Determine platter damage based on sound analysis

[TerraNova]

How often you change HDA and find out that one of the the platter (middle platter) has damaged and killed your good donor head?

Sound analyser, analysis the normal HDD start up and calibration sound and compared to the faulty drives based on HDD brands and models. The waveform is then compared and determine the platter/s are damage or in good conditions.

Currently i am doing some research on different hdd models and brands and found that there are ways to identify the platter/s are damaged. Damaged head has its on unique signature compared to the platter damage. But there are still many hdd to test and compared.

Anyone in this forum interested in this research or have some experience with it please give some comments?

Andrew

[rchadwick]

I'm interested. I actually purchased an Acoustic guitar pickup to use as a HD Microphone (Has a nice suction cup on it), and intended to make a library of drive sounds for similar reasons. Never got to it tho. Soldering the right jack on to it is still on my long to-do list. I'll move it up in priority.

I'm surprised you can actually tell paltter damage from just sound, unless if you're talking about when the head hits the damaged area.

[TerraNova]

Hi rchadwick , you need a very sensitive microphone in order to pickup the signal difference.

Top and bottom platter damage you can tell by visual inspection.

Let's see the scenario, lets say there are H0, H1, H2 and and H3 , 2 platters HDD, H1 and H2 are located in between platters which is not visible or very hard to see the damage area unless on the outer edge of the platter.

Of course you can try different methods of inspection, such as examining the HDA under microscope thats just another way.

[rchadwick]

I certainly see the need for examining platter damage between platters easily. I'm just wondering how it can be done by sound. I vaguely remember Salvation's products let you position the heads manually. If so, I'm sure a sweep of the surface would make sound when hitting damage, and I bet it would be pretty reliable. I suppose it's possible the surface damage might whistle while spinning, but can that be detected reliably? I'd imagine that if you listen close enough, you'll hear manufacturing differences, even between the same model.

Either way, I'm putting together a DigiKey order, and I'll look for a sensitive microphone to play with. The pickup I have now I got more for convenience (Suction Cup).

[beto]

Hello terranova friend

Im interesting on your research´s, at some cases, i tried to recorded the sound generated from x hdd the most cases seagate using Nero , sound editor, and u can use it too for check some usefuls thing and problems , analyzing the wavefrom from the patient vs a good hdd

Best Regards fron Mexico

Sinceraly

Alberto

[rchadwick]

If sound turns out to be useful, maybe we can start a database of drive sounds. Diagnosing by sound would be much easier and more productive if we had something to compare the sounds to. Also, a quick listen is much easier than opening the cover, and I'm all for something that makes things quicker and easier

[dobrevjetser]

Agree.
I will look into it also.

Dobre

[Odiferous]

This is something we've been working on for some time now.

However, to date we have only been able to discern anomalies where the platter has been severely damaged - for example, a deep scratch.

I don't believe it is possible to use sonar alone to pick up platters that have been damaged by minor head crashes.


Duncan

[TerraNova]

Hi Guys

In order to compare accurately we must all use the same sound analysis software, same microphone and same location for the probe.
I will post photos shortly.

Thank you all for your interest.

[pepe]

Hi,

I think frequency domain analysis could give some results.
the main problem (as always) here is the signal to noise ratio, in this case the signal is the noise and the noise is the sound of the spindle itself

pepe

[dmarques]

We've made a few recordings as well some years ago, but as it raised so many questions and few exact answers, we leave beside, but I'm willing to try again too.