Simple D.C. Noise detector on powersupple rail

Thread Starter

Mrdouble

Joined Aug 13, 2012
107
I'm building a DC power supply tester for a PA system. Part of the testing process, I would like to detect ripple and noise in 500 to 15k hertz range.
Was thinking a simple led if noise or ripple is detected.

Example, one specific power supply, the manufacturer forgot to put a high uFcapacitor on load side of +24v regulator so the power rail transferred the regulator ramp signal to the pa amp which transferred the noise across the PA system.

If I can get some suggestions on where to start, that would be great. I'm not splittering hairs on this one so brute and simple.

Thanks in advance
Micheal
 

wayneh

Joined Sep 9, 2010
16,400
To some degree you've answered your own question. Use the PA system – or any other inexpensive audio gear – as your test rig. Expensive gear will isolate itself from power supply noise.

An oscilloscope allows you to see the ripple instead of hear it, so that's a good option but not as cheap as you probably want. It'd be my only option at 15kHz, since I can't hear that high anymore.

Another approach would be to place a capacitor (0.1µF, for instance) in line with the power supply and a 10k to 100k resistor. Apply that signal to the input of any audio amplifier. Pure clean DC will pass nothing and be quiet. Any ripple will be coupled across the capacitor (and resistor) and appear as noise that gets amplified. You could also use a transformer to couple the power to the audio amplifier. You'd want a resistor on both the primary and secondary to protect both sides. The important thing is that in both coupling techniques, DC cannot pass.
 
Last edited:

#12

Joined Nov 30, 2010
18,210
Was thinking a simple led if noise or ripple is detected.
Noise is always a continuum from nanovolts per radical Hertz to whole volts. Decibels and all that. You must decide on a threshold of noise which is acceptable and how much is unacceptable. Only after you nail that down does the design become achievable.
 

Reloadron

Joined Jan 15, 2015
5,667
You may want to give this a read. The link is to the instruction manual for the old HP 334a Distortion Analyzer. They were really a pretty cool tool of their day. Take note of what distortion consist of (Harmonics, Hum and Noise) and how the unit using a modifies wein bridge tunes out the fundamental leaving the distortion and how the remainder is measured using an oscilloscope.

Ron
 

Thread Starter

Mrdouble

Joined Aug 13, 2012
107
To some degree you've answered your own question. Use the PA system – or any other inexpensive audio gear – as your test rig. Expensive gear will isolate itself from power supply noise.

An oscilloscope allows you to see the ripple instead of hear it, so that's a good option but not as cheap as you probably want. It'd be my only option at 15kHz, since I can't hear that high anymore.

Another approach would be to place a capacitor (0.1µF, for instance) in line with the power supply and a 10k to 100k resistor. Apply that signal to the input of any audio amplifier. Pure clean DC will pass nothing and be quiet. Any ripple will be coupled across the capacitor (and resistor) and appear as noise that gets amplified. You could also use a transformer to couple the power to the audio amplifier. You'd want a resistor on both the primary and secondary to protect both sides. The important thing is that in both coupling techniques, DC cannot pass.
I hadn't thought of passing it off to an amplifier but was thinking about passing it to a cap with a op amp hanging off other end
 

Thread Starter

Mrdouble

Joined Aug 13, 2012
107
To some degree you've answered your own question. Use the PA system – or any other inexpensive audio gear – as your test rig. Expensive gear will isolate itself from power supply noise.

An oscilloscope allows you to see the ripple instead of hear it, so that's a good option but not as cheap as you probably want. It'd be my only option at 15kHz, since I can't hear that high anymore.

Another approach would be to place a capacitor (0.1µF, for instance) in line with the power supply and a 10k to 100k resistor. Apply that signal to the input of any audio amplifier. Pure clean DC will pass nothing and be quiet. Any ripple will be coupled across the capacitor (and resistor) and appear as noise that gets amplified. You could also use a transformer to couple the power to the audio amplifier. You'd want a resistor on both the primary and secondary to protect both sides. The important thing is that in both coupling techniques, DC cannot pass.
Thank you for your response, you got me thinking about some op amp hanging off a cap which turns on led. Here's the thing, I can troubleshoot the power supplies but there's like 1000s of them so trying to speed up the testing
 

wayneh

Joined Sep 9, 2010
16,400
You can certainly build a peak detector into it so that you don't need to actually listen to it, and this could detect frequencies higher than we can hear.

Now that I think about it, I'd still use an audio amplifier to make the noise audible and then I'd use my smartphone to do an analysis of the sound produced (since there could be noise I can't hear, but the phone can). You could even do a fourier transform on it and recognize all the frequencies where noise is being produced. By standardizing the test procedures, you could quantitate the ripple level.
 
Top