PSU/Voltages Question

[Deckham]

Need a considered reply by people in the know - not just a generic 'get a new PSU'.

Ok - here's the scenario

System:

Intel Pentium 4A, 2866 MHz (20 x 143) (orig. 2.66, stock/air)
ECS 865PE-A / 865PE-A7 / PF1 / PF3 Extreme / Syntax S865PE
BIOS Type Award (12/24/04)
NVIDIA GeForce4 Ti 4600 (128 MB) GPU Clock 331 MHz
Creative SB0090 Audigy
Maxtor 6Y120P0 (120 GB, 7200 RPM, Ultra-ATA/133)
WDC WD1200JB-00CRA1 (111 GB, IDE)
ASUS CRW-4816A (48x/16x/48x CD-RW)
HL-DT-ST DVD-ROM GDR8162B (16x/48x DVD-ROM)
Realtek RTL8139/810x Family Fast Ethernet NIC
DIMM1: GC6CD404 512 MB PC2700 DDR SDRAM (2.5-3-3-7 @ 166 MHz) (2.0-3-3-6 @ 133 MHz)
DIMM3: 64MX64U-33B 512 MB PC2700 DDR SDRAM (2.5-3-3-7 @ 166 MHz) (2.0-3-3-6 @ 133 MHz)
Memory Clock 351 MHz

PSU: VTECH 480W
*** This PSU was purchased seperately around 2 years ago.
Report:
Voltage Values
CPU Core 1.52 V
CPU Aux 1.55 V
+3.3 V 3.38 V
+5 V 5.16 V
-12 V 4.92 V
-5 V -7.71 V
+5 V Standby 4.68 V
VBAT Battery 3.42 V

Now, this machine has been working fine for 2 years. The only strange issue, is that I usually need to hold the main pwr switch in for about 10 secs, release, then power on every morning.

My concern is the 12v rail. It, erm - doesn't show up on any reporting tools.
Also the -5v at -7.71V

I cannot seem to find this PSU on the net.

The question is - do I have a problem with this unit?

 

[PhoenixOrion]

If machine has been running fine for 2 years with that psu then i don't see a problem with it.....not affecting stability or performance.

If you really need to know your voltages, I'd borrow a multimeter from somebody as some windows app report tools, as you experienced, do not work.

have you tried most reporting tools from pcstats.com and majorgeeks.com?

 

[Deckham]

Thanks Pheonix.

Well, I'm using Everest atm - which seems to report most things fine.

However, Speedfan tells me nearly the same thing. 5.9V for the 12v rail.
Really weirding me out.

I adhere to the policy 'if it ain't broke, etc', but I like to understand the problem regardless.

Edit:
SiSoft Sandra -
Voltage Sensor(s)
CPU Voltage : 1.54V
Aux Voltage : 1.55V
+3.3V Voltage : 3.38V
+5V Voltage : 5.16V
-12V Voltage : 4.92V
-5V Voltage : -7.71V
Standby Voltage : 5.19V
Battery Voltage : 3.42V

 

[Deckham]

Anyone else have ideas?

 

[helpmeout]

I like Everest also, and it's voltage report matches my MOBO's utility. Does your MOBO have a monitoring utility you can activate, or does the MOBO manufacturer have one on its site you could download? If not, I think I'd replace the PSU. It's a lot cheaper than replacing other hardware.

 

[Match]

Originally posted by: PhoenixOrion
If machine has been running fine for 2 years with that psu then i don't see a problem with it.....not affecting stability or performance.

If you really need to know your voltages, I'd borrow a multimeter from somebody as some windows app report tools, as you experienced, do not work.

have you tried most reporting tools from pcstats.com and majorgeeks.com?

^ What he said.

Also, I know of two ways (there may be more) that software can incorrectly report voltages.
1) The software detects the chip that accesses the voltage sensors, but there are no actual sensors attached. This is the case with my ECS motherboard, an L7S7A2. In my case, I can only monitor the cpu and memory voltages correctly. All other voltages show up in software as fluctuating numbers (i.e. 12.3 -> 5.4 -> 9.7 -> -3.3 -> etc).
2) Another way is if the monitoring software uses the wrong algorithm to calculate the voltage from the raw data. I believe this is the case when a motherboard uses a typical monitoring chip with non standard resistors in the circuit that includes the sensor. As I understand it, software needs to know the values of these resistors to correctly compute the voltage. When non-standard value resistors are used, it throws off the calculations.

The gist of this is, it may be the fault of the software, not your psu. If your pc hasn't given you any problems over the past 2 years, don't worry about it.

With regards to the boot up procedure you mentioned. I seem to remember the older ECS K7S5A sometimes having a similar problem. The workaround was to either try another psu or put a small capacitor in series with the power button. I'll see if I can dig up a link for you.

Edit: Sorry, I couldn't find that link. Perhaps I was making it up. The only references I could find to people fixing their booting problems with that other board involved replacing either the cpu or the motherboard.

 

[Deckham]

Thanks, guys.

Seems the two issues aren't related, which was my main concern.
I can live with the 10sec power-up