How to test for bad cables?

[johnnyx]

If I ordered bad cables from Directron.com, I would pretty much know as soon as I plugged them in right?

If an orange was driving a racecar would it peel out? www.jxfiles.com

 

[HammerBot]

Yes, I guess. If a cable is bad, its probably not working at all. Never seen any cable that was 'partially' bad as in only being able to perform UDMA66 etc.

 

[aznguru]

I've had a 'partially' bad one where it would mis-detect the drive like 3/4 times. But then, that was my first attempt at making a custom cable...

-Dan
"I like to play with magnets..."

 

[lhgpoobaa]

I found a novel way, back when i was having IDE problems. eventually it turned out to be a controller problem, but i dont think any test can differentiate between controller problems and cable problems.

the test:
1) download and install ARJ32.EXE - its a DOS based archiver, much like ZIP.
2) gather a directory of random files, big and small, doesnt really matter. make sure you have alot (over 500mb). for ease i just use a few movies or mp3 files.
3) compress all the files into an archive using the command: ARJ M TEST.ARJ
4) then 'test' the archive for integrity using the command ARJ T TEST.ARJ

so what this does is do a test extraction of the big archive you have created, and checks the output with the archive for integrity. Corrupted files easily show up as CRC ERRORS.
In a normal system there should be ZERO errors.
If you are getting errors it means data is getting irredemably corrupted somewhere along the line. By replacing cables and swapping drive you should be able to work out whats going wrong.

hope this helps

<b>I am an Emotional and Intellectual Strumpet! </b>

 

[johnnyx]

thanks for the advice, but i think i got it figured out. there was nothing wrong with the cabled except for a slightly loose and figity molex cable.

Thanks again though

If an orange was driving a racecar would it peel out? www.jxfiles.com

 

[lhgpoobaa]

ahh. good to know. lucky you

<b>I'm Toms Hardware Guide Official Forum Strumpet! </b>

 

[Scotty35]

I have, I had an over length round IDE cable that performed badly. Data loss.

<A HREF="http://www.btvillarin.com/phpBB/index.php" target="_new">A better place to be</A>

 

[HammerBot]

Well, Im not really surprised by that. I have a couple of times raised concern about the quality of round cables in general. 80-wire cables were invented to reduce crosstalk by adding 40 groundwires interleaved with the existing 40. Round cables somewhat defeat that attempt by cramping all the wires close to eachother. Further, the increased length only makes it worse.

 

[lhgpoobaa]

sounds logical.
at the moment im using 2 18" silver EM shielded rounded cables... u know with the copper sheath over the wires.
as well as a plain 80wire cable.

in the past i did briefly use 24" unshielded rounded IDE cables. There were no data problems at stock speeds, however if i raised my FSB above 137 i started getting more and more data corruption.

in conclusion: longer = worse, no shielding also possibly worse.

<b>I'm Toms Hardware Guide Official Forum Strumpet! </b>