Crosstalk in the SpectraStrip cable

05/23/03

Signal source: Discriminator LeCroy 4413

Cable: SpectraStrip, length: 90 ns.

Load: ECL/NIM convertor LeCroy4616

Oscilloscope: high impedance probes.

Fig. 305231752. Red- single pair8, blue: pair8 + inphase pair7, green: pair8 + antiphase pair 7

Fig. 305231759. Red- single pair8, blue: pair8 + inphase pair9, green: pair8 + antiphase pair 9

The pair 8 senses a crosstalk from the pair7 but not from the pair9

Fig. 305231816. Signal on pair8 with antiphase pair7 and pair9. Blue: pair8 is disconnected from the source and open, green: pair8 disconnected and shortened, red: pair8 is disconnected from the source and terminated using 100 ohm.

Summary

The same signals have been applied to the neighboring pairs.

• The rise time (from minimum to maximum) is 7 ns
• Signal on pair 7 affects the signal on pair 8
• Signal on pair 9 does not affect the signal on pair 8
• Similarly, talking pairs: 1-2, 3-4, 5-6, 7-8,…. Not-talking pairs: 2-3, 4-5, 8-9,…
• If we take not-talking pairs and shift the signal on the aggressor line by 5ns then the victim signal will be distorted.

A possible explanation could be due to different wire lengths of pairs 2 and 3 relative to pairs 4 and 5 but the directly measured difference in delay between talking pairs is not detectable.