Accuracy of Cable Test with MS

[wey2go]

Anybody has weird results from the MS cable test?

It seems to me that if the end devices are "Meraki MR" devices, it seems to be always accurately (did not measure actual length but reported length looks reasonable to me). If end devices are PC (mainly Intel NUC), also the case.

However, I am getting wrong figures if the end device are Sonos AMP and Dahua NVR. Eg have a Dahua NVR which is in the same rack connected via a 2m patch cable, but cable test reported it as 19m (1Gfdx and all 4 pairs OK).

Anybody having this issues?

[BrandonS]

Yes, I have had "weird" and inconsistent results as well.  I don't consider it much more than a base line test and don't completely trust it.  That said, I think it is still a useful starting point, but I would not take the results and tell a cable contractor they need to go fix their cable necessarily without closer inspection.

[PhilipDAth]

The cable length is determined using TDR (Time Domain Reflectometry):

https://en.wikipedia.org/wiki/Time-domain_reflectometer

It measures the time for a signal to be reflected due to tiny imperfections in the cable termination.  The signal travells at a constant speed, so the length can be calculated.

If you find that a particular kind of device is constantly mis-reporting length it may also be that the device is messing with the impedance of the cable solution as well.

[joopv]

On MS120-8P and MS120-24P switches i see that wire pair 4 is not tested on ports that are in 100FDX mode.

• • Port Link Length Status Pair 1 Pair 2 Pair 3 Pair 4

    17	100fdx	1 m	OK	ok	ok	ok	-

    Is it a bug or a feature ?

[Rylo]

I have found the cable test tool pretty useful. I understand the slight inaccuracies of TDR, true testing really needs a device on both ends. But this tool makes for a great quick analysis. 

I am curious about a result I am seeing. One of the pairs shows "XXX" I couldn't find a document referencing the different results of the feature.

[PhilipDAth]

What was XXX?  Was that a number of some piece of text?

[Rylo]

It was literal...

[HoneybadgerIT]

I am also wondering what XXX means for a pair, as well as "unknown" and "short".

The Meraki documentation on this is sparse unless I am looking in the wrong place.

[cmr]

Crossed over?

[Bossnine]

I had something similar show up just yesterday. The remote end of the cable had a netgear switch attached, mind you it was a defective Netgear switch.

once removed it tested fine

[Holli69]

Maybe XXX indicates a high amount of crosstalk over the pair.

https://documentation.meraki.com/MS/Monitoring_and_Reporting/Using_the_Cable_Test_Live_tool

[AaronS]

Not trying to necro a post, but I came across this thread trying to figure out an XXX result that we received from a cable test on pair 2 when troubleshooting a device not connecting to the switch and this post was the only reference I found to it. I contacted support, and they gave me the following answer:

"XXX means pair 2 are failed to report voltage for unknown reason."

Hope this helps if anyone else comes across this. We're pretty sure the contractor installed two bad cables (the other one shows open on pair 4, so we have to force it onto 100Mbps), so we believe the result is likely accurate.

[charlie130]

I was testing the cable length from a POE port on the MS225 to a Cisco Catalyst 9100ax AP and it was wildly inaccurate. It displayed 25.25 m or approximately 83 ft. in the Meraki dashboard. I then used a Fluke MicroScanner2 and the cable length displayed was 32 ft. So a 51 ft variance is a massive inaccuracy. Bottom line I will never use the cable length measurement from the Meraki device for any reason. Sad face emoji.

[cmr]

I've found it to often be accurate within a metre or so, what cabling type were you testing?  We generally use shielded cables from 5e to 6a.