IDE, SCSI and the rpm

Ways IT TIES, Ultraties, UltraDMA, UDMA...

Material to test... and the "incident"

You try yield


Released page

Juan Herrerías Rey


Articles about storage

To install a hard disk

principal page
what is...
to update
Contact Us
forum and chat



The distribution of the present pages is prohibited without the express assent of the authors.

All the marks are registered by his respective owners.

For more legal details, pulsate in the following icon


ATA100 + 7.200 rpm vs. ATA66 + 5.400 rpm


To buy a hard disk is usually a task of insured success, especially because there are no many manufacturers and that there is sound, in general, of a very homogeneous quality... although there are manias and rumors, skylight.

Also, in this case yes it usually happens that the most expensive thing is better (something that in other products as the CPUs not always is like that, how there know the lucky holders of an AMD Duron). But: does he know really what he buys? Does he know if he is worth paying more for a disc of "7200 rpm" (and what sound the rpm)? And between ATA66 and ATA100? And that thing about the UltraDMA?

Since we are going to try to clarify little these questions, so that the next time is not right "of coincidence".


IDE, SCSI and the rpm

The first thing that should be evaluated when he goes to acquire a hard disk there is the interface of connection that it will use: IDE or SCSI. Nowadays this question is easier to solve that it does a time, thanks to the big advance that there have experienced the discs IDE (or EIDE, Enhanced IDE) in the last years.

Formerly, it was not doubt: the discs SCSI were more rapid, especially thanks to the fact that the checkers SCSI are "intelligent"; that is to say, that take charge for his account on behalf of the operations of entry / exit of information, unloading of this work the CPU. (Although, of course, there are always more "intelligent" checkers - and faces - than others...) Manual activation of the way DMA in Windows

Nevertheless, thanks to the application on the discs IDE of the ways UltraDMA (or, more correctly, than the bus mastering), the discs modern IDE do not block so much how earlier the CPU during the writing / reading of the information. It has not gone over (it will not even go over) to the efficacy of SCSI, but the advance is notable and for many uses more than sufficient. Even cases limit exist in which, in PCs with only 1 hard disk and scarce use of the multitask, it can be a little more rapid to use IDE that SCSI, since there is avoided the layer of commands that he allows to SCSI to handle simultaneously up to 15 devices.

In short: if he does not think to use his PC for highly professional "applications" (graphic applications, audio / video edition, servants...), does not need SCSI; a good hard disk IDE will be more than sufficient.

And what is a good hard disk IDE? Well, of answering to that is the intention of this article, but we can already be moving forward a fact: a good hard disk will have a speed of rotation of 7.200 rpm (revolutions per minute). The "normal" hard disks, "of consumption", turn normally 5.400 rpm, and the best of the market (SCSI of the highest services) to 10.000 rpm (a speed so high that it generates serious problems of noise and warming).

Good, this is quite on the porqué of IDE or SCSI, rpm and the others, but if he still wishes more clarifications on this matter pulsate here.


Ways IT TIES, Ultraties, UltraDMA, UDMA...

The first thing, be done to the idea of that, commercially speaking, all these terms recount to the same: the modern ways of access to hard disks IDE that they allow to liberate to the microprocessor on behalf of the work of the transference of information.

Access way

Theoretical maximum transference


UltraDMA33 (ATA33 or UltraDMA way 2)

33,3 MB/s

First of massive use; the standard IDE for years; "normal" cable

UltraDMA66 (ATA66 or UltraDMA way 4)

66,6 MB/s

Quite common; it uses a cable of 40 pines and 80 drivers

UltraDMA100 (ATA100 or UltraDMA way 5)

100 MB/s

Of recent implantation; it uses a cable of 40 pines and 80 conduct.

The new one UltraDMA100 (ATA100, or...) is a simple extension of the UltraDMA66. Since it will do little more than one year we already explain the characteristics of the UltraDMA66, we are not going to repeat them here; only to remember that to use this transference way (or in his case the UltraDMA100), it was necessary to fulfill the following requisites:

  1. that the hard disk is UltraDMA66 (or UltraDMA100), of course;
  2. that the cable is 40 pines with 80 drivers instead of the "normal one" of 40 pines with 40 drivers; this requisite is simple to expire and a bargain sale;
  3. that the operating system has support for DMA; Windows 98 and 2000 and the last versions of 95 and NT have it;
  4. that the checker IDE supports UltraDMA66 (or UltraDMA100).
The last one is the factor more limitante, since at present the support of discs IDE offers the chipset the motherboard; but it has solution: to buy another motherboard (ejem... skylight) or a card checker IDE, like the ABIT Hot Rod 66.

Cable classic IDE (left) and cable IDE for UltraDMA (right) - Source: Digital western

To use the way UltraDMA, the corresponding option will have to be activated in the BIOS of the PC (it it is usually for defect), and ALSO be activated in Windows, automatically well for the drivers of the chipset or going to Control panel-> System-> Manager of devices-> Disk drives; if there it does not find the pigeonhole "DMA" (to see top image), the fact is that the drivers take charge automatically (or that the disc does not support it, if it is ancient).

Finally, bear in mind that:

  1. the ways UltraDMA/UltraATA are compatible "backwards" with the most ancient (even with the already most ancient IT CHIRPED);
  2. the speed of transference of 33, 66 ó 100 MB/s is the theoretical maxim between hard disk and checker, not the disc; the last one depends on his mechanics, and especially on the rpm.
... Since we will see soon, it means the above mentioned that a disc UltraDMA66 ó 100 does not have porqué to be more rapid than one "only" UltraDMA33.


Material to test... and the "incident"

At first we think of proving all the hard disks that we had to hand, but soon we realized that he had not greatly sense: almost they all were giving results almost identical to those of the same characteristics (rpm, way of access and cache memory of disc or "buffer"), therefore we have centred on two discs:

  • Quantum Fireball CR of 8 GB (5.400 rpm, UltraDMA66, I caught of 512 K)
  • IBM DTLA 307030 of 30 GB (7.200 rpm, UltraDMA100, 2 MB cache memory)
  • Computer: Pentium III 733, 128 MB PC133, motherboard with chipset 815E

And what we have done is to prove them thoroughly, in diverse situations and with more possible applications. The Quantum is a high quality disc of manufacture (personal opinion, but credited by multiple sources), although of "normal" characteristics; the typical trustworthy but economic disc, let's say. Hard disk IBM

As for the IBM, it is supposed that it is a disc of exceptional characteristics, the most rapid, trustworthy, of prestige (used by Intel and AMD for his tests, for example)... and big part of this has been demonstrated in the tests, but the case is that it got damaged during the tests. Fatal error, since it is said usually.

The disc had scarcely 40 hours of use when it began failing, therefore we believe that simply it belonged to the familiar one 5 % (or 3 %, or what is) of "due defective discs, in normal conditions"; bear in mind that in any production process it thinks that there will go out defective products of inevitable form, and the only thing that can be done (or the economically profitable only thing) it is to realize a statistical testing of a certain number or even to prove all of them... but for days, skylight.

Anyway, it was in guarantee (logician, with 40 hours...), we have no had problems to change it and his substitute has not failed, but it is a sorrow. Any way do not believe that we have scared this as to change mark: we think to continue for the time being with IBM, although it should be only because in the last years there us have "died" discs of NEC, Maxtor, Samsung, Fujitsu...: if we were changing mark whenever it happens, we would not have left any for trying!!

Does it understand now why are the safety copies essential?


To return

Pulsate in the picture desplegable or in "Following"

Following - Sitemap - Contact us Best website of 2009 & 2010 voted by us are:
, , , , , ,,