Why SATA power connectors?

[yankeeDDL]

Hello,

I just bought a new Thermaltake 430W power supply and notices the "SATA" connectors.
Since I bought SATA HD I used that as power supply, however, me being me, I then switched to the "standard" 4-pin connector (which can also be connected to the SATA HD).
I noticed no change whatsoever.
Now I wonder, what's the difference between the two connectors (apart from the shape, obviously :roll: )

Just curious.
Thanks.

 

[cowboytech]

an exerpt;

The Serial ATA standard also specifies a power connector sharply differing from the four-pin Molex connector used by PATA drives and many other computer components. Like the data cable, it is wafer based, but its wider 15-pin shape should prevent confusion between the two. The seemingly large number of pins are used to supply three different voltages if necessary — 3.3 V, 5 V, and 12 V. Each voltage is supplied by 3 pins ganged together (and 5 pins for ground). This is because the small pins cannot supply sufficent amps for some devices, so they are combined. One pin from each of the 3 voltages is also used for hotplugging. The same physical connections are used on 3.5-in. and 2.5-in. (notebook) hard disks.

 

[GTMan25]

The pins on a SATA power cable (and data cable?) are staggered. The idea is that the ground always connects first. This allows for hot-swapping (where the cables can connect/dis-connect while the power is on) without damaging the circuit boards or components.

 

[yankeeDDL]

Thnak you. That makes sense. Not that I would even hot swap the SATA drives of my PC.

I currently have attached both SATA drives to the SATA power lines, but I noticed that my power supply (Thermaltake 430W) has only one line with 2 SATA connectors.
I usually try to use as many power lines as possible (as opposed to connecting multiple devices with one power line) but I have no choice in this case.
I'm wondering is I should use the "standard" connector on one of the HD so that I could use a separate power line ...

 

[joex444]

That's an interesting theory. I assumed that since the PSU puts out a certain number of amps on each voltage rail, that it didn't matter which physical connector you used, it all went back to the same source. I've opened up a few PSUs and noticed the connectors actually went into a grid, with all the 3.3v lines in a column, all the grounds, all the 5v, 12v, etc. They then just made connections from this grid and formed the physical lines.

On the other hand, the lines do offer some resistance, which assuming it is ohmic, would mean that as you increase amperes, the voltage lost in the wire increases. Assuming the lines were made of aluminum, the resistivity is approximately 2.75E-8 (ohm-meters)...if you guess the length from inside the PSU to the end is 1m and the radius of the inner wire is 2mm, and you have a really large load -- say, 30 amps -- you're talking about a voltage loss across the wire of approx 0.039V. That would make your 12v rail 11.96V. That's not going to be a cause of anything bad. Of course, 30Amps would cause a higher temperature increase on that line, which could be bad... (FYI, 2.75E-8*1m/(pi*.002*.002) * 30A = 0.039V, adjust the numbers if you wish).

The pertinent benchmarks, of course, would be a full load test of the voltage on each rail with everything connected as evenly as possible versus everything connected to as few lines as possible. Temperature of the lines, PSU, and case would also be useful. Alas, that's a lot of work.

The SATA drives themselves work equally as well on a molex or SATA power connector, connect them as you wish, since you aren't hot swapping.

 

[blueeyesm]

That's an interesting theory. I assumed that since the PSU puts out a certain number of amps on each voltage rail, that it didn't matter which physical connector you used, it all went back to the same source. I've opened up a few PSUs and noticed the connectors actually went into a grid, with all the 3.3v lines in a column, all the grounds, all the 5v, 12v, etc. They then just made connections from this grid and formed the physical lines.

On the other hand, the lines do offer some resistance, which assuming it is ohmic, would mean that as you increase amperes, the voltage lost in the wire increases. Assuming the lines were made of aluminum, the resistivity is approximately 2.75E-8 (ohm-meters)...if you guess the length from inside the PSU to the end is 1m and the radius of the inner wire is 2mm, and you have a really large load -- say, 30 amps -- you're talking about a voltage loss across the wire of approx 0.039V. That would make your 12v rail 11.96V. That's not going to be a cause of anything bad. Of course, 30Amps would cause a higher temperature increase on that line, which could be bad... (FYI, 2.75E-8*1m/(pi*.002*.002) * 30A = 0.039V, adjust the numbers if you wish).

The pertinent benchmarks, of course, would be a full load test of the voltage on each rail with everything connected as evenly as possible versus everything connected to as few lines as possible. Temperature of the lines, PSU, and case would also be useful. Alas, that's a lot of work.

The SATA drives themselves work equally as well on a molex or SATA power connector, connect them as you wish, since you aren't hot swapping.

Heh - 'ohmic' ("Dude - he's being soooo ohmic about it all")

In other words, using all the feeds for device each, or using every connector per line won't make a significant difference, except in tidiness and airflow.

Some PSUs have the luxury of being able to remove those cables and have less clutter, then later add more as needed (which, as far as I'm concerned, should be standard).

 

[yankeeDDL]

Thanks, yes, indeed the number seems pretty small, however, my concerns were a bit more complex.
Voltage drop is one thing, but there what about parasitic signals due to activation/de-activation of a device?
It's not my case but let's say I hot-swap one of the two sata HD.
The capacitance related to that power cable change. I guess that's why the SATA connectors need to have the ground as the last cable to be disconnected, but couldn't this generate an impulse on the power line and affects other devices connected?

When I plug-in my laptop in the power outlet there's always a spark. What if, even if with smaller power/voltages, something like this happens on the other HD?

I know I'm being anal, but 500GB in Raid 0 are a lot to lose and I like to minimize my risks.

 

[GTMan25]

If you are worried about stuff like that then you are going too far. You could also worry about running a vacuum cleaner next to your computer or whether your neighbor is overloading the local circuit with his grow-op. The world is full of dirty power and good components are designed to be fairly robust.