Changing voltages.

**stj** · 03-30-2016, 01:21 PM

Re: Changing voltages.

bad caps will cause voltage drift like that.
another sign is fluctuating fan rpm in the bios.

**Per Hansson** · 03-30-2016, 01:37 PM

Re: Changing voltages.

And a third is that voltage monitors built in to mainboards are notoriously bad.
Use a DMM to check the voltages instead and ignore what the BIOS thinks

**Spork Schivago** · 03-30-2016, 02:08 PM

Re: Changing voltages.

Originally posted by stj

bad caps will cause voltage drift like that.
another sign is fluctuating fan rpm in the bios.

The fan RPM is fluctuating as well. I thought maybe it was just trying to adjust to the temp. So with a fluctuating fan RPM and the fluctuating voltages, definitely bad caps?

**Spork Schivago** · 03-30-2016, 02:13 PM

Re: Changing voltages.

Originally posted by Per Hansson

And a third is that voltage monitors built in to mainboards are notoriously bad.
Use a DMM to check the voltages instead and ignore what the BIOS thinks

Thanks Per Hansson. Even though I've worked on computers for a very long time and know a good deal about them, I'm afraid that is beyond my current knowledge. What exactly would I be measuring on the board with the DMM? For example, the Vcore, 3.30V, 5.00V, 12.00V....are there normally test points that say Vcore for me to measure? Or would I need a schematic or something to find these test points?

For the fan RPMs, I believe I can just use my Hz mode on my DMM. From what I've googled, I just measure on the tach wire and multiple by 30. Does that sound right?

If it is bad caps, would you think they'd be the caps in the VRM circuit?

**stj** · 03-30-2016, 02:21 PM

Re: Changing voltages.

scope the power connector on the board.

the monitoring chips arent accurate, but they are stable.

**keeney123** · 03-30-2016, 06:01 PM

Re: Changing voltages.

Originally posted by Spork Schivago

Hello,

I have a friend who brought me his PC. He says it's been acting up. It takes around 10 minutes to shut down and it's running Windows 10. He updated it from 7 to 10, 64-bit. It has 4 gigs of RAM, running at 800MHz. It has an E5200 Dual-Core @ 2.50GHz. The onboard video seems to be a bit cheap. Total memory in the BIOS shows 4096MB with 8MB shared memory and 1MB GTT memory. I was planning on replacing the thermal paste, cleaning it real good, backing up his data, formatting and doing a clean install of 10. I went to the hardware monitor in the BIOS and noticed some that I wanted some advice on.

I see the Vcore voltage is changing. 1.208V - 1.232v. I also see the +3.30V changing. 3.344V - 3.60V. THe 5.00V changes as well, 5.040V to 5.064V. Finally, the 12.00V changes. 12.144V - 12.196V. The ranges might be a bit different, I got tired of watching them. Is it normal for a PC to change voltage a little bit like that? I know it's not much of a change, but I'd think the voltages would be stable. All I can think is maybe the board is doing some sort of auto-overclock or something and changes some of those voltages automatically...

If not, do you guys know what might cause this? It's an AsRock G31M-S board.

Thanks!

So I am wondering is the Ram 800 MHz the speed of the Ram or the speed of the Bus? I would think the Ram speed needs to be faster?

**fzabkar** · 03-30-2016, 10:34 PM

Re: Changing voltages.

Some of the voltage readings don't make sense to me, and that's not because of any inaccuracy in the hardware monitor IC.

AFAICT, the Super IO chip is an ITE IT8718F:

https://cdn.badcaps-static.com/pdfs/...45f7d6b508.pdf

The chip's ADC is 8-bit with a 4.096V reference. This means that it can resolve voltages in increments of 16mV (= 4096 / 256). The lower voltages (+3.3V, +2.5V/Vmem, Vcore, Vbat, 3V3_STBY) can be measured directly, but higher voltages (+5V, 5V_STBY, +12V) need to be stepped down via voltage dividers. Hardware monitor software typically has problems measuring the higher voltages because it is unaware of the voltage divider ratios. The motherboard's own BIOS, however, shouldn't have such a problem.

In the OP's case the Vcore readings do not differ by an integral multiple of the ADC resolution.

1.208V = 75.5 counts x 16mV per count
1.232V = 77 counts
3.344V = 209 counts

If we assume that the scale factor for the 5V readings is 1.5, then the resolution would be 24mV. This is consistent with the readings.

If the OP's 12V readings are accurate, then we can assume that the difference of 52mV corresponds to a single count.

So 256 counts x 52mV per count = 13.312V

... and the scale factor is 52 / 16 = 3.25

Note that many (most?) PC PSUs regulate the output voltages by sensing the weighted average of the +5V and +12V rails. This means that when the +5V rail drops (eg by -20mV), the +12V rail rises (eg by +50mV).

**Per Hansson** · 03-31-2016, 08:44 AM

Re: Changing voltages.

Originally posted by Spork Schivago

Thanks Per Hansson. Even though I've worked on computers for a very long time and know a good deal about them, I'm afraid that is beyond my current knowledge. What exactly would I be measuring on the board with the DMM? For example, the Vcore, 3.30V, 5.00V, 12.00V....are there normally test points that say Vcore for me to measure? Or would I need a schematic or something to find these test points?

For the fan RPMs, I believe I can just use my Hz mode on my DMM. From what I've googled, I just measure on the tach wire and multiple by 30. Does that sound right?

If it is bad caps, would you think they'd be the caps in the VRM circuit?

Don't overdo it, just shove a DMM probe into an empty HDD molex connector from the PSU or whatever else is convenient.
VCore is usually quite easy to measure via a coil or something close to the CPU, but that's overkill for now, you just want to confirm if the reading on the mobo is bullshit or not.

**Spork Schivago** · 03-31-2016, 11:29 AM

Re: Changing voltages.

Originally posted by stj

scope the power connector on the board.

the monitoring chips arent accurate, but they are stable.

The 24 pin one or the 4-pin one that's for the CPU? Would I see the changing voltages on either of those? If the PSU is good, wouldn't those voltages be normal at that point in the circuit and not be fluctuating until after the bad capacitor? Still in the progress of learning how electricity works.

If I have a 12v power supply on a circuit and then say a few resistors in there, in series, wouldn't the voltage be 12v before the first load (resistor), and then after each resistor it'd get lower until there were no loads left in the circuit, in which case it'd be 0v?

I'd think if the voltage was fluctuating because of a bad cap, before the bad cap, it'd be normal. Am I wrong in thinking that? Thanks!

**Spork Schivago** · 03-31-2016, 11:33 AM

Re: Changing voltages.

Originally posted by keeney123

So I am wondering is the Ram 800 MHz the speed of the Ram or the speed of the Bus? I would think the Ram speed needs to be faster?

The board says the bus is 1600 MHz. Whether it's running at 1600 MHz or not, I'm not sure. The RAM is DDR-1066 with a CL5-5-5-15 (2.0 - 2.1v), G.Skill F2-8500CL5D-4GBPK. The base frequency of the FSB would be 400MHz, right? The BIOS only has options to run the RAM at 667MHz or 800MHz, but my understanding is, unless someone's overclocking the FSB, there's no real point in running the RAM at a faster speed. Probably why the board only gives the option for 667MHz or 800MHz. Am I right in thinking that?

**Spork Schivago** · 03-31-2016, 11:40 AM

Re: Changing voltages.

Originally posted by fzabkar

Some of the voltage readings don't make sense to me, and that's not because of any inaccuracy in the hardware monitor IC.

AFAICT, the Super IO chip is an ITE IT8718F:

https://cdn.badcaps-static.com/pdfs/...45f7d6b508.pdf

The chip's ADC is 8-bit with a 4.096V reference. This means that it can resolve voltages in increments of 16mV (= 4096 / 256). The lower voltages (+3.3V, +2.5V/Vmem, Vcore, Vbat, 3V3_STBY) can be measured directly, but higher voltages (+5V, 5V_STBY, +12V) need to be stepped down via voltage dividers. Hardware monitor software typically has problems measuring the higher voltages because it is unaware of the voltage divider ratios. The motherboard's own BIOS, however, shouldn't have such a problem.

In the OP's case the Vcore readings do not differ by an integral multiple of the ADC resolution.

1.208V = 75.5 counts x 16mV per count
1.232V = 77 counts
3.344V = 209 counts

If we assume that the scale factor for the 5V readings is 1.5, then the resolution would be 24mV. This is consistent with the readings.

If the OP's 12V readings are accurate, then we can assume that the difference of 52mV corresponds to a single count.

So 256 counts x 52mV per count = 13.312V

... and the scale factor is 52 / 16 = 3.25

Note that many (most?) PC PSUs regulate the output voltages by sensing the weighted average of the +5V and +12V rails. This means that when the +5V rail drops (eg by -20mV), the +12V rail rises (eg by +50mV).

Although I don't understand everything you said, parts of it makes sense. I wanted to add that the voltage monitor in the BIOS, when the voltages change, they're always changing to the same numbers. Just as an example, the 3.30V would be 3.344V or 3.60V. Never 3.350V, etc. That's just an example though. I remember for some of the values, there would be maybe 3 or 4 numbers it would change too.

The 12V one was the one that I watched the least. It was at the bottom of the list and I got tired of waiting for it to change and just jotted it down real quick like. If you'd like, I can start the machine back up and be more thorough with the voltages. I can list exactly what voltages it's changing to, instead of listing the lowest voltage it changes to and the highest voltage it changes to.

Would you like me to do that? If I do, I'm sure we could figure out that voltage divider ratio for certain.

**Spork Schivago** · 03-31-2016, 11:42 AM

Re: Changing voltages.

Originally posted by Per Hansson

Don't overdo it, just shove a DMM probe into an empty HDD molex connector from the PSU or whatever else is convenient.
VCore is usually quite easy to measure via a coil or something close to the CPU, but that's overkill for now, you just want to confirm if the reading on the mobo is bullshit or not.

Okay. I also have a PSU tester. I can pull the 24 pin + the 4 pin CPU power supply, plug it into my tester and see what the voltages say, if you'd like. I think it only goes out to two significant digits though (ie, 3.31).

I'll grab the pinout for the 24-pin connector and measure the various voltages using the DMM and post back.

**Spork Schivago** · 03-31-2016, 11:46 AM

Re: Changing voltages.

OMG! There's this terrible smell inside! I'm not trying to be mean here, but sometimes old people smell like this. I think it might be like a mothball type of smell. The friend is definitely a bit older and his cloths smell just like this, but usually not as strong.

**stj** · 03-31-2016, 12:10 PM

Re: Changing voltages.

refering to the psu i hope?
time to get it open for some visual recon.

**Spork Schivago** · 03-31-2016, 12:13 PM

Re: Changing voltages.

Using the 24-pin connector, I've measured, with my DMM, the 3.3v, 5v, 12v, and -12v pins. There's four 3.3v pins, but they all ready the same. That's true for all the voltage pins I've measured. So instead of posting all four 3.3v readings, I'll post one.

Code:

3.3v = ( 3.370v through 3.376v) -- changes frequently.
 5v = ( 5.08v )        -- seems steady.
 12v = ( 12.08v through 12.10v) -- jumps to 12.10v and stays at 12.10v.
-12v = (-11.78v through -11.80v) -- seems to jump between the range.

**Spork Schivago** · 03-31-2016, 12:14 PM

Re: Changing voltages.

Originally posted by stj

refering to the psu i hope?
time to get it open for some visual recon.

No, the whole computer inside smells like this, not just the PSU. I opened up the case to the PC and it's awful!

**stj** · 03-31-2016, 12:22 PM

Re: Changing voltages.

open the psu, those outputs should be rock-stable.
specially as a psu tester has no real load in it.
you should have metered it connected to the motherboard atleast.

**Spork Schivago** · 03-31-2016, 12:27 PM

Re: Changing voltages.

Originally posted by stj

open the psu, those outputs should be rock-stable.
specially as a psu tester has no real load in it.
you should have metered it connected to the motherboard atleast.

I did meter it connected to the board, instead of the PSU tester, because the DMM has better resolution. Should I check the voltages hooked to the PSU tester as well?

**stj** · 03-31-2016, 12:39 PM

Re: Changing voltages.

no, open it.

Changing voltages.

Changing voltages.

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Related Topics