How long can an IDE cable be?

[headbrace]

As above really, how long can they be?

I would assume that power leads from the psu can be extended quite a long way, but im not sure about the signal length from the ide, as USB cna only be extended so far

[Agent]

Its about 18 inches iirc.

[Korky]

as far as u want it to be i seen an adapter to join two and continue the chain a few weeks ago, cannot remeber at all, and havent seen it on any site since.

[Agent]

Originally posted by Korky
as far as u want it to be i seen an adapter to join two and continue the chain a few weeks ago, cannot remeber at all, and havent seen it on any site since.

Not unless you want your data degrading on the channel its used on

[Dorza]

They shouldnt be longer than 18" really, after that data signals can be lost or become degraded.

So says my trusty Upgradeing n repairing PCs 15 edit book

[Saracen]

Originally posted by Korky
as far as u want it to be i seen an adapter to join two and continue the chain a few weeks ago, cannot remeber at all, and havent seen it on any site since.

There's a problem with any cable in which signals are transmitted in parallel format. Impedance varies in relation to frequency, and the longer the cable, the greater the chance of problems. When you split a byte into individual bits, and transmit the bits on separate lines, you can find that the rate of change of signal on each separate line (i.e. the frequency) varies according to the data you are sending, and therefore how often the state of each line changes. In extreme cases, this can mean that a bit from, say byte 100, might be delayed so long that it arrives at the same time as the rest of the bits from byte 101. Result - data corruption.

There is, therefore, a maximum safe length for any parallel-type cable, and the higher the frequency, the shorter that length. It applies to printer cables, too.

However, the maximum specified is usually WELL within what you'll actually get away with - most of the time. For a printer, it might cause a page corruption. For a hard drive, though, data corruption could be much less obvious (initially) and much more important.

[Korky]

very true, just saying it can be done

[5cupa]

i;pve run usb with a 10m cable no problem, but i woudlent like to compromise my ide data.

[Saracen]

Originally posted by Korky
very true, just saying it can be done

Yup. It's also possible to jump out of an aeroplane using an umbrella instead of a parachute, but that doesn't mean it's a good idea

[Number5]

Originally posted by Saracen
Yup. It's also possible to jump out of an aeroplane using an umbrella instead of a parachute, but that doesn't mean it's a good idea

lol

[TooNice]

the higher the frequency, the shorter that length

I am not so sure about that..
It may be true in cases where you exceed the recommended length, but I don't think it necessarily apply if you stay within.

Take the following example:

Ultra Wide SCSI (20Mhz, 40MB/sec) has a max cable length set to 1.5M (pretty restricted).

LVD, that is, Ultra2 SCSI (40Mhz, 80MB/sec) has the max cable length set to 12M.

I am assuming SCSI is also parallel (wasn't sure before), otherwise, there would be no talk of "Serial SCSI which I've heard a few times now.
But anyway, I guess that there is something about the IDE interface which limits the length the data can be transfered before corruption.

Still, shouldn't be be too much of a problem.. IDE devices are internal, so the chance of needing more than 18inch is slim...

[Bindibadgi]

Originally posted by Scupa
i;pve run usb with a 10m cable no problem, but i woudlent like to compromise my ide data.

That's cause USB is serial, as are SATA which also can use unlimited cable length in theory.

SCSI depends on the cabling and also the format/numbering cause you can have narrow, wide, UW and 1,2,3,4 bleh, 68, 50, 80pin! etc

[bluephi1914]

OK speaking of cables.. does it matter wich way your primary and master are positioned on the cable?? i hear some say it matters and some say it doesnt.

[Agent]

Originally posted by bluephi1914
OK speaking of cables.. does it matter wich way your primary and master are positioned on the cable?? i hear some say it matters and some say it doesnt.

Not unless your using the cable select mode on either of the drivers. Otherwise just set one to Master, now to Slave.

[Korky]

IDE devices are internal

true but u can run them outside the case if u want really.

[Saracen]

Originally posted by TooNice
I am not so sure about that..
It may be true in cases where you exceed the recommended length, but I don't think it necessarily apply if you stay within.

Take the following example:

Ultra Wide SCSI (20Mhz, 40MB/sec) has a max cable length set to 1.5M (pretty restricted).

LVD, that is, Ultra2 SCSI (40Mhz, 80MB/sec) has the max cable length set to 12M.

I am assuming SCSI is also parallel (wasn't sure before), otherwise, there would be no talk of "Serial SCSI which I've heard a few times now.
But anyway, I guess that there is something about the IDE interface which limits the length the data can be transfered before corruption.

Still, shouldn't be be too much of a problem.. IDE devices are internal, so the chance of needing more than 18inch is slim...

Yes, standard SCSI is parallel, and the same criteria about frequency and length still hold true - but they're not the ONLY criteria. Other things, like cable construction, shielding, etc also have an effect.

The original SCSI was what was referredto as single-ended (SE) and the signal for each bit was driven down the data line with ground as the reference. Thus the difference between line and ground is the signal. When voltage-differential SCSI came out (originally high, then low voltage, for HVD and LVD SCSI respectively), the game changed. Each bit is now driven down a pair of lines, andthe signal is the difference between the pair of lines. This sysyem is myuch less prone to noise interference, and is largely why the cable length went UP, despite higher signal frequencies.

LVD also uses a lower voltage, (3.3v as opposed to 5v) and the decreased voltage also has implications for length in terms of driving it down a cxable with a given impedance.

Originally posted by TooNice
I am not so sure about that..
It may be true in cases where you exceed the recommended length, but I don't think it necessarily apply if you stay within.

It holds true within the maximum length, but is one of the factors that determines the maximum length, so is not a problem provided you stay within recommenmded lengths.

Think of it this way. The higher the impedance the slower the signal is. It's a bit like you trying to swim through water or treacle. The further the distance you have to swim, the bigger the time difference will be.

So, imagine an 8-lane swimming pool with each lane separate, and alternate lanes filled with water and treacle. Every 30 seconds, 8 new swimmers dive in.

If you are swimming in treacle, there's a good chance that the guy in the next lane from the 8 swimmers BEHIND you will finish before you do as he's swimming through water, and you're in treacle.

If the pool is only 10 feet long, you'll probably make it to the far end in less than 30 seconds, before the next lot of swimers dive in, but if it's 100m long ???

Exactly the same logic applies to data down parallel cables, but how thick the treacle is depends on the rate of change of signal on the line, and THAT varies according to the exact data you send.