Other Parts Discussed in Thread: MSP-EXP430F5529LP, USB-2-MDIO,
Dear Team,
Our customer work on Proof Of Concept (PoC) with DP83867E PHY. he'd like to ask us for help with TDR cable diagnostic feature. I use DP83867ERGZ-S-EVM kit with USB-2-MDIO (MSP-EXP430F5529LP) to write/read PHY's registers.
DP83867E datasheet mentions that the chip can measure a cable length with ±1-m accuracy, but the results I got were pretty inaccurate.
He followed SNLA253 (https://www.ti.com/lit/an/snla253/snla253.pdf) with a suggestion from https://e2e.ti.com/support/interface-group/interface/f/interface-forum/825997/dp83867cr-tdr-peaks-location-calculation/ (We looked in further and found that the factor to substract for DP83867 is 17 instead of 8.) to translate LOC to meters.
He measured 3 cables (cat 5e 2 m, 3 m, 10 m) in the "Open Circuit Cable" connection. For the 2 m cable and the 3 m cable LOC translated to a length 0.94 m. For the 10 m cable LOC translated to a length 7.77 m.
Here is an example for the 10 m cable :
1) LOC of peak 0 (channel A) for the 10 m cable was 0x1C -> 28
2) IL = (28 * 0.8621) – 17 = 7.1388
3) FL = 7.1388 + ((70 - 7.1388) * 0.01) = 7.767412 m
Does the mentioned accuracy +- 1m apply to shorter cables as well? If so, would you please suggest a better algorithm to translate the LOC values to meters?
Also, DP83867E datasheet mentions :
For all TDR measurements, the transformation between time of arrival and physical distance is done by the external host using minor computations (such as multiplication, addition, and lookup tables). The host must know the expected propagation delay of the cable, which depends, among other things, on the cable category (for example, CAT5, CAT5e, or CAT6).
Unfortunately, SNLA253 doesn't say anything about the cable category.
Do we have any documents describing how to project a cable category to the LOC - >m calculation?
Or could you guide us how how to project a cable category to the LOC - >m calculation?
Many Thanks,
Jimmy
