Page 1 of 1

Mediacenter + PicoPSU

Posted: Wed Aug 11, 2010 11:27 pm
by Sr.Agaporni
Hi.

I am thinking in upgrading my mediacenter with several new parts. My main concern is about the PSU, which I have read is the key to low the power consumption. I am not sure about the capabilities of these kind of psu to manage with several HDD, so I ask for some help about this problem.

My computer has a Point of View mobo with Atom 330 + Nvidia 9400 (ION). I plan to use 1 x HDD 2.5" 1TB (main HDD with OS, software, etc...) + 3 x HDD 3.5" 2TB + 1 x BlueRay + 3.5" USB reader + USB Wifi. I use the computer as a media / data storage and as a mediacenter playing HD videos, photos, youtube, etc...

Do you think the picoPSU can handle all that without problems? I am not sure about the power to the HDD when the graphic card is playing MKV or any other HD videos which need hardware acceleration.

On the other hand, in case that no problem would appear. What model do you recommend me? I was thinking in 120w picoPSU, which should be more than enough. At this moment, the computer is at 67 watts when playing videos, and 50 watts when iddle (should improve with the picoPSU and new main 2.5" HDD)

Any help welcome.

Thanks!

Posted: Wed Aug 11, 2010 11:33 pm
by faugusztin
The main problem is not idle or load power usage, but the startup power usage. Each of your 3,5" hard drives will take ~2A on 12V rail to start up = 6A of 10A peak load (max 60 secs) spent just on hard drives. For WD20EADS it is 1.75A for example.

So you are right at the edge of what is acceptable for 120W picoPSU.

Posted: Wed Aug 11, 2010 11:43 pm
by Sr.Agaporni
faugusztin wrote:The main problem is not idle or load power usage, but the startup power usage. Each of your 3,5" hard drives will take ~2A on 12V rail to start up = 6A of 10A peak load (max 60 secs) spent just on hard drives. For WD20EADS it is 1.75A for example.

So you are right at the edge of what is acceptable for 120W picoPSU.
I have seen the specs of the Samsung Ecogreen, and it says it uses the 5V power. Does it mean something? The WD also appears to be able to get 12V or 5V power, and it seems more efficient using the 5V option, at specs it shows a 300mA need against the 2A of the Samung. I must doing something wrong, but that is what I read here:

http://www.wdc.com/en/products/products.asp?DriveID=773
http://www.samsung.com/global/business/ ... del_cd=513

So...

Posted: Wed Aug 11, 2010 11:57 pm
by markshu
faugusztin wrote:The main problem is not idle or load power usage, but the startup power usage. Each of your 3,5" hard drives will take ~2A on 12V rail to start up = 6A of 10A peak load (max 60 secs) spent just on hard drives. For WD20EADS it is 1.75A for example.

So you are right at the edge of what is acceptable for 120W picoPSU.
While the rated maximum inrush current draw of the WD20EADS is 1.75A/drive, the actual is usually well below this. WD has included soft-start routines in their green drive series which does a halfway decent job of limiting inrush to well below the maximum rating.

I checked this out with my old-but-still-accurate Fluke77 last week just out of curiosity, and hooked up a single WD20EADS to a test PS, and then a pair of them for a second round. I subjected each setup to a good 20+ power-up cycles. Each drive ramped up to a peak of around 1.05A @ 12VDC inrush before falling down to their normal average operating current. While this is nowhere even close to being definitive, it was remarkably consistant and I suspect that it will be fairly typical of the current generation of green drives. It certainly parallels the readings the Kill-O-Watt gives for my current WHS build on startup.

If I was still a hardware designer in the Industrial Controls marketplace, I would certainly always design to "worst-case spec". However, a one-off home-brew build can afford to design around the individual components at hand instead of worst-case spec, with the understanding that adding or changing components can invalidate an existing design if you play in the margins. <g>

-MWS-

Posted: Thu Aug 12, 2010 3:07 am
by Sr.Agaporni
Great, that means that WD are better HDD when computer is starting, but... What about the continuous use? Because the computer is going to be nearly 100% of the time ON, and WD are worst at idle and working consumption than the Samsung Ecogreen (all 2TB drives), at least, I read it at Tomshardware:

http://www.tomshardware.com/charts/2009 ... ks,50.html
http://www.tomshardware.com/charts/2009 ... ,1027.html

However, If I want to upgrade to more HDD with the same computer, using a PCIe SATA expansion card with four more HDD (similar to Samsung Ecogreen), It seems to me that the picoPSU won't be able to handle it. Am I wrong?

Posted: Thu Aug 12, 2010 3:39 am
by faugusztin
I don't get it how you got from those charts that WD is the worst - it's exactly the opposite. 3.51W for RE2-GP 2TB vs 4.10 for F2 EG 2TB, and btw, at wdc site you have the specs for power consumption.

higher peak load for HDD startup

Posted: Thu Aug 12, 2010 4:13 am
by charonme
some pico psu models can withstand higher load ("peak load") for a short period of time (~60s), consult the manual for your model.

Posted: Thu Aug 12, 2010 10:18 am
by fwki
You could start with one Winmate 130 watt supply and a 220 watt brick and add second Winmate once you add HDD's. That's how I am powering my HTPC. The 12v supply from the Winmate is higher quality than a pico psu in any case. See this link for a how-to:

viewtopic.php?t=55353&postdays=0&postorder=asc&start=0

Posted: Sun Aug 15, 2010 11:59 pm
by Sr.Agaporni
faugusztin wrote:I don't get it how you got from those charts that WD is the worst - it's exactly the opposite. 3.51W for RE2-GP 2TB vs 4.10 for F2 EG 2TB, and btw, at wdc site you have the specs for power consumption.
That version of WD cost nearly double than the samsung. Are not in the same league. The other WD (similar price to Samsung) need more watts in that chart.

Posted: Mon Aug 16, 2010 12:40 am
by Sr.Agaporni
fwki wrote:You could start with one Winmate 130 watt supply and a 220 watt brick and add second Winmate once you add HDD's. That's how I am powering my HTPC. The 12v supply from the Winmate is higher quality than a pico psu in any case. See this link for a how-to:

viewtopic.php?t=55353&postdays=0&postorder=asc&start=0
That is a nice idea! I didn't though in adding more power supplies. The problem is how would it impact in the power consumption of the hole system. After reading the review of the Winmate PSU, I think it is more interesting than the Pico PSU because of the 12V control. I am going to acquire one and test it. The problem would be with the AC/DC converter as far as I have understood.

Posted: Mon Aug 16, 2010 1:53 am
by lm
Forget picopsu and get the fanless seasonic that was just reviewed here. It has great efficiency at low loads and is a complete solution that will make your troubles disappear.

Posted: Mon Aug 16, 2010 4:37 am
by Sr.Agaporni
lm wrote:Forget picopsu and get the fanless seasonic that was just reviewed here. It has great efficiency at low loads and is a complete solution that will make your troubles disappear.
That is a nice PSU, but my system is 50w idle, and that is less than 20% of this PSU, so the efficiency would be ver low. At this moment, the top I have measured under "heavy" load (playing a 1080p video) is 65w. Do you really think that the Seasonic x400 would be the best choice? Even after upgrading the system with 4 more HDD, the load would be under 90W permanently, or that is what I think.

Posted: Mon Aug 16, 2010 5:50 am
by lm
The seasonic is 80.4% efficient at 42.6w DC load and already 84.5% efficient at 65.9w DC load. What numbers do you expect from your other choice?

And how many Watts would the difference save you? Must be less than 5w in any case.

Posted: Mon Aug 16, 2010 6:56 am
by neon joe
lm wrote:Forget picopsu and get the fanless seasonic that was just reviewed here. It has great efficiency at low loads and is a complete solution that will make your troubles disappear.
My $0.02...

I have to agree with lm; you're right on the edge of potentially having problems running that system with a pico psu.
The Seasonic should work very well for you.

Just out of curiousity, which case are you using for your media center?

Posted: Mon Aug 16, 2010 7:48 am
by MikeC
lm wrote:The seasonic is 80.4% efficient at 42.6w DC load and already 84.5% efficient at 65.9w DC load. What numbers do you expect from your other choice?

And how many Watts would the difference save you? Must be less than 5w in any case.
At 42.6W output, dif. between 80.4% and 85% is just under 3W. By 90W, the x400 is >87% (as per 80+ gold requirements).

Posted: Mon Aug 16, 2010 8:24 am
by Sr.Agaporni
Well, I think that 5w of difference is really noticeable when we are talking of 50w, because it is a 10% of the total. However, I must be mistaken about the numbers if you are right, I will re-check that. However, I am planning in buying the PSU with other components next week. Do you know where can I buy that PSU in Spain? Also any web with international shipping will be welcome, and it should get more expensive and that is against it.

At this moment I have a case MCE701, with touchscreen. But I am not very happy with it. I have had this case for two years but I plan to move to a Antec p180, which is supposed to be able to host 8 HDD plus a 5.25" blu-ray unit without problems, has good cooling and is as compact as mine (not as beautiful, but enough).

Posted: Mon Aug 16, 2010 8:59 am
by MikeC
Sr.Agaporni wrote:Well, I think that 5w of difference is really noticeable when we are talking of 50w, because it is a 10% of the total.
That's not an effective way of thinking about power/heat differences. Looking at only % differences, it's very easy to forget the reality: 5W is 5W -- never mind what the % difference is, 5W is the same amount of power and heat whether you're looking at 1000w, 50W, or 5W as reference points. 10% of 1000W, however, is 100W.... while 10% of 5W is 0.5W.

Stay with the W numbers and there's less possibility of confusion. In reality, 5W is a very small amount of power/heat. Over an hour, it is 0.005 kwh of energy. Over the course of a year, run 24/365, it represents 43.8 kwh. (As a reference point, an old fridge will use ~2000 kwh during the same period. A new high efficiency model might draw 600 kwh.)

as for efficiency numbers...

42.6w output at 80.4% efficiency means 52.9W AC input
42.6w output at 85% efficiency means 50.1W AC input

To get a 5W savings (47.9W instead of 52.9W input), efficiency would have to be 89%.

Posted: Mon Aug 16, 2010 10:31 pm
by Sr.Agaporni
MikeC wrote:
Sr.Agaporni wrote:Well, I think that 5w of difference is really noticeable when we are talking of 50w, because it is a 10% of the total.
That's not an effective way of thinking about power/heat differences. Looking at only % differences, it's very easy to forget the reality: 5W is 5W -- never mind what the % difference is, 5W is the same amount of power and heat whether you're looking at 1000w, 50W, or 5W as reference points. 10% of 1000W, however, is 100W.... while 10% of 5W is 0.5W.
I appreciate a lot your comments because I know your knowledge is above mine, but at this pont, that is your view, not mine. You can't compare different situations if you stay at relative numbers. You must proceed to compare absolute numbers to see the nature of the improvement. In my view, the 10% (even if it is "only" 5W) is the point of the question, because If I didn't matter it, I wouldn't be spending my time investigating and trying to make an "ideal" computer. Instead of that, I would buy the cheapest available to fit my needs.

However, at this moment, it seems impossible to get the Seasonic x400 at Spain, I haven't been able to find it on the common retailers. But I have found another PSU that seems to be nice choices, but it is hard to find reviews of some of them and the specs don't show efficiency or 80plus rating:

http://www.alternate.es/html/categoryLi ... =0&order=1

What do you think of them? (All the <400w). There are another Seasonic PSU, like the S12II with 330w, but I don't find information of it.

PD: I found a review on Anandtech, it seems to be a 85% efficiency PSU:
http://www.anandtech.com/show/2327/1