Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[waltergmw]

@ Everybody,

May I remind you all of one example where one of our local OR engineers achieved some improvement after fitting two RF3 filters on adjacent lines feeding a Thames Water kiosk.
It was a pure guess by him at first as we observed there was RF emanating from the area.
In this case the RF3 must surely have been blocking, or at least reducing, signals which otherwise were seriously affecting the EU's line presumably via O/h cable cross talk.

The 21CN ADSL1 line we were trying to improve - i.e. make work at all in a stable state now has a BRAS of 648 Kbps and a BT recorded line length of 7308 m.
The last record I have is on 7 May 2011, which is attached, and has a downstream attenuation of 74.2 dB. Even with the filters the mid-frequency bit loading is poorer than the upstream below tone 30 and downstream above tone 64.

The configuration is a 100 pr O/h cable splits at a DP 2 spans away with a 20 pr going on to the local DP. One pr on a thicker 1 pr drop wire 12 goes directly into an ADSL Nation Filtered faceplate on the NTE5 wall assembly. The two cross talk pairs are routed back one span into a BT 66 box where the RF3 filters are fitted and then the 2 pr drop wire goes underground from that pole into the kiosk.

Kind regards,
Walter

PS Although I wouldn't describe the interference as REIN (random noise) it was almost certainly RFI giving us all significant trouble to effect a cure.

[attachment deleted by admin]

[razpag]

Great feedback .... the plot obviously thickens now Walters posted up that particular instance ??

B*Cat -- If I could find any info at all on RF3's (which would be akin to wading through treacle), I very much doubt I'd be able to post it on here. BT/OR always have a disclaimer at the top of their pages stating 'For internal use only'.   

[waltergmw]

@ BKC,

You'll find pictures right at the bottom here:-

http://www.kitz.co.uk/adsl/btsockets.htm

The only active device is called a series inductor which has a double wound pair of wires and one I have has an identifier B302 C
The back of the PCB has no components and just links A to A and B to B via the inductor.

http://www.run-it-direct.co.uk/BT80BRF3.html

N.B. in the picture below I note the wire colours indicate it has been cross connected.
(This one was left by a litterbug probably because it was dropped from the top of the pole !)

Kind regards,
Walter



[attachment deleted by admin]

[burakkucat]

Great feedback .... the plot obviously thickens now Walters posted up that particular instance ??

I remember Walter's thread recounting that story. It has resulted in someone studying each pole that he now passes, if time permits. I wonder, having re-read that thread, if it could have been RF2 filters that were fitted to the two lines to the Thames Water pumping station?

B*Cat -- If I could find any info at all on RF3's (which would be akin to wading through treacle), I very much doubt I'd be able to post it on here. BT/OR always have a disclaimer at the top of their pages stating 'For internal use only'.   

Understood. If, for example, you had an RF3 in the back of your van and you had sufficient time to take a series of photographs, I wonder if that would provide sufficient clues to electronics Wizards amongst us?

Of course the idea situation would be the loan of an RF2 and an RF3, via an anonymous padded bag, delivered by Royal Mail, for an electronic "prod & poke" by one of the above Wizards. Then, once analysed, the devices could be returned and the circuitry discussed here.

[waltergmw]

@ BKC,

They were definitely RF3s as I watched the BT O engineer unpacking them from the poly bags.

I didn't bother trying to measure DC Ohms as I realise we would need a variable frequency signal.

Incidentally I've suddenly realised that Homer (AKA Martlesham Heath) has nodded ! The BT engineer had assumed the terminals were connected the same way round** but the result of his wiring he's swapped the A & B wires although I don't think that is of any consequence on the upside of the NTE5. ** Even though they are marked correctly on the PCB.

If anybody wants to borrow my second hand unit they are welcome to.

Kind regards,
Walter

[attachment deleted by admin]

[Shoei]

Hi,

Just switched to O2 Business BB on a very long wire so have been looking at interference issues hence the interest in this thread.

I think the RF3 is a common mode choke so it will reduce interference imposed equally on each wire of the pair while having little effect on the ADSL signal.

The modem will reject a huge amount of common mode anyway but it may not cope with high level interference. In that case an RF3 could reduce the interference to a manageable level.

There's some discussion of the I-Plate along with a limited schematic here.. http://www.jarviser.co.uk/jarviser/btiplate.html

[waltergmw]

Hi Shoei and welcome,

Just for interest here's an O2 service on a line of about 7140 m.
Note this EU had decided to leave the service on ADSL2+ and surprisingly it is so clean that it hasn't even dropped back to ADSL2.

Kind regards,
Walter



[attachment deleted by admin]

[Shoei]

Hi Walter,

Thanks for the welcome and interesting stats. That line looks similar to mine, 7.1km including a stretch of aluminium and the last 200m sharing poles with the mains.

It's holding a solid sync at 1375k on ADSL2+ with a 6dB margin. Sounds lame compared to most here but it's much better than my previous fixed 512k, which just held on by the skin of its teeth.

I won't go off topic with details of the setup and the interference issues/solutions but will post something when I've got my head around what works best.

Cheers,

Tony.

[burakkucat]

There's some discussion of the I-Plate along with a limited schematic here.. http://www.jarviser.co.uk/jarviser/btiplate.html

Thank you for that link, Shoei. The circuit diagram was a revelation to me -- I had always thought that the I-Plate was just a means of inserting an inductor into the "bell" wire.

[burakkucat]

Walter -- With your two excellent photographs of both sides of the RF3's PCB even these tired eyes can deduce and reproduce the wiring diagram. All we need is for you to examine it closely to see if the inductance of the coupled chokes is marked on the windings or core. Once that value is known, surely someone could calculate the . . . er . . . cutoff frequency . . . frequency band passed . . . or whatever.

(Muses: I wonder if Eric knows the correct formula to use? For some reason I have a feeling that he may be a Radio Amateur . . . )

[roseway]

(Muses: I wonder if Eric knows the correct formula to use? For some reason I have a feeling that he may be a Radio Amateur . . . )

Thanks for the thought, but no, I wasn't.   I was an electronics engineer though, so I ought to be able to answer the question. The thing is that a choke on its own doesn't have a cutoff frequency. For that you need a tuned circuit with coils and capacitors. Of course, coils have their own self-capacitance, so the RF3 may rely on this, but the value is unknown.

[Shoei]

There's an old thread on thinkbroadband.com where someone measured the devices on an I-Plate..

http://forums.thinkbroadband.com/technical/3392856-where-can-i-buy-a-bt-iplate.html?fpart=5

They measured one arm of the common-mode part and got 4.6mH/1.2Ω, the bell wire inductor showed  22mH/38Ω.

As Eric said those numbers don't help very much on their own, the behaviour of the choke will depend on a lot of factors including the quality of the core - presumably a ferrite, stray capacitance, load impedance etc.

It's only a guess but the RF3 seems originally to have been for situations where RF breakthrough was so extreme it affected a regular phone.

[burakkucat]

Some interesting information, above. Thank you, everyone.

I am now giving serious thought to purchasing a BT80B-RF3 and (unofficially, of course ) swapping it for my current BT80A at the junction of the drop-wire & the CW1308 internal 3-pair cable, which runs to the NTE5A (fitted with an NTE2005 filtered face plate).

Once my self-imposed testing phase is complete, I would then decide my next course of action based on the results obtained.

[Ezzer]

there is a plug in version of an rf3 which looks like a typical plug in splitter or a smaller dsl filter. The rf filters are usualy fitted if the normal voice telephony was picking up some interference. Typical things would be radio broadcasts (one village I can think of where everyone gets radio 5 clear as a normal conversation,great if you want to know the scores on a saturday, awfull if your trying to hold a conversation)
The rf2 were designed before the advent of broadband, I cant remember the attenuation rf2 caused in the dsl range but in was in the mid 40db's (ok I know thats mathematicaly speaking bad grammer) So you could only get dsl sync on a 0.5 copper line up to about 2.2km cable length. the rf3 lets the dsl range through. It sometimes cleans up a little of any possible interfering signals in the backround on a tetchy connection.
In fact on my old test nte I rigged up an rf3 via a 8pdt switch so at a flick of a switch I could compair results with and without an rf3. By the looks of the rf3 found I would say thats what an engineer had in his kit for just such a perpose and has been accidentaly dropped.
The quick way of telling the difference between an rf2 and rf3 if you cant see the print on the circuit board is an rf2 has 2 small black cylindrical objects inside, where as the rf3 has a small black square block.
There are some areas where due to local issues more properties have rf's fitted.I loved finding rf2's because that was such an easy job to fix (apart from where the rf2 has become very well hidden, lke the attic I once went in only to find a new roof had been built over the old one, the old one still in place and the rf2 was somewhere in between the 2  ) Norfolk/suffolk coast can have quite a few.
Just the fact an rf is fitted dosnt mean to say ther had been an issue in the past with interference. It may have been that an engineer only had rf's instead of a plain block terminal on the van at the time. Although a big clue there has been an issue is if theres an rf fitted right next to an nte.
If you've had an old alrm system connected to the line then there may be an old 15a filter. Its in one of the older style connection blocks that begin to resemble a bar of soap. With a 15a dsl will not work, even straight off the frame at the exchange. I've tried it.   

[burakkucat]

Thank you, Ezzer, that was an interesting read.

With regards to my own situation, I do just one thing at a time and then spend a while observing the results so obtained. At present this is how things currently are in b*cat's home (which was built in 1986). The drop-wire from the DP is attached to the usual fitting on the facia board, under the gutter. It then enters, via one of the soffit board vents, into the roof space and runs in a neat straight line to a point just above where the NTE5/A is situated, about eight feet below. Here there is a BT80A where the transition occurs from the drop wire to the three-pair CW1308 spec. internal cable which runs vertically down, through the ceiling, to the NTE5/A. The previous owner of the property had a wall mounted telephone at an appropriate height above the NTE5/A and had fitted a typical 1980s/1990s plug-in, flat, non-twisted four core telephone extension. Before I had a broadband service, I made use of that extension wiring for my modern 21CN telephone and my fully refurbished 1950s 300-series telephone.

Now with a broadband service, I have replaced the EU front plate with an NTE2005 and by judicial application of the heavy duty side-cutters, reduced the plug on the end of that extension kit to the four wire cores. The wire was then fed into the surface mounted back-box of the NTE5/A and the blue pair was extended by four inches by means of two soldered "Britannia" joints and insulated sleeving. The blue pair was then connected to IDCs 2 & 5 and the NTE2005 fitted to the NTE5/A. Currently the ADSL signal is fed to my modem/router via a plug-in modem extension cable (RJ11 plug to RJ11 socket) that is run as a temporary measure.

My first experiment has been to leave the modem/router powered up 24/7, as suggested by RazPag in another thread. There has been a positive outcome. The sync speed has improved as the target SNRM was gradually reduced by my CP's DLM system.

My second experiment, yet to be done, is to swap the existing BT80A for the BT80B-RF3. (Naughty cat!)

The third experiment -- really a task, not experiment -- is to replace all my wiring from the NTE2005 (a 6-way IDC version). I intend to have the filtered voice pair, run from 2 & 5, to a new pair of telephony only extension sockets and the unfiltered ADSL signal, run from A & B, to a new RJ11 socket. I have a reel of CW1308 spec., three pair cable and intend to use the blue pair for the filtered voice signal and the green pair for the unfiltered ADSL signal. That will eventually be done, when I finally manage to get "a round tuit".