I was wondering if you can verify/certify speeds on Cat5,5e, etc cables using either 2 PC's or one PC and a loopback?
Those ridiculously expensive verifiers basically just measure packet flow and speed, from what I can tell, so this shouldn't be too hard to do with a computer or 2.

What am I missing?

I've seen PCs "link up" at 10mbit or 100mbit or 1000mbit but not able to pass any traffic due to a bad connector, miswire, short, open, etc. So that's the reason those testsets are made. They test more than just link speeds.

Thanks for the info. I have a small tester that shows if the wiring is correct via LED's on it, but this is not the kind of tester I'm talking about.

I'm talking of the ones that test your cabling to see if it passes the speeds to meet the standards, like cat5,5e,6 etc. I don't see why a PC couldn't do the same thing. Am I still way off?

There is much much more to what a cable certifier does than I can explain here..

Packet flow and speed are not really a part of the certifiers job.

Lets take a look at what a Fluke DTX series certifier does.

It tests everything to the following standards.

TIA Category 3, 5e, 6 & 6A per ANSI/TIA-568-C.2
TIA Category 5 (1000BASE-T) per TIA TSB-95
TIA Category 6 per TIA/EIA-568B.2-1
TIA TSB-155 (DTX-1800 only)
ISO TR 24750 (DTX-1800 only)
ISO/IEC 11801 Class C, D, and E, Ea & F(Ea, F and Fa DTX-1800 only)
EN 50173 Class C, D, E, EA & F (Ea, F and Fa DTX-1800 only)
ANSI TP-PMD
IEEE 802.3 10BASE-T, 100BASE-TX, 1000BASE-T
IEEE 802.3an 10GBASE-T

It support the following tests

Wire Map
Length
Propagation Delay
Delay Skew
DC Loop Resistance
Insertion Loss (Attenuation)
Return Loss (RL), RL @ Remote
NEXT, NEXT @ Remote
Attenuation-to-crosstalk Ratio (ACR-N), ACR-N @ Remote
ACR-F (ELFEXT), ACR-F @ Remote
Power Sum ACR-F (ELFEXT), PS ACR-F @ Remote
Power Sum NEXT, PS NEXT @ Remote
Power Sum ACR-N, PS ACR-N @ Remote
Power Sum Alien Near End Xtalk (PS ANEXT)
Power Sum Alien Attenuation Xtalk Ratio Far End (PS AACR-F)


In the end, all of these test can be traced back to a common test standard. WHy is this part important? It allows you to compare results between different manufacturers equipment.

The final reason is it provides a hard copy of the tested result for certification and archiving.


Standard NIC don;t have the hardware circuitry to even do the basic tests the testers can do.

When all else fails and you don't have a certifier, try to pull a large file from one location to another. Use FTP if you can to get rid of the overhead from SMB shares. If you can get 90-95% of the link speed, usually you won't have many problems.

The Broadcom suite on most servers can do cable length but that's about it. There is no substitution for a certifier if that's what the customer wants.

As others have said, just because it links up at "X" speed doesn't mean it will do it. Remember all those times you'd get a super high connection speed (like 48.8!!) on dialup only to get the same 2KB/s you always got?

Thanks guys! I see why a normal NIC can't pull this off, and now see a LOT more goes into it than just speed/flow. Fortunately I've never had to verify speeds on cabling, as most of what I do is residential and small business at the moment. Just wanted to know if it could be done via PC.

I guess if you developed some special hardware that either interfaced via USB, PCMCIA, etc, and used a program on the PC, it could be done, but would probably be easier to buy the tester and save some hassle.

Thanks again!:)

Have you tried IPERF testing? Set up one PC as the Server, and one as the Client.