Kitz Forum
Broadband Related => Telephony Wiring + Equipment => Topic started by: chazzo on August 17, 2019, 10:39:46 AM
-
Is it reasonable to run a filtered DSL signal (from the A+B terminals on a Mk4 faceplate) on a spare pair within extension cable that is also carrying the filtered phone signal (from terminals 2+5)?
At the moment my ADSL modem is fed from a microfilter at the other side of the house from the master socket. I'm preparing to upgrade to FTTC and I like the idea of a filtered faceplate, but the modem needs to stay where it is and I can't install a Cat5/6 extension.
I've bought an NTE5c with a Mk4 faceplate. I assume the point of the A+B terminals on the faceplate is that you can run a DSL extension over ordinary phone cable instead of Cat5/6 – is that correct?
And if it is, would it be OK to run the DSL connection on terminals 1 and 6 of my extension wiring, leaving the phone on 2+5?
At the extension socket where the modem is located I'll obviously need a new faceplate with BT and RJ11 sockets, or perhaps I could get away with a non-standard BT-to-RJ11 modem lead (if anyone can tell me which pins to use on the RJ11 side I'd be grateful).
Many thanks
Charles
-
I had a similar issue when Openreach installed my FTTC. What they did was to wire the new master using my existing extension wiring and then run the extensions back through the same cable . As my internal wiring is all 3 pair CW1308 and they are star-wired this was easy to do.
Previous to this I had a filtered faceplate on the master and was using A and B to run just to the extension with the modem. I used a dongle type filter on there so I could also have a phone.
There is certainly no issue with using a spare pair to run the unfiltered A and B connections back to a new socket - it should work just fine. You could use a standard extension socket for this and use a UK-USA adaptor plug to give you a modem socket.
One of these:
-
Assuming that your current master socket is an NTE5, then you own all of the extension wiring and can use it any way you think fit.
To use one pair for the telephony service and a second pair for the xDSL signal is acceptable. There is one proviso . . . once the wiring has been configured in that fashion, the xDSL port at the master socket should never be used. The reason being that if a modem was connected to the xDSL port at the master socket, it would "see" the pair (running to the extension socket) as a bridging-tap and the service would be significantly degraded.
At the extension socket, ideally you would install something like:
- a faceplate surround (https://www.run-it-direct.co.uk/euro-faceplates-backboxes/doubleeurofaceplate/)
- a "snap-in" 6P6C (a.k.a. "RJ11") socket module (https://www.run-it-direct.co.uk/bt-euro-modules/RJ11module/)
- a "snap-in" telephony (secondary) socket (https://www.run-it-direct.co.uk/bt-euro-modules/btsecondarymodule/)
Other suppliers are also available . . . I used the above examples for convenience and not as an explicit recommendation.
The above will fit into a standard UK single electrical back-box (either recessed or surface mounted).
-
Is there actually any point in doing this though? If that extension wiring is going to all the same places, using a different pair would surely introduce no less noise than using the main one?
-
Not a lot of point unless there is other extension wiring to be filtered off at the master socket by the filtered faceplate. If there's no other extension wiring, it's just about the filtered faceplate having a theoretically slightly better filter than a plug in microfilter.
-
Not a lot of point unless there is other extension wiring to be filtered off at the master socket by the filtered faceplate. If there's no other extension wiring, it's just about the filtered faceplate having a theoretically slightly better filter than a plug in microfilter.
Agreed. In my case the extensions are star wired so it is important that they are filtered to prevent bridge tap issues.
I would have left my wiring as I had it originally (filtered faceplate at master with modem extension unfiltered with a dongle filter on it.) but the Openreach chap thought it best to change it.
:)
-
Many thanks, all. I did the wiring last night and it's working as expected.
I've ordered the faceplate and modules that burakkucat was careful not to endorse, and in the meantime lashed up an adaptor at the socket serving the modem (wires 1+6, remember, so an ordinary BT-to-RJ11 is no good). I also ordered some plugs and will try making up my own modem cable. Using round cable I think that will work OK, and will mean I don't have to mess with the socket.
"Not a lot of point", said ejs and Alex Atkin UK. I agree, in the sense that I wasn't trying to improve connection speeds. I'm happy with my ADSL performance, which hasn't changed with the new wiring arrangement: 20 Mbps sync at 3 dB SNRM. I hope it continues to work fine when I move to fibre.
No, the only point was to get rid of the original dangling microfilter and possibly to benefit from good-quality components in the Mk4 instead of whatever cheap filter my ISP chooses to send me. I originally thought about putting a filtered faceplate on the extension socket serving the modem, but couldn't find such a product.
But having read what burakkucat and tubaman said about bridge taps, I'm thinking there might be another advantage. I can see how a modem plugged into the master socket could suffer from reflections off the extension wiring, but surely that's true as long as the modem is plugged into an extension that is not the last on the line (as is the case here)?
If so, I guess I should try plugging my modem into the test socket. If that improves the connection I could then think about cutting wires 1+6 on the "downstream" side of the modem's usual socket. Since that would allow the downstream sockets to continue working for phone service, it would be an argument for keeping the DSL signal on a separate pair instead of filtering at the point of use.
Charles
PS: although the filter section of the Mk4 has a quality feel, isn't the rest of the NT5c a piece of junk? I can now see why people here recommend sticking with previous versions. The faceplate is attached by the skin of its teeth, and I don't like the snap connectors -- mine were really stiff and it's hard to get both wires in straight. It's as if they looked at Wago connectors and thought "what's the worst copy we can make?". OK I bought it on Amazon, but it has Openreach branding and looks genuine.
I made the mistake of pulling the faceplate apart so that I could look at the filter section. What with the socket covers dropping out and those long pins on the "bridging" plug, it's hard to put back together :rant:
-
: although the filter section of the Mk4 has a quality feel, isn't the rest of the NT5c a piece of junk? I can now see why people here recommend sticking with previous versions. The faceplate is attached by the skin of its teeth, and I don't like the snap connectors -- mine were really stiff and it's hard to get both wires in straight. It's as if they looked at Wago connectors and thought "what's the worst copy we can make?". OK I bought it on Amazon, but it has Openreach branding and looks genuine.
It very much is junk (the NTE5c). The MK4 SSFP appears to be manufactured better but as it only fits on an NTE5c that's not much help.
I'd have gone with an NTE5a with an MK3 SSFP myself.
It's probably not worth the effort of changing it all though.
-
. . . having read what burakkucat and tubaman said about bridge taps, I'm thinking there might be another advantage. I can see how a modem plugged into the master socket could suffer from reflections off the extension wiring, but surely that's true as long as the modem is plugged into an extension that is not the last on the line (as is the case here)?
Yes, you are absolutely correct. That is a disadvantage of using distributed microfilters and not one centralised filter.
If so, I guess I should try plugging my modem into the test socket. If that improves the connection I could then think about cutting wires 1+6 on the "downstream" side of the modem's usual socket. Since that would allow the downstream sockets to continue working for phone service, it would be an argument for keeping the DSL signal on a separate pair instead of filtering at the point of use.
b*cat nods in agreement. If you have a three pair, CW1308 specification, cable already in-situ then split the signals at the master socket (with a centralised filter) and send telephony down the first pair and the xDSL service down the second pair. At the modem's preferred location, install an appropriate pair of sockets. The 6P6C ("RJ11") socket will be the end of the line for the xDSL pair. The telephony pair can continue to meander around your property feeding a daisy-chain of sockets.
-
If you have a three pair, CW1308 specification, cable already in-situ then split the signals at the master socket (with a centralised filter) and send telephony down the first pair and the xDSL service down the second pair. At the modem's preferred location, install an appropriate pair of sockets. The 6P6C ("RJ11") socket will be the end of the line for the xDSL pair. The telephony pair can continue to meander around your property feeding a daisy-chain of sockets.
Thank you. I will do exactly that.
I know nothing about transmission lines, but shouldn't it be possible to put a termination resistor at the end of the stub?
-
I know nothing about transmission lines, but shouldn't it be possible to put a termination resistor at the end of the stub?
It is true the that metallic pair carrying the xDSL signal is, essentially, an RF transmission line.
In the case of a sub-optimal wiring situation, where one or more bridging taps exist due to star-wiring within the property, I would imagine that the value of a resistive terminator, appropriate for one stub, could be calculated. Assuming that an exact match for a resistive terminator could be applied, what would happen when the modem/router is moved to a different socket? If the situation was the worst possible star, i.e. all extension sockets are fed directly from one central point, the whole process of calculating and then fitting terminating resistors to all the stubs would need to be repeated. The theory is good but in practice it would be impossible to achieve the perfect result. Hence there will always be impedance mis-matches and resulting frequency dependent attenuation.
It is far simpler to remove the bridging tap(s) than to attempt to mitigate the problem that such tap(s) would cause.
-
The theory is good but in practice it would be impossible to achieve the perfect result. Hence there will always be impedance mis-matches and resulting frequency dependent attenuation.
Once my new sockets arrive I'll content myself with cutting the downstream wires :)
-
New faceplate with dedicated phone and data sockets installed in the extension box, and wire cut so that the downstream sockets are no longer part of the DSL circuit.
The terminals on the faceplate modules are quite tightly packed, so punching down the wires is challenging for the unpractised, even with a half-decent tool. Still, the job is done and it all looks neat. No change in ADSL sync rate.
-
Excellent. :)
Just one query . . . What is the status of your current master socket? An NTE5/A plus a SSFP? Or have you installed that NTE5C and its SSFP?
-
This is with the new NTE5C and SSFP. In fairness, the faceplate seems to clip on OK now though I wouldn't trust it near vacuum cleaners, dogs or small children.
Just in case, I did order an SSFP for the original NTE5a, so I could go back to that if necessary (it hasn't arrived yet). I also have an old ADSLnation faceplate but I think – I hope – I read somewhere here that it's not likely to be good for VDSL. If that's not the case then I have wasted my money twice.
This is a rented house, so I need to bear in mind that when I move out I might have to put things back to how they were. I don't think anyone will notice the change of master socket, and I figure that as an incoming tenant I'd appreciate any solid improvements to get the best possible connection.
The house is 14 years old and was built by a small local builder who knows what he's about. But why they could not have put in Cat5 instead of phone cable is beyond me.
-
Thank you. All understood. :)
-
Well, thank you. And everyone else who has taken the time to post. It’s really much appreciated.
-
Well, thank you. And everyone else who has taken the time to post. It’s really much appreciated.
b*cat blushes as he performs one of his Japanese-style bows. :blush:
-
Pretty sure my main line is the old ADSL Nation (marked as ADSL 1.0), it performs no different to any of the dangling microfilters supposedly designed for VDSL in my experience. Although to be fair, I don't have bridge taps, just a single extension leading to the DECT base station so that it not right next to the modem.
My second line I think has the Mk3, the first line has better stats although they are both syncing the same lately.
-
Pretty sure my main line is the old ADSL Nation (marked as ADSL 1.0), it performs no different to any of the dangling microfilters supposedly designed for VDSL in my experience.
Interesting but perhaps not surprising? The ADSLNation filter was supposed to be one of the best in its day, wasn't it? And I have the (uninformed) impression that dangling filters range from good to poor, at least in theoretical terms. I've no idea whether changing filters makes much difference in practice, and if it does, which design works best will probably depend on the line. This post from 2012 (https://forum.kitz.co.uk/index.php/topic,11859.msg227707.html#msg227707) compares the ADSLNation with presumably the BT Mk3 SSFP and found no difference in sync speed.
I don't even know how or why the filter affects the DSL signal. I presume the input stage of the modem ignores anything at speech frequencies, so we're only concerned with how it might muddle or attenuate the RF. A quick search of Google Scholar (https://scholar.google.co.uk/scholar?hl=en&as_sdt=0%2C5&q=ADSL%7CVDSL+POTS+filter&btnG=) turns up some references to active filters (ass opposed to active components), which I guess could be the future – unless we just give up on wired telephony altogether.
I had a quick look inside a Mk3 and it looks quite similar to the Mk4, with four large inductors. Bcat suggested they may be identical (https://forum.kitz.co.uk/index.php/topic,18631.msg334757.html#msg334757).
-
The excellent ADSL nation active filters cope better when there are many devices plugged in to the phone line or awkward devices with unusual impedance. The active design is more sophisticated in order to cope better with the different requirements of the on-hook vis off-hook situation. If you have a single ‘normal’ phone or no phone at all then you won’t see an improvement from an active type filter. I have seen very slightly better performance from a passive BT Pressac type filter but almost non existent, in the limits of what can reliably be measured only after many tests. For an explanation of the design rationale see the textbook by Golden and Dedieu - I think it’s the second volume which has a chapter on filter design. One of the authors worked on the only other active filter I’m aware of, made by a French company whose name escapes me.
If you have a situation where DSL performance goes bad while a telephone call is in progress, then you need a better filter, or even better a better filter and dump the telephone which is causing problems in favour of a different model, or get rid of telephones altogether.