PSU +5VSB output too high

[Idol]

Yesterday my gaming PC started shutting down on me suddenly; and I mean suddenly like I snatched the cord out of the wall kinda suddenly. The first couple times I hit the "Start" button and it came back up no problem. Tonight it decided that restarting was just not an option. I can get it past the post beeps and through BIOS no problem. Beyond that when I get to the options menu I've tried Normal, Safe Mode, and Last Known Good Configuration. Farthest I've gotten past that was using Last Good I can get to the scrolling green Vista loadup/start bar; the one before the Vista splash screen. Then it cuts off again.

I took it apart and gave it a good cleaning. That did nothing. So I plugged in my PSU tester as that seemed like a good place to start troubleshooting. I have an Antec SP-500 (500 watt) power supply and used an Antec 77204 PSU tester. I got an "HH" alarm on the +5VSB rail. I tested it using both the 4 & 6 pin mobo connectors and various SATA power cables to troubleshoot my troubleshooting. Same result no matter what I plugged in/didn't plug in.

My question is, since it appears that my +5VSB output voltage is too high, what are my remedies? Obviously replacing the PSU is one. Are there any other fixes? I'm not using any Wake-On-LAN options so to the best of my knowledge the SB rail is only used by the "Start" button on the front of my ATX tower. Would it be possible to disable/disconnect the SB rail and power-up some other way? Or am I just better off dropping the $$ for a new PSU?

Money really isn't an object. I priced new 500W PSU's already and haven't found one yet that wasn't well within my budget. I just want to avoid something like sending my new car to the scrapyard because one of the radio knobs fell off. Thanks for your help.

If replacing the PSU is my best option, I'm more than willing to entertain suggestions as to a replacement. If it's any help, here's a link to my PSU/Case

 

[xosniper]

easiest fix is to replace the PSU. They are cheap and easy to find.

 

[Master Gunns]

First off, is there a way you can get another PSU to test and see if the issue resolves?

If it's been sending too much voltage it may have fried the motherboard, which caused the issue.

Are you able to boot via the Windows DVD? Mayhap your harddrives got corrupted or something as well during a power loss.

All in all what I'm getting at is while a new PSU may be within your budget, there is more that could be causing the problem then just that. If you're able to troubleshoot other things (borrow a PSU or something etc) before spending money, all the better.

 

[Idol]

UPDATE: I just pulled the PSU out of the case (for complete isolation goodness). I took my trusty analog voltmeter to +5VSB circuit and got a whopping 8 volts! I'm realitivly certain I have found the problem. I've been doing some other readings and apparently I was wrong about not needing the SB circuit. Turns out that systems with a soft-power button (which is just about every ATX case I know of) use that circuit to detect you trying to turn it on. So I'll be buying a new PSU.

 

[Angellas]

For future reference, a volt meter will NOT give you a proper reading on any power supply that is not under load. +8vdc on a +5vdc rail is well within acceptable params. for most power supplies that are not under load. You can research Voltage drop if you would like to know more about why this is.

You mentioned a PSU tester. If this tester places the PSU under load (like any decent tester should) and gives you an overvotage warning then you have assumed correctly that your PSU is to blame. However, if the PSU tester does NOT place the PSU under load you may want to look into your motherboard as Master Gunns suggested.

I realize this is too late to be of great help to you, but maybe it will help you or others in the future.

 

[Idol]

Angellas, first off thank you. And you're right, it is too late. I went ahead and ordered another PSU. However, if you find this, I wouldn't mind if you could elaborate on why my voltmeter reading is wrong?

I'm familiar with voltage drop. I'm an electrical lineman. And that may be the problem. I routinely work with over 13,000 VAC at over 800A. So when it gets down to minuscule voltage levels (at least to my frame of reference) and DC systems, I'm not intimately familiar with the nuances.

While I understand the need to test under load, and I think my PSU tester does this, I thought load testing would affect amperage and not voltage. All my experience (and recent wikipedia readings) attribute voltage loss to inherent resistance in any conductor, thus making this loss a function of conductor length. How is this different in low voltage DC circuits?

Also, for me to get 8VDC on the 5VDC rail means that placing the PSU under load drops voltage by 37.5%. Whoa! The tolerance I'm used to is +/- 10%. That's why an electrician calls household wiring 110VAC while a lineman swears it is 120. When I test that newly installed transformer on the pole it should read between 108-132. So dropping from 8 to 5 is a bit out of scope for my thinking.

Please don't take any of this as me arguing with you. Based on your response, your sig, and your "Tech Help Guru" award I think you know what you are talking about. I'm just trying to figure about how to change what I know about primary voltage AC lines to apply to low voltage DC lines.

 

[Angellas]

Are you kidding?!?! I LOVE people that respond with honest and warranted questions and/or discussion!!!

Knowing that you're a Lineman makes this really easy to explain! Just like you mention, voltage drop is caused by a resistance (a load) being placed on a power supply. Regardless of voltage and type of load/resistance, the same rules apply.

Let's have a slice of Ohm's PIE to do some explaining. Unfortunately, the simple P=IxE formula does not include all the variables needed. It does, however, sound quite cool. The Ohm's law formula we need is P=E?/R.

*For those who may read this that do not know Ohm's Law:
P= Power, measured in Watts
I= Inductance, measured in Amperes
E= Energy, measured in Volts
R= Resistance, measured in Ohms.

Using the formulas of Ohm's law and assuming a 750w power supply, we plug-in the numbers:

P=E?/R becomes 750watts=E?/R if we're crazy enough to think that all of that PSU's power is in the +5vdc rail! More reasonable numbers would be 200watts=E?/R.

NOW! Working backwards to achieve proper and understandable numbers, we need to achieve +5vdc on a power supply output with a fixed resistance. Since some guy picked 200watts as a number, 5vdc should be easy enough to achieve!

200watts=5vdc?/R . That will give us 200watts=5vdc?/.125 Ohms .

.125 Ohms on a motherboard would give you a proper 5vdc.

Now, let's remove the PSU from the motherboard. We're not changing the power potential (the wattage), we're only changing the resistance by removing the load (the motherboard). Let's look at our formula now.

200watts=?vdc/R^i

We still have 200watts of power but our resistance has changed. The change in resistance will, of course, have a direct affect on the voltage. The question is....how much did the resistance change? We haven't reduced the resistance to 0 have we? Of course not. That would be impossible AND would create a serious problem parsing a formula against the numerical representation of infinity or infinite nothing (0). This is where the R^i representation comes into play. R^i, or intrinsic resistance, is the measurement of the inherent resistance of conductive material regardless of cross-section width or wire length. This is more in the realm of Electron Physics than electrical engineering, but, well, diversity in knowledge FTW!

I'm not going to school you on intrinsic resistance in this post, but here are a few hard numbers to work with.

Intrinsic Resistance of 2 metals used in PSUs:

Copper = 1.72x10-8
Gold = 2.24x10-8

Now, using the R^i against the formula R=p(L/A) we can get the internal resistance of your power supply if we knew the cross-section (AWG) of the conductors and the length of he wire.

R= Resistance
p= Intrinsic resistance
L= Length
A= cross-section area

Since there's no way to really know the ammount of wire used in your PSU and your PSU's transformers, I'll just throw a number at you.

200watts= 8vdc?/R ... Gives you 200watts=8vdc?/.32 Ohms.

Knowing the resistance differences of .32 Ohms for 8vdc and .125 Ohms for 5vdc, you have your tolerance levels down to numbers that should make sense to you.

I would go into the details of why resistances can go "backwards" but that would be a bit lengthy for this post. I trust you understand.

I hope this has been informative for you and welcome any questions or criticisim!

 

[ShadowBlade72]

This post doesn't have any point other than to say... holy crap.

 

[Idol]

Ah ha! I understand my mistake. Learned all about the myriad functions of Ohm's Law back in my apprenticeship days but never use anything except P=I*E at work. My error in over-simplifying the situation. Thank you for the detailed explanation.

 

[Angellas]

Bump, for friend to find/read.