Hitachi Shows Off New 4TB Hard Drive

Now only if these would come to market. Seagate announced their 1TB platters a few months ago yet no products at all.
 
Edel: You link seems to be broken although it most likely is a problem with storage review. For those that can't read the article, mind sharing any details it contains?
 
I'm be a little skeptical about the speeds claimed in the article that drescherjm linked to. Drive density makes a huge difference but for a single platter drive to get those transfer rates is a little suspect.
If those speeds are correct than I can't wait to see what the 4TB drives are capable of. It is too bad that the 4TB drives will most likely only be available as very expensive external drives forcing us to either wait till they are available as bare drives or simply buy the external drives and take them apart.
 
I'm be a little skeptical about the speeds claimed in the article that drescherjm linked to. Drive density makes a huge difference but for a single platter drive to get those transfer rates is a little suspect.
If those speeds are correct than I can't wait to see what the 4TB drives are capable of. It is too bad that the 4TB drives will most likely only be available as very expensive external drives forcing us to either wait till they are available as bare drives or simply buy the external drives and take them apart.
So what speeds were they?

The reason the article is down is because they "botched the NDA".
 
They mention a 1 platter 1TB drive with media rates of 228MB/s and the 5400 RPM drive 180MB/s. Those are quite fast.
Whoa, yeah, that is unusually fast. :eek:

2TB WD Green EARS drive can do a sustained 125 MB/s read and write, but Hitachi's numbers there are insane.
 
I see 150+ MB /s from Hitachi 7200 RPM 2TB drives and 140 MB/s from 5400 RPM Samsung 2TB F4s. However these are 666 GB platters.
 
What I really expect from 4TB drives are affordable 3TB drives !
 
Platter density can be a big thing in terms of "speed". If they just push the tracks closer to each other, then not much speed is gained. However, if they push the space between bits closer together, then there is a lot to be gained even with the same rotational velocity. My guess is some other magic is going on as well...but damn, them be some nice numbers der.
 
I need some 5400rpm 3TB drives to replace my 1TB 7200rpm WD Black's.
 
Sorry for the link guys, anyway, if i remember correctly this was the image:

kia8I.jpg


Hitachi has also released the single-platter 7K1000.D (transfer speed of 227MB/s)
 
3tb drives *are* affordable. the 5400s have been priced at $109 on amazon. the 7200s have been at $150, directly in line with the 2tb 7200s offered at $100.
 
3tb drives *are* affordable. the 5400s have been priced at $109 on amazon. the 7200s have been at $150, directly in line with the 2tb 7200s offered at $100.

Its not just the money. I prefer 3 platters for the noise, power and reliability benefits.
 
1. How much louder is a 4 platter vs a 3? Are there any measurements that show at least a few dB difference when inside a typical case? I can see if you are trying to build a really quiet HTPC. But if you are norm just building a file server, the fans will dominate.

2. Is there any real proof that a 3 platter is significantly more reliable than a 4? I mean, if you go by Google's data, a colder HDD is actually less reliable. People can make all the correlatioons they want, but until there is solid causation with it...meh.

3. Furhtermore, drive statisics are based upon large populations. I doubt there is a large varience in the MTBF between a 3 and a 4 platter (not that I have been able to find on consumer drives). Even if that varience was 10%, a drive that can hold 33% more data, overall is still better.

4. As for saving power, are people telling me that a 4 platter consumes 33% more power than a 3 platter assuming the same density platter?

Slice it anyway you want, less drives will result in a reduced chance of failure overall. The only thing that is worse is the loss of data per failure is larger.
 
WOW! They mention a 1 platter 1TB drive with media rates of 228MB/s and the 5400 RPM drive 180MB/s. Those are quite fast.

Be careful what you believe. I suspect that StorageReview (or possibly Hitachi's PR wonk) has messed up by publishing a "media data transfer rate" in units of mega-BYTES-per-sec instead of the accepted mega-BITS-per-sec. My bet is that Hitachi engineering spec'd that number as 1816 Mbits/sec and some pseudo-techie performed a gratuitous divide-by-8.

There is a fair amount of "overhead" between the media rate and a for-real data transfer rate.

Bottom line: expect a "Typical sustained transfer rate [outer zone]" of ~180 MBytes/sec for the 7K1000.D (7200rpm) model. Also, I'll speculate that the 5K1000.B model is actually ~5700 rpm.

But ... I hope I'm wrong:).

-- UhClem
 
Be careful what you believe. I suspect that StorageReview (or possibly Hitachi's PR wonk) has messed up by publishing a "media data transfer rate" in units of mega-BYTES-per-sec instead of the accepted mega-BITS-per-sec. My bet is that Hitachi engineering spec'd that number as 1816 Mbits/sec and some pseudo-techie performed a gratuitous divide-by-8.

I actually thought about that when I saw such a high number..
 
Well it would make sense that higher density = higher throughput.

Think about it like this: pretend that the platter isn't moving at all, but the head is active. At a single coordinate, I think the heads are capable of a significant throughput. The reason we can't get that throughput is for one big reason: sectors and tracks. The second reason, though not as significant as the first, is because the platter is moving. The third reason is because even if the platter were still, you could get a higher throughput if you got rid of that medium of air between the head and platter and went straight to solid state where everything is one electronically cabled circuit.

Now, let's pretend there's no such thing as sectors and tracks. Let's just imagine that we have a spinning platter with a head that can move (but we're going to visualize it as just hovering over one particular spot). Pretending that the head is like a laser and we aren't bound to coordinate specific sectors and tracks, that head should still be able to read and write with huge throughputs... just probably not as great as if the platter weren't moving.

Now add in the fact that we have tracks and sectors that are at specific coordinates and axis. The head is limited to these, and this adds latency two-way. First way is that the head has to adjust itself to the correct track and move from track-to-track, and the second is the distance between sectors on a track (or from track-to-track).

So if we had a higher density, there is less latency, and we get closer to the picture of a platter not bound by tracks and sectors (but still moving).
 

I find it odd that two different manufacturers come out with 4TB drives within 24 hours of each other. It is almost like one of the companies had a 4TB drive for a bit and was simply waiting for the other company to catch up. Either that or they are both getting the 1TB platters from the same manufacturing plant.
 
I find it odd that two different manufacturers come out with 4TB drives within 24 hours of each other. It is almost like one of the companies had a 4TB drive for a bit and was simply waiting for the other company to catch up. Either that or they are both getting the 1TB platters from the same manufacturing plant.
That is the tendency in the real world. There are many products and technologies out there that aren't even accessible and known to consumers and the majority (if not all) enterprises and who knows, perhaps even governments (mere speculation), solely on the basis of turning a profit at the right time.

You simply don't let innovation and new technological findings enter the market when everything is grand and turning a pleasant profit. You move up the technological ladder (in terms of publicizing and making available to enterprises, governments, and consumers) slowly and gradually.

It is exactly like masturbating over the course of an hour using slow and gentle strokes. More profit!
 
Last edited:
Considering the 7K3000 has a media trasfer rate of 207MB/s yet only achieves 150MB/s, these new 7K1000.D will probably be ~165MB/s and the 5K1000.B ~130MB/s real-world.

In other news, Hitachi finally made the switch to 4K Advanced Format sectors for these 1TB platter Deskstar drives.
 
3tb drives *are* affordable. the 5400s have been priced at $109 on amazon. the 7200s have been at $150, directly in line with the 2tb 7200s offered at $100.

You're right, if you need only one or a couple, they're now affordable (I'm only referring to 5400rpm ones).

But the price per GB is 20% higher than for 2TB drives.
 
But the price per GB is 20% higher than for 2TB drives.

How do you figure? The Hitachi 3TB Coolsin is easily had for $110 while the 2TB drives average $70 sometimes a bit less but we don't count rebates.

Even if the 3TB was slighty more per GB on the average sale you are getter better throughput and more space per drive so it's worth it.

The 3TB drives have been worth it all summer.
 
Any word on whether these will be 4K sector drives or did Hitachi manage to eek a little more life out of 512b?
 
How do you figure? The Hitachi 3TB Coolsin is easily had for $110 while the 2TB drives average $70 sometimes a bit less but we don't count rebates.

Even if the 3TB was slighty more per GB on the average sale you are getter better throughput and more space per drive so it's worth it.

The 3TB drives have been worth it all summer.

I used a mix of your figures and mine, I shouldn't have. I'm in France and 2TB drives have been at 60€ for months, while the cheapest 3TB is 120€ (prices include VAT). Hitachi drives aren't widely available for some reason, and their prices are not competitive at all. In Germany where prices are often a little better and drop more quickly, it's not very different, 3TB drives at 110€ but 2TB at 55€. In the end, the GB is quite cheaper in 2TB drives, and I don't care much for throughput.
 
Metaluna: The rumor is that Hitachi has finally gone to 4K sectors with this drive.

I bought six Hitachi 3TB drives from amazon in retail boxes for $106 shipped each. At that price they were the same GB/$ as a $70 2TB drive. In my opinion, this makes 3TB drives (at least the 5700RPM ones) cheap. To get the same GB/$ ratio, 4TB drives would have to sell for $140 shipped which I don't think we are going to see for a long while.
 
Metaluna: The rumor is that Hitachi has finally gone to 4K sectors with this drive.

I bought six Hitachi 3TB drives from amazon in retail boxes for $106 shipped each. At that price they were the same GB/$ as a $70 2TB drive. In my opinion, this makes 3TB drives (at least the 5700RPM ones) cheap. To get the same GB/$ ratio, 4TB drives would have to sell for $140 shipped which I don't think we are going to see for a long while.

Port cost should be factored in as well...something often forgotten.
 
Considering the 7K3000 has a media trasfer rate of 207MB/s yet only achieves 150MB/s, these new 7K1000.D will probably be ~165MB/s and the 5K1000.B ~130MB/s real-world.
If your 7K3000 (2TB) gets a max transfer rate of only 150 MB/s, maybe you got a lemon. Hitachi's spec is 162, and my tests confirm that. I stand by my prediction of ~180 MB/s for the 7K1000.D (several posts before yours). And, I expect ~150 MB/s for the 5K1000.B.

Heck, the 5K3000 (2TB) is spec'd at 136, and my tests support that also. Again, this assumes you didn't get a lemon.

-- UhClem
 
I don't own a 7K3000, lemon or not. I was only going from memory of early benchmarks I'd seen online of the 3TB variety. I forgot that Hitachi actually had numbers published in their OEM specification guides, so you may be correct. Add the missing +8% and you do get ~180MB/s for the 7K1000.D and ~150MB/s for the 5K1000.B.
 
Last edited:
Back
Top