60 Hz interference

60 Hz interference, a forum discussion on Cleverscope. Join us for more discussions on 60 Hz interference on our Questions forum.

Back to Forum Index : Back to Questions   RSS

27 May 2012
Posts: 19

I am measuring a broad band noise signal and there is a lot of interference at 180 hz (3X 60Hz).
I have trigger = Auto and Y-axis = V^2/hz

So much interference that the side lobes in neighboring bins pick it up, and the measured power is much greater than the physical noise.

Would the spectrum setting hanning, blackman, matter ?

Options I am considering:
turning off non-essential AC items.
Faraday cage

Notch reject/comb filter

Would there be a way to use the external trigger to center the 60hz in one bin, so that all the other bins would have 0's on the 60hz harmonics?

Years ago I used an FFT such that the spectral bandwidth always had the 60hz centered in one bin, so the nearby bins were not affected.

2 Jul 2012
Posts: 395

Sorry about the delay in responding.
I am assuming that the interference is in your signal. If not, and the 60 Hz, and third harmonic are coming from a nearby source suggests that you should try and stop this source getting into your signal.

Some possible causes are a ground loop becuase the Cleverscope acquisition unit, the PC and the unit under test are not sharing the same ground. I would make sure you have a good solid connection to mains earth, and then distribute power/earth form the same power connection.

If there are power sources nearby (such as transformers, switch mode power supplies, displays etc) I would seperate these from your signal leads, as the linkage is probably magnetic at these low frequencies. If you are not actually measuring the 60 and 180 Hz, then a low frequency common mode choke might help (wind a few turns of the signal lead around a low frequency ferrite or soft iron core).

On the signal measurement front, make the scope graph an integer number of 60 Hz cycles across in time. This means there will not be any end-point discontinuities to generate higher frequency spectral products. In addition set sufficient time width to get frequency resolution you want. Frequency resolution is roughly proportional to 1/scope graph time width. So for example with a 200 msec wide scope graph, you will capture 60 x 0.2 = 12 cycles, and have a frequency resolution of 5-6 Hz. Having an integer number of cycles does result in 0's at the harmonics, as you want.

Section 4.5.3.6 Spectrum discusses the window specifications. You want a window with low lobe width, and high side lobe roll off. Two such windows are HAnning and Blackman. These windows do suffer from reduced accuracy as the signal frequency moves away form the Bin centre frequency. Use the table to estimate the maximum error (for example with Hanning, the -6dB main lobe width is 2 bins. Thus one bin either side of the main lobe is 6 dB down. At half a bin away from the center you are down about 3 dB - which represents the error of teh measured signal as the frequency sweeps from -1/2 bin to +1/2 bin.

You can see the effects of the windows by capturing a pure tone, and then bringing up the Settings/Spectrum dialog, choosing a window and clicking Apply.

So in summary - go for an integer number of interference cycles, make the time duration long enough to get the frequency resolution you want, and use a Hanning or Blackman window for maximum spectral discrimination.

Please let me know if you have further questions.
Back to Forum Index : Back to Questions   RSS
You must be logged in to post a reply



You need to Register or Log In before posting on these forums.

×

Your shopping cart is empty.