Link to home
Start Free TrialLog in
Avatar of sllouw
sllouw

asked on

How can I verify e1 clocking setup using an ISDN analyser?

I am hiring an ISDN analyser which I am assured "does L1". This I agree would be necessary to independently verify the clocking setup of an E1 trunk between a Cisco IOS voice gateway (AS5350) and an NEC PBX (Model not sure).

The AS5350 is set to drive clocking/provide clocking at it's end. The NEC is set (I am told) to derive it's clock from the line. In fact there are 3 E1 links between these 2 devices, all 3 have same clock setup.

We have symptoms (intermittent faxing issues, slow couting up of errors on AS5350 end) which influence me to want to independantly verify the clock setup at both ends. Industry colleagues tell me that that an ISDN analyser properly equiped should be able to tell from the signal it examines from each  end if that end is clocking or being clocked by the line. I have researched the matter thoroughly but no one seems to know much and there is not much documentation out there.

This suggestion that an analyser would be able to tell me this information seems to be coroboroated(sp?) by the theory which in my understanding is as follows. The line encoding in our case is HDB3. HDB3 like other encoding methods have measures designed to deal with the 1's density problem. In the case of HDB3 a long series of 1s does not present a problem. A "violation" code is appearantly inserted where there are 4 or more 0s to be transmitted.

Presumeably it is this violation code that a L1 capable analyser will see, which would imply that the line is being clocked from that source.

Any comments appreciated, including those dealing with these questions
- How often might this (clocking) information be propagated by a clocking end?
e.g. Do I need to be lucky in having the analyser on the line when there is a clocking signal? Or are these signals likeyl to be sent often?
- Is this a routine matter for an analyser? My industry colleague tells  me he has seen this feature on an HP analyser. I am hiring a NetTest Lite 3000E.
- Might this or any other analyser be able to examine the quality of the clocking? etc
L3000E-basic-product.pdf
Avatar of tvman_od
tvman_od
Flag of United States of America image

I'd say that regardless of clocking source this rule of 4 consecutive zeros will be followed by both parties.

Do you have straight cable from PBX to AS5350 or it's connected over some sort of transport network?
Very often the network will clock all the terminals, so PBX should be configured to receive clock from the line and your Cisco should receive clock from the line as well. Any professional ISDN tester connected instead of terminal can give you clocking parameters such as jutter, code violations, frame violations and so many other interesting things.
Avatar of sllouw
sllouw

ASKER

Thanks. From your comment I infer that I can be more confident about the assumptions
- You have a clocked and clocking end in this scenario for E1 lines (And T1 I would imagine also); and
- 4 zeros treatment is indeed a method (the method in my case) by which the clocking end allows the clocked end to maintain synchronisation. This to accommodate what I understand as the 1s density problem.

Interesting you say both clocked and clocking end would make the same use of 4 zeros treatment. I was thinking that for traffic from the clocked end would not bother because there would be no objective achieved in terms of the model I have sketched out so far. Put another way why would the clocked end want to fix the 1s density problem when it TX traffic to the clocking end? Happy to accept what you say of course. Your comment does throw a spanner in my theory for now. Back to the books.

We indeed have a straight cable from the PBX to AS5350. L1 only, the devices are in the same comms room and the length of copper is about 4m or so.

I am about to go and fetch my analyser so perhaps will post questions in context of it's output.
As I understand HDB3 is a standard line code and I never seen that line code will be different for clocked and clocking sides. I might be wrong, I never tryed to check it out myself but according to books this 4 zeros rule is a part of the method regardless of clocking source. What king of connector do you have? If it's a short line, you may want to play with gain settings and specify that your line is less then 100 ft (30 m).
Cisco controller command
cablelength short {133 | 266 | 399 | 533 | 655}

http://www.cisco.com/en/US/docs/routers/access/3800/3810/software/command/reference/E1T1ref.html#wp32108
Avatar of sllouw

ASKER

Thankyou. Connection is
NEC PBX---(Copper)-----RJ45male > RJ45 socket(1 of 8 sockets)---(Think copper aggregation cable)---(SlimlineConnector on slot card. I believe part no is AS54-DFC-8CE1)

cablelength command looks useful. I have just got analyser and (sigh!) only BNO presentation no RJ45/48. Will open a separate question for clarity about how if possible I can convert between physical peresentations.
ASKER CERTIFIED SOLUTION
Avatar of tvman_od
tvman_od
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of sllouw

ASKER

Thanks tvman od great chat. I am hiring an analyser and I hope someone who knows how to use it, so I will be finding out a few answers. Would you like to know the outcome? I am not sure how to contact you. I have posted a quesiton to the community advisor group to see if possible. If not I will post something... Thx again
You can send me a message to ee {at no spam} tvman dot us