?
Solved

power supply problem - 12 v line

Posted on 2005-03-15
28
Medium Priority
?
488 Views
Last Modified: 2010-04-26
so i have two computers... one uses a 450 watt psu and one has a 220 watt psu
in the bios... the 220 watt psu runs at 11.91 amps

HOWEVER.. the 450 watt psu runs at 11.55 amps... and my graphics card brings a "not enough power" message intermittedly.... so...i just bought the power supply yesterday....
  should i take it back?  (its a 3 hour drive....)

  i am concerned cause the tech support guys at the company for the card (leadtek) said it should be very very close to 12 A on the 12 v line!  they said that 11.6 is way too low... and is something i should be concerned about on my whole system... not just my card.  are they right?  i dont know anything about this... but i am learning slowly... and i want to make sure i learn right.  i am running a geforce 6800 gt card by the way... its a hog for energy!

  i switched the power supplies... just to run in bios... and test... same result.  so... this psu is only churning out a 11.55 amp power on the +12v line.  

  also... is there something like... if you increase the watts... you also increase errors?  i mean... the one that runs nearly perfect to 12 A is a 220 W psu.  since this psu jumps to 450 W... isn't it possible with all that extra energy... some gets dissipitated in quality ? (amps)  i am asking cause people say... its better to go with lower watts and better quality... also... i know cache in system memory also works this way... the higher the energy output... the more chances you have to make mistakes... (look at the P4 HT EE 3.46 GHz cpu... so much power... and with all that cache... it constantly misses its memory location targets... err... thats another topic, sorry).... so what i need...
is someone tell me if i am going down the right path on this.... should i get a new power supply?
  what about a multimiter?  i dnt know what one looks like... never even seen it!  so how do i use one?  i was thinking it might help if i did a physical test of the amps.... any ideas on that.... and how to do it... details please... i dont know how to do this stuff!!!!

0
Comment
Question by:alenknight
[X]
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 9
  • 5
  • 4
  • +5
28 Comments
 
LVL 6

Expert Comment

by:tapkep
ID: 13543646
What brand your 450W PSU is?
Most likely, you need *better* (not higher rated) PSU. Try Antec, Chieftec, InWin PSUs.
0
 
LVL 14

Assisted Solution

by:tmj883
tmj883 earned 400 total points
ID: 13543783
The tech guys at Leadtek are correct. You need a minimum of 20A or more on the 12V rail for use with this video card...to calculate your system needs more closely:
http://www.adecy.com/psu/
Use the above tool not only for total watts needed, but attention to the individual power rails is also need...the demand on any given rail cannot exceed the supply.
For a 450 watt ps, very poor output in 12V...this is what is deceptive about only using the total watts as an indicator of performance...yes, I would get a new power supply...but this time use buy with more information...
a multimeter can be used to check voltage but is too limited to check amperes especially a high watt power supply. Overpower? Not a problem...Excess watts...good....will not hurt anything...a high quality power supply is a better choice but neither is any good if it does not meet your system power requirements.
T
0
 

Author Comment

by:alenknight
ID: 13544044
tmj883... just a quick question... ummm... is it possible to get 20 A????  i just go into the asus PROBE software... and it has 11.5... i assumed this is AMps... no?  cause its the same on both power supplies.  my bios also says 11.5... next to +12v .... 11.5 is the reading... i assumed this is amps... but maybe i am wrong... so....

 is 11.5 v too low ?
0
Veeam Disaster Recovery in Microsoft Azure

Veeam PN for Microsoft Azure is a FREE solution designed to simplify and automate the setup of a DR site in Microsoft Azure using lightweight software-defined networking. It reduces the complexity of VPN deployments and is designed for businesses of ALL sizes.

 

Author Comment

by:alenknight
ID: 13544050
and if so.... why is 11.9 not too low?
0
 
LVL 69

Assisted Solution

by:Callandor
Callandor earned 500 total points
ID: 13544149
I seem to recall mentioning to you in another question that the current on the +12v rail is what separates the good power supplies from the cheap ones.  I think your card is right at the edge of what that cheap power supply can provide, and it's actually providing a little under.  If you can afford a 6800, you should not skimp on the power supply.  Some good brands that can deliver the current are Enermax, Antec Truepower, PC Power & Cooling, Zalman, Thermaltake Purepower and now OCZ, I think.

Take a look at the specs of these:

Enermax 460W power supply $92.25
http://www.newegg.com/app/viewproductdesc.asp?description=17-103-488&DEPA=0

Antec 480W $90
http://www.newegg.com/app/ViewProductDesc.asp?description=17-103-914&depa=0
0
 
LVL 69

Expert Comment

by:Callandor
ID: 13544168
By the way, your BIOS readings are volts, not amps.  The fact that they're not close to the rated value is an indication of a cheap supply also.
0
 
LVL 14

Expert Comment

by:tmj883
ID: 13544179
You are reading voltages in the BIOS not Amperes. The rating of a power supply is on the sticker affixed to the unit itself. Peak/Constant Watts and the individual rails ampere ratings are on the sticker. By the way, the voltage is normal.
http://www.firingsquad.com/guides/power_supply/default.asp
More about choosing power supplies.
T
0
 

Author Comment

by:alenknight
ID: 13545390
callandor and tmj883 both helped me out on similar issues... but different problems (those have been resolved... thanks)
anyway....
tmj883... you are saying the voltage is normal???  so... you mean its normal to get 11.55 v on the 12V rail???  so... there shouldnt be any stability issues?  sorry... i am confused... if that is the case... then the psu is not the problem....

callandor... so... 11.55 is NOT close ?  i am asking cause i dont know... it sounds like its only 5% off... which doesnt sound crazy to me... but i seriously am just learning this stuff now! :)  11.91 on the other power supply is acceptable then?  

  just want to make sure, cause i did spend a good bunch of $$$ on this power supply... over $100... but obviously that didnt make a difference... yes... if you guys confirm my suspicions on the two questions i just commented on here... i guess the next step is to get a bigger name brand supply....

thank you
0
 

Author Comment

by:alenknight
ID: 13545515
also... i read about the 12v rails... and the Amps that should be on each 12V rail... the antec true power 550W brings in... 18A i think.... ... however... mine brings in 20A at 450W.... however... if 20A is not enough to power an athlon 64 3500+ cpu AND GEFORCE 6800 GT card ... why would an 18A be enough?  sorry... just... getting confused....
  also... i do put them on different rails... so... does that 18A mean PER rail?  or is it for the whole power supply altogether that 18A goes through (on the ANTEC model)?  if its on the whole... then i cant see how it will work... as i am using a 20A now... and that might be causing instability..... anyone know how much a athlon 64 cpu and geforce card takes up in power?  

sorry.... to confuse you guys... just i am soooo green on this stuff....
0
 
LVL 12

Accepted Solution

by:
GinEric earned 500 total points
ID: 13545600
Reading Amps would require you to work "in line" from the power supply to the rails with a multimeter.  Not so easy to do.  A clamp on ammeter can be used, but you'd have separate the 12V wire and clamp around it, also not so easy to do.

If the supply is rated in Peak/Constant Watts, the true wattage should be something like
.3535 times the rating, i.e., a 400 Watt supply is really about a 120 Watt RMS or True supply.  Same for the Amps, 12 is about actually 4 Amps.

Rails [voltage supply bars] on printed circuits have limits on current because of area, pinching, and skin effect.  They can only carry so much.  20 Amps is a lot to carry, no matter what the Voltage, so problem exists for a lot of designs.  Personally, I do think the Amperage, overall, is starting to get too high for small computers.

The better power supplies mentioned have beefier components, SCR's, Regulators, Darlington Pair Power Output Transistors, diodes, bridge, and transformer, which is why they last longer and can put out more instantaneous power on demand.  Cheaper ones skimp on these components, leading to failure and flakey operation.

Callandor's suggestions for good power supplies applies.

Just for reference, Watts are not Energy, Current [Amperage] is Energy.  From a typical stereo amplifier example, 1000 Watts is better than 500 Watts and as long as the speakers can take the current, you're better off as far as noise ratio.  Computers also have a noise ratio, though more technical, but it's a basic indication of increase in noise on any signal with an increase in current consumption.  So, 400 Watts is better than 200 Watts, providing the current is not allowed to get too high.  A hot running board can fall into this problem though with higher wattage supplies.  The thing is, with higher wattage you need more cooling to keep the heat from the amperage down.

The real problem, as you noted, is the memory requiring so much current.

Cache memory usually refers to the microprocessor on board cache memory, the system memory is RAM, or external memory.  It is the external memory [RAM, in any of its variations], and the higher energy consumption [current], that generates heat; the hotter it gets, the more errors it generates.

Try adding another fan if you're going to double the power supply specs.

0
 
LVL 69

Expert Comment

by:Callandor
ID: 13545879
> 11.55 is NOT close?

It's acceptable (withing 10%), but if you paid over $100 for it, I would have expected better performance.  Is it your +12v that is rated at 20a?  I would expect that to be sufficient, but I would say you shouldn't be getting this intermittent "not enough power" warning, so take it back.  It should outperform the 220 watt PSU.
0
 
LVL 13

Assisted Solution

by:Watzman
Watzman earned 400 total points
ID: 13547039
First of all, you are measuring volts, not amps.  It's impossible for you to easily measure the current being drawn (amps), but you want a supply that COULD supply 20 amps at 12 volts IF the circuits wanted that much current (20 amps).  The supply determines the voltage, the circuit (video card, motherboard, CPU, disk drives) determine the current, BUT there is a maximum current that the supply can provide.  You want a supply that COULD provide UP TO 20 amps (if not more).

[Water analogy:  Volts is pressure, amps is volume moved at that pressure.  A soda straw and a fire hose might both have the same pressure (voltage) and shoot a stream of water 10 feet, but there is a lot more water (amps) coming out of the fire hose.]

Ideally you would like to be within 5%.  At 12 volts, that's 11.4 volts as a minimum.  I'd say that 11.55 volts is acceptable, but it's on the low side of acceptable, and I wouldn't be completely comfortable with it.  At the same time, it's probably not a reason that things are not working; most circuits would work 10% low (10.8 volts) but that is clearly "too low" and you will have problems.

For a modern system with a high-power video card, you do want a PS that can supply at least 20 amps at +12 volts.

Note, however, that the reason that a voltage is low isn't necessarily because the current being drawn exceeds the ratings of the supply (although that can certainly cause this).  But there are tolerances in all of the parts that go into the regulator, and some power supplies are even adjustable (at the factory), which has a tolerance also.  There is a statistical distrubution of the voltage that finished products will produce, that is, of the voltage that you will get even when the current draw is WELL WITHIN the capacity of the supply.

And, finally, one other thing:  The problem might not be the supply at all.  Just as the supply has a tolerance, so does the measuring circuit that is telling you that it's only putting out 11.55 volts.  Before you conclude that you have a problem, one thing that you should do is get a resonable quality multimeter and independently measure the +12 volt output yourself (easy to do, it's the yellow wire of every disk drive connector).  The problem might not be the voltage, but rather the "voltage guage".  Maybe it really is 12.1 volts, but your motherboad is wrong, and is telling you (incorrectly) that it's 11.55 volts.  The chances that the "meter" is wrong are very bit as high as the chances that it's really low.

0
 
LVL 23

Expert Comment

by:sciwriter
ID: 13547167
This may have been said above, and I apologize if it has been -- but measuring an "amperage" (are you sure you don't mean voltage?) on a DC line requires you to have an in-line load tester.  Are you sure this is inline, and that the true current draw is 11+ amps?  I don't see how you could be drawing that much current with a normal PC setup, unless you have oodles of CD/DVD roms and HDDs attached.  Current can only be measured by the draw, not a static condition.  I apologize if you already know this, but many people confuse amperage with voltage and don't realize that amperage can only be measured inline when all of the devices are drawing power.
0
 
LVL 12

Expert Comment

by:GinEric
ID: 13547809
I'm sure he means Voltage.

There is a trick way to measure Amperage, but I don't know how good the programmers are that wrote the meter programs.

You can take a measure of Voltage, and a measure of the phase, and using the Power Factor method, calculate the Amperage.  Gee, trying to remember that is like digging for gold, but it's something like the Inverse Cotangent of something or other.  It's all done with matrices so it applies quickly to computer software, but all based in some form of linear differential equations, which I simply don't remember at this point.

But, it is a second hand measure, unlike the inline which is a direct measure.  The margin of error increases thereby.

Sure he mans Voltage though.
0
 
LVL 23

Assisted Solution

by:sciwriter
sciwriter earned 200 total points
ID: 13548014
Thank you GinEric -- that's what I suspected -- in which case, the VOLTAGE should approximate 12 volts.  Most PSs have logic built in to deliver constant 12V +/- 5-10% up to 10 amps or more -- therefore the values he is measuring are within tolerance specs.

The bottom line is this -- unless you have VERY specialized equipment, you cannot truly test a PS for current delivery.  You can DEBUG it, adding devices, to see when it fails, but you cannot really truly test its "reliability".

That is why you just replace them, and get on with life.  For $15 per PS -- you are just kidding yourself if you think that measuring the voltage out of the leads is any useful test at all.  It is what the devices draw that is crucial.

So forget trying to debug it an buy a new one.
0
 
LVL 5

Expert Comment

by:J-A-L
ID: 13548406
Your video card may also say it doesn't have enough power if it requires an additional power connector on the video card, and it is not plugged in.  My 6600GT for example will run without the extra video power connector... but just runs in a lower power mode and probably slower.

As for power supplies... you can try the Antec Neopower 480W with active PFC... very good supply.  Or Zalman makes good supplies and so on.  Spend $50 or more ;-)

Jeff
0
 

Author Comment

by:alenknight
ID: 13550220
J-A-L... i use a 2 molex adaptor that comes with the card to connect

sci, gin, and watz... thanks for your ideas... definitely gonna switch psu's though... as i plan to add A lot OF EXTRA hardware next month..... have to try ANTEC TRUE550... its the only one available in this area....

before i close this question... SPEAKING OF AREA.... i live in japan... the power here is 110v... coming from the wall... as opposed to 120 v from usa... but... i bought these tools in japan... however... they are all in english... so i am assuming the design specs were taken straight off the american versions....

so... my last question is... could the difference in voltage from the wall be affecting anything?

 ALL appliances run fine however... and it was never a problem... i mean... again... these were bought in japan... so they should be made to work at the local voltage.....
0
 

Author Comment

by:alenknight
ID: 13550286
one more thing... HOW DO I USE A MULTIMITER???

there was a quick reference to a yellow wire... but... no idea wich one... can someone give me a more detailed explanation?  i would like to get one for future use... might as well test all power systems i get as a benchmarking tool from now on... and i woudl like to learn.... how?
0
 
LVL 13

Expert Comment

by:Watzman
ID: 13550470
The wall voltage is not going to be a factor.  Switching power supplies are amazingly indifferent to their input voltage.  They bring it in, rectify it to DC, and do some very crude filtering with a couple of caps, and from then on, it's a DC-to-DC converter, and they don't much care what the input level is (it's not difficult to build a switcher that will work over a range of 80 volts to 260 volts).

Multimeter .... if you don't know how to use one, I'd really suggest a course in basic electricity and electronics, but for this purpose .... put the black lead to ground (the metal chassis of the computer case, also the black leads of disk drive connectors) and the red lead to the voltage to be measured, and have the multimeter on a "voltage" scale (of at least 15 volts if it's not auto-ranging) before you do this, and measure the voltage.  When you ask a question like "how to use a multimeter" and also switch volts and amps, it's a suggestion that, if you are working on computers, you would benefit (greatly) from even a very small amount of education in electricy & electronics.  Computers are electrical/electronic, after all.  BE VERY CAREFUL NOT TO SHORT THINGS (unintentionally touch things together) when making measurements.

0
 

Author Comment

by:alenknight
ID: 13550848
watz... i agree... i would do well to learn by a course in electronics... however... my background is in graphic design and web programming... particularly ASP and Flash environment... but... my old system crashed... wanted to build a new beefy one... so... I drove to the nearest town (i live in japan's mountain side... barely 1,000 people in my village....) and the closest town is hiroshima... a good 3-4 hour drive... and the computer shops there... carry only second-hand fudgy stuff.... or new low end stuff... so... i got the most expensive psu i could find... this one... by TERCO (japanese company i think) for $150 equivalent... and... well... when i set everything up... it didnt work.  so... in order to do my graphics work... i find myself learning as quickly as possible about PSU's... and yes... i am learning a lot!  more than i hope to ever know again....
  it is hard to replace things here... as i lose a total of 8 hours just to drive back and forth... so... before i do that... i need to make sure that i find exactly what the problem is....
  there is a small hardware shop in my little town... they dont sell volt-meters... but the owner has one he said i can borrow.... .... so thats why i asked how to use it.... he doesnt know either... at least not on a computer....

thanks for everyone's help... time to close this question... i think i sucked you all for what you know... thanks again for your patience... and comments...
0
 
LVL 13

Expert Comment

by:Watzman
ID: 13550930
You have my sympathies.
0
 
LVL 23

Expert Comment

by:sciwriter
ID: 13551066
obvious, by the points assignment, that this Q-er has no idea of what is going on.... Oh well.....that's life
0
 
LVL 14

Expert Comment

by:tmj883
ID: 13551828
Still following this thread...
Voltage specifications +/- 5% is considered tight specification. While loose spec is considered +/-10%. Tight spec is better but loose remains acceptable.
You have confirmed that your power supply has a 12V rail(s) with 20A available.
Try this:(from personal experience)prior to replacement...
Isolate the power to the video card...solo on a single power rail, no other device connected on that power cable to the video card, the error can be caused by interference from another device on the same power cable. The error is false as the amperage is available but will cause the video card to throttle down anyway.
Worth a try before having to loose all the travel time...T
0
 

Author Comment

by:alenknight
ID: 13554889
tmj883... yep... i did a multimeter test on the rail... teh graphics card is getting adequate power... HOWEVER... i tested the 12v lines... and sure enough... they are going at about 11.90 (a little tiny tiny bit low eh)... but... the connection to the motherboard.... is going at 11.40..... which is way low... however... my mobo is reading it as 11.60v... which is also way off.... so it could be a combination
  i borrowed a friends power supply... and... it is pushing the right amounts now... straight 11.95 to 12.00v... right into the motherboard... but... the motherboard is still reading it as 11.78... still low... but... not as bad.  could be the mobo is also wrong (like some people posted on here).. with this power supply... the fans also spin off the mobo... so... i will get a new power supply... and if it continues giving me trouble... must be the mobo... i took out the graphics card... and unplugged all connections... ust cpu, mobo 20pin and the 4pin .... and same results... bios reading it wrong.  (even with the new power supply)... so... it must have been an unlucky combination.  however... with my friend's supply... it seems to run much more stable.  

sciwriter... sorry you feel that way.  and yes... i dont know anything about this.  learning as i go... and awarding points based on effort.  I dont have the capacity to test who is right since it will take longer than acceptable (i will get into town to get a real power supply next month again....)... so... i have to give points based on who spent effort and what i feel sounds logical to my simple mind (again... dont know anything bout electronics).  my mistake i think was i should have made that clear in the post... i tend to do that with the hardware questions... as it takes an awfully long time until i can test things out.  will try to post it to the question so that people know what they are getting into.... again... sorry for the dilemma.
0
 
LVL 69

Expert Comment

by:Callandor
ID: 13555205
No problem, alenknight - we're all volunteers here, and the points don't let us buy anything.  Just treat it as a learning experience, because we all have things we can learn from here.
0
 
LVL 13

Expert Comment

by:Watzman
ID: 13555307
Inside the power supply, all of the 12 volt lines go to the same place -- there is, inside the power supply, only one 12 volt supply.  The only way you can measure 11.90 volts at one 12 volt connection and 11.40 volts at another connection is if there is tremendous current being drawn from the 11.40 volt connection (causing a voltage drop of 0.50 volts in the wire carrying that current).  Even 11.40 volts is (barely) within (or at) a 5% tolerance, and should be ok.

If the connection to the motherboard is through a "string" of output connectors, then put the motherboard on the FIRST connector, with as little wire as possible between the motherboard and the power supply.  The real fix here is to use a heavier wire, but that's impractical.  Also, if your motherboard has multiple 12 volt connectors (some have both the square ATX12V connector and a standard hard drive connector), you might use them both, to provide more "wire" and a lower voltage drop in the wiring.
0
 

Author Comment

by:alenknight
ID: 13561415
watzman... that is an interesting point... about the distance of the wires... i never new... that having the first connector makes a difference as opposed to the last connector... i understand... your saying that it SHOULD NOT... but... hey.... if i can... doesn't hurt to try eh?  thanks for that info.... make sure i do that with the new power supply....

 
0
 
LVL 13

Expert Comment

by:Watzman
ID: 13561702
If you have one of these "octopus" strings with 3 to 6 connectors on them, you ideally want to place the major load(s) relatively close to the power supply, all of the loads will be better off by doing that.  There is a voltage drop caused by current flow through the resistance of the wire.  There are two issues at work here, the amount of current flow and the resistance of the wire.  Relatively, the current drawn by the motherboard is likely to be much larger than the current drawn by the other individual loads (typically disk drives and fans).  So if you have a choice (and recognizing that quite often you don't), put the major loads first.  The whole purpose of the square ATX12V connector is to provide a dedicated line for the 12 volts used by the motherboard, and to lower the resistance (and voltage drop) by providing TWO wires.

[No, I'm not saying that it should not make a difference, in fact, if the wires are "too small" one would expect that it would very much make a difference.]
0

Featured Post

Bringing Advanced Authentication to the SMB Market

WatchGuard announces the acquisition of advanced authentication provider, Datablink, with one mission – to bring secure authentication to SMB, mid-market, and distributed enterprises with a cloud-based solution, ideal for resale via their established channel & MSSP community.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

This is about my first experience with programming Arduino.
In the modern office, employees tend to move around the workplace a lot more freely. Conferences, collaborative groups, flexible seating and working from home require a new level of mobility. Technology has not only changed the behavior and the expe…
Michael from AdRem Software outlines event notifications and Automatic Corrective Actions in network monitoring. Automatic Corrective Actions are scripts, which can automatically run upon discovery of a certain undesirable condition in your network.…
Have you created a query with information for a calendar? ... and then, abra-cadabra, the calendar is done?! I am going to show you how to make that happen. Visualize your data!  ... really see it To use the code to create a calendar from a q…

777 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question