Can I use a scope to measure powerline noise?

[Saurav]

I've used my scope to look at noise on a power amp's rail, but that's DC, so it's easy to set the scope to AC and just look at the mV noise signal even though the DC component can be hundreds of volts (for a tube amp). What do I do if I want to look at noise coming in from the wall though? Use a capacitor so the 60Hz component is strongly attenuated? I tried it a couple of days back, and of course if I set the scope for 10s of ms and the highest voltage range, I can see the basic AC signal going way off the scale of my scope. I tried increasing the amplitude resolution while fiddling with the horizontal and vertical position controls to try and keep the trace on the screen, but that didn't work very well.

Just wondering if there is any reliable way to look at that.

Thanks,
Saurav

[SY]

A twin-T notch filter set at 60 Hz can do wonders.

[Saurav]

Like this?



Taken from the bottom of this page:

http://ourworld.compuserve.com/homep...wden/page9.htm

Thanks. My experiments with a surplus Topaz isolation transformer didn't work, the transformer's mechanical hum was louder than the 60Hz hum from my speakers. So now I'm debating trying for a DIY power cleaning/balancing solution, something I've wanted to stay away from. And before I do that I'd like to have a way of seeing the results of any experiments/products, instead of relying on my ears.

[Swedish Chef]

Yes you can do that. I guess you probably already know this but the ground of 99.9% of the scope's out there is connected to protective earth (mains ground). So DO NOT connect ground on the probe to any mains leads. If you pick the wrong one you are going to short the mains through your probe and you will have created a nice little firework and one wasted probe if the fuses in your house aren't fast enough. This might sound trivial but it happens ALL THE TIME at our student's lab with fresh people and since I am responsible for keeping all the instruments in shape, I'm not exactly too excited about it.

A very good thing in your measurement arsenal is what is called a "sniffer". It is basically a few turns of wire (to form an air-coil) connected to a coaxial cable. With this you can pick up even very weak signals by "sniffing" through your design. It's particulary great in RF design since you can for example measure the frequency of an oscillator without loading it with the capacitance of a probe. Do a google on "sniffer" and see how you can build one yourself.

The sniffer is very good with a spectrum analyzer to basically sniff through a power supply for audio and have a look at the induced noise. An oscilloscope isn't exactly a precision instrument and the sensitivity is usually not enough. But you can give it a try.

/Magnus

[SY]

An active twin-T is much better. See, for example, http://www.analogzone.com/avt08062.pdf for some design guidelines. You'll have to attenuate the line voltage to accomodate an opamp, but you could conceivably implement the circuit with some discrete HV parts. Personally, I'd just attenuate, extract the residue, then amplify it. Opamps are your friend.

[Saurav]

I guess you probably already know this but the ground of 99.9% of the scope's out there is connected to protective earth (mains ground). So DO NOT connect ground on the probe to any mains leads.

Wow, I didn't know that, and I did have my probe connected across hot and neutral when I tried my experiment. I didn't see any sparks, and I think my probe still works, but I haven't tried measuring anything after that. So I should just connect the main probe lead and leave the ground floating, and rely on the connection back to safety earth to provide the reference?

Oh I see what you're saying - I got lucky and put the probe's tip on the hot line, if I'd reversed them, I'd have seen what you described. As my wife says, "Which of your lives did you use up this time?"

Thanks for the advice on the sniffer, I'll look that up. I'm not into this quite deep enough to invest in a spectrum analyser yet

Personally, I'd just attenuate, extract the residue, then amplify it. Opamps are your friend.

Except that I don't trust that my opamp circuit wouldn't be adding its own noise residue. I guess if I ran it off batteries and didn't use one of the fancy 100MHz opamps that it would be fairly quiet. And I could check that on the scope too, actually.

Thanks for the link.

[Mr Evil]

Originally posted by Saurav
...So I should just connect the main probe lead and leave the ground floating, and rely on the connection back to safety earth to provide the reference?..

No. The mains supply should be floating (unless someone near you has miswired something), so you won't be able to get a meaningful measurement between live and earth.

[jackinnj]

it's even more interesting if you look at it with a spectrum analyzer ! -- (with a notch filter so that you can get the most out of its dynamic range) -- TI has a 60Hz notch filter (with adjustable Q) in their audio circuit compendium

[Swedish Chef]

No. The mains supply should be floating (unless someone near you has miswired something), so you won't be able to get a meaningful measurement between live and earth.

Well, I don't know how you Brits do it but at least here the neutral conductor and earth is tied together, usually at a central tap for a big building or group of houses. Sometimes you have to run all the way to the power station but they are tied together.
But I agree that for a proper measurement you have to measure between the hot and neutral conductors. The best way to do this is with a true differential module (available as plug-ins for many older Tektronix scope's). But you can easily build one yourself with an opamp (or preferably three opamps in the "instrument" amp configuration). If you choose a low-noise opamp the noise is going to be negligible compared to your scope. With a spectrum analyzer it is a different story.

But you would be amazed just how good a simple sniffer can be. Make a preamp for it if your scope isn't sensitive enough.

/M

[Mr Evil]

Originally posted by Swedish Chef


Well, I don't know how you Brits do it but at least here the neutral conductor and earth is tied together, usually at a central tap for a big building or group of houses....

They are definitely separate here. I had thought it would be the same everywhere, since connecting them together surely defeats the purpose of having three pins on a domestic socket.