Please see the attached graph, which shows an overlay of two identical drive models (WD3000GLFS) on the same machine measured with HDTune (64k block size), both w/o any data on it.

The brighter line in the foreground is the drive in question. As you can see, the graph oscillates wider and more irregular until the 40% mark. It starts slow and then builds up, and has a downward spike at the 10% mark.

This pattern is 100% reproducible, even when running on different controllers (Areca ARC1210, Nvidia Nforce).

My question:

1) Does this mean the HD is damaged ?
2) What exactly could be the cause for this ?

Thanks for you answers!

Could be.

Try something like MHDD.

_________________
You don't have to backup all of your files, just the ones you want to keep.

Seems that the drive has some 'more defective' zone. Rma or ask for repair. P.s. Before you ask, you -user- cannot repair by yourself.

Thanks a lot drccsc and BlackST.

Here's a screen of the HDDScan report (I stopped it after 30 min or so, but it shows the zone in question anyway). What is interesting is that when I scroll through the "map", the zone in question shows a band of light and grey (and a few green) blocks that repeats itself on every second screen or so, whereas this pattern does not show up on the rest of the disc.

Seems like it is RMA time.

And BlackST, no, I am not crazy enough to try and repair that myself, just want to ask for your pro's opinion whether this is a real damage or just my paranoia.

most every new one that comes of the factory floor has defective
im using a western digital with no problems at all
but it shows up defective.

Ok, just to keep you guys in the loop - and thanks again for your help/opinion.

I have returned the drive in exchange for another WD3000GLFS. Seems like this was the right decision - see below for comparison.

Just being my curious self - what are the technical reasons for this? Is this a damage of the platter surface (i.e. bad/slow blocks), a weak servo, damaged heads ?

My opinion (strictly technical) :

1) The "pattern" below OLD is perfectly regular. Drives are divided into ZONES and each zone has different recording density because angular speed is different if you measure it at the inner diameter and at the outer. If you use for instance MHDD on a Maxtor Calypso drive you'll see that different zones have different "patterns" on the screen, but the pattern in different areas of the same zone are the same, and regular.
What REALLY means are :
- AVERAGE access speed
- Number of relocated sectors at factory production level (this is related to the quality of the surface)
- Production "adaptives" like servo track writing techniques and/or equipment used, differences in heads (manufacturer) , differences in platters.
- Other variables like firmware revisions, SMART collection etc.
Let's analyze :
- Average access speed : the angular velocity is different at OD and ID so we have slightly different block access speed due to different density. It's physical. Otherwise, we should use different speeds depending on the head position to keep density equal. It's not practical on HDDs so the rotational speed is kept CONSTANT and we play on zone density.
- Number of relocated sectors: a 100% perfect surface is impossible. Defective areas are isolated and there are spare sectors AT TRACK LEVEL. Obviously there is a table to "translate" the relocated sector in transparent mode, so the user doesn't have to handle this problem. But this is a time-consuming (even in microseconds) process. You have to check the table, extract the relocated sector and seek it, while keeping the other processes alive.
During the life of the HDD bad sectors are generated every moment and relocated at the same time, the drive is capable to handle the entire process.
Of course there are drives with few or more defects, still perfectly in working order, but this will slightly affect performance.
WHAT YOU SEE is a drive with ZERO defects because the real defects are isolated by the drive itself.
- Production "adaptives" : different platters or different heads may behave differently. The same for improvements in firmware. The same for the equipment used for servo writing. What you see is the result.
- Other variables : newer drives usually have improvement in performance due to improvements in firmware and/or manufacturing process. I mentioned SMART : the SMART updating is likely to happen at regular intervals or when a relocation occurs or when an alarm is triggered - "when" it depends on the firmware code. But SMART updating is another time consuming process that CAN affect performance.
There could be other instances, but I should see the whole system and analyze it to make a diagnose.
Hope this has clarified something, I tried to be less "techie" as possible.

Hi BlackST,

wow, thanks for your exhaustive answer. So basically what you say (if I understood you correctly) is that whilst the "old" drive was probably still within the defined parameters, because of the analogue nature of hard drives there are always differences between two drives (i.e. platter/head quality, number of reallocated blocks at track level etc.) and the new one looks better ?

@all: thanks for your input, issue resolved!

Exact. Drive within specs but the newer was better or had improved firmware/characteristics. Where performance is critical maybe it's better to make a test before. P.s. Drives wear out during time,so their specs may vary.