Why 12v reading only 11.84 ??!!

S

Super KW

Limp Gawd
Joined
Feb 12, 2006
Messages
137
hi

i just got my new silverstone 750w with quad 12v 18a each (60A combined) and i see in the bios reading is 11.84 and 11.78 ?? is that normal ?
 
Super KW said:
hi

i just got my new silverstone 750w with quad 12v 18a each (60A combined) and i see in the bios reading is 11.84 and 11.78 ?? is that normal ?
Click to expand...

Software/bios reads can be off.

Do you have a multimeter to test?

mine say 11.8x and mm tests 12.0x

But those numbers look about right
 
mikeblas said:
Multimeters can be off, too.
Click to expand...

Even the cheap ones are accurate within about .05v, which is MUCH better than using software.
 
chinesepiratefood said:
they are far more accurate than any software readings though
Click to expand...
I've seen DC accuracies which are less than the specs for the monitor chips. $20 specials from Home Depot and Radio Shack, for instance.

There's also user error. A person with a multimeter almost always measures a different part of the circuit than the monitor chip does. Comparing fractions of a percent of accuracy when measuring two different voltages is meaningless.
 
chinesepiratefood said:
they are far more accurate than any software readings though
Click to expand...
We can only wish that this was true in all cases.

Take this $20 model from Radio Shack. It's only 3.5 digits, so 12 volts has to be read on the 20V scale, meaning it'll show up with only four digits of precision: 12.00.

The manual says that the meter has a basic linear accuracy of +/- 0.8%. On a 12-volt reading, that's an error of +/- 0.096 volts. The manual further specifies the meter has an error of +/- 3 counts on the least significant digit. If you connect the meter to 12.00 volts, you might read 11.77 volts, for a total error of nearly 1.93%.

The reason most people think the motherboard chips are so inaccurate is that they're measuring a different voltage, someplace else in the circuit, than the motherboard monitor chip is. The chip's input is connected directly to the 12-volt line near the switching regulator on the motherboard. It might be on the EPS connector, or on the ATX connector; nobody knows. People who measure with a meter often just grab a spare drive connector, which might not even be on the same rail as the supply given to the motherboard. Even if it's the same rail, the voltage drop between all the connectors and wire is not negligible.
 
I don't think it would be THAT OFF even with a $20 Radio Shack DMM, but you do make a very valid point.

The motherboard sensor chips measure off of the ATX, while at the same time getting readings off of the 3.3V, 5V, -5V, -12V and +5VSB leads. And you still have to take in consideration that it's a single 12V lead, going through traces, with the additional load of fan header, a few cards in a couple slots, etc. So there's bound to be some additional resistance there.

And you're right that grabbing a Molex is bound to give you a different reading because resistance is going to be different resistance and possibly a different rail, but since the 12V reading isn't really taken right off of the 12V line near the VR for the CPU, you're STILL going to get a different reading if you want to see if the +12V is good enough for your CPU.

IMHO, ultimately you'd want to have a DMM with sharp probes that you can stick down into the tops of the wires as they enter the ATX connector, or even the 4-pin/EPS connector.

Flukes have probes so sharp you can draw blood. :D

OP: If your voltages were 11.78 and 11.84, you're still fine. Is the computer working well? Yes? What are you worried so much about?
 
Buying a $20 multimeter and expecting a quality measurement is like buying a $20 power supply and expecting it to actually put out the 500W it says it can.
 
jonnyGURU said:
I don't think it would be THAT OFF even with a $20 Radio Shack DMM, but you do make a very valid point.
Click to expand...
It certainly can be THAT OFF; check the specifications yourself. You don't know if it's THAT OFF or not. It might be THAT OFF today, and not THAT OFF tomorrow. That is the nature of error and tolerance.

Probes are replaceable, so any meter an have sharp or dull probes.

Problem is, people do buy inaccurate multimeters and expect good measurements. You need to be sure that your meter is accurate enough to meausre the output of your supply within reasonable tolerance -- and that you're measuring something reasonable.

The upshot is that measuring the output of the supply with a meter and comparing it to the measurements that the monitor sensors give is rather pointless. Believe the one you think that's most accurate and use that information to diagnose your power supply.
 
fireburster said:
mines at 11.54 in bios and in mm. Thats not good right? Its fsp 530w.
Click to expand...

There's no reason for Motherboard Monitor to read differently than your BIOS. They're both reporting the same value from the same sensor.

The ATX Spec says that there's a +/- 5% tolerance for the +12-volt line. So anything between 11.40 and 12.60 is acceptable.
 
larrymoencurly said:
I think that the $3 ones are off by just 1-2% when measuring DC voltages.
Click to expand...
Where can you get a multimeter for $3?
 
mm= multi meter in my case. 11.54 is low though right? i opened the case and turned up the 5v pot and my 12v is reading 12.5v with a multi and in bios.
 
fireburster said:
mm= multi meter in my case. 11.54 is low though right? i opened the case and turned up the 5v pot and my 12v is reading 12.5v with a multi and in bios.
Click to expand...
The ATX Spec says that there's a +/- 5% tolerance for the +12-volt line. So anything between 11.40 and 12.60 is acceptable.

12.5 volts seems high, though it's in range. (It's just as high as your original reading was low, in fact). Why not go for 12.00?
 
I did not want to start a new thread for my question so I am going to ask it here. Using speedfan my +12V is reading 11.13V the highest I have seen it go is 11.19V. Does this mean that there is something wrong with my power supply? I have not encountered any problems thus far.
 
I don't think a BIOS reading alone is enough to make a conclusion about your supply.
 
It's from speedfan so it can't be off by that much. Can someone let me know if this reading is a sign of a problem?
 
mikeblas said:
Where can you get a multimeter for $3?
Click to expand...

Harbor freight. I bought an armload of them, because I end up dropping mine/loosing them. So when I drop one/lose it, im only out $3.00.

They arnt all that accurate if your trying to do real crap, but if your just poking around saying "yep, ive got about 12v here and 10v there and...", they work fine. Ive compared them to the insanley expensive handheld fluke unit my school had, it was "close enough". Far off enough that you shouldnt use it for important things (repairing things with tight voltage/resistance tolerances), but its fine for automotive and computer crap when troubleshooting if something has/does not have voltage.
 
bob said:
Harbor freight. I bought an armload of them, because I end up dropping mine/loosing them. So when I drop one/lose it, im only out $3.00.

They arnt all that accurate if your trying to do real crap, but if your just poking around saying "yep, ive got about 12v here and 10v there and...", they work fine. Ive compared them to the insanley expensive handheld fluke unit my school had, it was "close enough".
Click to expand...
Applications where 20% error is tolerable are rare indeed, and for me don't include automotive troubleshooting. If my fuel pump is getting 10 volts instead of 12, I'm not going to win many races. There are many meters which have decent accuracy which aren't "insanely expensive".
 
mikeblas said:
There are many meters which have decent accuracy which aren't "insanely expensive".
Click to expand...

Would you mind pointing out some for the financially challenged? (such as myself)
 
I'm not sure why you can't find them yourself, but I guess I can give it a shot when I have time. What's your definition of "not insanely expensive"? What minimal functions are you looking for? Which features are required? Which form factor?
 
mikeblas said:
Applications where 20% error is tolerable are rare indeed, and for me don't include automotive troubleshooting. If my fuel pump is getting 10 volts instead of 12, I'm not going to win many races. There are many meters which have decent accuracy which aren't "insanely expensive".
Click to expand...

Where do you see that the multimeters have a 20% tolerance?
 
bob said:
Where do you see that the multimeters have a 20% tolerance?
Click to expand...
In your post, where you're comparing a measurement of 12v to a measurement of 10v.

I don't see any $3 meters at Harbor Freight; this one is cheapest, and it's $10. Its manual specifies an 0.5% basic accuracy for VDC, which isn't bad for the price—but still probably not good enough for testing PC power supplies. Note that the meter has a 0.5% accuracy for its 2V range, but a 1.5% accuracy for all other ranges -- including the 20V range which you'd need to measure the 12V output of your supply.

In this range, the meter has a resolution of only 10 mV, which means your reading of 12V carries an inherent inaccuracy of more than 0.8%. This is the problem with most super-cheap meters; they're only 3-1/2 digts, and that reduces their resolution for useful measurements.

Sure, for packing into your bump-around toolbox, this kind of meter is better than nothing. But I don't think it's too useful for anything else.
 
mikeblas said:
I'm not sure why you can't find them yourself, but I guess I can give it a shot when I have time. What's your definition of "not insanely expensive"? What minimal functions are you looking for? Which features are required? Which form factor?
Click to expand...

Something that will give me an accurate reading for testing power supply readings to get an idea whether or not the thing is a piece or if it's going to blow, if it will truly give enough juice on the various rails to power my equipment, etc.. Something that's also easy to use for a novice such as myself, who is also not sure as to what you mean by form factor. While I'm into computers, I haven't delved into the deeper part of it such as doing things with a multimeter.

So all in all, something about $100 or less, max budget ceiling being $130.
 
mikeblas said:
In your post, where you're comparing a measurement of 12v to a measurement of 10v.
Click to expand...

So you assume that since I mentioned measuring 12v and 10v, that the meter is off by 20%?

mikeblas said:
I don't see any $3 meters at Harbor Freight; this one is cheapest, and it's $10.
Click to expand...

Hmm, I dont see the one I bought from them on that website at all, looks like they no longer carry it. The store was selling them for $3.00, and it wasnt a sale/clearance price. Oh well.

mikeblas said:
Its manual specifies an 0.5% basic accuracy for VDC, which isn't bad for the price—but still probably not good enough for testing PC power supplies. Note that the meter has a 0.5% accuracy for its 2V range, but a 1.5% accuracy for all other ranges -- including the 20V range which you'd need to measure the 12V output of your supply.
Click to expand...
Its good enough to check "We have about ____ volts here, we dont have voltage there", which Is what I was trying to get at. 1.5% of 12v is 0.18v, which may not seem like much, but 12v could read anywhere from 11.82 to 12.18 volts. And its fairly trivial, considering its near what the voltage tolerance of some PSUs are (around 1% for decent ones). Just 4/100ths of a volt off from some PSU's specification. 4/100ths of a volt could be caused by a drop in voltage across the power leads on the ATX PSU, phasing of the moon etc.

mikeblas said:
In this range, the meter has a resolution of only 10 mV, which means your reading of 12V carries an inherent inaccuracy of more than 0.8%. This is the problem with most super-cheap meters; they're only 3-1/2 digts, and that reduces their resolution for useful measurements.

Sure, for packing into your bump-around toolbox, this kind of meter is better than nothing. But I don't think it's too useful for anything else.
Click to expand...

I consider them mildly usefull for anything that does not have a very tight specification on voltages/resistance. But yes, you shouldnt have any buisness repairing/checking things if your trying to use a $3.00-$20.00 multimeter and expect accurate results. Ill have to get out my O-scope and see what its specs were, and compare them to these pile of meters I have. One thing I dislike about them, if the battery voltage is low they just go crazy with readings.
 
techie81 said:
It's from speedfan so it can't be off by that much.
Click to expand...
Then my meter must have been reading the 100% too high in the case of the +12V rail of my ECS K7VTA3 v. 8, but I doubt it because the hard drive was spinning. Even MBM5 was off, but only by about 12%.

The worst inaccuracy for a digital meter should be around 5%, but that's only if the meter has just 2.5 digits and the reading starts with a "2" because 2.0V could be 2.1V or 1.9V. With a 3.5-digit meter, that error would drop to 1/2%.

Sometimes Sears.com has deals on meters, but many of their models use annoying AAA cells.
 
bob said:
Its good enough to check "We have about ____ volts here, we dont have voltage there", which Is what I was trying to get at. 1.5% of 12v is 0.18v, which may not seem like much, but 12v could read anywhere from 11.82 to 12.18 volts. And its fairly trivial, considering its near what the voltage tolerance of some PSUs are (around 1% for decent ones).
Click to expand...
I'm looking at the math the other way. If I have a meter with 1.5% tolerance, and I'm trying to test a power supply to see if it is within ATX spec or not, I'm going to have to throw out a lot of power supplies that are marginal, but good. A supply with 11.40 is still just within the +/- 5% specified tolerance, for example. But I have to see 11.57 volts on my meter, which might be reading 1.5% low, to be sure.

SamuraiBlack said:
Something that will give me an accurate reading for testing power supply readings to get an idea whether or not the thing is a piece or if it's going to blow, if it will truly give enough juice on the various rails to power my equipment, etc.. Something that's also easy to use for a novice such as myself, who is also not sure as to what you mean by form factor. While I'm into computers, I haven't delved into the deeper part of it such as doing things with a multimeter.
Click to expand...

You didn't answer my question about form factor, so I'll assume you're looking for a handheld unit.

I think that ease of use is irrevelant, since you need to know about electronics no matter what meter you're using. I don't think any meter will help you figure out if a power supply will "truly give enough juice"; to do that, you need to load test it and that involves building a pretty complicated test jig.

The Tenma 72-7760 has 0.05% basic DC accuracy, and lists for $150. Seems like you should be able to find it discounted to fit your $130 budget. If you can't, I'd consider their 72-7750 model, which has 0.5% basic DC accuracy, and lists for $70.

Trolling eBay, you can often find much better used Fluke meters in this price range. You might even find a good, but much older, Agilent or HP benchtop unit.
 
Super KW said:
Get my new mobo and my +12v reading is 12.34 :D plus my GPU's are much cooler now
Click to expand...

You're happier with .34 over than you are with .22 under?

No offense, but you are clueless.
 
jonnyGURU said:
You're happier with .34 over than you are with .22 under?

No offense, but you are clueless.
Click to expand...

lol

i did't replace my mobo for .22 under reading !! i was having issue's with my asus mobo thats why i replace it

my gpu temp was idle 55 now 48
my cpu temp was idle 38 now 28
my MB temp was idle 54 now 48
 
Super KW said:
my gpu temp was idle 55 now 48
my cpu temp was idle 38 now 28
my MB temp was idle 54 now 48
Click to expand...
How are you measuring the temperatures?
 
You must log in or register to reply here.
Back
Top