Caviar Blue WD10EALS platter size?


#1

Is this hard drive using the 500GB/platter design or the older 333GB/platter? Or does it depends on manufacture date?


#2

According to this specification sheet …

http://support.you.gr/catalog/82/82F27B5B331C464290CA917902848843.pdf

… the WD10EALS has a maximum sustained data rate of 126 MB/s.

We can compare this against HD Tune’s read benchmark results:
http://www.hdtune.com/testresults.html

Unfortunately your model is not listed, but we can estimate the capacity per platter by comparison with other drives.

For example, Seagate’s ST3500641AS model is a 3.5", 7200 RPM, SATA drive with a capacity of 500GB distributed over 4 platters. This works out to a data density of 125 GB/platter. The max data rate for the ST3500641AS is 65 MB/s.

A drive that has 500GB per platter would have four times the data density (bits per square inch). This means it would have twice as many tracks per inch, and twice as many bits (or sectors) per track.

The data rate is …

(sectors / track) x (bytes / sector) x (rotations per second)

This means that a 3.5", 7200 RPM drive with 500GB / platter would have an expected transfer rate of 130 MB/s (= 2 x 65 MB/s).

OTOH, a drive with 333 GB/platter would have a transfer rate of only 106 MB/s (= 130 x sqrt(333/500)).

So ISTM that the WD10EALS has 500GB per platter.


#3

That’s a very technical estimation  :slight_smile:

Actually not to compare with Seagate, but comparing with WD Black series: http://www.wdc.com/wdproducts/library/SpecSheet/ENG/2879-701276.pdf

WD2001FASS | WD1002FAEX | WD1001FALS | WD6402AAEX

138MB/s | 126MB/s | 106MB/s | 126MB/s

I know that the WD2001FASSand WD1002FAEX are definitely 500GB/platter and WD1001FALS is 333GB/platter, which matches quite right with your estimate. However, I’m not sure the platter size of WD6402AAEX.

Also, when you look at all the other Blue series drive, even the 80GB version has 126MB/s, but it’s not possible that it’s using a 500GB platter.


#4

That’s very interesting.

If you ignore the capacities and just go by data rate alone, then a data rate of 138 MB/s would correspond to a capacity per platter of …

(138 / 126) ^ 2 x 500 = 599.8 GB

Similary, a data rate of 113 MB/s gives …

(113 / 126) ^ 2 x 500 = 402.1 GB

These results seem too close to 600GB and 400GB per platter to be a coincidence.

So my guess is …

WD2001FASS - 2TB - 4 platters - 8 heads - 300GB per head - 138 MB/s
WD1002FAEX - 1TB - 2 platters - 4 heads - 250GB per head - 126 MB/s
WD1001FALS - 1TB - 3 platters - 6 heads - 166GB per head - 106 MB/s
WD7501AALS - 750GB - 3 platters - 5 heads - 166GB per head - 106 MB/s
WD6402AAEX - 640GB - 2 platters - 3 heads - 250GB per head - 126 MB/s
WD6401AALS - 640GB - 2 platters - 4 heads - 200GB per head - 113 MB/s
WD5001AALS - 500GB - 2 platters - 3 heads - 200GB per head - 113 MB/s

Since all the above models rotate at 7200 RPM and have the same 3.5" platter size, this begs the question, if the above calculations are correct, do some of the drives have shorter strokes than others, ie less tracks, and if so, is this reflected in better full stroke seek times?


#5

To my surprise, this is the response from WD, even after I gave them the serial number of my hard drive:

“I sincerely apologize, but we change the parts in the drive as we build them, therefore, there is no set or specific platter size for this hard drives.”

“I apologize, but we do not have specific information on platter size on our hard drives, this hard drives are submitted under several revisions and changes that makes it difficult for us to keep on track on this information.”

That sounds to me like they don’t know what they’re making.


#6

I’ve had a closer look at the Caviar Blue spec sheet.

As you have observed, all the models show the same max sustained transfer rate of 126 MB/s. However, this is clearly not possible for the PATA drives since they are limited by a buffer-to-host rate of 100 MB/s, ie UDMA 5.

Furthermore, here are two threads which show HD Tune results of 71.6 MB/sec and 72.0 MB/sec, respectively, for the WD800AAJS:

http://ut2004.titaninternet.co.uk/forums/showthread.php?t=18410
http://www.geekstogo.com/forum/Slow-Hard-Drive-Speeds-t163564.html&st=15

Based on the data rate, one would expect a capacity per platter of …

500 x (71.6 / 126) ^ 2 = 161 GB

Therefore I expect that the WD800AAJS has 1 head. The 160GB, 250GB, and 320GB models probably have 2,3, and 4 heads, respectively.

This URL quotes a max data rate of 100.2 MB/s for the WD5000AAKS:
http://www.tomshardware.com/forum/246584-32-barracuda-7200-st3500320as-caviar-se16-wd5000aaks

That would suggest a platter size of …

500 x (100.2/ 126) ^ 2 = 316 GB, ie 320GB per platter

Here are two threads which show HD Tune results of 113.8 MB/sec and 108.4 MB/sec, respectively, for the WD6400AAKS:

http://forums.storagereview.com/index.php/topic/27104-640-gb-seagate-st3640323as-vs-wd6400aaks/
http://www.tomshardware.com/forum/245302-32-western-digital-drive-wd6400aaks-640gb

That suggests a capacity per platter of between 370 and 408 GB, probably 400GB per platter.


#7

See this article:

Western Digital serves up 320GB on a Platter:
http://www.anandtech.com/show/2455

“It does not come as a surprise that the first drive from WD to utilize the 320GB per-platter technology is the Caviar SE16 320GB WD3200AAKS. What is surprising to us is that WD is not changing the product designation on drives that feature this new technology. … Ordering a WD3200AAKS could land you the new drive or the older design with two 160GB platters. Since only the part number changes in this case, the one you want is WD3200AAKS-00B3A0.”


#8

Thanks for the follow up. However, I’m only interested in the WD10EALS 1TB model :slight_smile:

From this blog: http://yertech.blogspot.com/2010/02/wd1001faes-and-wd1002faex-appear.html, it suggests the WD10EALS should have the 500GB/platter design as the other new WD hard drives, but I just want to have a more concrete answer, which WD itself was unable to provide.


#9

Well, it was an interesting intellectual exercise. ISTM that hard drive specs can’t be trusted because the spec sheets have plenty of errors, and because the specs are being continually improved.

BTW, my assumption that the number of tracks per inch increases at the same rate as the number of bits per track is questionable. For example, in the case of Seagate’s 7200.11 and 7200.12 families, the numbers increase from 1090 kbits/in max and 150 ktracks/in avg to 1413 kbits/in max and 236 ktracks/in avg. That’s 1.3x and 1.57x, respectively. The areal density increases from 219 Gbits/in2 avg to 329 Gbits/in2 avg, which is 1.5x. This is less than the 2.04 (= 1.3 x 1.57) one would expect. Furthermore, since the 7200.12 series had 500GB platters, then one would expect that the 7200.11 family would have had 333GB discs (= 500 / 1.5). Yet the 7200.11 models were limited to 250GB per platter. Strange …