Cable Diagnostic accuracy?

[jluthman]

Everything here is on *unpowered devices*. I'm talking about the cable itself, not plugged in to anything.

But saying the only real concern is a cross-short, this looks to be the only thing to worry about which I have yet to see.

[sirhc]

Everything here is on *unpowered devices*. I'm talking about the cable itself, not plugged in to anything.

But saying the only real concern is a cross-short, this looks to be the only thing to worry about which I have yet to see.

The cable itself not plugged into anything should ALWAYS report OPEN on all 4 pairs.

If you get something like a short or abnormal termination on a CABLE WITH NO DEVICE then either the cable or end is bad or the port Ethernet Transformer is damaged.

You can test this by running the Diagnostics on a port with no cable and once again you should get all OPEN but if not then that is an indication you damaged your Ethernet Transformer.

Keep in mind that damage to Pair 3 or 4 will still allow a 10/100 link to happen as 10/100 only uses pair 1 and 2 for data and pair 3 and 4 only for power.

[jluthman]

If you get something like a short or abnormal termination on a CABLE WITH NO DEVICE then either the cable or end is bad or the port Ethernet Transformer is damaged.

This isn't the case as after adding the device it both powers up and does gigabit with port 2 24vh and port 8 24v.

[sirhc]

If you get something like a short or abnormal termination on a CABLE WITH NO DEVICE then either the cable or end is bad or the port Ethernet Transformer is damaged.

This isn't the case as after adding the device it both powers up and does gigabit with port 2 24vh and port 8 24v.

Well I think I answered your question.

When running a cable diagnostics on a device that is not powered results will vary based on the device design as the cable diagnostics send basiclly a signal down the wire across magnetically couple connections and based on how the device is designed different results will occur.

Once the device is powered on you should always get all OK on 1G devices (Surge Suppressors will mess with this)

10/100 POE devices often cut corners and do not use 4 channel transformers and direct connect pair 3 and 4 to the board forming an electrical connection.

The main thing to worry about is seeing CROSS-SHORT which is always BAD.

You can Google and read up on Ethernet Transformers and Cable Diagnostics and learn more about what is going on here.

[jluthman]

>When running a cable diagnostics on a device that is not powered results will vary based on the device design

The cable is *NOT* plugged in to a device.

>If you get something like a short or abnormal termination on a CABLE WITH NO DEVICE then either the cable or end is bad or the port Ethernet Transformer is damaged.

In my situation, there is a CABLE WITH NO DEVICE and both the cable and Ethernet Transformer are in fact good.

[sirhc]

Then you have bad cable that is allowing cross talk between wires.

Or if a cable was kinked or stretched during installation it can cause crosstalk and may also result in high Ethernet Errors under heavy load.

Look, do some testing.

Use a long cable on the ground that you know is good and not damaged and run the test.

You should NOT get a short on any pair on a cable with no device on the end unless 1 or more things is bad:
1) Cable is bad or damaged
2) End is bad or crimped wrong
3) Transformer - But in this case I do not think so as it would show up on a test with no cable and you would not get 1G connections.

That's it, nothing else.

This is not a software feature we wrote, it is built into the switch core software we just call it and it returns the values and we display them.

I am not sure where I can be of any further help here, you need to do some testing.

[jluthman]

>Use a long cable on the ground that you know is good and not damaged and run the test.

I am on the ground, in the office. The cable is my test ~300 footer and has worked on many other things functions just fine on the WispSwitch. It's just the cable diagnostics that come back bad (either powered off device or no device at all).

My initial question is on the reliability of the cable diagnostics, of which I believe I have my answer.

[sirhc]

>Use a long cable on the ground that you know is good and not damaged and run the test.

I am on the ground, in the office. The cable is my test ~300 footer and has worked on many other things functions just fine on the WispSwitch. It's just the cable diagnostics that come back bad (either powered off device or no device at all).

My initial question is on the reliability of the cable diagnostics, of which I believe I have my answer.

The cable diagnostics has been very reliable to me as far as length and telling how each pair is terminated. It is by no means an expensive Fluke Tester and can not be used to sweep a cable but it is a decent hammer (crude tool)

I would like you to use another cable to see as I have never seen this before except with damaged cable that had been stretched or kinked which causes cable cross talk between wires which would show up as a short.