Talk:SATA

Inaccuracy in article

Speed

The article contains this passage:

"Users with a SATA 1.5 Gbit/s motherboard with one of the listed chipsets should either buy an ordinary SATA 1.5 Gbit/s hard disk, buy a SATA 3 Gbit/s hard disk with the user-accessible jumper, or buy a PCI or PCI-E card to add full SATA 3 Gbit/s capability and compatibility."

You cannot truly be 3 Gb/s capable with a PCI slot since it only is capable of 133MB/s. —Preceding unsigned comment added by 151.197.29.186 (talk) 21:29, 20 September 2008 (UTC)

Read your own terminogloy. Gb/s MB/s the b and B are different. Another factor to remember the spindle speed of the disk they havea maximum speed anyway, i think it might be refering to ssd which is nearing the 3GB/s speed--Andrewcrawford (talk) 19:44, 25 September 2008 (UTC)

133 MB/sec would still only be a bit less than 1.1 Gb/sec. However, "PCI" could include "PCI-X" which can of course go faster. As for spindle speed, it does not put a limit on transfer rate unless you assume the bits/inch on the tracks never increases with later technology... and of course it has. Jeh (talk) 22:44, 25 September 2008 (UTC)

Spindle speed does put a limit on the disk it cn only read at teh speed of the disk which si effect by spindle speed.--Andrewcrawford (talk) 14:38, 12 November 2008 (UTC)

I have no idea what you're trying to say. But it's quite clear what Jeh is trying to say, and he/she is correct. Spindle speed doesn't really put a limit on the transfer rate since as platter density increases speed also increases which is what we see all the time (we've had 7200 RPM drives for a very long time and they've increased in speed a lot). Of course there is an upper limit. Clearly bit's can't be smaller then an atom wide. And in reality, the limit is likely to be a lot higher then that because of other effects. But whatever the limit it's definitely higher then 1.5 Gb/sec for a 7200 RPM hard drive since modern hard drives are nearly meeting that. I'm confident enough to say it's higher then 3.0 Gb/sec. It is of course true that until recently, the transfer rate of a single hard drive didn't reach the limit of PATA or an unshared PCI channel but what matters there is the maximum sustained transfer rate (which is a function of spindle speed, platter density and platter size/location), not simply the spindle speed Nil Einne (talk) 16:15, 22 January 2009 (UTC)

Read a few raw drive specifications. Spindle speed is still king for sustained reads and writes with most drives falling far short of 750Mbps sustained. That is why all modern drives have buffer RAM. The 3GBps is in fact burst speed which with similar buffering a PCI bus board could handle. However, most chipsets no longer optimize for PCI bus master burst cycles and exchanges so you would choke heavily at the at the PCI bus bridge which does not buffer for more than a few hundred bytes in a burst and cycles only in terms of scores and hundreds of PCI clocks. This buffering is not directly part of the PCI bus spec although there are certainly changing chipset design suggestions over time. That is there is no longer a true direct DMA path for PCI bus to main memory or any other path that is not severely restricted by interface chip design. If there was it would choke CPU and memory performance to a standstill. PCI is a dying backward compatible auxiliary bus which does not even connect at its own full potential to the rest of the system anymore.

However, a 1.5Gbps motherboard probably doesn't need to worry about full 3.0 throughput in that sustained speed based on spindle and data density of tracks have not increased some much in the last 2-3 years as to even fill 1.5Gbps with sustained data. It will merely take a bit longer to send large file data in slower bursts. Small files may not even show a difference due to rotational latency unless the application and OS are really above average in read ahead prediction.

69.23.124.142 (talk) 08:28, 24 March 2009 (UTC)

Cost

As far as inaccuracies go...the implication that the cost savings is in the number of pins and cabling and connectors is a farce. 12 cents per foot PATA ribbon cable is not that much more expensive than 4-5 cents per foot SATA and the connectors are pretty similar in price. Some savings is made on chips due to removal of serial to parallel data convertors and all those line drivers probably 0.50 to 2.00 USD per drive. But basically much of the savings comes in drive testing of only a handful of data and control lines versus a score plus the labor in assembling and inspecting complete systems. 69.23.124.142 (talk) 08:48, 24 March 2009 (UTC)

3.3V and hotplugging

The article says "However, without 3.3 V power, the SATA device may not be able to implement hotplugging as mentioned in the previous paragraph." but does not give any citation, nor does it explain why 3.3V power may be important for hotplugging. Does anyone have any more information? Plugwash (talk) 00:37, 12 February 2009 (UTC)

No idea. Sounds like B.S. to me; each supply voltage has one early-mate pin. Deleted. 71.41.210.146 (talk) 11:50, 10 March 2009 (UTC)

don't need a pic

you dont need the pic showing hdmi in the eSATA part! change it to just the SATA and take of the sky comment. —Preceding unsigned comment added by 84.252.231.222 (talk) 13:20, 5 March 2009 (UTC)

10G Fibre Channel

The comparison table includes 10G Fibre Channel which is used solely for inter-switch links (ISLs) in a fabric. There is no such thing as a 10G FC HBA...as far as I know. It may be more appropriate to include 8G Fibre Channel. It's a bit pendantic I know, but I though I'd mention it anyways. Wxb2744 (talk) 11:48, 10 March 2009 (UTC)

You are correct, FC is currently only at 8GB/s, but you can use 10 gig fiber (and copper) to connect to a server, NAS, or iSCSI device. I think it'd be more approptiate to change Fiber Channel to 10 Gb Ethernet. —Preceding unsigned comment added by 67.175.134.167 (talk) 16:24, 27 April 2009 (UTC)

photo quality

I don't usually split hairs re photo quality, tonal range, etc. for photos intended as illustrations rather than art as long as they get the job done, but when a photo apparently intended to illustrate the differences between two connectors has insufficient shadow detail to make both those connectors visible it's not getting the job done. 68.80.95.189 (talk) 16:07, 13 March 2009 (UTC)

Which image(s) are you talking about? If you mean the one comparing SATA and eSATA connectors, the differences are quite visible on my monitor. — Aluvus t/c 17:45, 13 March 2009 (UTC)

Master - Slave

The page is missing any info on the Master - Slave drives. Is this done on the Motherboard by ports or by Hard Drive jummpers? Telecine Guy 18:52, 4 April 2009 (UTC)

There is no such thing as a master or slave drive under SATA, unlike "parallel" ATA. It would be a good idea for the article to note that fact. — Aluvus t/c 21:27, 4 April 2009 (UTC)

PATA

Can we add a misnomer paragraph to cover the misuse of "PATA" please? After reading the term "PATA" over and over again in this article, I've got to comment. ATA is by its very nature parallel, there is no need to add an extra P on the front to denote that it isn't SATA. I realise that many computing amateur refer to ATA as PATA now, but that doesn't make it correct. —Preceding unsigned comment added by 81.107.183.201 (talk) 07:13, 29 April 2009 (UTC)

Accepted Term

As I understand it, with the advent of SATA technology, "PATA" has become the accepted term for what we used to call "ATA." Maverick Solutions (talk) 15:11, 22 May 2009 (UTC)

Splitting into a new topic of eSATA

Agreed.

71.16.143.71 (talk) 17:39, 14 May 2009 (UTC)

esata/usb combo

i recently bought a laptop. it has a combo esata/usb port. you can plugin either usb or esata. should i include this somewhere in the article. if yes, where? Perryizgr8 (talk) 13:05, 19 May 2009 (UTC)

SATA Power Connector Photo

I remove the following writer's comment from within the page and moved it here to be addressed: "need better photo showing , index pin, this one is bad. (sorry)" It referred to the 15-pin SATA Power Connector Photo. Maverick Solutions (talk) 15:08, 22 May 2009 (UTC)