Personal computing discussed

Moderators: renee, morphine, Steel

 
etilena
Gerbil Jedi
Topic Author
Posts: 1674
Joined: Wed Jun 12, 2002 7:43 am
Location: .ozziefied.

40 and 80 pin IDE cable difference?

Mon Feb 24, 2003 6:08 am

Other than the physical differences, will it affect performance if I stick a 40 pin cable to a 133ATA drive?
*yawn*
 
BigMadDrongo
Gerbil Elite
Posts: 909
Joined: Mon Apr 29, 2002 12:57 pm
Location: London, UK
Contact:

Mon Feb 24, 2003 6:55 am

It'll force the drive into UDMA2 mode, i.e. ATA33. ATA66 or higher requires the 80-pin cable. I'm not sure what the current word is on to what extent that will actually affect performance (i.e. how much real-world difference there actually is between ATA33 and ATA133).
 
Buub
Maximum Gerbil
Posts: 4969
Joined: Sat Nov 09, 2002 11:59 pm
Location: Seattle, WA
Contact:

Mon Feb 24, 2003 11:26 pm

Other than the physical differences, will it affect performance if I stick a 40 pin cable to a 133ATA drive?


Um, yeah... that's the entire reason they invented the 80-pin cable! :-)
 
highlandr
Gerbil Elite
Posts: 575
Joined: Thu Dec 27, 2001 7:00 pm
Location: Somewhere in downstate IL
Contact:

Mon Feb 24, 2003 11:31 pm

If your drive is old, it won't make much difference. With my results (on an ATA66, a while ago) my nero speed tests - the cutting edge of precision in hard drive speeds - went from about 27 megs/sec to 35. The drive will still work, but you probably will get an error every time you boot the machine. You can easily ignore this error, if that's the only cable you have.
 
mac_h8r1
Minister of Gerbil Affairs
Posts: 2974
Joined: Tue Sep 24, 2002 6:57 pm
Location: Somewhere in the Cloud
Contact:

Tue Feb 25, 2003 1:41 am

The 80-wire cable has 2 wires per pin, which acts as both a ground and as EMI protection. It is necessary because at speeds over ATA33, the wires will cross-talk, resulting in bit-rot and frequent catastrophic failure and data loss.

Small price to pay, you've got the UDMA/66+ mobo, the UDMA/66+ HDD, so why not pay the extra dollar for the cable, 2 for a round one, and get your money's worth because you'll actually see the drive perform at its rated speed.
mac_h8r1.postCount++;
Chaos reigns within. Reflect, repent, and reboot. Order shall return.
Slivovitz owns you.
 
etilena
Gerbil Jedi
Topic Author
Posts: 1674
Joined: Wed Jun 12, 2002 7:43 am
Location: .ozziefied.

Tue Feb 25, 2003 2:26 am

Okie thanks, just another question. Do optical drives transfer data at speeds fast enough to suffer from data errors, like say a 48x burner?
*yawn*
 
mac_h8r1
Minister of Gerbil Affairs
Posts: 2974
Joined: Tue Sep 24, 2002 6:57 pm
Location: Somewhere in the Cloud
Contact:

Tue Feb 25, 2003 9:20 am

Fast enough? as in fast enough to overload the bus? a 48x burner will overload ATA33, but not UDMA 66. One thing that the manufacturers have done as a precaution is the new buffer underrun-proof drives adn software. By loading the drive with anywhere from 2 to 8 MB of cache, the more the merrier, and monitoring software, it maintains a constant stream of data to the drive. When the buffer drops below 25%, the computer bursts data across the IDE channel. It keeps at least 10-40 megs of data in RAM, so once again, the more the merrier.

if you're running an ATA/33 bus, it's time for a complete hardware renovation, unless you want to run at like 12-16x. Get a nice new SOYO Dragon board, almost anything in stores today will support your needs/ Fry's usually has a few sub-150 deals, mobo and processor.

Good luck, i gotta get to school
mac_h8r1.postCount++;
Chaos reigns within. Reflect, repent, and reboot. Order shall return.
Slivovitz owns you.
 
pattouk2001
Gerbil Jedi
Posts: 1903
Joined: Thu May 30, 2002 10:44 am
Location: Birmingham, UK.
Contact:

Difference

Tue Feb 25, 2003 9:37 am

Hi. Here's the essencial difference:

40pin IDE = Up to ATA33
80pin IDE = Up to ATA133

Basically, assuming your board also supports ATA133 too, coupling an ATA33 cable to an ATA133 drive/board, will hinder the drive/board's ability to maintain a peak data rate of up to 133mb/s, the ATA33 cable will simply act as a bottleneck within the system.
 
Steel
Global Moderator
Posts: 2330
Joined: Wed Dec 26, 2001 7:00 pm

Tue Feb 25, 2003 12:16 pm

Um guys, a 48X CD burner only transfers data at 7.2 MB/sec which is well within the 33MB/sec that ATA-33 provides. We have a few 48X Lite-On's here at work attached via ATA-33 cables and have no trouble at all burning at full speed.
 
pattouk2001
Gerbil Jedi
Posts: 1903
Joined: Thu May 30, 2002 10:44 am
Location: Birmingham, UK.
Contact:

Agreed, sorta

Tue Feb 25, 2003 12:39 pm

Um guys, a 48X CD burner only transfers data at 7.2 MB/sec which is well within the 33MB/sec that ATA-33 provides. We have a few 48X Lite-On's here at work attached via ATA-33 cables and have no trouble at all burning at full speed.


.....Yes, I agree, a ATA33 cable would be fine coupled with a CDRW/CDROM, but that isn't the issue here. An ATA133 drive transfers data at an average of 100-110mb/s, and has a peak of 133mb/s, as a HDD is a couple of hundred times faster than a CDRW/DVD/CDROM. What we/I are saying, is that if etilena couples one of these cables to a ATA133 capable HDD and Motherboard, he's going to be crippling the performance of both parts, and as I say, this causes a bottleneck within the system.
 
mac_h8r1
Minister of Gerbil Affairs
Posts: 2974
Joined: Tue Sep 24, 2002 6:57 pm
Location: Somewhere in the Cloud
Contact:

Tue Feb 25, 2003 5:37 pm

ugh, you're right....my math is shot at 6 in the morning...thinkin megabits not bytes.


maybe it's time for me to enter the world of coffee dependence!

nah...just 16, i'm good 'til college.
mac_h8r1.postCount++;
Chaos reigns within. Reflect, repent, and reboot. Order shall return.
Slivovitz owns you.
 
eckslax
Gerbil Elder
Posts: 5320
Joined: Sat Oct 05, 2002 4:22 pm
Location: Vast Right Wing Conspiracy HQ

Tue Feb 25, 2003 7:14 pm

Pattouk, I have never heard of an ATA133 drive actually averaging anywhere near 100-110mb/s, let alone peaking. To my knowledge, I don't think any IDE harddrive has even broken ATA100 transfer speeds outside of marketing lies. Could you please clarify with some links to some tests that show transfer rates this high? Not that I am trying to stomp down your statement, I just want to know if I am off base here.
"God created man. Samuel Colt made them equal."

"Government does not tax to get the money it needs; government finds a need for the money it gets." - Ronald Reagan
 
pattouk2001
Gerbil Jedi
Posts: 1903
Joined: Thu May 30, 2002 10:44 am
Location: Birmingham, UK.
Contact:

Ok....

Wed Feb 26, 2003 6:10 am

....Ok, fair enough. The issue here though is that a ATA33 cable coupled with a ATA133 drive/board will cripple its performance and act as a "bottleneck". Simple as that.
 
eckslax
Gerbil Elder
Posts: 5320
Joined: Sat Oct 05, 2002 4:22 pm
Location: Vast Right Wing Conspiracy HQ

Wed Feb 26, 2003 6:22 pm

Agreed. Sorry if I sounded a bit critical or argumentative. That was definately not my intention.
"God created man. Samuel Colt made them equal."



"Government does not tax to get the money it needs; government finds a need for the money it gets." - Ronald Reagan
 
Buub
Maximum Gerbil
Posts: 4969
Joined: Sat Nov 09, 2002 11:59 pm
Location: Seattle, WA
Contact:

Wed Feb 26, 2003 11:48 pm

Not exactly a "bottleneck". Using a 40-pin cable with a properly designed ATA-133 controller will force it into ATA-33 mode. ATA-66 and higher requires the 80-pin cable.

But yes, eckslax is also right. Even the fastest hard drive on the planet, the SCSI-320 Seagate Cheetah 15k.3 is only capable of blasting a maximum of around 70MB/s of sequential data per single drive. For any IDE drive, ATA-66 is the fastest it can do in real world performance. The extra bandwidth WILL make a very slight difference for command overhead and already cached data, but really, you would be hard pressed to measure any meaningful difference between ATA-100 and ATA-133 with current drives.

It's a hope-for-the-future thing. In time, drives will be able to burst 133MB/s or more, but that is a long ways away yet for IDE drives.
 
pattouk2001
Gerbil Jedi
Posts: 1903
Joined: Thu May 30, 2002 10:44 am
Location: Birmingham, UK.
Contact:

Don't worry

Thu Feb 27, 2003 3:00 am

Agreed. Sorry if I sounded a bit critical or argumentative. That was definately not my intention.


.....No worries pal.
 
etilena
Gerbil Jedi
Topic Author
Posts: 1674
Joined: Wed Jun 12, 2002 7:43 am
Location: .ozziefied.

Thu Feb 27, 2003 4:05 am

If an 80pin IDE cable can transfer theoritically 66MB per sec at least, why is it a problem to hook a CD ROM and a CDRW together? Even at 48x you can only burn 7.2MB/s.. which is a lot, lot less than what the cable can transfer.
*yawn*
 
Steel
Global Moderator
Posts: 2330
Joined: Wed Dec 26, 2001 7:00 pm

Thu Feb 27, 2003 9:51 am

etilena wrote:
If an 80pin IDE cable can transfer theoritically 66MB per sec at least, why is it a problem to hook a CD ROM and a CDRW together? Even at 48x you can only burn 7.2MB/s.. which is a lot, lot less than what the cable can transfer.
It's less about the data transfer rate and more about the fact IDE doesn't multitask well. In a master/slave configuration only one drive can be active on the bus at a time. Normally that's not a problem but when you do a CD to CD copy, sometimes the drives can trip over each other and cause it to take longer, or in extreme cases, coaster the CD. That's one reason why serial ATA is so nice, each drive is on its own channel (though MB makers need to start putting more than two ports on the board... :evil:)

Who is online

Users browsing this forum: No registered users and 1 guest
GZIP: On