[SOLVED] Is it possible to check DC ripple of a PSU without an oscilloscope?

[Veloci]

PC is acting a bit strange as of late. Sometimes, the system unexpectedly shuts down and when I touch the PSU, it feels quite hot, even though the weather is quite chilly here nowadays.

I want to check the DC ripple of its rails, namely the 12v, but unfortunately, I don't have a $500 Oscilloscope lying around in my basement!

So, are there any other alternatives?

I did check its voltages with a mutimeter and they're well within tolerance (±5%). I just want to make sure that the ripple is within tolerance as well (<120mV @ 12v).

Thanks!

 

[Darkbreeze]

No, there is no way you can check ripple without a scope. Also, it's nearly impossible, or at least very difficult, to check your voltage readings under load conditions, using basic tools such as a volt or dvom meter, so what you see without a load doesn't necessarily reflect an accurate picture of the condition of the unit, but it IS a good baseline to know if it's definitely bad. It's just not a solid way of determining that it IS definitely GOOD either.

What is the EXACT model number of the power supply, how long has it been in service and what are your full hardware specifications for the rest of your hardware?

 

[BFG-9000]

It actually is possible to do this, you just set the multimeter to check "AC" as you probe the DC voltage while things are under load.

This is one of those situations where an analog display on the multimeter may be preferred because a digital one may jump around too quickly to read.

 

[Veloci]

What is the EXACT model number of the power supply, how long has it been in service and what are your full hardware specifications for the rest of your hardware?

I've a Dell Optiplex 755 MT with a 305W standard PSU running a Core 2 Duo E8400 and 2 x 2GB DDR2 800MHz RAM modules.

Quite old and shabby, really.

Since the PSU always feels hot after the unexpected shutdowns, I think its caps are worn out and causing high ripples, but that's just a wild guess.

It actually is possible to do this, you just set the multimeter to check "AC" as you probe the DC voltage while things are under load.

I just tested the PSU with your method and the results are strange:

12v (DC): 11.67.
12v (AC): 23.6.

That's 23,600mV whereas ATX standard specify a maximum ripple of 120mV, if I'm not mistaken.

Out of curiosity, I also tested the voltage of a AA cell. Same weird result:

DC: 1.51v.
AC: 2.6v.

Any idea?

 

[BFG-9000]

Reverse the leads (that is, black on +12v and red to GND) and try again!

 

[Veloci]

Reverse the leads (that is, black on +12v and red to GND) and try again!

Man... You're a GENIUS!

It worked! That darn trick worked! I can't believe it. You just saved me $500, not that I was about to buy an oscilloscope but still...

The ripple is jumping between 130-150mV @ 12v rail with mild load (Tested with a 12v 0.6A fan) so it's time to replace that darn thing!

Thanks a ton, mate!

 

[Darkbreeze]

That is news to me. Nice trick BFG.

I've never known you could do this without a scope meter.

 

[BFG-9000]

LoL, that's actually a pretty standard quick&dirty way to check if a rectifying diode pack is bad, such as in an alternator.

While it does tell you the average ripple/noise which was enough for the OP, things could actually be much worse than that because a scope can also show the presence of brief voltage spikes that the multimeter won't. The scope can also differentiate between ripple at one frequency vs. noise to make it easier to guess where the problem is, for repair.

Now that the people who know how to use them are dying off, decent old scopes are showing up at local electronics recyclers for ~$20. If you can get a deal like that it's highly recommended for tinkering!

And to the OP, I am looking at a 305w Dell PSU and besides some extremely short but standard cables, it also has this odd 6-pin proprietary connector hanging off of it. Hopefully your Optiplex doesn't use any connector like that.

 

[Veloci]

While it does tell you the average ripple/noise which was enough for the OP, things could actually be much worse than that because a scope can also show the presence of brief voltage spikes that the multimeter won't. The scope can also differentiate between ripple at one frequency vs. noise to make it easier to guess where the problem is, for repair.

Well, that'd be the WORST case scenario. In my humble opinion, the voltages and DC ripple will go beyond tolerance long before the PSU start showing alarming spikes in voltages.

But even still, an oscilloscope would be a nice gadget to have, no doubt about it! Speaking of which, what are your thoughts about this one:

Oscilloscope-Kit Fully-Assembled - AliExpress

And to the OP, I am looking at a 305w Dell PSU and besides some extremely short but standard cables, it also has this odd 6-pin proprietary connector hanging off of it. Hopefully your Optiplex doesn't use any connector like that.

Yes, the 24-pin is EXTREMELY short, probably a cost cutting measure by Dell, and apparently one of their favorite. Almost ALL Dell PCs I've seen have the same problem, let it be Precision servers or Optiplex PCs.

But other than that, everything is 100% ATX compliant without any proprietary connectors, thankfully. Besides, I'm thinking about building a custom i5 Ivy Bridge machine soon so it doesn't matter much.

Thanks!

 

[mdd1963]

Quibbling over output ripple is something the tech sites do in PSU reviews, as though lack of or an additional .02V vs. a .06 V ripple is going to matter...

If you suspect you need a new one (such as when trying to install an actual GPU in a Dell crippled with a 250-300 watt PSU!), just replace it with a decent 500-650 Watt unit and move on