How do you all feel about rounded calbes?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
How do you all feel about rounded calbes?
What I have read is rounded cable can contribute to data corruption, but has anyone else felt this way or had this type of bad luck. I know this is a paltry modding maybe not worth the money, but maybe everylittle bit helps eh?
-
- Posts: 210
- Joined: Sat Jul 19, 2003 1:40 pm
- Location: Kortrijk, Belgium
-
- Posts: 226
- Joined: Sat Sep 06, 2003 5:59 am
- Location: Finland
Any hard drive that is running on a UDMA-66 or faster channel must have an 80-conductor FLAT IDE ribbon cable to adhere to the ATA standard.
To exceed the 33 MB/s possible with UDMA-33, some way of controlling electromagnetic noise had to be introduced, as EM interference was the primary limiting factor in terms of speed. Of the 80 conductors in UDMA-66 cables, every other wire is a ground wire, to reduce crosstalk between conductors. When cables are rounded, this benefit is removed. Thus, you should be no more comfortable using a rounded cable as using a 40-conductor IDE cable for your hard drives. It would be like using Cat3 telephone cable instead of Cat5e Ethernet cable to connect your 100Mb/s network. Sure, it would sort of work... but it wouldn't be up to snuff, and speed/reliability could suffer.
Rounded cables are fine for CD-ROM drives, since most of these drives are ATAPI and limit the IDE channel to PIO Mode 4 (~16.6 MB/s I think) anyway. And of course, floppy drives certainly won't notice a difference with rounded cables! But for hard drives, if your flat cables are really bothering you, it's time to upgrade to SATA, or put your hard drive's data at risk of corruption.
I would also like to agree with those who feel that with a properly-organized case, flat cables can still be installed effectively and attractively without compromising airflow significantly.
To exceed the 33 MB/s possible with UDMA-33, some way of controlling electromagnetic noise had to be introduced, as EM interference was the primary limiting factor in terms of speed. Of the 80 conductors in UDMA-66 cables, every other wire is a ground wire, to reduce crosstalk between conductors. When cables are rounded, this benefit is removed. Thus, you should be no more comfortable using a rounded cable as using a 40-conductor IDE cable for your hard drives. It would be like using Cat3 telephone cable instead of Cat5e Ethernet cable to connect your 100Mb/s network. Sure, it would sort of work... but it wouldn't be up to snuff, and speed/reliability could suffer.
Rounded cables are fine for CD-ROM drives, since most of these drives are ATAPI and limit the IDE channel to PIO Mode 4 (~16.6 MB/s I think) anyway. And of course, floppy drives certainly won't notice a difference with rounded cables! But for hard drives, if your flat cables are really bothering you, it's time to upgrade to SATA, or put your hard drive's data at risk of corruption.
I would also like to agree with those who feel that with a properly-organized case, flat cables can still be installed effectively and attractively without compromising airflow significantly.
Bahh, that problem with rounded cables was real when they were first released. Though I am sure that some of the rounded cables available now still have these problems, many do not.
With my own tests I saw about a 5% increase in performance. Not to mention the RD3XP's look sexy and really helped air-flow. Flat cables were a no go since folding them out of the way would have shortened their length so much that they would have come up sort. Rounded cables really do have
Flat cables are nice and all but they are dead. Look at all the new data cables out now. SATA, USE, FireWire, Serial SCSI, most of which look just like the RD3XP's do on the inside.
Grounding wire, mesh cover, booted ends and each data wire surrounded by 4 ground wires. What more could you ask for?
http://www.pcabusers.com/reviews/gladiator/rd3.gif
http://www.ioss.com.tw/web/English/RD3XPGladiator.html
With my own tests I saw about a 5% increase in performance. Not to mention the RD3XP's look sexy and really helped air-flow. Flat cables were a no go since folding them out of the way would have shortened their length so much that they would have come up sort. Rounded cables really do have
Flat cables are nice and all but they are dead. Look at all the new data cables out now. SATA, USE, FireWire, Serial SCSI, most of which look just like the RD3XP's do on the inside.
Grounding wire, mesh cover, booted ends and each data wire surrounded by 4 ground wires. What more could you ask for?
http://www.pcabusers.com/reviews/gladiator/rd3.gif
http://www.ioss.com.tw/web/English/RD3XPGladiator.html
I'm sure a qualified electronics engineer or a particle physicist could answer better, but I'll try my layman answer.al bundy wrote:What about all the intricate folding of flat 80-wire IDE cables that some members around here do, in order to clean up the interior of their cases?
Does folding 80-wire IDE cables also perhaps allow data corruption?
8)
With high enough signal frequencies in ATA cables one must use a 'ground' or 'null' cable right next to each signal cable in order to cut down interference. Interference can cause all sorts of signaling problems eventually leading to data corruption.
Rounded cables can be made in various ways.
The homebrew method of striping a flat ribbon cable and rounding it is perhaps the worst.
It provides no systematic interference shield for each signal line.
Twisted pair type cables where each signal line is paired and twisted around a null/ground cable is probably amongst the best options for a rounded cable.
Various types of shieldings / insulators can also be used (aluminium foil, teflon, etc) to cut down on interference.
The trouble is that all these additions (twisting, shielding, more expensive materials) cost.
When one consideres that the average consumer has grown accustomed to paying 2-5 bucks for an IDE cable that 'just works', it's hard to justify a price of 20-50 USD for a cable that has superior signal integrity (even up to 90 cm lengths).
Unless of course, the cable is thin as hell (no shielding, minimal insulator, no twisting = crap in terms of signal integrity) and comes with UV-reactive paint and a sticker that yells 'for l33t modders only' :)
If you want good ATA cables (parallel) in terms of signal integrity, shop at Granite Digital:
http://www.granitedigital.com/catalog/p ... able&a.htm
They are not rounded though and they cost.
cheers,
Halcyon
-
- SPCR Reviewer
- Posts: 8636
- Joined: Sat Nov 23, 2002 6:33 am
- Location: Sunny SoCal
Yes, that's true but there's a nifty twist to that with Win2000 and WinXP, and I've been bit by it. That's why I haven't used rounded cables since then.Zhentar wrote:You should be pretty well protected from data corruption, it willl hurt performance though. Every time an error is detected, it has to resend that data a second time. I still use rounded cables though.
What happens is that 2000/XP allows for a certain amount of CRC errors and once they go over whatever the threshold is, the operating system drops the speed of that IDE channel down a notch (i.e. from DMA-5 to DMA-4) to "protect" the data. If the errors conitnue, the OS drops the speed another notch and this continues until the IDE channel is running in PIO mode (PIO-4, IIRC. That's 16MB/sec, yecch!).
I discovered this when I was investigating the cause of a slowdown on my system. I noticed an error in Event Viewer that said something about errors on IDE(0) so I did some MSKB research and discovered this nifty "feature".
I deleted the problematic IDE channel in Device Manager and rebooted and it found it and reinstalled it on the next boot. It was running at full (DMA-5) speed so I kept my eye on it to see if it started slowing down again and sure enough, it did. I started troubleshooting at that point and found that I had a bad rounded (Coolermaster) cable. I RMA'd it and the new one did the same thing.
I decided to stick with the stock flat ribbon cable instead and never had any more problems with that IDE channel. That's when I had to teach myself how to do cablegami, before that I used rounded cables to obtain better airflow.
-
- Posts: 6
- Joined: Sun Aug 24, 2003 6:55 am
Zhentar got it right.
If you run linux you will get pages of errors that look like this:
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
Newer kernels behave like winxp and auto-drop the UDMA mode after a few errors.
I've seen it happen even with flat cables. Sometimes you just get a bad one. Round, nonstandard cables are just asking for trouble.
If you run linux you will get pages of errors that look like this:
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x84 { DriveStatusError BadCRC }
Newer kernels behave like winxp and auto-drop the UDMA mode after a few errors.
I've seen it happen even with flat cables. Sometimes you just get a bad one. Round, nonstandard cables are just asking for trouble.
I quite like rounded cables. I don't use them for HDD's anymore though since I got my SATA drive. Still, rounded floppy and CD-ROM cables are nice for neatness purposes. That said, a well done flat cable fold looks really good.
I just wish there was an optical replacement for SATA - that's would rule all solutions.
I just wish there was an optical replacement for SATA - that's would rule all solutions.
-
- Posts: 153
- Joined: Mon Oct 27, 2003 5:35 pm
- Location: Arlington, Virginia
- Contact:
I've just been thinking about that. Seeing the popularity of firewire hard disks and external enclosures. I nearly bought a shell for my new big HD but decided that it was an expense I didn't need and I would have had to run it USB2.0 with my current MB.
However, I was thinking how that might evolve someday to storage devices connected by wifi. My little USB wifi adaptor is only 2 x 3" and you could use bluetooth or wifi type links to connect things now connected by USB. Wifi speakers, wifi LCD, wifi HD's, wifi multi cpu, who knows how far it could go with the right kind of wirless network. You could have parts of your computer scattered all over the house!
About the round cables.
I am very concerned about data loss from my own experience. I know what it is to have the registry corrupted on a weekly basis. To replace three drives in a row, installing windows over and over on a weekly basis.
On the other hand, I decided to replace my ribbon cables with round ones because of the trouble I've had with them. I use 18" or 10" round store bought (COMPUSA) IDE cables in my system for HD and CDROM. They have clear insulation (my preference) and are shielded copper braid. I have not found any IDE errors reported by Windows XP since using them. My A7V133-A only supports ATA/66/100 UDMA-5. I have seen no reduction in data rate or falling back to PIO4 mode.
I can imagine that the early attempts at rounded cables would result in data errors when applied to drives above ATA33 given they were made out of the ATA33 cables.
My InWin A500 case, which is a small mid-tower with the PSU hanging down in front of the CPU has very close quarters. I would say the flat ribbon IDE and floppy cables block about 90% of the area beneath the CDROM bays in front of the front case fan. The IDE flat cable hangs directly in front of the front case fan (which is trapped in a plastic cage) like a door. The round cable does not block the fan. How much this improves airflow I cannot say, but this case needs every bit of help it can get.
I have had a ribbon cable go bad from the very tight twists required to get from the drive to the mainboard connector. It is not always possible to elegantly route or fold IDE cables in all cases. My case and mainboard combination is a good example, where the mainboard IDE connector is oriented 90* to the drive connectors. I never found a way to avoid a kink nor a way to avoid blocking the front intake fan. There are just too many cables ribbon and power in too small a space with opposite orientations and other things, bays, fan covers in the way. I think these kinks must have some electrical effect too.
I think round cables can be very helpful in a small ATX case with a limited airflow. I realize the limits of my present case and search for a new one that better fits many criteria important to me. I don't care how much the cables cost. They're not that much more than an extra ribbon cable cost a few years ago at a office superstore.
Here are some references I found.
http://www.sysopt.com/articles/cabround/
http://www.casejunkiez.com/reviews/round_ide.shtml
http://www.ramelectronics.net/html/ata-66_cables.htm
http://www.techtv.com/screensavers/show ... 23,00.html
http://dansdata.com/rcables.htm
http://www.3dvelocity.com/reviews/akasa ... ata133.htm
http://www.a1-electronics.net/PcHardwar ... bles.shtml
Many of the sites talking trash about round cables sport google ads selling them.
It seems to me that with the litigious nature of US society that if there were significant problems with round cables that major retailers would avoid selling them lest the end up in a class action lawsuit.
I could find no warnings against using round cables on the major HD manufacturer sites, maxtor, seagate and wd. If the were a significant problem, their call centers would hear about and they would place an insert (as they have for other problems) in the box to not use them.
I could not find any reports of verified data error from use of modern store bought shielded cables 18" in length.
I found a lot of speculation but not many facts on the web. I found a dearth of any scientific evidence that modern store bought round cables are causing data errors. It seems in the years they have been available it would have been easy for a review site to observe the signal on an oscilloscope for flat and rounded cables and post the results.
Steve
However, I was thinking how that might evolve someday to storage devices connected by wifi. My little USB wifi adaptor is only 2 x 3" and you could use bluetooth or wifi type links to connect things now connected by USB. Wifi speakers, wifi LCD, wifi HD's, wifi multi cpu, who knows how far it could go with the right kind of wirless network. You could have parts of your computer scattered all over the house!
About the round cables.
I am very concerned about data loss from my own experience. I know what it is to have the registry corrupted on a weekly basis. To replace three drives in a row, installing windows over and over on a weekly basis.
On the other hand, I decided to replace my ribbon cables with round ones because of the trouble I've had with them. I use 18" or 10" round store bought (COMPUSA) IDE cables in my system for HD and CDROM. They have clear insulation (my preference) and are shielded copper braid. I have not found any IDE errors reported by Windows XP since using them. My A7V133-A only supports ATA/66/100 UDMA-5. I have seen no reduction in data rate or falling back to PIO4 mode.
I can imagine that the early attempts at rounded cables would result in data errors when applied to drives above ATA33 given they were made out of the ATA33 cables.
My InWin A500 case, which is a small mid-tower with the PSU hanging down in front of the CPU has very close quarters. I would say the flat ribbon IDE and floppy cables block about 90% of the area beneath the CDROM bays in front of the front case fan. The IDE flat cable hangs directly in front of the front case fan (which is trapped in a plastic cage) like a door. The round cable does not block the fan. How much this improves airflow I cannot say, but this case needs every bit of help it can get.
I have had a ribbon cable go bad from the very tight twists required to get from the drive to the mainboard connector. It is not always possible to elegantly route or fold IDE cables in all cases. My case and mainboard combination is a good example, where the mainboard IDE connector is oriented 90* to the drive connectors. I never found a way to avoid a kink nor a way to avoid blocking the front intake fan. There are just too many cables ribbon and power in too small a space with opposite orientations and other things, bays, fan covers in the way. I think these kinks must have some electrical effect too.
I think round cables can be very helpful in a small ATX case with a limited airflow. I realize the limits of my present case and search for a new one that better fits many criteria important to me. I don't care how much the cables cost. They're not that much more than an extra ribbon cable cost a few years ago at a office superstore.
Here are some references I found.
http://www.sysopt.com/articles/cabround/
http://www.casejunkiez.com/reviews/round_ide.shtml
http://www.ramelectronics.net/html/ata-66_cables.htm
http://www.techtv.com/screensavers/show ... 23,00.html
http://dansdata.com/rcables.htm
http://www.3dvelocity.com/reviews/akasa ... ata133.htm
http://www.a1-electronics.net/PcHardwar ... bles.shtml
Many of the sites talking trash about round cables sport google ads selling them.
It seems to me that with the litigious nature of US society that if there were significant problems with round cables that major retailers would avoid selling them lest the end up in a class action lawsuit.
I could find no warnings against using round cables on the major HD manufacturer sites, maxtor, seagate and wd. If the were a significant problem, their call centers would hear about and they would place an insert (as they have for other problems) in the box to not use them.
I could not find any reports of verified data error from use of modern store bought shielded cables 18" in length.
I found a lot of speculation but not many facts on the web. I found a dearth of any scientific evidence that modern store bought round cables are causing data errors. It seems in the years they have been available it would have been easy for a review site to observe the signal on an oscilloscope for flat and rounded cables and post the results.
Steve
-
- SPCR Reviewer
- Posts: 8636
- Joined: Sat Nov 23, 2002 6:33 am
- Location: Sunny SoCal
Referencing my post above, I've had data error from 2 different sets of Coolermaster 18", 2 device rounded cables. Once I discovered what was giving me errors in Event Viewer and knocking my transfer speeds down to PIO I was able to swap flat ribbon cables into the system and the errors disappeared. When I put the rounded cable back into the system the errors reappeared. I returned that rounded cable and got a replacement. The replacement gave me errors too.canthearyou wrote:I could not find any reports of verified data error from use of modern store bought shielded cables 18" in length.
That might not be good enough for some, but it sure was for me. I have always used flat ribbon cables since then and have never had any more errors of that type for the past 1.5 years.
-
- Posts: 153
- Joined: Mon Oct 27, 2003 5:35 pm
- Location: Arlington, Virginia
- Contact:
I would be wary after an experience like that. The HD is connected by a 10" (minimum length) single cable, which may be a factor in my not having seen any errors. Also, I would be hesitant to put the HD on a Y cable, just as I am hesitant to power it from a Y cable, which I have had trouble in the past. I keep the CD's on a Y cable.
From what I have read, I would be cautious about using round cables at ATA100 or above. I might use them, but would monitor them closely.
I did notice in one of the early round cable articles, a tech from WD did not have any objections to the idea of round cables.
I will be more cautious about them in general, however.
Steve
From what I have read, I would be cautious about using round cables at ATA100 or above. I might use them, but would monitor them closely.
I did notice in one of the early round cable articles, a tech from WD did not have any objections to the idea of round cables.
I will be more cautious about them in general, however.
Steve
One aspect that hasn't been mentioned yet, but should be included, is to always use as short a cable as possible. This is especially true for rounded cables, where the incidence of crosstalk increases with length, but it applies to flat cables as well.
Don't use a 24" cable where an 18" would reach. Don't use an 18" where a 12" would do. The shorter the better.
Going from a 24" cable to a 12" cable will yield real, measurable, performance gains.
Don't use a 24" cable where an 18" would reach. Don't use an 18" where a 12" would do. The shorter the better.
Going from a 24" cable to a 12" cable will yield real, measurable, performance gains.
-
- Posts: 153
- Joined: Mon Oct 27, 2003 5:35 pm
- Location: Arlington, Virginia
- Contact:
I want to add that what made me switch to round cables was that it got to the point where I could not reach into my case because the tangle of ribbon cables. It was not the cool look or modding. I could not see what I was doing installing nearly anything. Without removing the MB and all. The ribbons were blocking nearly any view or entrance into my case. If you ever played with a "Chinese finger puzzle" you will know what I mean trying to get my hand out of those ribbons!
That's quite a good idea, but I was thinking more along the lines of EM radiation, or the lack thereof. An optical solution would be immune from interference from other components and would also not generate it's own. Coupled with an ultra thin cable and the potential for ultra highspeed datarates give the spectrum available, optical connections would be awesome.Zyzzyx wrote:Wireless. :pI just wish there was an optical replacement for SATA - that's would rule all solutions.
As cool as wireless storage would be, it's not really practical given the electrical noise inside a PC case.
We're missing an obvious solution to the whole "ribbon vs round" debate: Do away with IDE altogether.
Sure SATA is a nice improvement; thin cables up to 3 meters long, higher transfer speed, etc etc.
But may I suggest an even better choice: Fiber Channel
For those unfamiliar to this SCSI replacement here's a blurb from Sun:
Sure SATA is a nice improvement; thin cables up to 3 meters long, higher transfer speed, etc etc.
But may I suggest an even better choice: Fiber Channel
For those unfamiliar to this SCSI replacement here's a blurb from Sun:
Yes folks that's right, 10km cable lengths. Sure the drives are screaming loud 15000RPM SCSI drives, but who cares when you can mount them in the next county?Sun Microsystems Whitepaper wrote:Fibre Channel is an industry-standard, high-speed serial data transfer interface that can be used to connect systems and storage in point-to-point or switched topologies. Fibre Channel Arbitrated Loop (FC-AL), developed with storage connectivity in mind, is a recent enhancement to the standard that supports copper media and loops containing up to 126 devices, or nodes. Like SSA, FC-AL loops are hot-pluggable and tolerant of failures.
The FC standard supports bandwidths of 133 Mb/sec., 266 Mb/sec., 532 Mb/sec., 1.0625 Gb/sec., and 4 Gb/sec. (proposed) at distances of up to ten kilometers. Gigabit Fibre Channel's maximum data rate is 100 MB/sec. (200 MB/sec. full-duplex) after accounting for overhead.
In addition to its strong channel characteristics, Fibre Channel also provides powerful networking capabilities, allowing switches and hubs to enable the interconnection of systems and storage into tightly-knit clusters. These clusters will be capable of providing high levels of performance for file service, database management, or general purpose computing. Because it is able to span up to 10 kilometers between nodes, Fibre Channel will allow the very high speed movement of data between systems that are greatly separated from one another.
About fiber channel : now, you can get quite easily 'old' entreprise HDD (9-18 Go) for cheap, same for PCI cards for it or SCSI-FC-AL adapters for disks.
For mass storage, just use a file server in a gigabit network and you can have really fast {internal disk}less workstations.
That's what I would do if I wasn't living in a 2-rooms appartment
For mass storage, just use a file server in a gigabit network and you can have really fast {internal disk}less workstations.
That's what I would do if I wasn't living in a 2-rooms appartment
Last edited by dago on Tue Nov 04, 2003 8:04 am, edited 1 time in total.
Crosstalk is from inductive coupling between parallel conductors. Ground wire between signal wires is to reduce coupling. When wires cross at a high angle (i.e. 90 degrees) there is little or no inductive coupling. So keep your bends sharp, avoid doubling your cables, and you should have no problem.al bundy wrote:What about all the intricate folding of flat 80-wire IDE cables that some members around here do, in order to clean up the interior of their cases?
Does folding 80-wire IDE cables also perhaps allow data corruption?