Print

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

[waltergmw]

@ BKK,

As well as the two circuits taken off BEFORE the unused master socket they contained two unfiltered phones as well as an unfiltered red care alarm system.

It's little wonder poor performance was observed nor was the EU fully aware of what should be expected of the magic HH5.

Kind regards,Walter

[NewtronStar]

Walter[/b]'s Wheelbarrow performed the wiring optimisation) were present, the circuit was sufficiently out of balance for noticeable RFI to be seen on the QLN plot.

Just very curious as to out of balance pairs omhs is used to see the difference on each pair what would be an unexceptable balance to see RFI on the QLN graph is it 1 omh to 20 omhs ?

[burakkucat]

It would the AC balance of the circuit that is the most critical, though a lack of DC resistive balance would also signal a circuit abnormality.

I believe that if you search the forum for the phrase "AC balance", you will find more than one post where B*Sheep has mentioned the percentage AC balance for "good" circuits . . . I have a feeling it is something like greater than 60% (or was it 65% ?).

[NewtronStar]

I believe that if you search the forum for the phrase "AC balance", you will find more than one post where B*Sheep has mentioned the percentage AC balance for "good" circuits . . . I have a feeling it is something like greater than 60% (or was it 65% ?).

Yeah but what are these percentages from the JDSU display when converted to ohms as omhs is used to show the resistance on the pairs and if one leg on the pair has more resistance (HR) than the other leg on the pair it is out of balance.

[burakkucat]

Perhaps Black Sheep will be able to quote a few typical circuit parameters for you? 

There is no way that you could measure the DC resistance of each leg. For resistive fault location (I know there is a menu option for such on a JDSU HST-3000 or an Exfo AXS-200/635) the circuit would need to be disconnected at each end and, depending upon which sort of test is to be applied, the pair would need to be looped at the far end.

Every cable is manufactured to a specification. I attach three snippets, below, showing some of the physical characteristics for CW1128, CW1224 and CW1326 cables. A technician performing resistive fault location on a particular length of cable would identify the cable type, look up its characteristics to determine the parameters to used and then make the measurements.

It is no secret that I own an Ohmmeter 18C and that device can be operated in a number of modes. If you are interested, I could scan the relevant manual and send a copy in your direction . . .

[NewtronStar]

Perhaps Black Sheep will be able to quote a few typical circuit parameters for you? 

There is no way that you could measure the DC resistance of each leg. For resistive fault location  the circuit would need to be disconnected at each end and, depending upon which sort of test is to be applied, the pair would need to be looped at the far end.

That sounds like a job for two engineers one at the PCP cab and the other at the premises NTE5a plate

[burakkucat]

That sounds like a job for two engineers one at the PCP cab and the other at the premises NTE5a plate

Indeed it does. Perhaps that illustrates just why some of the "best of the flock" look so hassled . . . trying to be in two places at the same time.

[waltergmw]

Gentlefolk,

Meanwhile this morning a further pleasing result.

25/01/2016 03:32 - RESYNC detected (DS 19998 Kbps, US 6000 Kbps), AS = 35, Retrain Reason: 1
26/01/2016 11:09 - RESYNC detected (DS 19998 Kbps, US 6000 Kbps), AS = 902, Retrain Reason: 0
27/01/2016 05:52 - RESYNC detected (DS 22399 Kbps, US 6000 Kbps), AS = 41, Retrain Reason: 1
28/01/2016 03:29 - RESYNC detected (DS 22399 Kbps, US 6000 Kbps), AS = 58, Retrain Reason: 1
29/01/2016 03:36 - RESYNC detected (DS 25000 Kbps, US 6000 Kbps), AS = 26, Retrain Reason: 1
31/01/2016 03:35 - RESYNC detected (DS 27398 Kbps, US 6000 Kbps), AS = 58, Retrain Reason: 1

Kind regards,
Walter

[Black Sheep]

Not 100% sure in my head what it is being asked here ?? I shall endeavour to answer what I think you mean ??

If we wish to test from lets say the EU's premises, or the DP etc etc, back towards the Exchange, then obviously nominal voltage will be applied (circa -50v), and as such only a TDR (HR Detection) test, and AC Balance test can be applied.
To obtain readings against a 'Dead pair' we simply access the Remote Unit Emulation (RUE) via our phones and can access a variety of requests such as ............. remove voltage, replace voltage, apply CPI tone, loop line .......... we can then run a variety of tests including resistance tests (ohms). Ergo these tests can be performed on a singleton worker basis.

The perfect circuit will of course have an equal ohms reading on both wires, however this is very rare and so our edict initially was to get the 'pair' within a maximum of 3 ohms difference. This caused mayhem as jobs were taking a ridiculous amount of time to repair and often unachievable with E-side cables having lots of 'joints' which were causing the impedance mis-match, no matter if you tried a new E-side pair, the same readings would often be obtained.

So, a re-think of the situation now have us working at a maximum of 13 ohms difference between wires (or 'legs' as we call them). Again, the problem with issuing a blanket statement such as this is that it can be applied innacurately. If for example you live next door to the Exchange and their was an impedance mis-match of 10 ohms, then there is definitely something amiss. Apply that same reading on a rural route of 9km, and I would leave it be, if all the other readings were satisfactory and the circuit was performing as it should.

 

[burakkucat]

I think you have answered all the points of which N*Star would be interested. Thank you. 

[burakkucat]

Looking at the circuit's resynchronisation log, posted by Walter above, it seems as if a banding is in place on the US. My understanding is that the service is provisioned as 40/10, so there is the potential for "more to come".

[WWWombat]

I bet it is still banded DS too. Look at the sequence:

20, 22.4, 25, 27.4.

Almost exactly 2.5Mbps between each one - perhaps there is an aversion to using "x.5".

Any bets on the next one being 30Mbps?

^{Edit:
Ah. That wasn't the full sequence. Adding the first two, it seems to have been:
10, fix, 18, 20, 22.4, 25, 27.4.}

[NewtronStar]

A most upto date look at the pbParams would help along with the current SNRM state then i'll make a judgment wether this line which is under investigation will hit 30Mbps

[burakkucat]

A most upto date look at the pbParams would help along with the current SNRM state then i'll make a judgment wether this line which is under investigation will hit 30Mbps

Will the attached montage, with data from 0335 hours this morning, be of any use?

[WWWombat]

Hang on... the attainable went up in the last one!

My previous estimate, of 32Mbps, was based on the fact that the attainable was over-reading because DLM had intervened old-style - so would drop.

However, DLM has now removed downstream interleaving, so now it is using much less bandwidth for the FEC alone (which is still present).

I'm now guessing at somewhere around 35-37Mbps.