets many previous requirements to also provide low impedance (not just low resistance). Cable TV and satellite dish make that 'less than 3 meter' connection using only wire; no protector required.
GDTs (gas discharge) or semiconductor based protectors are for telephone service to provide that low capactiance requirement. Semiconductors are preferred. Telephone protectors are so inexpensive as to be installed on every subscriber interface in North America for free. However, these protectors sell in the UK for surprisingly high prices. Protectors are installed where phone cables enters a building. Each wire in each cable is connected short to the nearby building ground via a protector. Protection for all types of surges - longitudinal mode, normal mode.
1.5 meter copper clad earthing electrode is too short. Earth rod should be at least 2.5 meters. Best earthing is provided by the bottom half of that rod. Additional electrodes are installed if ground is a poor conductor (ie sand) or is often dry. Examples of how a building's earth ground is upgraded to also provide surge protection.
-
One thing to consider with surges is the amount of energy recived from the surge. if its lightning causing the surge then this can be a real variable. If enough energy comes down the line then no protector is going to prevent damage; the only sure way to protect yourself is if there is a hint of a storm comming your way then physicaly unplug any sensitive equipment fron the mains, telephony and aerial. even though the house may not suffer a hit, the mains and telephony can act like a casting net bringing the energy of a strike home.
If you're in an urban enviroment then the tendency is the energy gets disipated amongst a number of properties so each recives less energy. in rural areas a property may take a larger proportion of the energy hence rural dwellings tend to have a greater history of lightning issues. bear in mind this is just a tendency not a sure thing.
Different parts of the country have differing rates of lightning damage dependant on weather patterns, conductivity of the soil. If the land is flat it dosn't mean (dispite popular belief) your less likely to suffer lightning damage.
-
Just to confirm Ezzer's data, I was once in a Herefordshire valley close to the Black Mountains. I was standing in a farmhouse doorway watching a storm when a lightening strike went some way up overhead, but not very near. I heard the new PVC guttering creak as it does in sunshine. I.e. even wet insulating material can absorb lightening energy from quite distant lightening.
Kind regards,
Walter
-
One thing to consider with surges is the amount of energy recived from the surge. if its lightning causing the surge then this can be a real variable. If enough energy comes down the line then no protector is going to prevent damage; the only sure way to protect yourself is if there is a hint of a storm comming your way then physicaly unplug any sensitive equipment fron the mains, telephony and aerial.
Which is why BT disconnects all phone service with every thunderstorm? Because no protector can provide protection from those 100 surges? BT's computers are connected to wires all over town - and BT never disconnects their computer. Instead, BT uses protection that was that effective from direct lightning strikes even 100 years ago.
Disconnecting for protection is very poor protection. Humans are asleep or away over 16 out of every 24 hours. A human can only protect electronics typically two out of 24 hours. That is near zero protection. And that assumes the human knows a surge is coming. Most destructive surges are unknown until after damage happens.
No protector is going to prevent damage if one believes a protector is protection. When the protector is simply a connecting device to earth, then energy from direct lightning strikes are harmlessly dissipated in earth without damage. 100 surges to every BT switching center. How many towns are without telephone service for four days while they replace that computer? Ezzer must ignore 100 years of experience to post a popular urban myth.
Correctly noted: conductivity of earth is fundamental. Lightning more often strikes a mountain side rather than its highest top. Geology is why lightning more often may strike the valley; ignore the mountain. Same is why a protector is only as effective as its earth ground. Even with 40 direct strikes annually to electronics atop the World Trade Center, all those electronics work - no damage. Routine is to have direct lightning strikes - and no damage - if earthing is properly installed.
-
I said "if there's a hint of a storm comming your way."
Otherwise as you say properly fitted lightning protection helps when your not there to take action your self. I'm not ignoring past experience, other wise I thought unplugging is a myth until I became a bt engineer. Unplugging means the equipment is isolated from the most likey path of energy
-
Of course BT will employ surge protection and correct and effective bonding arrangements to ground, as will many businesses and agencies who require that level of protection, however the costs to protect the equipment and maintain system integrity, particularly for critical systems, will be commensurate with the level of protection required.
Clearly a risk assessment would be needed to analyse the probability and severity of potential incidents, but for the average ADSL user I would speculate that the costs to implement such a system would not be worthwhile. As has already been stated taking the precaution of simply isolating any sensitive equipment should be sufficient to mitigate any risk.
Also, while conductivity to earth can and is important, there are other scenarios and methods of protection, e.g. H.I.R.F. and lightening protection for aircraft, that can demonstrate other methods of suitable protection.
-
Incidentaly Kitz when you took the photos in the exchange of the mdf blocks did you get any which show the left side of the blocks with the colourful protection modules as they're part of the surge protection system
-
I said "if there's a hint of a storm comming your way."
If a protection system is so inferior as to require unplugging, then why waste good money on that protection? Why be dependent on a solution that has little knowledge of surges until after the event?
We don't unplug and don't have appliance damage even when a 33Kv wire fell on local distribution. Hundreds of meters even exploded from their pan. How would unplugging protect from that unexpected event? Earth protection so that even that does not cause appliance damage. So that unreliable unplugging is unnecessary.
Should we disconnect smoke detectors, dishwasher, and furnace when a thunderstorm approaches? Of course not. Did telephone operators remove headsets and leave the room when thunderstorms approached? No. Solution was that well proven and that effective even that long ago.
Early 20th Century ham radio operators would unplug their antenna and put the lead into a mason jar. And still damage occurred. Damage stopped when the antenna lead was earthed. A solution that well proven even that long ago.
Telco wire surge protection at each subscriber has been standard in North America for generations because it is so inexpensive, so easily implemented, and so effective. Even protectors installed in the fifties are sufficient for DSL operation - if still properly earthed.
-
Just because it's appropriate for big companies like BT to spend out on full lightning protection, it's no reason to assume it has any relevance to the home ADSL user.
One big reason for that is that the home insurance policy will already cover lightning damage. That, with whatever small print there is, will generally replace any damaged equipment. Such insurance would be much more complex for the likes of BT since the cost of lost equipment would be the least of their worries, compared to lost income from service provision while the equipment gets repaired. So BT (or BT's Insurers) will probably mandate extra protection.
Anothert point that (I don't think) has been mentioned yet... last time I shopped for an extension socket I found the usual crop of 'surge protected' extensions at extra cost. Reading the label, in the small print, many of them said 'does not protect against lightning strikes'. I'm aware that there are other forms of surge apart from lightning, but I do suspect there's a whole 'surge protection' industry springing up that is simply cashing in on ill-founded fears of an event that, in any case (thanks to normal home) insurance, most folks really needn't worry about unduly.
-
Such insurance would be much more complex for the likes of BT since the cost of lost equipment would be the least of their worries, compared to lost income from service provision while the equipment gets repaired. So BT (or BT's Insurers) will probably mandate extra protection.
Insurance is irrelevant. BT must work without failure - always. How much will an insurance company pay for your lost time and lost data? How much do you really get for a three year old computer? Almost nothing. How much money does it cost to file those claims? Less expensive is to eliminate damage rather than accept failure as an option. Insurance is irrelevant since damage is so easily aleviated.
How much are protector components? Pence. A protector to effectively earth a subscriber’s ADSL line? A quid - if that much – to actually protect from lightning.
A protector that does not protect from lightning does what? It protects from surges made irrelevant by existing protection. So what are retail stores selling for how much?
How much for protectors that *do* protect from destructive surges? Compare that to an inflated price for protectors that do not. Appreciate why BT probably wastes no money on insurance or plug-in protectors. Appreciate why BT spends less money for the effective solution. Same solution can be installed by a homeowner.
Destructive surges are rare events especially in the UK. A frugal and well protected homeowner spends nothing on plug-in protectors, and can earth one 'whole house' type protector. For ADSL, that is a low capacitance protector for everything on phone lines, and one AC mains protector earthed short to the same electrode. When damage must never occur, an earthed protector is the effective solution that also costs maybe tens or 100 times less money. Rather than accept failure, spend little to avoid damage.
-
Hi Westom,
It would perhaps be helpful if you could provide some links to protection equipment suppliers in the UK and detailed installation instructions if they are available on-line.
Kind regards,
Walter
-
How much will an insurance company pay for your lost time and lost data?
Data loss can occur for many reasons other than lightning strikes. Surge protection doesn't remove these other risks, and doesn't remove the need for regular backups.
Same solution can be installed by a homeowner.
In England and Wales we have something called Building Regulations Part 'P' that stops the homeowner from doing much more than changing a lightbulb without calling in an Electrician. I'd imagine (anybody confirm?) that protection on the AC mains would need to be installed at the incoming electric feeds, and so I''d be surprised if Part 'P' allowed the homeowner to do it himself.
And don't forget the TV arial as a conduit for surges, it's not just phone line and mains.
-
It would perhaps be helpful if you could provide some links to protection equipment suppliers in the UK and detailed installation instructions
That was provided in posts on 15 and 16 May. One paragraph started:
> As noted, house grounding must be part of the earthing ...
Even if you don't personally do the work, confirm work is performed properly; especially since these principles are little understood where damage is so infrequent or considered acceptable. First and fundamental point, earthing is the central feature of a protection system. The solution costs so little and routine when damage is not acceptable. Obviously, costs are near zero if implemented first when concrete footings are poured. Solution is always about diverting to where energy may be harmlessly absorbed. A solution that is routine whereever damage cannot happen. Any recommendation that does not divert energy into earth is not an effective solution.
Of course, the most important part of a useful answer are reasons why it works.
-
Curiouser and curiouser
After 10 days uninterrupted service at 4704kbps down (SNRM 13.5db) and 448kbps up (SNRM 16db), synch was lost this evening and the stats are back to like they were before I removed the apparently offending surge suppressor. The stats are now 4256kbps down (SNRM 16db) and 256kbps up (SNRM 7db).
Total mystification ???
-
It looks as though the surge suppressor was a red herring then. Something else is apparently causing low frequency interference.
-
Curiouser and curiouser
The stats are now 4256kbps down (SNRM 16db) and 256kbps up (SNRM 7db).
As roseway notes, this problem is not due to unfiltered devices. Something is probably imposing low frequency noise (25 to 130 Khz) on you DSL circuit (or their DSLAM is defective). Those previous suspects (or an intermittent wire connection) would have diminished ownload signal strength.
-
It was an amazing coincidence that the problem seemed to be resolved as soon as the surge suppressor was removed. The disappointing thing is that things were working so well for so long. I did however notice that three days before things went haywire, there were two intervals each with a lot of errored seconds but no disconnects. I had a spike of errors at around the same time (c. 11.20am) about two days previously.
During this golden period, the downstream SNRM moved up and down between 13.5db and 15.5db. The upstream SNRM remained constant at 16db, then BANG!. It's down to 6db again.
?
[attachment deleted by admin]
-
Thanks westom. 25+kHz would not be audible to humans on the quiet line test I think. I will have to borrow a dog to listen for me.
-
What make/model of router do you have?
-
Thanks orainsear. It is a BT Voyager 2100 (firmware version 2.18.01.17i_a023g1.d15g).
I am beginning to think that this is an impossible quest. Lacking any devices which can measure and record the characteristics of my ADSL connection so that frequencies in the 25kHz to 130kHz range can be revealed, I am hitting the buffers. Presumably, stumping up the £144 for a BT visit would bring such equipment into play.
Does BT possess systems capable of monitoring each copper loop to facilitate identification and diagnosis of problems? BT seems to work on a presumption of its own infallibility and that problems can only occur on subscriber's premises (and no doubt many do as I have found previously on my own premises). Since almost the entire cable run to and from the exchange is outside the subscriber's control, there would seem to be a degree of unreasonableness in this attitude often found with monopolies, such as appear to be held by BTw over its copper loops.
I imagine that the only way I could push this is to move to an ISP with LLU arrangements at my exchange but I do not wish to leave PlusNet who give excellent satisfaction in other respects such as tech support and facilities. I am therefore stuck with BTw.
-
I imagine that the only way I could push this is to move to an ISP with LLU
Sadly, you imagine in vain. It's still BT's copper and the same people will look at it irrespective of your ISP. :(
-
So the UK digital economy is being held to ransom by a state protected private monopoly. >:D
The same high-handed attitudes used to prevail when it was part of the GPO. Nothing has changed.
-
I'd suggest trying out DMTtool by following this guide (http://www.kitz.co.uk/routers/voyager2100_DMT.htm). When you get it running we'll be able to get some graphs of your frequency spectrum
Don't play around with the SNR margin though - just get the line stats.
-
Thanks orainsear. It is a BT Voyager 2100 (firmware version 2.18.01.17i_a023g1.d15g).
I am beginning to think that this is an impossible quest. Lacking any devices which can measure and record the characteristics of my ADSL connection so that frequencies in the 25kHz to 130kHz range can be revealed, I am hitting the buffers. [...]
I notice from your earlier post that you seem to be using DMT Tool v3.21 (with an exciting German interface...).
Perhaps you could try v8.07, which specifically lists your router -- and which might let you have a look at those lower frequencies. The DMT Tool version for my SpeedTouch 546 (see screenshot) shows the bitloading (though not the SN Margin) for the upstream connection. If you've got massive low-frequency interference, it /might/ show up as uneven bitloading... you never know.
The 30 - 300 kHz frequency range encompasses LW radio broadcasting, as well as things like RF ID tags and heart-rate monitors. You could hunt around the bottom end of the LW broadcast spectrum with an AM radio, to see if there's any hideous crackling, or a 1megawatt local radio transmitter 9km up the road -- which as you can see from the screenshot I have, broadcasting in the 960-970kHz range (not to mention R5Live on 909kHz...). However, I think they are pretty much all in the MW band rather than LW.
Good luck, anyway!
rwm32
[attachment deleted by admin]
-
Thanks rwm32. I have wandered round the house with a portable receiver and scanned LW over the complete range the tuner allows me (upward of 150kHz). I have to confess that I don't know what I am trying to hear.
These are the stats from the router: -
Line Mode G.dmt Line State Show Time
Latency Type Interleave Line Up Time 00:19:52:30
Line Coding Trellis On Line Up Count 1
Statistics Downstream Upstream
Line Rate 4256 288
Noise Margin 14.8 dB 6.0 dB
Line Attenuation 38.0 dB 24.0 dB
Output Power 12.4 dBm 19.8 dBm
K (number of bytes in DMT frame) 134 10
R (number of check bytes in RS code word) 16 16
S (RS code word size in DMT frame) 1 8
D (interleaver depth) 32 4
Super Frames 4208557 4208555
Super Frame Errors 2 0
RS Words 286181892 35772717
RS Correctable Errors 150961 56
RS Uncorrectable Errors 17 0
HEC Errors 2 0
OCD Errors 0 0
LCD Errors 0 0
ES Errors 0 0
I have attached an output from DMT v8.07. Hope this tells you something.
[attachment deleted by admin]
-
Well, I suppose it was worth a try with the v8.07 DMT Tool -- but the 'bit allocation not available' says it for me: it's not going to help. Sorry!
LW radio: if you de-tune your receiver from (say) R4 on 198, you should simply hear a hissing noise. If there's something generating RF interference, you'll hear an unmistakeable crackling hum, totally distinct from the 'de-tuned' hissing. However, upwards of 150kHz is taking you outside the suspect range, anyway.
Below 150kHz there are some defined bands for amateur radio operators (hams), plus a number of bands reserved for maritime navigational aids (Decca, LORAN). So /if/ there is low-frequency interference hitting your upstream bandwidth, it might be because one of your neighbours is a ham radio operator... or perhaps there's a Decca or LORAN transmitter down the road (unlikely unless you're pretty near the coast)?
I'm rambling now, so I'll stop.
Rom
-
On the contrary the DMT tool is showing dB per tone which actually looks fine; bit allocation tends to follow that pretty closely. ;)
It tells us that you're probably still ona 15dB target SNRM though. :(
You could try Routerstats of course, see if that shows any more detail.
-
25+kHz would not be audible to humans on the quiet line test I think.
Any noise on a POTS phone may also be heard at higher frequencies. Furthermore, diminish low band DSL signals need not be due to interior appliances. For example, calcium has been seen building a bridge to a wire. This 'connection' may become more problematic with increased humidity. Some signal creating problems (such as a calcium bridge) can be frequency limited. But generally, noise in those low frequency DSL band would also make noise at lower audio frequencies. Faint - but distinctive.
Making a connection directly to the telco master socket with all other interior wires (temporarily) disconnected may report something useful. Diminished signal during a direct connection to BT means the problem is 100% on their side - can only be fixed by BT linemen or the ISP DSLAM. Means you cannot do anything to fix what is their problem to solve.
-
On the contrary the DMT tool is showing dB per tone which actually looks fine; bit allocation tends to follow that pretty closely. ;)
It tells us that you're probably still ona 15dB target SNRM though. :(
You could try Routerstats of course, see if that shows any more detail.
... But the DMT plot only shows /downstream/ dB per tone, as far as I can see, whereas I believe the OP's question was about how to monitor 25-130kHz frequencies, which are used for upstream bandwidth.
I agree about Routerstats --- that's a good call; I'd forgotten that the more recent versions show bitloading for some modems.
-
I believe the OP's question was about how to monitor 25-130kHz frequencies, which are used for upstream bandwidth.
Certainly one of the issues and DMT would indeed appear to be useless there, but I've just realised:
Line Rate 4256 288
Have we been here before? Looks like sort of a fixed rate setup, not ADSL-Max? :'(
-
Have we been here before? Looks like sort of a fixed rate setup, not ADSL-Max? :'(
This is a question born of ignorance, but... wouldn't a 4mbps fixed rate be 4096 down / 384 up?
-
I think it's more likely to be just coincidence that the upstream is 288. Hake's connection suffers from variable upstream rates.
-
Yes, on looking back he had 256k upstream as well.
It's very odd though, given the downstream doesn't look too bad. Maybe time to revisit the filters?
-
All microfilters are Exelcus passive. I intend to replace the ADSLNation filtered faceplate with one like Clarity sell. At the same time, I will connect straight to the test socket. Unfortunately, this is obscured by a piece of heavy furniture. It will have to wait until my Dad has recovered from illness.
The advice I have received is much appreciated and very helpful.
-
All microfilters are Exelcus passive. I intend to replace the ADSLNation filtered faceplate with one like Clarity sell.
Making a connection directly to the telco master socket with all other interior wires (temporarily) disconnected may report something useful. IOW it completely eliminates questions about inferior microfilters AND numerous other problems such as frequency sensitive bridging by calcium growth, and other reasons which are even more obtuse and not listed.
Calcium bridging was offered as an example of strange things that are known only after tests identify the problem. Swapping parts will not provide useful (defintive) answers. A temporary disconnect test - only modem connected to the master socket by a separate and temporary wire) will.
-
Calcium bridging was offered as an example of strange things that are known only after tests identify the problem.
Is this when the calcium in damp concrete/building material forms a path between the telephone wire and ground?
-
Damp is not a problem at my house, including the point of entry of the telephone cabling.
-
Damp is not a problem at my house, including the point of entry of the telephone cabling.
Calcium bridging is an example of strange things that can happen due to normal dampness in every basement, and problems due to desert dry conditions. All 73 problems tested for by temporarily connecting only an ADSL modem directly to the master socket with all other connections temporarily removed.
IOW dampness can create problems just like rain only once a month or temperature changes. Point is to identify the problem at a half way point. Which side of the master socket causes low band signal degradation? No reason to even ask why yet. First, which side of the master socket is the problem?
-
First, which side of the master socket is the problem?
Well, I guess the very first thing of all is to wish hake's father a speedy recovery... :-\
-
Thanks for those kind wishes.
-
Hi
just tried Google with "80KHz"(about mid point on UL frequencies ) various bits of equipment use that frequency including MONITORS . Is it possible that there could be some interference affecting the UL frequencies ?
Regards Jeff?
-
Is it possible that there could be some interference affecting the UL frequencies ?
Yes. ADSL wire is twisted pair so that 80 Khz interference does not exist. However if someone accidentally connected two phone wires to different (not paired) twisted wires, then ADSL interference could exist. Just another of those 73 problems that get tested for by temporarily connecting a failing ADSL connection directly to the master socket - to learn if that problem (or so many others) still exists or disappears.
-
Using a portable wireless on LW, I found that one of my PC's PSUs was spewing out a horrible noise at 210kHz. It has been binned. Isolating that faulty equipment has not, however, affected the upstream problem.
Are there any radio receivers which can scan the whole of the frequency range of an ADSL signal? The LW range of available receivers only seems to cover 150kHz to 300kHz.
-
Are there any radio receivers which can scan the whole of the frequency range of an ADSL signal? The LW range of available receivers only seems to cover 150kHz to 300kHz.
If a device creates noise at 100 Khz with enough strength to interfere with twiste pair ADSL, then it is also making noise in the 150+ Khz range. However relevant noise would more likely be on the ADSL pair from another pair of wires - not noise in the air.
Another in the long list of 73 problems include an inadvertant grounding of one wire in the pair. Until you have determined which side of the master socket is the problem, well, a 30Khz noise interference is only one of too many possible reasons. Rather than take on all 73 problems one by one (which is what you are doing), first is shorten that list massively by determining which side of the master socket has that problem.
-
Hake,
I'd not worry tooooo.... much about the exact frequency of the noise. If the interference signal were a perfect sinusoidal waveform, then it would only be detected at it's own resonant frequency. But it's awfuly difficult to generate such a pure sinusoidal waveform even when you're trying to do so, and it's most unlikely to occur by accident.
Much, much, much , more likely it's a spurious waveform, full of harmonics, that happens to be based on some particular frequency, like - say - 80 Khz. That being the case, you'd detect signals of various strenghts at 2x, 4x, etc the basic frequency.
If you can 'home in' in some noise on your LW radio, that co-incides with the problems, then I'd not worry about the exact frequency. However, bear in mind that LW radio signals tend to follow the curviture of the planet. It's quite normal on a summer's eveing to detect crackle emanating from many (hundreds of) miles away. Make sure the onset (&drop) conincides with the problem before taking too much notice off it.
-
Gotcha! I heaved the big chest of drawers out of the way to gain access to the master socket. I connected direct to the test socket via an Exelcus passive microfilter and, bingo, the upstream bitrate is 224kbps (a few weeks ago it was 448kbps).
I am therefore confident that any problem is not in the cabling that I have installed.
What should I do next?
-
I connected direct to the test socket via an Exelcus passive microfilter and, bingo, the upstream bitrate is 224kbps (a few weeks ago it was 448kbps).
What are your 'signal to noise' or 'signal strength' dB numbers during good and bad operation? Numbers are always important.
Step back. Just to be sure I understand what you posted. First, connection to the master socket is without anything else in the house connected.
Second, apparently you don't understand what the filter does. Everything else in your house 'eats' DSL signals. Your DSL modem must make a direct connection to the providers DSLAM. Not through a filter. That filter is normally between BT phone line and everything else in the house - to separate everything else from eating DSL signals. Correct operation is ADSL modem connected directly to master socket and ISPs DSLAM. Your test has nothing else connected to the master socket.
Is an ADSL filter defective? Your test - only modem connected and nothing else; not even a filter - is the test.
Again, dB signal numbers are important. However it sounds like you have a problem that must be solved by the ISP. Any problem on the BT side of a master socket is 100% their problem. Best you can do is have so much information that they cannot deny the problem. Those dB numbers would help make your case (and might have made my reply more useful). If a modem loses signal strength when it is the only item connected, then your problem can only be solved by the ISP.
Now, is any (even the slightest) noise heard on other POTS phones at any time? Any crackling, pops, or snaps – even faint? And when? If so, that is critical information (a symptom) to help identify and to confirm that the problem exists and when the problem is finally solved.
-
I recollect that the upstream SNRM during a good spell (upstream speed 448kbps) was around 16db. The quiet line test is just that: quiet. There is no audible noise and none has been heard during the many quiet line tests I have done over the past few weeks.
PlusNet's support has now taken the issue on board.
I have had variable line performance over the three years I have been on MaxADSL. Disconnects were becoming more frequent over the past six months and the stats wobbling around more. My main worry was intermittency and the difficulty of firm evidence. Now, at last, I have a firmly demonstrable problem which was previously not the case.
-
HI
quote My main worry was intermittency and the difficulty of firm evidence.
As suggested in an earlier post by hpsauce quote You could try Routerstats of course, see if that shows any more detail.
you can get it from here http://vwlowen.co.uk/internet/files.htm
Regards Jeff
-
I recollect that the upstream SNRM during a good spell (upstream speed 448kbps) was around 16db.
You now have a benchmark. Failure would probably be indicated by 10 dB. Your ADSL would still work - but has failed. An example of why the numbers are so important. A benchmark that says when Plusnet has 'really' fixed the problem and how to solve problems before cuaing ADSL speed degradation.
Also a method to determine if anything else in your house (when reconnected to the master socket) contributes to DSL signal degradation.
My experience: you are about to deal with the hard part. Get them to finally send out someone who really knows how ADSL works. Some ISPs will do anything to avoid rolling a truck. Some (if not many) understand nothing more than disconnected wires. Good luck.