What technological improvements do you think we will see next on consumer 2.5" HDD?

[cbn]

Besides increased areal density here is what I think is fairly likely:

1. Lower RPM (comparing spec sheets on the 2.5" Barracuda and 2.5" Barracuda Pro) I estimate 4200 rpm would lower idle power from 450 mW to 300 mW and ~3300 rpm would be ~200 mW.

2. Helium to increase number of platters in 7mm thick drive from 2 to 3 while lowering active and idle power at any given RPM by ~20% (see mechanical savings in picture below as a reference point).

I also wonder if it could also reduce failures due to humidity if the disk controller is also enclosed in helium.

https://www.zdnet.com/article/heat-doesnt-kill-hard-drives-heres-what-does/

High relative humidity causes disk failures largely due to controller/adapter malfunction.

Disk controller/connectivity failures are greatest during high relative humidity.

High temperatures are not harmless, but are much less significant than other factors.

3. Multi-actuators. 1.) Even at low RPM sequential performance should be much better than any 7200 rpm single actuator 2.5" drive. 2.) Sequential performance per watt at the edge of the platters at 5400 rpm with three actuators (hypothetically in air, not helium) should be better than a typical SATA SSD based on the 2.5" Seagate Barracuda spec of 1.7 watt 140 MB/s read vs. 129.8 MB/s Sequential read performance per watt of a Crucial MX500. (reason: 3 actuators at 140 MB/s = 420 MB/s which at 1.7 watts* would be 247 MB/s per watt. 247 MB/s per watt is 90% better performance per watt than the MX500's 129.8 MB/s per watt)

EDIT: 1TB MX500 Sequential Read performance per watt is better than the 500GB MX500 at 148.2 MB/s vs. 129.8 MB/s. Triple actuators (5400 rpm, air) at 247 MB/s Sequential Read/watt works out to have 67% better performance per watt than 1TB MX500.

*This assuming the controller for a 3 actuator HDD doesn't use any more power than a single actuator controller. However, with that noted, even if the triple actuator controller did use more power I expect Helium to be used and the RPM of a triple actuator 2.5" HDD to be lower than 5400 rpm....both of which should separately increase performance per watt.

4. Larger NAND cache for SSHD versions.

 

[nosirrahx]

SSHD is an interesting case. I am not sure that SSDs won't just take over the market for them. With huge SSDs getting cheaper and multiple actuators potentially making the drives themselves much faster there may be no SSHD market at all in a few years.

An Optane like cache and controller on a SSHD could make for a compelling product but the prices would need to come down a lot.

 

[cbn]

An Optane like cache and controller on a SSHD could make for a compelling product

I think that is a great idea.

As a reference point here is what a SATA Express SSD (not SSHD) looks like installed in a laptop:

https://www.mobile01.com/topicdetail.php?f=240&t=5036019

Power, PCIe x2 and up to 2 x SATA Gbps go through a connector the same size as regular SATA 6 Gbps data and power connector:

which is a subset of the U.2 (SFF-8639) data and power connector:

So an Optane SSHD based on SATA Express could potentially be the same size as a 7mm thick SATA HDD or SSD which is, of course, smaller than 7mm thick HDD or SSD + separate M.2 2280 Optane or M.2 2242 Optane.

 

[cbn]

SSHD is an interesting case. I am not sure that SSDs won't just take over the market for them. With huge SSDs getting cheaper and multiple actuators potentially making the drives themselves much faster there may be no SSHD market at all in a few years.

Two advantages I see with SSD vs. SSHD:

1. Capacity. SATA SSD can go as high as 8TB with current controllers (eg, Phison S12 and the Marvell controller used in the 7680GB 7mm thick Micron 5100 ECO). 2.5" SSHD capacity (currently 2TB maximum for 7mm thickness) is limited by the number and capacity of platters due to lack of Helium and HAMR or MAMR. With this noted, I wonder if having (much) lower RPM would allow the HDD companies an increase areal density and/or increase platters separately of Helium and HAMR/MAMR.

2. Idle power consumption. Crucial MX500 500GB (used as an example in the opening post) has idle of 79 mW according to Anandtech...with other SATA SSD having even lower idle.

https://www.anandtech.com/show/12263/the-crucial-mx500-500gb-review/8

With that noted, I extrapolated using 700mW idle for 7200 rpm 2.5" HDD and 450 mW idle for 5400 rpm 2.5" HDD that it would take 1725 rpm for an 2.5" HDD to hit 75mW idle.

While 1725 rpm sounds really low I think with multi-actuators such a low RPM HDD would be plenty fast in Sequential (particularly if the very low RPM allowed super thin platters to be used....maybe even allowing 4 platters (and thus 2 or 4 actuators) in a 7mm thick enclosure with Helium). IOPs, of course, will be much lower because of the very low RPM (re: the increased rotational latency will dominate over seek time for access time) but that is where the Solid state part of the SSHD comes in to help out.

 

[hojnikb]

I only wish capacity/density goes up (especially on 2.5" formfactor) and prices go down.

 

[nosirrahx]

I only wish capacity/density goes up (especially on 2.5" formfactor) and prices go down.

I think in 2.5 inch form factor SSDs have already reached 8TB, that was an option on my Eurocom laptop.

2.5 inch HDDs have a lot of work to do if they want to catch 8TB but at least the price will be reasonable.

 

[rsutoratosu]

with ssd so cheap, I doubt much can be done for the mechanical part except to increase storage size that isnt viable for ssd yet.. i dont even see much 2.5 drive space in newer laptops anymore, they either move to smaller m.2 or full size.. no room for sata drives which is sad, I have a laptop with m.2 boot and a 2tb 2.5 for storing data.. those are slowly being phase out

 

[cbn]

In the 2.5" Seagate Barracuda Pro thread I wondered about the potential of the little 7200 rpm drive for recording DNxHR HQX at 4K30---> https://forums.anandtech.com/thread...-rpm-with-7mm-z-height.2539964/#post-39340008

.....but then looking at the write speed for the 2.5" 2TB Barracuda (slower at 5400 rpm but using the same 1TB SMR platter) as a reference point I reckoned 7200 rpm will still not be enough to handle the 104 MB/s 4K30 DNxHR HQX write requirement. This even for the short stroked 500GB version after factoring the need for consistency in write speed---> https://forums.anandtech.com/thread...-rpm-with-7mm-z-height.2539964/#post-39340856

With that noted, triple actuators would change that as the write speed would 3x faster at any given RPM. (EDIT: It might be that double actuators at 7200 rpm also works....even at the inner most part of the platter)

 

[cbn]

Did some research on Seagate Barracuda SATA SSD and according to this review it uses Phison S10 and 64L Toshiba 3D TLC......so my guess is that a 2.5" Seagate Firecuda refresh will use Phison S11 (a very good DRAM-less SATA controller with 32MB SRAM and strong 4K QD1 Read) and Toshiba 3D TLC (maybe even a multi-bit version of Toshiba XL NAND at a later point*).

*Would be very interesting to see a SSHD with maximized 4K QD1 read for legacy applications.

 

[cbn]

With consoles (including the very latest ones) supporting 2.5" x 9.5mm maybe 9.5mm makes a comeback?

(Normally I wouldn't mention 9.5mm.....but multi-actuator would strongly benefit from it. Reason: even with air they could get 3 platters in a 9.5mm and four with Helium.)

 

[nosirrahx]

With consoles (including the very latest ones) supporting 2.5" x 9.5mm maybe 9.5mm makes a comeback?

(Normally I wouldn't mention 9.5mm.....but multi-actuator would strongly benefit from it. Reason: even with air they could get 3 platters in a 9.5mm and four with Helium.)

In laptops, no chance, 2.5 inch drives in general will eventually get phased out on most laptops.

Drive form factor directly competes for space with batteries and drive thickness directly effects the thickness of the entire device.

 

[cbn]

Drive form factor directly competes for space with batteries and drive thickness directly effects the thickness of the entire device.

That is true, but by making the drive 9.5mm vs. 7mm the HDD manufacturrer can put another platter in there.....which also means another actuator.

3 actuators vs. 2 actuators (or 4 actuators vs. 3 actuators) means the 2.5" HDD can spin at lower RPM while still having more IOPs and more Sequential.

So while the drive is thicker (and thus reduce battery capacity all things being equal) it would save on power.

 

[VirtualLarry]

3 actuators vs. 2 actuators (or 4 actuators vs. 3 actuators) means the 2.5" HDD can spin at lower RPM while still having more IOPs and more Sequential.

And when the laptop experiences G-forces from falling, that's four heads that need power from the platter spindle motor to retract / rampload, and not just one.

 

[cbn]

And when the laptop experiences G-forces from falling, that's four heads that need power from the platter spindle motor to retract / rampload, and not just one.

Virtual Larry do you mean 8 heads? Because 4 heads is already happening in existing 7mm 2.5" HDD like this one (ie, single actuator and two 1TB SMR platters).

In any event, the energy required to park the heads (via magnet) will be greater anytime the mass of the actuator arms and heads is greater. (This mass is independent of the number of actuators involved* but can be affected by drive thickness because a thicker drive means more heads and actuator arms are possible.)

*assuming the actuator arms and heads in a multi-actuator design weight the same as a the actuator arms and heads in single actuator design.

 

[cbn]

Here is a history of the 2.5" 7mm HDD:

2010= First 2.5" 7mm HDD. This one had only 1 platter. At the same time (2010) I noticed that 9.5mm had 2 platters and 12.5mm had 3 platters.

2015= First 2.5" 7mm HDD with 2 platters---> https://www.kitguru.net/components/...duces-unique-2tb-2-5-inch7mm-hard-disk-drive/

The new 2TB HDD from Seagate comes in 2.5” form-factor and is only 7mm thick. The drive is based on two 1TB SMR platters with industry-leading areal density as well as new high-spatial efficiency mechanics, including new head, preamp, channel and other components. The usage of new mechanics allowed Seagate to pack two platters into a 7mm drive and reduce weight of the HDD to 3.17 oz. The drive uses Serial ATA interface, but Seagate does not disclose its spindle speed, cache capacity and so on.

Using the figure of .381mm planned for 3.5" glass platters as a reference point, could it be up to four thin glass 2.5" platters are possible in a 2.5" 7mm chassis? This assuming lower (or much lower) RPM............A RPM low enough that the four platters could even work in air?

My guess is that the current limiting factor to putting more platters in 7mm 2.5" is the space needed between the actuator arms rather than actual platter thickness. Lowering RPM and using Helium helps decrease this distance without needing to decrease platter thickness.

However, I do wonder at what point they would have to consider decreasing shock resistance resulting from having to use thinner platters? (eg, from hard drive being dropped) Not from the heads hitting the platters......But rather the platters being so thin that even with the heads parked there is increased risk of shattering during impact? Hopefully by the time this happens they have a new platter material which can be thinner but still maintain the necessary shock resistance.