Signatures Explained

Cirris cable testers use a mathematical algorithm to derive a six-character “Cable Signature” for each learned wire lists. Although the details of this concept are incomplete without a long discussion of the math, you will find below a “basic” explanation of how signatures are derived.
Each test point used in a wire list is assigned a prime number. After the wire list is generated, “learned” by the tester, the prime numbers of the test points used in that wire list are added up. This results in a unique “signature” for that wire list. Thus it could be expressed that, the signature of a wire list, is the “sum of the prime numbers”. In addition, the algorithm adds into this mix the unique signatures of each adapter used in the wire list as well as their position on the tester. The algorithm uses a hexadecimal numbering system which is why signatures consist of digits (0 – 9) and letters (A – F).

Adapter Signatures?
Adapter signatures are generated using a similar algorithm. The number of position in the connector, the gender of the connector and how the strapping information on the board is configured all play a role in the generation of these unique adapter codes.

What does this all add up to for the customer? There is one chance in 16,777,216 that different wire lists will end up with the same signature. In the unlikely event that this occurs the tester will still catch all cable errors because the tester tests against the information found in the wire list. It does not test signature information against signature information!

Single errors will always have different signatures.
It may be helpful to know that the least significant bit of the first character of the signature will change even/odd with the number of connections under test. This means that a single open or short will add or eliminate respectively a connection, causing the signature to change. Also, a miss-wire will always cause the signature to change. Hence, all single errors in a wire list will cause the tester to display a different signature. The risk of a signature on a good cable matching the signature of the same cable but with defects is extremely unlikely.

Duplicate signatures on different wire lists can happen however.
As mentioned earlier, the 6 character (12-bit) signature has one chance in 16,777,216 of matching the signature of another assembly when the wire lists are different. In 128 points, there are 3 X 10 213 possible different wire patterns. That means that there are 1.8 X 10 206 different wire lists on average for each signature. Clearly this means that there are many potential matching signatures for different wire lists. But the question is how likely will we see one? There is a reasonable chance that in a large number of different wire lists you could see a matching “Cable Signature”. The math indicates that there is a 1% chance of two cables having matching signatures in a group of 600 unique cables.

Cirris Recommendations:

1. Do not use “Cable Signatures” to identify cable assemblies. Someday you may end up with two cables identified with the same name (cable signature) and yet are different wire lists. (In the same way, we don’t use our birth date for a unique name)
2. Always use a “likely correct” cable to learn from. This presents negligible risk (no risk for single errors) of cable assemblies having the same “Cable Signature” even though the cables are different. It is always a good practice to save the learned “good” wire list to memory and to not learn an assembly each time you wish to test a run of cables. Another safe option is to print out the learned wire list and compare, line by line, the learned wire list to the saved wire list to confirm that each list is the same. (You may still compare signatures if you choose to manually inspect wire lists because our experience has been that people do miss differences much more often than once in 16,777,216 times).