AAM only effective during Hitachi test

mjcuk · Post by **mjcuk** » Mon Nov 13, 2006 2:31 pm

Hi all,

I've just got a 250GB Western Digital SE16. On installing it, it did seem to have a reasonably loud seek noise. This was to be expected as AAM was disabled.

I used the Hitachi feature tool to enable AAM. In the feature tool, I went into the AAM test and was hugely impressed when the seek noise went from being quite noisy to being even quieter than I expected as I decreased the setting value down to 128. At 128 the seek noise really was inaudible unless my ear was right at the drive.

I was rather disappointed when I rebooted to find that the drive was not half as quiet as it was during the Hitachi feature tool test. It was a little quieter than it had been without AAM but was by no means quiet. At first I thought that the AAM setting had been reset (as I had read about this issue appearing) but I then downloaded WinAAM and found that the AAM setting was still at 128. I tried putting it back to 256 and I did notice that the seek noise was slightly higher, but not much.

My question therefore is, why is it that AAM seems to be incredibly effective when I test it within the Hitachi tool but isn't anywhere near as effective in normal use? The drive is literally going from being astonishingly quiet in the test to being borderline unacceptable elsewhere.

kesv · Post by **kesv** » Tue Nov 14, 2006 2:40 am

mjcuk wrote:My question therefore is, why is it that AAM seems to be incredibly effective when I test it within the Hitachi tool but isn't anywhere near as effective in normal use? The drive is literally going from being astonishingly quiet in the test to being borderline unacceptable elsewhere.

One reason could be the access pattern used by AAM test. We don't really know what kind of pattern the test invokes on the drive, but it could be that it doesn't try to simulate a real filesystem.

Things like the fragmentation of your filesystem do affect how many seeks the drive needs to do to read a file.

mjcuk · Post by **mjcuk** » Thu Nov 16, 2006 2:40 pm

I guess it is possible that the test simply isn't simulating the real world very well.
I can't help but wonder if it's something else though. There's just such a stark difference in the level of noise produced in the test and that of the real world. The drive contains a fresh install of Vista which I would hope is reasonably fragment free (especially with the new installation method Vista has.) The drive seems to be noisy on any file access.
Have other people experienced the hitachi test being considerably quieter than the real world?

If this is as good as it gets with the drive (which I suspect it is) then I am somewhat disappointed with my SE16 unit (it's probably unfair to tar them all with the same brush.) It certainly doesn't live up to the rave review given on this site. I would say it's the loudest seeking drive I've got at the moment. The other drives I've got are Seagates - some with and some without AAM.

DonQ · Post by **DonQ** » Thu Nov 16, 2006 5:31 pm

mjcuk wrote: ...
If this is as good as it gets with the drive (which I suspect it is) then I am somewhat disappointed with my SE16 unit (it's probably unfair to tar them all with the same brush.) It certainly doesn't live up to the rave review given on this site. I would say it's the loudest seeking drive I've got at the moment. The other drives I've got are Seagates - some with and some without AAM.

My SE16 500GB is the loudest part of my computer. Then again the only other sources of noise are the 100mm fan in a Scythe MINE @ 990 RPM and a D12SL-12 Yate Loon @ 975 RPM. I'm researching enclosures...

Devonavar · Post by **Devonavar** » Fri Nov 17, 2006 1:27 am

I don't think it's the Hitachi tool that's at fault. As far as I can tell, the test pattern just seeks to random sectors in sequence. The noise it generates is far noisier than anything I've heard in actual use. Real usage is generally sequential, not random, and is often quieter than the tool. I can say this based on a year of testing drives with the tool and examining how they perform in actual systems.

I suspect that it may be a new revision that's at fault. Today I heard a fully assembled system that used a 500GB WD drive that was noticeably louder than our original sample. I'll try to check the revision number tomorrow. Mjcuk: Could you post the revision number of your drive? It would be good to know whether WD has changed something crucial, or if this is just sample variance rearing its ugly head.

mjcuk · Post by **mjcuk** » Fri Nov 17, 2006 6:58 am

The revision number of the drive is 00MJB0.

Devonavar wrote:I don't think it's the Hitachi tool that's at fault. As far as I can tell, the test pattern just seeks to random sectors in sequence. The noise it generates is far noisier than anything I've heard in actual use. Real usage is generally sequential, not random, and is often quieter than the tool. I can say this based on a year of testing drives with the tool and examining how they perform in actual systems.

Any idea why there's such a big difference then? I appreciate that the revision and sample variance of the drive could cause it to be more noisy, but why is it so quiet during the test?
During the test, it sounds as described in your review (almost inaudible from a metre away.) In reality the seeking is easy to hear 5 metres away.

Devonavar · Post by **Devonavar** » Fri Nov 17, 2006 9:57 am

Don't know. I'll get back to you after I test our original sample.

Felger Carbon · Post by **Felger Carbon** » Fri Nov 17, 2006 12:04 pm

I'm a hardware guy, so this may be a really dumb suggestion. Hitachi AAM works fine, and then you reboot and the AAM disappears? Is it possible the reboot is disabling the AAM?

A simple check: Use the Hitachi tool again. Reboot. Then run the Hitachi tool again, but do not set the AAM. Just test it. Is it still quiet under the Hitachi tool test? If not, the reboot is for some reason cancelling the AAM.

A dumb suggestion, but a really simple test.

mjcuk · Post by **mjcuk** » Sat Nov 18, 2006 1:36 am

Felger,

Thanks for the suggestion, it's not a dumb at all. When I saw the symtoms, the first thing that came to mind was that AAM was being disabled on Reboot. I downloaded WinAAM that lets you change the AAM setting within Windows. This showed that AAM was still enabled.
Going back into the Hitachi tool also showed that AAM was still enabled. So it looks like AAM isn't being disabled on reboot.

The Hitachi tool doesn't actually let you do the test you suggest. When you run the AAM test in the tool, it disables AAM (sets the value to 256) irrespective of what the hard drive had been set to. It then provides you with a sliding scale that allows you to incrementally enable AAM by letting you decrease the AAM value from 256 (loud) all the way to 128 (fully quiet.)

AAM only effective during Hitachi test

AAM only effective during Hitachi test

Re: AAM only effective during Hitachi test