This is what Cisco has found so far.. Looks like a quick fix might be to use cable pinned out with only 1,2 and 4,5.. don't use a full blown 4 pair cable..
Hi all,
I wanted to try to summaries what we are doing to try to achieve progress, forward momentum, resolution to the issues that you have been experiencing regarding the WIC-1DSU-T1-V2= module
01) We are able to recreate the failure condition consistently in a Cisco lab environment
02) This issue was simulated by connecting 2 "landcruiser" cards back to back with 3 unused connectors fused together - In this scenario we see no problem
03) We have identified a difference between how we terminate the unused pair between V1 and V2 card (which causes noise to be induced on the connector)
04) The working theory is that the failure is due to V2 card inducing noise on the receiver at the connector.
05) As soon as one end is open - we see the same type of errors as seen by SBC.
06) We are shipping reworked boards/ cables for use in testing
07) In addition to the above theories, Cisco is also investigating receiver noise immunity robust differences between V1 and V2. Engineering is reviewing the default setting on the V1 and V2. Cisco is working to also improve the framer firmware performance to mitigate this issue. Initial testing appears positive.
We are attempting to validate several possible fixes to this issue during the planned test window occurring on Friday 4 February.
Regarding the news forum thread provided by Roger, we can confirm that the only difference between Build of Materials pn#'s 80022193-01 and 80022193-02, is that
a label was added to 80022193-02. no other differences exist. Cisco's engineering team are working to investigate the statements made in the newsgroup discussion thread and will close this off shortly.
DETAIL OF WIC-1DSU-T1-V2= ISSUE RECREATION/THEORY:
Based on the report errors (LCV and PCV), engineering is able to repeat a specific scenario that could exist in the customer environment. This scenario is the combination of an incorrect wiring of the cable and the design of the V2.
Specifically, the V2 RJ45 connector connects unused pins (3, 6, and 7) together. If a straight through, 4-pair cable segment to the V2 is used such that differential signals are split among two pairs, one conductor of each pair, with the unused conductors wired through to the V2, the resultant end-to-end transmission is impaired. (more simply put, the V1 is robust against certain incorrectly wired scenarios using 4-pair cabling while the V2 is not.). At this point, it is too early to draw any conclusion. Engineering only has a working theory that we would to get the confirmation from the field. See attached test data.
Here follows a draft of the actions that are intended to occur on Friday
To test this theory, engineering has sent out two modified V2 cards and several special cables (pin3,6,7 disconnected). If the modified V2 works in the field, it would confirmed our theory. . Using the special cable on the production V2 has the same effect as the modified V2.
The special cable pin out is as followed
Pin 1 <--- twisted pair A ----> Pin 1
Pin 2 <--- twisted pair A ----> Pin 2
Pin 4 <--- twisted pair B -----> Pin 4
Pin 5 <--- twisted pair B ----> Pin 5
