Jump to content
SuperW2

2nd Power Supply?

12 posts in this topic Last Reply

Recommended Posts

I got a hand me down 500w Thermaltake Purepower 500 Power supply that I would like to add to my 550w Corsair PS in my UnRaid box... my Motherboard only has one primary PS plugin, so I was curious if there was something I needed to do to use the 2nd power supply. My 550 seems to power the system OK (16 drives), but It would make the cabling cleaner and neater with a 2nd PS.

 

Wade

Share this post


Link to post

I got a hand me down 500w Thermaltake Purepower 500 Power supply that I would like to add to my 550w Corsair PS in my UnRaid box... my Motherboard only has one primary PS plugin, so I was curious if there was something I needed to do to use the 2nd power supply. My 550 seems to power the system OK (16 drives), but It would make the cabling cleaner and neater with a 2nd PS.

 

Wade

 

If your primary PSU is running your system, why add another one?  It will use more electricity.  The largest draw of electricity is when the system is powered up and all drives spin up at once.  If your PSU can handle that surge, you are good to go.  PSUs also need a certain amount of power draw to work.  If you are hooking one PSU to drives only and all of the drives are spun down I'd be nervous that the PSU would have problems.  (Maybe someone else could comment here, I've never had a 2 PSU system).

 

If cabling is the issue, be creative.  I had a very messy cabling setup and would up making a long PSU power cable with multiple SATA connectors.  I pryed connectors off of several splitters I had to make this long cable (only certain connectors can be used in this way).  The case went from a rats nest to very clean inside.  I was a little worried that this was too much current for one cable, but have carefully monitored and it produces zero heat.  The wire is heavy gauage (modified modular cable from an 850watt PSU, not that thin crap used on many splitters).  I soldered and taped all the connections wires to ensure solid connections and no shorts.

 

10disksatapowercablewa1.jpg

 

Share this post


Link to post

Nice idea - but could someone with electronics background tell us if posses any problems? (for example feeding the last disk vs. feeding the first disk)

 

 

Share this post


Link to post

(for example feeding the last disk vs. feeding the first disk)

 

Electronics don't work like that... because they are in parrallel, the voltage to each of the drive will be the same (ignoring the very small difference in the resistance of the different length of wires in because in camparison to the much larger resistance in the disks, they are negligable).  The current needed by each disk would be additive.  And the total power would be the total current times the voltage.

 

Voltage is a measure of potential... it's like temperature, a larger temperature difference between two bodies and the larger the flow of thermal energy from one body to another... in voltage, the higher the potential between two connectors the more current will flow given the same resistance between.

 

And because voltage is "predefined" by the power supply (12 and 5 volts), you just need to make sure that the power supply can in fact provide enough current to one connector you plan on putting all those drive on.

 

The drives WILL draw the needed power, it is a question of if your power supply will be able to give it to them or not.

 

Matt

 

 

Share this post


Link to post

Nice idea - but could someone with electronics background tell us if posses any problems? (for example feeding the last disk vs. feeding the first disk)

That would be me...   

 

The major source of voltage drops in most servers is the resistance in the connectors, not the wire itself.

With a typical configuration, there might be one or two splitters involved between the power supply itself and the drive.   Each splitter would add a tiny voltage drop.  As long as good connections are made, there will be no problems.  One poor quality splitter, or one poor connection between splitters, and the connected disks will not work consistently.

 

Typical drives draw anywhere from 1.5 to 3 amperes when spinning up. This is their peak current draw.

The drives are typically fairly tolerant of their supply voltage.  Seagate says +/- 10% is OK. 

 

The effect of a poor connection will show when a drive is spun up from idle.  At that point, it draws anywhere from 2 to 3 amperes of current.    The set of 10 drives shown in bjp999's post will draw anywhere from 20 to 30 amperes for the first few seconds when all spun up at the same time.  This is  360 Watts. (10 drives * 3 amps *12 volts)  He obviously has a power supply that can supply that on a single rail.

 

A typical 4 pin molex connector has a contact resistance of < .005 ohms, but let's look at a really poor connection first. 

 

If we assume a really poor connection of .1 ohms there would be a voltage drop of 3 volts (30 amperes * .1 ohms) ...  This would be bad for the drives ability to spin up... also, the contact will be dissipating 90 watts of power (3 volts * 30 Amperes) as heat.  Ever grab a 100 watt light bulb after it has been on for a while... pretty hot, isn't it., and part of its energy is given off as light, not heat..  The disks would see 9 volts instead of 12... Odds are this is outside of their normal tolerance and they might not spin up reliably.

 

A poor, but much better contact resistance might be closer to .01 ohms.  Now we're seeing a voltage drop of .3 volts when 30 amperes is being passed through the connector  (.01 ohms * 30 amperes)   The contact will dissipate 9 watts of power as heat. It will still get warm, but for the few seconds needed to spin up, it will deal with it just fine.  The disks would see 11.7 volts instead of 12 volts, well within their 10% tolerance.

 

A good quality contact of .003 ohms will have a voltage drop across it of .09 volts. (30 amperes * .003 ohms) and the power dissipated will be 2.7 watts (.09 volts * 30 Amperes)  It will be able to handle the heating just fine, as it is only for a few seconds.

 

When daisy-chaining splitters, it does not take too many "good" connections in series to be the equivalent resistance of a "poor" connection.  Minimizing the number of splitters is probably not a bad idea, but for most power supplies, you will need to use some.

 

As already mentioned, the biggest current draw is when spinning up all the drives.  If your existing supply is dealing with this, then no need for a second.

Your 16 drive array will draw 576 watts (assuming 3 amperes current draw at spin-up per drive.)    It probably is sized perfectly for your array, and since most drives draw closer to 2 Amperes when spinning up, is running a bit below its full capacity.

 

As far as the "first" disk vs the "last" disk in bdp999's wiring... The only added resistance is that of the wire.  It looks like about a foot or so...

Assuming he used heavy 18 gauge wire, the resistance of 1 foot would be .00639 ohms.   Now, the voltage drops differently the further from the power supply, as there are fewer drives drawing current.  The last inch or so of wire is only powering 1 drive at approx 3 amps of peak current draw.

 

Doing the math... (per 10th of a foot of wire) we get:

27 amps * .000639 = .017253 volts

24 amps * .000639 = .015336 volts

21 amps * .000639 = .013419 volts

18 amps * .000639 = .011502 volts

15 amps * .000639 = .009585 volts

12 amps * .000639 = .007668 volts

9  amps * .000639 = .005751 volts

6  amps * .000639 = .003834 volts

3  amps * .000639 = .001917 volts

 

Adding up all the voltage drops (per tenth of a foot of wire) I get a total difference of 0.0862650 volts between the first drive and the last in bjp999's array.    If the first drive had exactly 12 volts applied to it, the last would get  11.91374 volts.  Well within the 10% tolerance the drives are designed for.

 

Joe L.

(Electronics was my first love, computers came *much* later... I built my first computer when I was in the 7th grade.)

Share this post


Link to post

my Motherboard only has one primary PS plugin, so I was curious if there was something I needed to do to use the 2nd power supply.

You can't star a PSU to work without starting the MB.

Unless you're using

200001-001_800x640.jpg

Share this post


Link to post

AHA. I knew the power supply would not stay active without some sort of dongle. At least now I understand what it is!

Thanks!

Share this post


Link to post

If you still don't have a warm and fuzzy feeling about Bjp999's wiring then run another molex connector off the other end of that harness and feed another power supply connector in there too. Make sure you use a different cable coming off the power supply though if you're going to that much trouble, basically adds a redundancy factor to the wiring.

 

Peter

 

Share this post


Link to post

This might be a stupid question, but if i remember well my Kirchhoff's circuit laws, with each Sata in this custom cable, the Amperage will increase by at least 2A

I know that bjp999 mentions something about it:

 

... I was a little worried that this was too much current for one cable, but have carefully monitored and it produces zero heat.  The wire is heavy gauage (modified modular cable from an 850watt PSU, not that thin crap used on many splitters)....

 

But in other posts in the site i have seen mentions of cables even melting when 2 or 3 sata drive were connected on a sata cable

(for example: http://lime-technology.com/forum/index.php?action=profile;u=557)

Share this post


Link to post

Molex PSU connector is rated at 12V@11A for 30deg C rise.

 

Specifications

Width    21 mm

Height    6 mm

Electrical    Yes

    Max. voltage    12 V

    Max. current    11 A/pin (30 °C rise)

Pins    4

 

18 awg wire is rated at somewhere around 10A for 12V DC for automotive usage (this figure varies between manufacturers). So we are running a system well beyond spec but only for a very short period of time and only occassionally. A more realistic figure is to assume 10 drives @ 8W = 80W or approx 50% of the design capacity for the connector and cable used.

 

As Joe stated 2-3A inrush current is only momentary so isnt a cause for concern for the cable or connector. In some industry sectors it would be an unacceptable risk. 

 

I would of used heat shrink, not tape, it tends to come undone when heat cycled leaving bare wires.

 

I have one molex connector feeding each 4 HDDs on each of my home made splitters.

Share this post


Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.