Disappointed with the picoPSU...

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Disappointed with the picoPSU...

Post by jones_r » Fri Aug 24, 2007 10:41 am

Got the picoPSU-120 today, expecting big changes, but received none.

I bought it with the 110W ACDC brick.

My computer consists of:
- Athlon XP2400 @ 933mhz (1.15v)
- 7VM400M-RZ Gigabyte motherboard.
- Gigabit lan card.
- 1 stick of 512mb memory.
- 1 40GB desktop hard drive.
- 2 fans with zalman fanmate.

Here are the measurements with my old 300W PSU (the one which came with the CalmPC case):

- PSU not connected to the computer: 6.63W.

- Computer at Off state: 11.3W.

- Computer at idle: 40-42W.

- Computer at 100% CPU usage (Orthos): 54.5W.


And now the picoPSU results (I expected dramatic changes in all states):

- PSU not connected to the computer: 13.5W (!!). When the picoPSU-120 is not connected to the 110W ACDC brick, the brick fluctuates between 11.1W and 13.5W.

- Computer at Off state: 15.7W

- Computer at idle: 38W

- Computer at 100% CPU usage (Orthos): 50W.

I simply do not understand. Some People with low power consumption systems, reported 25W reduction after using the picoPSU, so where is it ?

I also don't understand why the picoPSU takes MORE power for itself, than my regular PSU. The 13.5W that it takes to itself, all turns to heat, what a waste...

Arvo
Posts: 294
Joined: Sat Jun 10, 2006 1:30 pm
Location: Estonia, EU :)
Contact:

Post by Arvo » Fri Aug 24, 2007 11:27 am

I don't think that PicoPSU takes more power to itself or is less efficient. I think it is ACDC brick, which consumes about 12W. Your old PSU consumes 6W itself, thereby we can assume that brick consumes 6W more.

Looking at your previous threads you estimated old PSU efficiency about 65%, what makes real power usage of your system at full load about 0.65*55W=36W (I think this 65% includes PSU 6W).

Using PicoPSU and correcting it by ACDC brick power usage, you get PicoPSU efficiency about 36W/(50W-12W)=95% (+/- 3%, considering approximate numbers). Not bad at all, although this efficiency seems almost entirely to be eaten by ACDC brick.

I hope my thoughts were not complete nonsense.

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Fri Aug 24, 2007 12:34 pm

I tested two more PSUs with my system:

1. Cheap Chinese 500W PSU (the cheapest they come)

- PSU not connected to the computer: 13.5W (I had to short the green and black wires in order for it to draw power).

- Computer at Off state: 8W.

- Computer at idle: 44W.

- Computer at 100% CPU usage (Orthos): 56W.


2. Seasonic SS-550HT (an "80 Plus" PSU, one of the very best, which I use to power my main system)

- PSU not connected to the computer: 11.2W.

- Computer at Off state: 12.5W.

- Computer at idle: 43W.

- Computer at 100% CPU usage (Orthos): 54W.


Conclusions:
It seems that whether I use a cheap chinese PSU, or an expensive "80 Plus" Seasonic PSU, or the picoPSU, the results all fall ~5W of each other, which makes the differences meaningless.

I don't care if the Brick is to blame for the picoPSU's mediocre results, it's not like I can run the picoPSU without a brick, and the brick I used is the one that the picoPSU's manufacturer sold me!.

Something else, when Silentpcreview.com did the review of the picoPSU, they said, and I quote:
TEST RESULTS: picoPSU-120 from Mini-box, 120W brick:

Power Draw with no load applied: 4.7W
So how come I'm getting 15W ???, could it be that my brick is faulty ?, or maybe they gave Silentpcreview an unusually and unrepresentative good brick ?

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Fri Aug 24, 2007 2:48 pm

Heh.....you can run a PICO without a brick. You simply hook the PICO to a 12V car battery, attach a solar panel to your roof to charge the battery......works perfectly. I'm doing it with my PICO-powered P4-2.4. :D

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Fri Aug 24, 2007 3:43 pm

Heh.....you can run a PICO without a brick. You simply hook the PICO to a 12V car battery, attach a solar panel to your roof to charge the battery......works perfectly. I'm doing it with my PICO-powered P4-2.4.
Good idea, but I'm living in an apartment.

drees
Posts: 157
Joined: Thu Jul 13, 2006 10:59 pm

Post by drees » Sat Aug 25, 2007 1:43 pm

The key to getting good efficiency with the PicoPSU lies with the brick you use.

It's fairly obvious that your brick is not very efficient for whatever reason.

Is there any reason you didn't get an 80w brick instead? Can you tell us what brick you actually have?

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Sun Aug 26, 2007 12:26 am

The Pico is most commonly teamed with EDac bricks, either the 80 watt fanless, or the 120 watt screamer. I'm not familiar with the 110 watt one you speak of.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Sun Aug 26, 2007 12:56 am

Eunos wrote:The Pico is most commonly teamed with EDac bricks, either the 80 watt fanless, or the 120 watt screamer. I'm not familiar with the 110 watt one you speak of.
It's the same.
spcr review wrote:Mini-box supplied two power bricks to test alongside the picoPSU, one rated for 80W, the other for either 110W or 120W (depending on which numbers you believe).

ddrueding1
Posts: 419
Joined: Sun Sep 19, 2004 1:05 pm
Location: Palo Alto, CA

Post by ddrueding1 » Sun Aug 26, 2007 10:55 am

So we need someone to test the efficiency of power bricks? Great...

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Sun Aug 26, 2007 11:11 am

jones_r......you need to be more creative. If you cannot install a solar panel to charge a 12v battery, there are other ways. Put the battery in a plastic marine battery carrier. Take it out to your car, and hook it up to the car's 12V connection. You'll only need to run the car a short time to re-charge the battery, and you'll only need to do this every few days.

Virtually free 12v power for your PICO.......no power brick needed. :lol:

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Sun Aug 26, 2007 12:20 pm

This is the brick that I got:

http://www.mini-box.com/110w-12v-8-5A-A ... ategory=13

Imo, this is exactly the same brick that SPCR checked with the picoPSU, which consumed mere 4.5W (in comparison to 13.5W on mine!)

Bluefront,
Carrying this heavy battery to the car, every several days, is not something I'd like to do...

drees
Posts: 157
Joined: Thu Jul 13, 2006 10:59 pm

Post by drees » Sun Aug 26, 2007 12:39 pm

ddrueding1 wrote:So we need someone to test the efficiency of power bricks? Great...
That's one thing that has kept me from getting a PicoPSU - not knowing how efficient the various power bricks are that are available.

It would be nice to get some efficiency numbers for more of the bricks out there.

I am disappointed to see that there may be a fairly big variance in the efficiency of some of the common bricks.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Sun Aug 26, 2007 12:42 pm

Most people/stores who sell the picoPSU also sell the same power bricks reviewed by spcr.

Bluefront
*Lifetime Patron*
Posts: 5316
Joined: Sat Jan 18, 2003 2:19 pm
Location: St Louis (county) Missouri USA

Post by Bluefront » Sun Aug 26, 2007 2:37 pm

jones_r......you seem to be interested in saving a few watts here and there. I'm giving you a few ideas. It certainly is easier to plug in a PSU and flip a switch. But if you want to avoid a power brick, you'll need to find a 12V source. There are automotive batteries that are fairly small......don't weigh much. :?

ddrueding1
Posts: 419
Joined: Sun Sep 19, 2004 1:05 pm
Location: Palo Alto, CA

Post by ddrueding1 » Sun Aug 26, 2007 3:15 pm

I live about 15 minutes drive from mini-box.com. If they were interested in submitting devices for testing, I'm happy to test.

fraz
Posts: 45
Joined: Sat Aug 16, 2003 9:52 pm
Location: Australia

12v bateries and the PICOpsu

Post by fraz » Sun Aug 26, 2007 6:37 pm

As far as i know the PicoPsu does not regulate the 12v line. So if your battery is fully charged and putting out 12.6v your components get that too. Charging could be even more of a problem with the battery connected.

Anyone know the usual tolerances of computer components on the 12 v line?

drees
Posts: 157
Joined: Thu Jul 13, 2006 10:59 pm

Post by drees » Sun Aug 26, 2007 10:28 pm

ddrueding1 wrote:I live about 15 minutes drive from mini-box.com. If they were interested in submitting devices for testing, I'm happy to test.
The nice thing about testing 12v bricks is that it should be very easy to run tests since you only have to worry about a single load...

ddrueding1
Posts: 419
Joined: Sun Sep 19, 2004 1:05 pm
Location: Palo Alto, CA

Post by ddrueding1 » Sun Aug 26, 2007 11:14 pm

Single factor with a relatively low maximum load.

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Tue Aug 28, 2007 3:11 am

Can people with the 120W or 80W bricks, who bought them from Mini-Box, please report the power consumption of the brick ?

b3nbranch
*Lifetime Patron*
Posts: 130
Joined: Wed Aug 09, 2006 6:10 pm
Location: Columbus, Ohio
Contact:

Post by b3nbranch » Tue Aug 28, 2007 6:23 am

I replaced a Seasonic S12-380 with a picoPSU in the rig in my sig.
I'm using the mini-box 110watt brick.

AC draw as measured by a kill-a-watt:

computer off: S12-380 1 watt, picoPSU 3 watts
computer idle: S12-380 52 watts, picopsu 43 watts
cpu burn: S12-380 85 watts, picopsu 77 watts

Although the absolute watt difference is modest, it's an
18% drop at idle and 10% drop at load. I'm pleased with it.

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Tue Aug 28, 2007 11:44 am

b3nbranch,

When you connect the 110W brick to the Kill-A-Watt device, without connecting the brick even to the picoPSU, how much wattage does it take ?

Btw, what is the voltage in your outlet, 120V or 230V ?, I'm starting to think that maybe this brick is less efficient with 230V outlets...

b3nbranch
*Lifetime Patron*
Posts: 130
Joined: Wed Aug 09, 2006 6:10 pm
Location: Columbus, Ohio
Contact:

Post by b3nbranch » Tue Aug 28, 2007 12:39 pm

jones_r wrote:b3nbranch,

When you connect the 110W brick to the Kill-A-Watt device, without connecting the brick even to the picoPSU, how much wattage does it take ?

Btw, what is the voltage in your outlet, 120V or 230V ?, I'm starting to think that maybe this brick is less efficient with 230V outlets...
I'll test the unconnected brick the next time I power down. For now, I
can tell you that I'm using a 120V outlet.

ddrueding1
Posts: 419
Joined: Sun Sep 19, 2004 1:05 pm
Location: Palo Alto, CA

Post by ddrueding1 » Tue Aug 28, 2007 12:42 pm

I've probably bought 20 or so of the 80W bricks from mini-box, but they are deployed to client sites...not really practical to test them...

Elixer
Posts: 520
Joined: Wed Jun 02, 2004 10:31 am
Location: Las Cruces, NM
Contact:

Post by Elixer » Tue Aug 28, 2007 4:27 pm

jones_r wrote: So how come I'm getting 15W ???, could it be that my brick is faulty ?, or maybe they gave Silentpcreview an unusually and unrepresentative good brick ?
What are you using to measure power? I know many of those things have problems measuring well at low loads, especially when the power drawn is highly reactive. Many power supplies, including many power bricks, have extremely poor power factors at no/low loads, which power readers don't read properly. It's likely that Mike C's power brick is drawing exactly the same amount of power from the wall as yours (+-10%), but Mike C's power reading is giving better results.

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Wed Aug 29, 2007 4:00 am

What are you using to measure power? I know many of those things have problems measuring well at low loads, especially when the power drawn is highly reactive. Many power supplies, including many power bricks, have extremely poor power factors at no/low loads, which power readers don't read properly. It's likely that Mike C's power brick is drawing exactly the same amount of power from the wall as yours (+-10%), but Mike C's power reading is giving better results.
Not likely.

My computer without the picoPSU (with a cheap, inefficient PSU), takes 42W. My computer with the picoPSU takes 39W. These numbers support the conclusion that the picoPSU is to blame, and not the Watt meter.

My current assumption is that the picoPSU was designed for 120V outlets, and performs horribly with 230V outlets. I will check this assumption using a 230V to 120V converter.

a1bert
Posts: 4
Joined: Mon Aug 27, 2007 5:29 am
Location: .cz

Post by a1bert » Wed Aug 29, 2007 4:23 am

jones_r wrote: My current assumption is that the picoPSU was designed for 120V outlets, and performs horribly with 230V outlets. I will check this assumption using a 230V to 120V converter.
IMHO picoPSU has nothing to do with AC, it must be ACDC brick problem....

take some car 12V bulb (20W,50W etc), connect it to your ACDC adapter and measure with kill-a-watt (to be more accurate, measure current and voltage on a bulb)......

b3nbranch
*Lifetime Patron*
Posts: 130
Joined: Wed Aug 09, 2006 6:10 pm
Location: Columbus, Ohio
Contact:

Post by b3nbranch » Wed Aug 29, 2007 6:43 am

jones_r wrote:b3nbranch,

When you connect the 110W brick to the Kill-A-Watt device, without connecting the brick even to the picoPSU, how much wattage does it take ?
It flickers between 1 watt and 2 watts. I don't know if the kill-a-watt
does rounding or truncation 8)

jones_r
Posts: 140
Joined: Fri Sep 17, 2004 3:27 am

Post by jones_r » Wed Aug 29, 2007 8:20 am

IMHO picoPSU has nothing to do with AC, it must be ACDC brick problem....
I meant the brick of the picoPSU, of course.
take some car 12V bulb (20W,50W etc), connect it to your ACDC adapter and measure with kill-a-watt (to be more accurate, measure current and voltage on a bulb)......
Thanks, but I don't need to connect anything. As I said more than once, when *nothing* is connected to the brick, it takes 13.5W !!!
It flickers between 1 watt and 2 watts. I don't know if the kill-a-watt
does rounding or truncation
That's unbelievable, mine flickers between 13.5W and 11W.

The only question now, is whether I got a bad brick, or as I said before, the brick is not efficient when running from 230V outlets.

ryboto
Friend of SPCR
Posts: 1439
Joined: Tue Dec 14, 2004 4:06 pm
Location: New Hampshire, US
Contact:

Post by ryboto » Wed Aug 29, 2007 9:40 am

my Dell 220W/Pico PSU draws 7W from the wall when connected to a powered down PC. I haven't tested the brick itself for power draw with no load. By comparison my FSP Zen 400W draws 5W from the wall when the PC is off. I've yet to actually test the Pico with my updated system though, I'll have more info when I do.

mimwdv
Posts: 110
Joined: Mon Dec 27, 2004 3:54 pm
Location: Sydney, Australia

Post by mimwdv » Thu Aug 30, 2007 6:43 pm

I've got 3 of the 80W bricks (model EA10953A), and I'm on 240V. From my figures, it would appear 230V isn't your problem. All figures measured with a killawatt. All 3 of my bricks measured within a watt of each other.

Power spike on plugging in only the brick ~10W.
Brick only draw after spike 1~2W
[Edit] Brick and Pico, no PC 1~2W
Attached to hibernating HTPC 3~4W
HTPC idle 51W
HTPC load (playing Xvid plus orthos) 80W

I'm planning on some testing over the weekend with my server of Pico vs Seasonic - I'll post numbers here when I have them.

Post Reply