Slashdot Mirror
Rounding Out Your IDE Cables
BrookHarty sent us a story that proves that sometimes it's the little things that are clever. Are you as annoyed as I am that those pesky IDE cables are big flat things that are hard to move around? Well, here's a HOWTO that explains, well, how to round them out! It won't solve global warming or change the world. But dang it, that's cool.
FYI - running the bus out of spec isn't going to bug SCSI. Look at your SCSI card - it has a on board crystal to deal with the SCSI clock rate.
I wouldn't worry too much about doing the mod on 50 pin SCSI - but I wouldn't touch 68. Most SCSI problems are cable problems, and if you're just getting some retries, you wont even necessarily know it, until it screws you.
I think what he meant was that with dull knives, one tends to use more force than you would otherwise, which means that when the knife sleeps you end up with a nice, deep puncture wound, instead of a nick from a sharp knife moving at lower speed that hits you. You're right about pressure and spoons, of course, but the general rule with knives is that ones that are a little bit dull are more dangerous than sharper ones. Dull knives encourage excesive force, and foolish habits that are painfully unlearned the next time you pick up a sharp knife.
"Sweet creeping zombie Jesus!"
Then I thought about it. The smaller the cut, the smaller chance for error. Objects follow the path of least resistance. If you just make a small cut into one of those grooves, then peel the cable apart, it is almost impossible to screw it up -- you are less likely to make a bad cut, and the thinner shielding in the grooves will ensure that the ribbon will split right down the middle.
WARNING: the above comment does not link to goatse.cx
(end comment) */ }
[an error occurred while processing this directive]
> I'll bet that these kinds of cables eventually become the standard,
> especially if the cooling requirements for x86 hardware is going to
> start requiring 1lb. heatsinks like the upcoming P4...
You bet they will - but perhaps not why you think. SerialATA, the Intel-created (or just backed?) upcoming standard for connecting low-cost harddrives (i.e. non-SCSI drives) have cables about the size of a CD-Audio cable. Very cool (but not as cool as IEEE1394b). Very nice for cooling.
Weren't we all supposed to be using FireWire harddrives by now? *sigh*
FireWire may be faster than SCSI but I don't believe that FireWire is faster than FibreChannel. A FibreChannel controller that I worked on did around 190 megabytes per second. Can any FireWire controllers push that kind of data? (not trying to be cocky, just asking)
-- soldack
The trick is to use your fingers (if you can) instead of a knife to avoid severing a cable.
--
Wooden armaments to battle your imaginary foes!
Places like 2CookTek, teamawe, and Case ETC have the same products with better explanations of what you are buying at more reasonable prices. I guess the disadvantage is that you don't get to have a site with fancy animated GIF's.
www.everything2.com/index.pl?node=karma%20whore
www.cybernothing.org/~holychao/karmaho.html
www.wirednews.com/wired/archive/8.07/mustread.html ?pg=9
http://slashdot.org/users.pl?op=userinfo&nick=sign al%2011
This place should give a lot of resources too:
http://goatse.cx
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
I had less luck with this manufactured round cable, and wound up removing it because it caused the SCSI bus to constantly reset.
WARNING: the above comment does not link to goatse.cx
(end comment) */ }
[an error occurred while processing this directive]
Just in case you didn't know, a 66MHz square wave is made up of several (ok, really unlimited) different sine waves. You need to look at the rising and falling times of the signal if you want to calculate the RF produced.
Also, look at an FM transmitter. One inch of wire is enough to transmit a signal!!! Even AM doesn't require that big an antenna while it operates at a blazing 1MHz. The 66MHz signal will most definitely be producing a ton of RF.
The ground lines are nothing more than shielding in this case.
Are you serious? You really think putting a wire next to another wire will provide shielding? No, sorry bud, that's just not how it works. Shielding can be provided by completely surrounding the transmission line with a ground wire - like with your TV cable or RCA wires. An extra wire does nothing to stop RF.
So why do people put ground wires by transmission wires? It's actually not usually like that. The two wires aren't positive + ground, they're a differential pair, like your ethernet cable. The idea being that one wire is more positive then the other. So then when a pile of RF comes along and jolts the voltage up in those lines by 20%, the data isn't lost because the relative voltage of the two lines isn't effected. The second line protects against data corruption, not RF production.
This reminds me of a cool lab I did back when I was in school. Our class wired several RS232-RS422 converters to connect to the serial ports of different PCs. We then connected them all to a single differential pair (ie, two wires). After writing some software we had our entire lab networked via the serial port! Ok, it sounds lame but was great fun to implement.
Willy
Hey, in the same store I found the FireWire drives, an 80GB 100/ATA is $320.
Yeah, I needed to get another controller... but let me see you chain a scanner, camcorder, PS2(whatever good that is), an 6 hard disks to your built in controller? That's the limitation of your puny built in controller ^^
The nick is a joke! Really!
GPL Deconstructed
63 devices, I think... maybe a repeater could allow for more devices?
The nick is a joke! Really!
GPL Deconstructed
IEEE1394a (current standard) tops out at 400Mbps (megaBITS). The upcoming IEEE1394b standard spans 800-3200Mbps.
SCSI drives are currently made with interfaces for 160MBps (megaBYTES), and the newest standard - don't know if it's finalized yet - is 320MBps.
Obviously, hard drives cannot transfer at anywhere NEAR those rates, BUT, there's still a reason for them: multiple drives! With the advent of RAID (ATA RAID controllers built-in to motherboards are all the rage now), then these speeds become much more justifiable.
The problem with FireWire hard drives is, as far as I know, there aren't any that have NATIVE FireWire interfaces on them - they're still ATA native devices with FireWire converters built-on to them (usually made by TI). From what I've read online, the current crop of TI chips aren't too efficient, but the next generation (due now or soon) are much better. Nevertheless, I'd feel better about harddrives (I prefer IBM these days) with native FireWire interfaces. Let's hope IEEE1394b gets here right quick.
The advantages of FireWire are numerous:
That's all I can think of right now - I'm sure there are more.
There are already 2 ways, but cost shaving does in both of them. One way is using shrouded headers (the part with the pins). The shroud has a notch in the middle on the odd number pin side (1,3,...39) and the Insulation Displacement Connector crimped onto the cable is supposed to have a corresponding "bump". But headers without shrouds are cheaper. The other way involves the absence of pin 20. It isn't there on the header and the corresponding hole on the plug body is filled up, but that adds at least one more step to the manufacturing process, which adds cost.
At least it's not as bad as those 10 pin headers for serial ports. Some of them have pins 1, 2, 3, 4, and 5 in one of the rows and pins 6, 7, 8, 9, and the never used 10 in the other row, whereas others have the odd number pins on one row and the even number pins on the other row. This means that when the 9 or 10 conductor ribbon cable gets to the DB-9 or DB-25 connector at the other end there are two choices of how to wire it and the one that was right for your old motherboard or controller card may not be right for the new one. I've even got an old Zeos 486 board where the Com 1 header is set up one way and the Com 2 header is set up the other, so that it isn't even consistent right there on the board.
I see even classic Slashdot is now pretty much unusable on dial up anymore.
Eventually you won't have to worry about rounding out your IDE cables. Low-end hard drives will go from parallel ATA to serial ATA. Serial ATA cables will be nice and thin with a lot few fewer conductors. You can read more about this at serialata.org, but the basic idea is that hard drives will transition to a software compatible, 1.5 Gb/s serial connection. Among the benefits promised are "easier routing of cables." A serial ATA drive has already demoed but they're not promising systems until 2002.
A 80gb FireWire drive is $369... A 45gb drive is $250
A FireWire controller is only $100...
Why aren't you using FireWire drives yet?
The nick is a joke! Really!
GPL Deconstructed
You don't have to be careful with ATA/33 cables. I cut them freehand with a razor blade, not being particularly careful. I mostly got them right, but sliced into two or three wires by mistake. I didn't sever them, of course, but you don't have to worry about stripping the insulation off a wee bit (as long as two such wires don't connect, of course.
High-speed Road Trip (18.000KPH)
To clarify, people seemed to miss it. I did only cut it a very small amount, then pull the rest apart. But you still have to be VERY careful when you cut into the cable.
-Julius X
-Julius X
remove "-whatkindofspamdoyoutakemefor-" from email to send
There is an electrical effect called the "transmission line effect".
A cable can be modelled (for high-speed signals) as a series of inductors, resistors and capacitors.
If a certain relation between the values of the inductors, and capacitors exists, the signal will surprisingly come out the other side, very similar to how it went in. Take a garden hose and shake it up and down. You can make the waves "walk" through the hose. That's the effect.
This effect is most clearly observed with coax cables. There are several "waves" of TV signal on the wire, even if there is only a few yards of cable between your wall-socket and the TV.
Now, in a ribbon cable, a signal that has a "ground" conductor on each side of a signal, this effect is noticable enought to be useful.
For UDMA33, and especially higher, the transmission line effect becomes noticable, and required for correct operation.
If you just pack a bunch of lose cables in a bundle with tiewraps or stuff like that, this will not guarantee that there are ground conductors next to the signal wires.
As an experiment: Try making two round cables. One with the odd and one with the even numbered conductors. You certainly shouldn't be able to run UDMA66 anymore.
I really don't know who you talked to at WD. They DO have people who know this, but apparently the guy you talked to didn't.
Oh, you will probably be able to get away with splitting the cable in pairs. But for the tranmsission line effect to really work, the fact that there is a ground line on the other side too really helps.
Roger.
2) I intentionally threw in a goatsecx link.
3) Proper moderation would be: (-1, Flamebait) or (-1, Offtopic)
Thank you.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
There are round cables going to the DSSI bus disks in the disk bay. If you got SCSI disks, you have round cables as well.
The driving force for PCs has been cost, and putting connectors on a flat cable takes seconds. With round cable you have to manually dress out the wires, put them in the right order by color, etc. before you can terminate on the connector.
Round cables for external disks (SCSI) are very common even for PCs.
use your fingernail to separate the cable is better than using an exacto knife. that will avoid cutting one of the wires in the cable. if you do, then you start all over. if you have to use a knife, just make a small inscision and tear down the length of the cable.
l es_howto/round_cable_howto _pg1.html
also, when you are separating, forget separating each individual strand. I'm not sure why this article said to do this. from my experience. just make an incision every 4-5 wires, and the cable will "round" (well, rectangle, really) just as well. less cuts means less chances to make a mistake!
lastly, i wouldn't even touch those ATA-66 cables. there are way more worries about possibly cutting one of the much thinner wires, and there is also mention of potential crosstalk interference with those high density cables. i have found that merely twisting the cable together so it "rounds" and then using zip straps works wonders. These ATA-66 cables are not as plentiful as the piles of 33 cables you have sitting around. don't split those!!!!
because i'm an [H]|OCP'er, check out their article on rounding too. at least it's only TWO freaking pages... fewer banner ads to distract you, i say. the [H]ard|OCP article also tells you about cable placements.
http://www.hardocp.com/articles/cooling/round_cab
Both 3M (Scotchflex) & Thomas&Betts sell round IDC cable by the roll, all the way from about 10 core to about 80 core.. However its not as simple as that, with round IDC cable, the cores have to be split from each other otherwise the round cable would buckle & also spring back too much, of course the more expensive twisted pair variety has to be split for the cores to twist too. However one also has to have a unsplit patch every 3 inches or so, of at least one inch in length, along the length of the core, underneath the outer insulation - otherwise it'd be a bugger of a job attaching IDC end clamp connectors to them. Consequently if you purchase a roll of the stuff, & you plan to make IDE or SCSI cables you have to cut them an extra 3 inches or so, because you might have to backtrack a bit to unraval a flat unsplit length to attach the IDC connector to each end. Twisted pair round IDC cable rolls are expensive. So if you see SCSI IDC round cables going cheap, they are either handmade jobs made from ordinary ribbon cable or its made from a roll of ordinary round IDC cable. The simple fact is, if you were to open up many brand name computers (such as the latest Compaqs), they actually have round IDE cable in them & many profesional workstations (such as some IBMs) now have round SCSI cable in them to. The reason is it aids ventilation in those cramped Compaq PCs & in those hot workstations. I agree with you on the proper art of folding & routing IDC cables, though.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
>Would'nt this have some interference?
For modern ATA-66 cables and SCSI cables, I would think so. These cables alternate DATA/Ground pins for sheilding/impedance purposes. When you separate these cables, you change the shielding and impedance properties of the cable.
To create a properly rounded ATA-66 or SCSI cable, you really need to separate it into pairs of Data / Ground, or else you change the properties of the cable. You'll likely get crosstalk and interference from the computer otherwise.
Oh, BTW: The 4 pairs (8 wires) in CAT 3/5 ethernet cable are there because in most standards you can use the other 4 wires for voice or another Ethernet connection. Also, the added redundancy means that if a wire turns out to be cut somewhere along the run, you are still OK (use a different pair).
Some ethernet standards actually use all 8 wires (like 100VG... that stuff's neat, doesn't use CDMA in the normal sense).
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
Rounding cables like this is a time-wasting method of destroying all that paid-for engineering. Try one of these instead:
1. Look in the Newark (or similar) catalog and buy one, for pity sake.
2. Take the long section (what you are most concerned with), carefully fold it up lengthwise, and slide it into a length of half-inch split loom tube (any good parts store for about fifteen cents).
Either of these will get you where you wanted to be, which was moving that cable out of the way. Also note that IDE and especially ATA cables must be cut to a certain length to avoid possible SWR problems. You might also look up the proper folding and routing of ribbon cables, which seems to be a lost art among PC manufacturers these days.
*whup* "Get along, little electrons. Heeyah!"
I just tried this, and it works way _way_ better if you use a Pizza cutter. (they were intended for cutting in straight lines) If yours it truely dull, consider getting a new one anyway -- crust is hard to cut with a dull knife. --Advice from the Pizza Hut professional I am--
External SCSI cables are carefully arranged to minimize the interference. The cable is arranged in layers. The ground lines are all placed on the outside layer, with the data lines on the next layer, so the data lines are all close to ground lines to minimize crosstalk, and so the ground lines protect the data lines from external interference. It would take a lot of hand cramping work to rearrange a ribbon cable to mimic this arrangement.
Meldroc, Waster of Electrons
Wow, thats, um... neat, yeah, neat....
Buut seriously, why not just roll the cable , or bend it in half a safe distance from the end?
Ands how often do you actually get inside your computer? Bout' once a month for me.
This would help some people though, who work on computers alot, because I have severed wires on computer shrapnel (sharp little corner and crap)
Here's place that sells them pre-rounded. You can get rounded cables for ATA33, ATA66/100, floppy, and SCSI 50-pin.
At 66MHz IDE cables are going to look like a transmission line and I assume that is why there are so many earth connections in parallel with the signal wires. If you separate them then you could end up with a mismatch, more RF interference and possibly data corruption.
If you are going to try it out then backup your data first and make sure that you have a spare cable in case you break something
You're right about that (just like SCSI is a luxury... or even a PC, to some extent.)
But within the next year, a PS2 will not be so far fetched (at least in the average US household), and a scanner at $180 is not too bad either. As a FireWire HD is only about 10% or so more expensive, their value add is pretty decent, for the flexibility. A camcorder, at about $1k, is still a little rich for most of America, but if Sony and Apple had their way, I'm pretty sure a FireWire'd household isn't too far off in the future.
The nick is a joke! Really!
GPL Deconstructed
Cross-talk interference in the cable will increase in both cases, but with IDE this may not be a limiting factor due to error correction.
*whup* "Get along, little electrons. Heeyah!"
Just push the tip of the Exacto knife in between the conductors until it penetrates through the ribbon, then separate the conductors by pushing the DULL side of the knife tip along. This will guarantee that you don't cut into the conductor shielding. You can then either separate the entire length with the dull side of the blade, or just a few inches and do the rest by hand.
well, ummm, not really. in the world of digital audio, signal degradation is of utmost importance. wound grouped cables are wound in such a way that they augment signal transfer. Say for instance a 300 foot long snake that runs from the front of house mixing console to the power back stage. millivolt signal distortion happens almost instantaneously when cables are improperly grouped. more relative to this subject would be cat5 for instance. 4 twisted pairs. all isolated, in four pairs which will group the signal over distance. there is an eqaution but don't ask me what the hell it is, because I don't know. In IDE drive cables, the short distance might be a saving grace, but low level signal distortion might me some thing to consider. A thought. :)
a/s/l here. Sorry, adding domain tags to your s
There's one more reason: it's cheaper. Ribbon cables use insulation-displacement connectors which are gang-crimped onto the cable in one operation. Easy, fast, and reliable. And since very few of us actually care, a manufacturer is going to go with whatever saves money.
From what I would assume, flat-ribbon cables are flat for a reason... it helps keep the signals isolated from one another. If you start cramming all those wires together, you wind up changing the impedance and capacitive characteristics of the cable.
However, I've noticed that the wires on my 86-pin UltraWide cable (came with my Tekram SCSI card) are... well.. interlaced. I originally assumed that it was merely to make the cable more flexible (which it does, nicely), but now that I think about it, might there be another more scientific reason for doing this?
Would'nt this have some interference? I mean it might be flat for a reason. Maybe like the way ethernet cable has 8 wires but only uses 4? Theres a reason for that.
I saw a similar article to this a few months ago, and did it. It's not easy work. An hour or so sitting at my desk carefully splicing these cables with an exact-o-blade isn't exactly fun. I did end up cutting into a couple of the cables, and needed to spend about $15 in new cable, but its not that bad. You just have to be really careful that when you cut into the cable that you hit directly into the groove between the wires.
But in the end, it all turned out well, and my system looks a lot less cluttered on the inside as a result, and my graphics card temperature dropped by a few degrees (my IDE cables were sitting over my GeForce fan).
So if you have the patience, go for it. Its definitely worth it I think.
-Julius X
-Julius X
remove "-whatkindofspamdoyoutakemefor-" from email to send
Floppy cables can be rounded (you can even remove the middle connector), the older IDE cables can be rounded, and 50 pin SCSI wires can also be rounded. Bind everything together at the end with zip ties and then wrap it up with spiral wrap from Radio Shack and similar (e.g. Ax Man in the Twin Cities).
ATA/66 and 68 pin SCSI cables are a different story. The wires are stiff, and if you don't do the cable exactly right, you have either trashed an expensive cable, corrupted the data on your hard drive. People seem to have mixed success with those.
If you want to mess with your cables, try some old junky ones, first. If you must use a knife or razor, make the incision as small as possible, then peel the cable apart in opposite directions -- those grooves are the path of least resistance, so it should be more reliable than making a long, perfectly straight cut with a small sharp object.
If you want rounded cables and don't want to take any risks, I know that Plycon has all sorts of high quality machine made cables of all types, albeit for a very steep price (just like everything else they sell).
I'm still not sure why this isn't the standard. Some PC manufacturers have been doing this in their servers and micro towers to improve airflow through the chassis. I'll bet that these kinds of cables eventually become the standard, especially if the cooling requirements for x86 hardware is going to start requiring 1lb. heatsinks like the upcoming P4...
WARNING: the above comment does not link to goatse.cx
(end comment) */ }
[an error occurred while processing this directive]