30" IDE Cables, Too Long ??

[Richard_Cranium72]

Discussion per cables longer than 24" has been that data loss or corruption may occur.

Is this still the case ?

http://www.cablesonline.net/3040pin80wir.html

[Nighthawk]

I would assume so, since the ATA specs haven't been rewritten AFAIK.

While I would have liked an IDE cable with maybe 6.5" between pri and sec connectors instead of the standard 6", I really don't think a 10" difference is going to work... Remember that it's the Primary's drive controller that's controlling both drives and a longer distance means more latency on the secondary, plus the higher chance of corruption or interference on the cable.

[Brangwen]

Very interesting. I sure would like a slightly longer IDE ATA100 cable. The ones at A2Z Cables (http://www.a2zcables.com/) appear to be up to 24". The Website does, however, have some interesting info and other niffty products.

Brangwen http://www.sysopt.com/forum/wink.gif

[surrealchereal]

Why does the hard drive manufacturer include cables with the new hard drive that is longer than 18"? To get more calls from angry customers? Actually this question is 2nd on my list of mystical computer questions. The 1st one is, how can there be software to copy your old hard drive loaded on your hard drive that enables you to install a new hard drive and have your old stuff? I have never heard backup connected with this idea,, hmmmm must mean network, another word I don't hear spoken with this concept.
sigh.. uh oh, I see I better touch up my blonde roots again..... http://www.sysopt.com/forum/smile.gif

[Goldwingnut]

This is an educated guess, not a definitive reply. I can tell you why there is a limit, but not what that limit is.

The wires run parrallel to each other, with insulation in between. So they act as inductors and capacitors simultaneously. This means that the longer they are, the better job they do of filtering high frequencies.

The high frequency component of square waves, which is what digital signals are, provides the fast risetime of the leading edge of the waveform. If the cable is too long, the leading edge gradually ramps up to its High state voltage. The digital circuitry is waiting for the rising voltage to exceed a certain level to read the voltage on the line as High or Low.

If the signal snaps decisively from High to Low or Low to High, all the lines in a parallel signal cable transition at the same time. 00101101 leaving one component reads as 00101101 on the far end.

If the high frequencies get rolled off, minor variations in the device that is waiting for the incoming signal at the other end may cause some of the bits to get read as High or Low a little later than others. So 00101101 going out might get read as 00111101 at the far end, because that bit got there a little TOO late.

Cable length is limited to a length that keeps the high frequency rolloff in a tolerable range.

These newer cables - ATA100, possibly ATA66, use coax instead of parallel lines. Less mutual induction, better shielding.

[This message has been edited by Goldwingnut (edited 06-07-2001).]

[Richard_Cranium72]

Coax is fine for shielding against spurious radiation within the case.

My question is about loss of signal due to cables too long.

The thread on "rounded" cables centered on inducted noise/corrupted data from one conductor to another from the IDE cable itself.

Brangwen
Please note , the link I provided shows 30" cables !




[This message has been edited by Richard_Cranium72 (edited 06-08-2001).]

[surrealchereal]

doc, what's the diff btw the 2 types they have there, what is "enhanced"?

[surrealchereal]

2. Not too scared to visit http://www.sysopt.com/forum/smile.gif
hehee! Now you're talkin' http://www.sysopt.com/forum/wink.gif

[surrealchereal]

Discussion per cables longer than 24" has been that data loss or corruption may occur.
Hmmm, I thought I remembered 18' from the book How to Repair and Upgrade PC's.
eh, what do I know eh?

[Ed_S]

For what it's worth, I've got a 36" cable that's been running error free for a couple years now. Plain 'ole 40 wire ribbon IDE, has 4 connectors spaced out. Of course you can use only two, but I've used tham indiscriminately.

Guess if I'd have known better, it wouldn't have worked, huh. http://www.sysopt.com/forum/frown.gif
Don't think I'll take it out just yet, though. Works too good. http://www.sysopt.com/forum/wink.gif

Have also made my own using connectors from a show supplier. No probs. (yet)

[Brangwen]

Richard_Cranium72: I noted the cable length difference before I posted and provided a link to A2Z Cables (http://www.A2ZCables.com) . I felt some of the cable info available on that link would provide some explanatory material for this discussion.

By the way, I ordered a 30" rounded enhanced ATA100 cable from CablesonLine which will be here next week. I'll give it a go and post results.

Brangwen http://www.sysopt.com/forum/wink.gif

[Richard_Cranium72]

Don't know, however I mailed them an invitation to visit here and explain for themselves.

Now we'll wait and see if they're
1. Not too busy to come by
2. Not too scared to visit http://www.sysopt.com/forum/smile.gif)

DrVette