0

I have a 15volt dc signal with an FFT on channel 1. I see a pulse on the FFT. Does this mean I have a voltage at a certain freg?

FFT

BKnight
  • 31
  • 1
  • 3
  • 1
    Why would you do an FFT on a DC 'level' (they arent really signals if they just stay at a certain voltage)? – KyranF Nov 21 '14 at 22:57
  • 1
    Switch your scope to AC coupled and zoom in on that signal at a few hundred mV with a sweep of maybe 10usec per division. – Spehro Pefhany Nov 21 '14 at 23:12
  • 6
    @KyranF, there's nothing wrong with looking for noise on your DC power supply. (HF noise can leak through opamp supply rails.. perhaps not what's going on here.) – George Herold Nov 21 '14 at 23:14
  • 4
    I totally agree with George. Doing a spectral analysis of a DC level is a totally valid test to see what interference it my be producing. Have you guys heard of EMI? – Andy aka Nov 21 '14 at 23:19
  • This is not looking like high frequency noise.....2.5Msamples means the SPAN will be 1.25Mhz --> Spike looks on the right hand half may be between .8M to 1.25M...This is not high frequency noise...there may be measurement issue..... – user19579 Jan 21 '15 at 05:33
  • You can see spikes on the the time domain signal so this will appear on the fft. Ac coupling the DC would make it easier. It prob is some dcdc hanging off that rail so either the carrier or the switching edges ASSUMING no measurement issue –  Jan 21 '15 at 09:36

1 Answers1

1

You see a spike. That means this frequency is very much present in your signal. All the other frequencies don't have a spike, but they're still present, though less.

But really, why are you doing fourier transform on a DC level, as KyranF points out?

On Wikipedia, Fourier transform is explained, if that's the problem. This image (from there) may help as well:

enter image description here

  • It's possible that it can help to show the amplitude/portion of noise at a certain frequency.. but yeah, strange. – KyranF Nov 21 '14 at 23:02
  • for example, if i'm reading the graph properly, there is a spike at the 200kHz frequency. This could be a 200kHz switching supply? – KyranF Nov 21 '14 at 23:03
  • @KyranF I read 225kHz, but yes. It can't be net voltage coming through, and I believe the mysterious frequency you see when you touch a probe with your hand (what's that called anyway) is much lower as well. Look at the other (recent) questions of OP though. I have a hard time figuring out what he's trying to do at the moment. But okay. –  Nov 21 '14 at 23:05
  • 1
    I work in Solar. 2014 NEC requires ARC fault detection on the module level. Modules produce DC. One of the input parameters to measuring arc faults is sensing noise on the DC line at a certain frequency. Noise can come from the inverter up stream as well as from a bad connector, etc. – BKnight Nov 22 '14 at 12:14
  • How did you read the 200kHz from that scope shot? – BKnight Nov 22 '14 at 12:18
  • @BKnight the power level shown in the FFT drops off from 0Hz DC slowly, and then after about 8-9 divisions (of 25kHz/division) there is a spike, meaning there is a lot of 'power' in that particular frequency. This is what Camil and I are talking about. – KyranF Nov 23 '14 at 01:27
  • Is there some kind of switched mode supply like an MPPT tracker connected to this DC supply? –  Nov 23 '14 at 20:47