How measure ripple/noise of psu

[cypherpunks]

Sorry to dig up such an old thread but I was wondering if a resistor (e.g. 1K) is needed before the capacitors to lower the Amps or the oscilloscope can cope with high Amps (with psu in full load)?

You must not add a resistor if you want to make an accurate measurement. 1 K and 10 μF give a corner frequency of 16 Hz, so everything above that will be severely attenuated.

The current would only matter if it were going through the scope (and even 1 amp would blow it up very reliably); generally oscilloscopes have 1 MΩ input impedance, so it'll draw 12 μA from the 12 V line.

What makes you imagine that load current not going through the oscilloscope would cause it the slightest difficulty?

[crmaris]

Thanks very much for the answer. I have understand it now. Indeed the current doesn't flow through the Scope when you measure ripple. However another question is raised now; if Ι would like to add a switch before the capacitors, in order to select the rail I want to monitor (e.g. 5V, 3.3V, 12V) what type of switch should I use, one that could handle the full Amps( e.g. 117A for 12V)?

I am little confused lately because of reading overdose. You see I am in the process of building my own Load Tester and I am trying to clear up some things.

[cypherpunks]

Thanks very much for the answer. I have understand it now. Indeed the current doesn't flow through the Scope when you measure ripple. However another question is raised now; if Ι would like to add a switch before the capacitors, in order to select the rail I want to monitor (e.g. 5V, 3.3V, 12V) what type of switch should I use, one that could handle the full Amps (e.g. 117A for 12V)?

Again, why would the switch care about current that wasn't passing through it? Do you need special extra-rugged faucets in houses near the Mississippi river?

You will have a very small surge as the capacitors charge, but that doesn't require much of a switch. And is a function of the capacitors, not the power supply.

[crmaris]

Again, why would the switch care about current that wasn't passing through it? Do you need special extra-rugged faucets in houses near the Mississippi river?

You will have a very small surge as the capacitors charge, but that doesn't require much of a switch. And is a function of the capacitors, not the power supply.

In the attached scheme I have added two switches before the capacitors to switch between the monitoring rails; so current will not pass through the switches, it will follow the resistors path?

[rafal_iB_PL]

A bit of necro post, but I have the question:

What are the consequences of not applying those two caps in parallel with load ie. what other waveforms may appear on the scope? I just got me the Fluke 97 scopemeter, and for starters, I wanted to check out the ripple on my PC's outputs. Probe grounded to the case, no parallel caps. As a result, on top of the (most likely) ripple waveform I see some major aperiodic spikes on +5V and +12V, much like those in oc3d's Cougar review. Is it just the interference form the computer, like OKW said?

[Makalu]

not happy with my transient load test again so maybe i can do some good here lol...yeah the caps are going to filter out a lot of the noise spikes there although there's always going to be a lot of noise on the PC...most of it seems to come from the fan motors near as I can tell. Anyway here's a couple of shots to compare:



This is on the load tester...same signal points...probe with caps on Ch 1 and Ch 2 is just a probe with the little spring clip and the ground lead (AKA interference antennae).



same setup again except with the 20mhz bandwidth limit set on Ch 2 which cuts down on a lot of the noise but still some little tails there and the main difference I guess is that the trace of the ripple waveform has more detail with the caps.

[rafal_iB_PL]

Once again, major thanks. Glad I joined here.

Wonder why icon highlighting new posts didn't appear since my last one.

[Per Hansson]

Holy thread batman revival time!
Just a small question on this theme: what sort of setup for the capacitors do you folks use?
I have always neglected to do it: but when testing to rebuild this Seasonic PSU discussed here I found out the hard way that it really matters!
Not satisfied with the result after recapping it with LOW esr caps (actually motherboard grade) I bought some general purpose caps like the originals seems to be.
But the result was worse, but then I put on a 10µF electrolytic cap and a 100nF film cap and most of the ripple I thought was there was gone, meaning it was mostly HF noise...
But holding those caps + probes and cables in close proximity made me feel there might be a prettier solution to this

[ashiekh]

You need an oscilloscope to measure ripple and noise.

How well would the following work?

http://www.ti.com/lit/an/snoa690/snoa690.pdf
(good to 500 kHz)

[Stefan Payne]

How well would the following work?

http://www.ti.com/lit/an/snoa690/snoa690.pdf
(good to 500 kHz)

No, you need 20MHz.