Is there a minimum length for IDE cables?

[grrl]

Has anyone shortened their ribbon cables to minimize the clutter in their case? I've read that between 18 and 24" is the maximum length for most IDE cables, but is there a minimum length too?

 

[bacillus]

no there isn't a minimum length.

 

[grrl]

So there would be no problem trimming them to get rid of the extra cable between drives?

 

[Rand]

Originally posted by: bacillus
no there isn't a minimum length.

If you stick directly to the IDE specification there most definitely is, specifically 10" in length.
Not a big deal to go slightly below that if you stick to PIO mode 1-4... but not many of us would be willing to do that.
One can generally get away with IDE cables as small as 8", but it's not recommended and doesnt comply with specifications.

If one were to stick directly to specifications then 18" is the longest and 10" is the shortest allowable.
You can generally go as long as 24", or as low as 8" however if your willing to accept marginally higher data error rates which the ATA interface's built in CRC error checking should usually be able to hande adequately.

You might be able to pull of as small as 6" if you really feel like pushing it assuming your using a modern ATA60/100/133 IDE HDD, your bound to have awfully high data error rates which CRC should find and retransmit but you'll take a performance hit. I wouldnt try that small with any sort of Optical Drive however.

 

[Auric]

If there is a specification for minimum length it is not necessarily related to signal integrity, unless the specification specifically specifies that.

It could just be for standardization in regards to device mounting locations and such so in lieu of complete and accurate information I think it is safe to fall back on ye olde logic and critical thinking abilities and be suspicious that shorter lengths could possibly cause a problem.

 

[KF]

It is my understanding that shorter lengths will not cause more errors, but fewer. Zero length would be optimum for minimizing errors.

Shorter lengths pick up less noise. Shorter lengths introduce less disparity in the delays of the various frequency components of a square wave, giving a cleaner signal. Ideally you'd want the signal to look identical at the end of the cable as the beginning, and that is what you would have for zero length.

 

[Auric]

After re-readin' what I wrote I can see how it could be misinterpreted. I should have wrote "...and be suspicious *of the notion* that shorter lengths could possibly cause a problem". To wit, go for it.

 

[grrl]

Rand, can you provide any links on this?

 

[Eug]

Originally posted by: grrl
So there would be no problem trimming them to get rid of the extra cable between drives?

How would you trim it? Why don't you just buy a shorter cable? Anyways, I use a 2" cable in my Firewire enclosure.

 

[corkyg]

I use a Trios II drive selector by Romtec. It comes with 12-inch IDE cables for three drives and the IDE controller. There are also 10-inch rounded cables available. That's the shortest I have seen on the shelf.

cables

 

[Auric]

Originally posted by: Eug
How would you trim it?

With sharp scissors or a box knife. To move the connector, unsnap the tabs with a small screw driver and when repositioned where you want it, squeeze together with pliers.

 

[Paperlantern]

I dont see a minimum for any IDE interface. Laptops use IDE interface drives, jsut like pcs, and they are either mounted directly on the board, or have an extremely short cable to hook them to the interface, so your looking at 0" to 1", cant get much shorter than that. *shrugs*

 

[Baileybbk]

I've seen 5 inch flat cables......never did find a short round cable.

Although the spec might state a minimum, I'd like to see a RWT with a modern controller that shows error rates.
I'd be surprised if the error rate increased.