Print

[spillage]

This is a copy of the post I just made on the Xbox Scene Forum. http://forums.xbox-scene.com

Today I purchased an 80 wire ATA133 cable for the EIDE interface of the XBOX. The length of the cable is 70cm, the extra length needed for my new custom case layout. I decided to add my 2 cents to the on going debate of the alleged speed increases some people have experienced using such a cable. The cable I used appears to be a standard flat profile EIDE ribbon with 80 conductors. Cost £4.00

(IMG:http://img.villagephotos.com/p/2003-9/392648/ATA133.gif)

The harddrive is an IBM Deskstar 120Gb with 2Mb cache.
The XBOX used for the test is a Version 1.1
Chip is an Xecuter 2.2 Pro loaded with 4983.06
Dashboard is UnleashX
Game Copying was achieved using UnleashX built in game copy utility.
As the table shows the tests were repeated three times to allow for human error.
To minimise errors the EIDE cables were changed after every test.

My conclusion is not in anyway to be interpreted as a means to put into disrepute the findings of others on the forum. My findings may not correspond to findings of others but are a representation of the results I obtained.
  <

[FarM_]

i tried that 2, but my standard cable was corrupted so i must buy ATA 100.
no big changes reason: Xbox have ATA33 HDD controller

But when i change my HDD from
Baracuda IV 80GB/2mb <--- to ---> Baracuda 7.7200 120GB/8mb
i see a differences, games load little bit faster now Halo 4 example.


 <

[Dan Wysocki]

Thats cool info, you rekon the lack of improvement was due to the extra length of the cable? Also, where did you get your cable from?
  <

[spillage]

I bought the cable today from a computer fair for £4.00. It came in a clear plastic bag with no identity. I bought it because I was under the impression that any 80 wire cable was ATA100 or higher.

A mate of mine, the guy I have the box on lone from whilst I work on mine, reckon his HDD does not support anything higher than ATA66 so I may not see an improvement. I will install my WD2500 and see if that gets any better. I am not sure if the lenght of cable makes any difference as I could have bought one that was 120cm.
 <

[opjose]

Yup this is consistent.

Usually the only reason people see an improvement to begin with is because of better noise immunity with the better cable.

That is their OLD configuration was problematic, hence the change.

But your findings are correct, the Xbox is limited by it's slow IDE interface rate.  <

[spillage]

Thanks.     <

[thedustycelt]

QUOTE

I was under the impression that any 80 wire cable was ATA100 or higher.

You are correct about this also. 66/100/133 all use the same 80 conductor cable.

Some might argue that better quality cables provide better results, but as far as I have seen, even the cheap ones are manufactured to a high enough standard.

  <

[opjose]

BTW:

EXCELLENT "Scientific" technique!  <

[Pizza Pizz]

Good post

At last peeps may now realise there is very very little if any when changing the cable and also that a 8mb cache isn't any big improvement ober a 2mb cache h/d

a 2mb cache & old ide works fine for me and I didn't see any improvement when i tried the go faster stuff  <

[heinrich]

hi,
I believe that the bios limits the transfer rate to udma2 (ata33), but can be changed to udma5 (ata100), (but requires an 80 pin cable of course)

btw, these are recent findings of mine, which dispute what I have said in the past on the issue..  <

[opjose]

QUOTE (heinrich @ Jun 2 2004, 10:12 PM)

hi, I believe that the bios limits the transfer rate to udma2 (ata33), but can be changed to udma5 (ata100), (but requires an 80 pin cable of course) btw, these are recent findings of mine, which dispute what I have said in the past on the issue..

Benchmarks DO NOT bear this out.

On PC's faster data rates are enabled by the bios when the hardware reports the presence of the 80 conductor cable.

AFAIK the Xbox has no such provisions.  <

[heinrich]

so the function call (have to change it considerably to post on the forums)
SetTransferRate(device, udma2);

is just for kicks?  
There are defines upto udma5, so it would seem rather silly if this could not be changed, while the code is there.  The again, I'm sure the the bios was developed before the exact hardware specs came about.

You are correct in that there is nothing to test for an 80 pin cable, hence the xbox always using ata33, regardless of limit imposed by hardware.

This post has been edited by heinrich: Jun 3 2004, 09:04 AM <

[opjose]

QUOTE (heinrich @ Jun 3 2004, 08:48 AM)

so the function call (have to change it considerably to post on the forums) SetTransferRate(device, udma2); is just for kicks?   There are defines upto udma5, so it would seem rather silly if this could not be changed, while the code is there.  The again, I'm sure the the bios was developed before the exact hardware specs came about. You are correct in that there is nothing to test for an 80 pin cable, hence the xbox always using ata33, regardless of limit imposed by hardware.

Haven't you hit it PRECISELY on the head though?

The calls are there but since there is nothing to detect the cable, so the higher data rate is never enabled.

Of course there is still the question as to whether the hardware actually supports it, which is unlikely.

You'd have to at the least assume that the IDE interface would bear the standard "marks" of an UDMA 5 capable interface.

The Xbox was manufactured after these were available and there would have been no reason to avoid putting in a UDMA 5 type IDE header (yeah I know there is no actually physical difference, but it would have cost MS nothing extra... as the OEM's were geared up for this anyway.).  <

[heinrich]

QUOTE (#ozxodus on efnet)

[18:13:47] [+h3inrich] Artifex: you around ? [18:14:15] [+h3inrich] got a quick question [18:14:35] [+h3inrich] "UDMA support (disabled for beta) for maximum HDD performance" = udma5 ? [refering to their feature list of XOS2] .... [18:41:09] [@Artifex] h3inrich: yes, with proper cabling and a touch of luck :-)

So it would appear that either the xbox hardware supports those speeds, or its just a empty feature.
just flat out changing udma2 to udma5 is simple, the detecting of the cable though.. no idea on that one :/

but it seems to throw out the argument that the ata controller was limited to ata33 (udma2).

Maybe spillage would like to do another test with a simple modbios of mine ?    

This post has been edited by heinrich: Jun 3 2004, 09:52 PM <

[jefferelli]

Has anyone tried a udma5 modified bios? Thanks  <