Need Help about PSU Voltages (3.3 V)

[FirzaVista]

Good Afternoon

 

 

Sorry for asking this short question. Is this still fine if my 3VCC is below 3.300 V?

 

Thank you.

[Srijan Verma]

your 3VCC is more then 3.00V

[mariushm]

It's OK. the standards allow for around +- 5% variation and most components will work just fine with some variation.

 

Any component that needs exactly 3.3v won't rely on the power supply but will take 5v from the power supply and use voltage regulators to produce 3.3v locally on the component.

 

edit:

Also, it's acceptable for some voltages to be higher than what they're supposed to be because there's some losses that happen between where that voltage is generated and where that voltage is picked up.

There's the well known Ohm's law which says voltage = current x resistance which tells you there's gonna be some voltage loss the longer the distance is between a supply and the component consuming power.

So for example you have the voltage regulator producing 3v and you have the component 3-5 cm away on the motherboard ... the wires on the motherboard have some resistance, so depending on how much current the component needs, there's gonna be some loss... the voltage regulator may output 3.1v but by the time this energy reaches the component, the component may "see" exactly 3v, because 0.1v are lost as heat in the motherboard traces (wires in the circuit board)

[--SID--]

Should be fine and the readings of the input voltage by HWmonitor is not accurate. Only a digital PSU and a multimeter show you an accurate measurement.

What's the brand and model of your PSU?

[Somerandomtechyboi]

14 minutes ago, FirzaVista said:

Sorry for asking this short question. Is this still fine if my 3VCC is below 3.300 V?

Yep

 

My p5qs seem to have a chronic issue of having the 3.3v be really loaded for some reason so 3.3v just starts drooping, even with my new (used) psu it still droops to 3.2 or 3.1v, though better than the previous 2.9-2.7v it kept drooping to using crappy oem psus

 

8 minutes ago, mariushm said:

It's OK. the standards allow for around +- 5% variation and most components will work just fine with some variation.

Only issue with slightly lower or higher volt it that components may get under or overvolted cause with a crap psu on my p5q the 12v rail goes to 14v and side affect of that is when i set 1.6v in bios it turns into 1.75v idle and the expected 1.54v when load (llc disabled), technically its a ~16% increase in volts without considering other losses

 

Though a +-5% is perfectly fine for most components so no worries, i might even take advantage of this effect and use a boost converter to feed 14v on my g31 eps connector for higher than 1.6v vcore due to the cpu volt controller being unknown and having no datasheet so rip voltmod the normal way

[FirzaVista]

15 minutes ago, --SID-- said:

Should be fine and the readings of the input voltage by HWmonitor is not accurate. Only a digital PSU and a multimeter show you an accurate measurement.

What's the brand and model of your PSU?

AeroCool Lux RGB 550W 80+ Bronze

 

And here's my PSU Tester result:

[--SID--]

1 hour ago, FirzaVista said:

AeroCool Lux RGB 550W 80+ Bronze

The Aerocool way: putting garbage in a fancy box.

1 hour ago, FirzaVista said:

And here's my PSU Tester result:

12,5v @12v is very bad.

[mariushm]

It's not that bad.

The thing is there's a few problems:

1. Those psu testers aren't that accurate. The measurements will be done in 20-50mV steps .. so if your power supply outputs 3.16v, the tester may round it out to 3.2v ... or if the tester measures 12.42v, it may show 12.5v

 

2. When you're using the tester, you're measuring without any load applied to the power supply.

A lot of group regulated power supplies are "tuned" to keep the 12v proportional to the 5v and maintain regulation with the amount of current consumed on 5v.  So without any load on 5v, you may see exactly 5.00v and 12.5v on 12v, but if your PC consumes let's say 10-20w on 5v, the 5v may drop to 4.95v and 12v will also drop a bit to 12.4v,  and with maybe 40-50w on 5v and 3.3v (normal for a regular computer), you would see 4.8-4.9v on 5v and 12.1-12.2v on 12v.

Easiest way to test would be to go to a car/bike parts store and buy a 12v incandescent light bulb and connect it to either the 5v or the 12v outputs on a molex connector. You should be able to find 12v 20w and 21v 40-60w bulbs for very cheap, under 2-3$. Then, you can use the tester while the power supply has some load applied to it, and you'll see the voltages drop a bit.

 

3. If a component consumes a lot of power, there's gonna be some voltage drop so even if power supply outputs 12.5v, you'll get 12.4v or less at the end of the cables on a video card, for example. A hard drive that only consumes around 10 watts (some on 5v and some on 12v) won't cause much voltage drop  but also a hard drive won't be that much sensitive about the 12v input, as that's only used to spin the motor.

 

The software measurements will also be affected by the accuracy and precision of the ADC (analogue to digital converter) chip, that chip will also measure voltages typically in 8...20mV steps. The chips are usually powered with 3.3v ...3.6v and they have a voltage reference inside that does 1.25v or 2.048v or 2.5v  .. so a voltage like 12v is reduced by 3-6 times using a couple of resistors with around 1% precision and then the chip measures the reduced voltage and that voltage is converted to a number (ex 256 steps of 8 mV = ~ 2.048v if an 8 bit ADC is used)  ..

The values will also be affected by where the measurements are actually performed. If you get a multimeter and measure by the 24 pin connector, you may get 12.5v , but if you measure right by the chip which does the measurements, you may see 12.3v, because the voltage has to travel all the way across the board.