cannot differentiate between interferance and the ADSL signal itself

I thought the idea was that the ADSL signal was balanced and flowing in both legs of the pair whereas the interference would be induced equally in both legs of the pair and have a return path via earth. Is this not the whole point of a common mode filter ?

RobertoS said:
Ah, do you mean a wired-in extension somewhere? If so then you are right.

It is indeed a wired-in extension that my router is plugged into and not a plug-in phone extension.

Edited by risk_reversal (Thu 03-Sep-09 19:43:05)

No you are thinking of what is know as a humbucker, that is a transformer that is used to reduce hum or interferance in circuits by breaking the DC path but allowing the signal to flow. a common mode choke is just two coils on the same former and is arranged to present a high impedance to a group of frequencys, signal or otherwise.
When running at audio frequencys, common mode chokes can be arranged to cancel interferance by connecting chokes wound in opposite directions on the transmission lines, the theory being that interferance induced in the lines would be cancled out, but the signal would not be affected. This only works at frequencys way below those used for ADSL.

Edited by deleted (Thu 03-Sep-09 20:26:20)

common mode chokes can be arranged to cancel interferance by connecting chokes wound in opposite directions on the transmission lines

that's what is in the iPlate.

http://www.butlerwinding.com/inductors/common-mode2.... is the best explanation I've found of the principle.

This only works at frequencys way below those used for ADSL.

It's only an inductor so it'll have an increasing attenuation with frequency, to a point.
http://www.yuden.co.jp/ut/product/inductor/TLF9UB302... shows an impedance / frequency chart, in an iPlate you wouldn't want a large inductance as it would attenuate the ADSL, but the common mode design should differentially affect interference over singal (voice or ADSL). The inductance of the series connection through the choke is 4.6 mH compared to 22 mH in the ring wire choke.

yes but the Common mode choke will only work over a fixed band of frequencys, and the one fitted to the I-Plate is set for telephony, not ADSL frequencys. the explanation is good as far as it goes, and explans how it would work at low frequencys, but the coils in a No3 filter (which is what is fitted in the I-Plate) are not wound anywhere near accurate enough for the cancelation to work at ADSL frequencys, they only work if the interferance on telephony is quite weak, on telephones close to an AM transmitter, No2 filters have to be used which block everthing but telephony, including broadband

Edited by deleted (Fri 04-Sep-09 17:04:43)

Hi,
Was doing some research on ferrite choke's for ADSL and found this thread.

Don't know much about electronics but I've found a few articles about supressing common mode (CM) inteference.

systemx - I've seen that ferrite cores are available in a variety of materials that appear to have differing attenuation effects on CM - i.e. you could choose from a variety of chokes that induce peak additional impedendance to your wiring at different points in the spectrum. Interestingly too I found an article which suggested things like a long, close fitting choke may work better than a loose one with windings - not sure whether this is good advice or not.

I seem to have a source of noise affecting my line during mid evening, and I was tempted to try cores to see if they would dampen the impact. Note - I'm aware that it may be mains-bourne inteference in which case perhaps a filtered power supply to the router might help? In truth I've only seen maybe one or two drop-outs per day but do have high levels of CRCs so if its not costly or a waste of time trying I'd like to eliminate these.

Do you think categorically there is no benefit in using these at all? Do you have any other suggestions - obviously which aren't overkill?

To clarify my set-up its a long line anyway with high attenuation - I've already fitted an XTE2005 and my extensions are wired on Cat5. An isolated one for the DECT phone which is upstairs and a ~2.5m run from the master socket to my study for the router.

Interested to see what people think anyway - thanks.

Edited by deleted (Wed 07-Oct-09 11:10:07)

It might be worth posting your line stats - you mention a long line but that doesn't really tell us too much.

NB. See the link in my sig. I found that changing routers can help a lot!

an i-Plate would cost less than a tenner to try and has a common mode choke. Strikes me that a wound choke is a lot more effective then a couple of twists of a phone cable around a core, but I'm no electronics engineer.

Hi b4dger!

I reckon 62-63.5 - hopefully real figures as the routers see it - not capped.
DG834GT reported it as 62, Billion 5200G reports it as 63.5.

I raised a fault with plusnet as recently I dropped to a 250kb profile and lost sync repeatedly even when not using the connection (way below the FTR - my MSR is 2072). This may/may not have been related to a couple of complete loss of service episodes where the copper was stolen from our exchange.

Plusnet put interleaving BACK onto the line (should have always been on it but was off for some reason). Sync returned to 1536kbps sync on the DG834GT - no tweaks. I monitored this 24 hours for several days with router stats and there were modest CRCs and only small changes with the margin staying around the 15db mark (no gradual reduction over time)

I decided to try the Billion 5200G with 6DB fixed target margin.
It will sync from 2848 to ~3104; just into the 2.5mb profile. I want to maintain this profile as a minimum if possible - this is why I want to try and mitigate the current noise problem if possible - its not widespread - just want to avoid spikes that force a resync the wrong side of 2848.

I've tried other routers from Speedtouch, Netgear, 2Wire with reasonable success, tweaking the Speedtouch / Netgear with DMT / UberGT. Given the current low occurence of noise intereference and general line stability (I think) this seems more sensible than for example using the Netgear with the percentage SNR tweak via ADSLCTL (UberGT / TeamGT tweak will reduce the 15db to 9/9.5 db at most which is not enough).

I know its more than you asked - but hopefully its enough information to show what I've been doing to try and coax the most from the line - hopefully without being too aggressive.

Colin

Hi Yarwell,

I thought you couldn't use an I-plate in conjunction with an ADSLNation XTE2005? That's what I have in place now - so seems from the thread all I'm missing is the choke?

Regarding the choke in the Iplate I refer back to some commments in the thread suggesting the choke that's included is designed to provide maximum CM protection at voice frequencies? The research I did into chokes suggested that they reject CM inteference with different strengths across the spectrum depending on the shape / size / material the choke uses. It also suggested use of more than one choke would increase the attenuating effect (I guess in much the same way that a well wound coil would).

Also as mentioned I saw some material implying close-coupling of the choke to the cable sleeve - and a longer sleeve might provide more protection than using a large diameter ferrite ring and winding turns about that. (Also on a side note I've used some structural Cat5e so it isn't practical to use that method on the cable as it currently stands - not flexible enough - plus I agree with you on precision - if I were to attempt that myself I doubt I woudl get the desired effect).

Anyway - not about to get shielded cable, conduit and glands and try to construct a Faraday cage! Looking to see if a few quid on choke's or maybe a mains filtering device might do something. I'm sure it'd be marginal but if cheap enough I'll have a go.

Colin
edited out a typo.

Edited by deleted (Wed 07-Oct-09 15:34:33)