How measure ripple/noise of psu

[seemps]

hi friends ;
I want know how measure power supply ripple/noise .
at first I want khonw : do we need a special filter beetwen psu and scope ? if yes .what is this filter?
and next quastion is : how measure ripple in scope screen. what are difference between ripple an noise .
thanks a lot.

[Oklahoma Wolf]

You need an oscilloscope to measure ripple and noise, and if you want to test to ATX spec, there must be two capacitors in parallel with the load: a 10uF electrolytic and a 0.1uF ceramic disk.

Ripple is lower frequency waves in the waveform, while noise is higher frequency spikes.

[HOOfan_1]

what causes the two different frequencies?

Is the noise at the switching frequency and the ripple at a lower frequency?

[seemps]

thanks. I test my power supply with this method . ripple was sensible but the noise spikes p-p were too high. you think what is this problem?

[Oklahoma Wolf]

what causes the two different frequencies?

Is the noise at the switching frequency and the ripple at a lower frequency?

It can be caused by a few things. Depends on the unit, and whether or not it's faulty.

thanks. I test my power supply with this method . ripple was sensible but the noise spikes p-p were too high. you think what is this problem?

How were you loading the PSU? You can't really test them in the computer, because then you get interference from the rest of the components.

[cypherpunks]

what causes the two different frequencies?

Is the noise at the switching frequency and the ripple at a lower frequency?

Ripple is at the switching frequency (that's pretty much the definition) and noise is higher frequencies. They're hard to separate exactly, so the spec just gives a total.

For a detailed discussion, see Linear App Note 101, "Minimizing Switching Regulator Residue in Linear Regulator Outputs".

Ripple is inherent in switching regulator operation; they operate in discrete cycles. Noise is basically the "click" as the various switches slam open and shut.

[HOOfan_1]

switching frequencies are usually in the 900Hz-KHz range right?

[cypherpunks]

switching frequencies are usually in the 900Hz-KHz range right?

No, switching frequencies are always well above the audible range. Higher powered supplies tend to be at lower frequencies; 50 kHz is popular with ATX power supplies. Smaller supplies are at higher frequencies in the 100–250 kHz range, and where space is at a premium, like cell phones, supplies can switch at MHz frequencies.

I just measured a super-generic Aspire PSU at 49.7 kHz.

[Kvar]

Power supplies can contain ripple at two frequencies.

The switching frequency, and the switching frequency modulated at 2x line frequency. (100 or 120Hz).

[crmaris]

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)?