Slashdot Mirror
Serial ATA for Mini Hard Drives Planned
Lord_Slepnir writes "Cnet is reporting on a consortium of companies that wish to develop a Serial ATA hard drive interface for Miniature hard drives called CE-ATA. The goal of these new drives would be to cut power consumption and use smaller connectors, not to provide an increase in speed. 'The purpose is to design a new interface tailored to the consumer electronics and handheld gadget segment,' said Intel's principal engineer for CE-ATA, Knut Grimsrud. The consortium consists of Intel, Hitachi Global Storage Technologies, Marvell Semiconductor, Seagate Technology, and Toshiba America Information Systems."
Yes you missed FTA:a new drive interface for miniature hard drives. Such drives are often used to store data in handheld consumer electronics, including devices such as Apple Computer's iPod music player.
I assume that by "use smaller connectors", they meant that SATA is smaller than the connectors currently being used in mini hard drives. While power usage may be limited more by the drive itself, size may not be. Take a look at the currect standard for 1 inch hard drives. It needs 52 pins on it. SATA in contrast, only needs 8, plus whatever is needed for power.
eclecti.cc
Er, what do you mean? I switched to sata and it cut my load times by pretty much half globally - startup, games, everything. It's one of the most amazing upgrades I've done simply because there wasn't much of a downside to it. At ALL. With CPU there's heat and more fans, with my new gamer case case there was going out and having to buy longer cables, with sata there was just computery geeky goodness. Yum!
ACs are modded -6. I don't read you, I don't mod you, I don't see you. Don't like it? Don't be a coward.
Any distro based on the 2.6 kernel series will support SATA (Mandrake 10 (my reccomendation), SuSE 9.1, Fedora Core 2, etc.). I wonder if this will make it harder for people to port Linux to mini devices, it took a while just for normal SATA support in the kernel.
I have a box running Fedora Core 2 and a Silicon Image 3114 based S-ATA controller, works like a charm, no extra drivers necessary.
And I found http://www.linuxmafia.com/faq/Hardware/sata.html to be a very interesting read, helped me decide on the Dawicontrol DC-154 controller.
SATA *does* specify a standard for a power connector and location (most still have the standard ATX power connector). You could feed it through the same cable through the motherboard in theory, though that would increase the power draw of the drive controller significantly. This is mainly for the possibility sane backplanes (including longer ground pins than other pins for hot-swap capability.)
XML is like violence. If it doesn't solve the problem, use more.
The issue here is not the connection to the PC. You can still use USB/Firewire/etc.
This technology is black box to the end user. Unless you look up the design specs, all you will know is that works.
The issue is the internal connection between the device hardware and the storage medium. If I am designing a device that is supposed to be small, what do I design for? Remember that the storage interface control IC for embedded device will not vary much in price unless the standard is proprietary. It costs the same to imbed a chip with SATA control protocols as ATA. Probably cheaper since you need less I/O lines and smaller board footprint is possible...
If I have a space constraint in the device on the MB, which am I going to go with? A total of 12 or less pins (assuming power embedded--IDE 44 is just IDE 40 with embedded power lines) or 44 pins? Particularly if the 12 pin connector is faster?
Also consider that unless the header connects directly to the drive, you have to use a flat cable. Very prone to interference and complicates the internal geometry for assembly. Believe me, there is nothing worse than a small computer like device with a LOT of flat cables commecting the internal workings.
Before answering, I'll point out that 2.5" ATA laptop hard drives have run power through the same cable with data for years with no serious problems.
At this point in time, it would be a BAD idea to supply power to the drives through the motherboard. For one thing, the motherboards would have to be able to handle the extra amount of current flowing through them.
That's a non-issue. All motherboards have a power layer and ground layer. Each layer is solid except for circles around the pins to which they don't connect. Huge amounts of current can pass through there. You can pass 3 amps at 12 volts through a distance of 10 inches using about a 0.2" wide trace. A SATA hard drive takes about 1.1 amps when operating and seeking and about 3 amps of startup current (though that could be mitigated by sequenced start-ups to prevent all drives from maxing out at once).
I think they might have problems as is supplying current to the Graphics card and CPU.
The CPU current isn't the issue. It's the heat that causes problems. On the graphic cards, it's the bus spec and connector that's the limiting factor. If you make an AGP card that exceeds the bus spec, then there will have to be an external power connector, but that doesn't mean that providing power through the motherboard is technically infeasible from a design standpoint.
Also, if you did do that, you run into two other issues with power as well. Heat dissapation in the cable and intereference with the data lines.
Do SATA power cables get hot now? Of course not. They won't get hot if fed from the motherboard either. Noticeable heat in a power cable is a sign that the wire is WAY too small. As to interference, AC and pulsed DC can cause interference, not straight DC. Even then, twisted pair cabling would resolve it. It's also a non-issue. Seen interference problems on USB? It runs power through the cable.
Also, if the power cable through the motherboard goes bad, you have to get an all new motherboard.
As someone who has designed PC boards, I can tell you that it's not going to happen. That's like worrying that your soup spoons will "go bad."
Keep the power supply supplying power to everything directly, it cuts down on complications that can crop up.
More cables, more connectors, more routing problems, more expense. It adds complication.
Right that it isn't the same connector as the data, it is adjacent and a bit larger, and yes, the connection cycles are reduced in comparison with the typical SCSI hot swap connector. For the sake of short term compatibility, I still have to lean toward having the power connector optional between ATX and the adjacent connector for a transitional time.
As far as controller draw, I was mostly referring to controller cards more, which have to pull power from the PCI slots and as such have to be mindful of the power limitations motherboard manufacturers have in mind. For example, have you ever seen a SCSI controller card that, itself, pushes power through its connectors? In every case I've ever seen, hot swap scsi backplanes in systems pull their power from somewhere else (either direct from the power supply, or from a port on the motherboard that could also be used with a different cable to power IDE devices). The SCSI drives with SCAs are typically hard to use with non-backplane systems, and vice-versa. SATA drives can easily be used with or without a backplane, and that is a strength in the desktop market.
XML is like violence. If it doesn't solve the problem, use more.