Episode 828 I show how to use the FFT function on a digital oscilloscope to show a spectrum. There are some tricks. Read about window functions: en.wikipedia.org/wiki/Window_... Be a Patron: / imsaiguy
Low frequency operation on a spectrum analyzer is possible with the use of an RF mixer. You enter an HF signal of 10 MHz and the LF input could be a sweep from 20 Hz to 20 kHz On the spectrum analyzer you set the frequency offest to -10 MHz. Set the start frequency to 0 Hz and the stop frequency to 20 kHz. Setting the sample/hold fucntion would allow to show frequency response.
Thanks for this vid. I have the Rigol and it was over my head as to get the FFT function and understand what I am looking at. MORE MORE MORE Please, Carl
The FFT resolution on Rígol's scopes can be improved by changing Memory Depth setting. By default the FFT is taken using only datapoints from the screen, which can be enlarged by setting the Memory Depth to 6 or 12 million points. Unfortunately, this setting does not "stick", and often volunterely switches off back to Auto.
WTAF. The important part of this tip isn't even the memory depth. It's that they only do the FFT on the portion in the screen. Well....actually, that kind of makes sense. I don't know how else you'd set the FFT length. I guess make it a parameter, which would make sense from a DSP POV, but maybe not to an EE using a scope.
this video has allowed me to actually use this function on my Rigol. Thank you, it is difficult but it does work and gives me access to another tool i didnt have. It also shows signals and harmonics way above the scopes bandwidth...
Thanks for this video. I thought I understood FFTs until I looked at my scope's FFT. The settings are indeed very weird, but now I know what knob does what familiar operation, despite the names.
I found that on DSOs that support it, one can switch the scale from dBV (that's too senstive at low amplitude) to DC and get rig of much of this noise. The spectrum diagram looks much more like that of the analog spectrum analyzer.
Thank you for this. I never had the patience for the FFT function in the Rigol preferring to reach for a real SA. It is a very powerful scope with a bewildering number of non intuitive features.
Thanks for reviewing a FFT function of the scope. Can the Rígol scope provide FFT with dynamic range over 50dB (8-bit)? Simply put, what is the FFT noise floor (dB) when a perfect sine is provided to the input?
Assuming linear 8 bit ADC, you have 8 doublings of voltage and each doubling is 6 dbv for a total of 48 dbv. The problem, or A problem (one of many) is that when displayed on a log scale, the least significant bit of data becomes the most visible part of the FFT. What intermediate voltages exist between 0 and 1 (LSB)? None! There are no intermediate voltages consequently this HUGE step increase produces low power harmonics from here to the moon. But because of the logarithmic scale, that noise in the least significant bit is over-emphasized. Also, that noise will randomly change the 0 to a 1, a 1 to a 2, and so 253 becomes 254. But the 0-to-1 transition, or 1-to-2 transition is a DOUBLING of voltage or 6dbv with NO INTERMEDIATE voltages. It is a really big step change. But from 253 to 254 is barely any change at all in terms of dbv. That's why, in log mode, it is really ugly at the bottom and rather useless as compared to a Real Spectrum Analyzer (tm).
I am so glad i bought a rohde&schwarz (hmo1202) although it was the cheapest one, its fft is really usable, it measures in dbm or dbv or whatever you want and accurate enough to give usable values, it is compared to the rigol a dream to use. Money well spend i know now.
robber576 I am thinking about making a video about this scope, exciting as i am not in the business of making videos ;-) , so i made yesterday some measurements on this scope, to give you some idea: it is a "100 mhz" model but measured with the leo bodnar 40 ps pulser it goes up to 185 mhz, my marconi 2955b confirmed that and yesterday i measured in fft mode a 120 dbm 30 mhz signal coming out of the marconi within +1,7 dbm accuracy, the scope is 8 bit but up to 16 bit in highres mode and apparently can do that reliable within +/ 2% accuracy down to -120 db with a noise floot @ -130 dbm, its quite impressive for something sold as the "cheapest model". Granted, at €1000 not in the same league as a €300 rigol but then what can you expect for €300? A decent multimeter costs more... As i said, i am glad i did spent that money.
I would add that unlike the trace averaging function, which produces noise-free output, there is no FFT averaging on Rígol's low-cost scopes. The FFT is simply taken on a single acquired trace no matter what settings are. This is a real pity, as such FFT averaging option could have been easily supplemented, greatly improving FFT spectrum S/N.
Looks like the Rigol is about the same as the Siglent as a spectrum analyzer. I haven't played with the FFT on my Siglent enough to get it all figured out, but I can usually play with the knobs and eventually get a useful display. I have a "real" spectrum analyzer here and I usually use that. But like most spectrum analyzers, my real analyzer won't go low enough to look at audio. Using the FFT function, I was able to measure the harmonics (better than -40 dB) of a "pure sine wave" DC to AC inverter I have, and was also able to look at intermod products of an SSB transmitter. Maybe some day somebody will add a traditional spectrum analyzer interface to their FFT spectrum analyzer and give us the best of both worlds.
More, More, More please! More door cracking light. Thanks for the look. BTW, have you checked out Joe Smith upload on the Nano VNA, looks like it would be something up your alley. It's over me but some light in.
Thanks for the demo. I've been needing a scope for electronics purposes and also for analyzing audio, and was thinking of using a Rigol DS1054Z for both purposes. But what I can see here is that it's not very useful for audio. Great for circuit debugging, but not for writing music.
@@qddk9545 A scope is a great tool for musicians, especially when using synthesizers. But it needs to be a scope which is designed for audio. Many digital audio workstations have one built in, but I was looking for a convenient solution for an all-hardware setup. Eventually I settled on just using a free phone app.
My FNIRSI Scope has FFT, But no selection menu at all. LOL. But I did get similar results to yours. (it looked the same, but there are no readouts or measurements).
Thanks for the demo. Rigol really made FFT a pain to use. I ended up buying a cheap USB scope as a backup to my Rigol just for FFT use. If you can show XY mode on the Rigol that would be great too. I can't ever get that to work right on mine.
@@IMSAIGuyA great video. Any chance of making a new video of FFT using the new MSO5000? I have that scope thanks to you (a good thing) It has a lot.. more/different...functions
If you have a computer with an audio input, you can use much nicer spectrum analyzers for low frequencies, such as the free voxengo span. there are apps for phones and tablets as well.
I second this, i use REW, it has a table that calculates the THD of the harmonics and has an accurate Oscope function with cursors. I say accurate because i checked it against my rigol 1102ze, Band width of REW is up to 30k which is plenty for audio. now would i like to own/have access to a nice Frequency analyser? of course but this seems to be a decent tool
I have a potentially stupid question. What method of square wave signal generation is used in this particular case so that it can be spectrally analyzed using FFT? I must admit I'm not quite knowledgeable and have deep gaps in my education. For instance if a multivibrator is used as a the source of the square wave output signal can it be spectrally analyzed in this way? However I must admit I'm so dumb that I dare ask another question: When using the spectral analyzer does it display the real composition of the square wave input as in this case or does it on the contrary shows a theoretical representation of its input signal based on the Fourier sine wave odd harmonics composition that would give us an equivalent square wave form be that sine wave odd harmonics composition actually implemented to generate the square wave form?
the multivibrator (555) is just fine. The display is the actual waveforms. The FFT is a mathematical transformation of signals in the time domain (oscilloscope) to the frequency domain, called a Fourier (four ee ay) transform. FFT is fast Fourier transform (a software algorithm)
I is very informative video, I hope it will push me to try FFT on my Rigol DS2072A. I wonder why all FFT videos only covers RF signals. My main use would be checking the noise in power supplies, and harmonic distortion in HiFi and guitar equipment. I think an overlooked issue is ringing in power transformers.
Younger folks may be coming at it from the reverse perspective--we're familiar with FFTs and RTSAs are a little confusing...or maybe my audio engineering background involves more FFT from DAWs.
GREAT VIDEO.WONDERING IF I CAN USE FFT AN OFFSET TO PRODUCE 2 SLIGHTLY DIFFERENT FREQUENCY FOR MY COMPONENT TESTING ? I HAVE A O SCOPE WITH A WAVEFORM GENERATOR,CAN CHANGE AMPLITUDE OF VOLTS AND FREQUENCY OF WAVE FOR CHANNEL 1,OFFSET FOR CHANNEL 2? THANKS JOHN
thanks for reply on the FFT.thought I might be mathematically able to change the frequency input to the component, that would show a phase difference on the scope.do you have any other suggestion? seems like component testers rely on a slight change of amplitude and/frequency,perhaps this is mainly done with the current limiting resistors? John
you need to check the max usable FFT frequency for the scope you have. This scope goes to about 10MHz: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-JbZ1-qFaVG4.html
I would say the FFT function on scopes is deliberately rubbish to get you to buy one of their spectrum analysers. This is not just Rigol. The way you have to make the time domain signal almost invisible to get a decent RBW is odd but part of the hardware I guess.
@@IMSAIGuy "you are easily amused" And a bit pedantic. Your editing is quite a bit better than "Kiss Analog" where I find myself skipping forward trying to find the "meat" of oscilloscope explanation; or cringing at 500 microseconds per division, two divisions is one second. Um, no, two divisions of 500 microseconds is one millisecond. But, whatever, he does eventually get some useful information posted. Struggles with triggering a 100 millivolt AC signal riding on 3.3 volts DC. Easy peasy, switch the trigger to AC; and since you don't care about the DC, rather than a huge offset, just switch the input to AC.
@@IMSAIGuy I' have a very nice Philips dual trace delayed sweep analog oscilloscope but it's big, a bit long in the tooth, but it FEELS and works like an oscilloscope with nice clicky detents on the knobs and genuine dual time base. upload.wikimedia.org/wikipedia/commons/e/ef/Analog_Oscilloscope_Philips_PM3267.jpg Rigol has a nice gradient fill appearance for two-tone SSB alignment, makes it look a lot like doing that on an analog oscilloscope. Rigol's FFT is by all indications horrible. I like the Keysight Bode plot thing, rather a lot like VNA but nowhere near the frequency range of NanoVNA *but* can do that sort of thing all the way down to audio frequency. Likewise the FFT can process audio frequencies. So at the moment I am leaning to Keysight and basically look for reasons NOT to buy Rigol, or Keysight, or whatever. None of them feel like the venerable Tektronix.
the newer Rigols have better FFT: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-JbZ1-qFaVG4.html it also has bode plots. if you want a cool instrument, Siglent has spectrum analyzer and VNA in the same box. SVA1000X
I think Rigol FFT can do for single shot, which Spectrum Analyzer couldn't. can Rigol show impulse response? Maybe? en.wikipedia.org/wiki/Impulse_response
actually dB is a relative measurement and so your comment is valid. However a dBm is an absolute measurement as is dBuV. They are reverenced to a milliwatt and microvolt. So there are very big differences.
@@IMSAIGuy True enough. I suppose the scope/FFT might actually be calibrated to read dBV! :-} But I gather it doesn't have 50Ω input ability, so dBm makes no sense. Why complain about it?
@@qwaqwa1960 I have an HP and Keithley multimeter which both support dBm directly. Rigol could do the same. I point out something Rigol could have done better. Many scopes also have an internal selectable 50 ohm input. Also a feature I wish Rigol would have added.
You video is very misleading as you are using a special digital scope, not a typical analog scope. Why not use a regular scope to demonstrate this? Better yet why use that scope at all since you already habe a spectrum analyzer
the reason to use this instead of spectrum analyzer is the frequency. a normal spectrum analyzer will only go down to 10kHz. The FFT allows frequencies below. so great for audio frequencies. I did a video on analog scopes: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-2ASidXr8Iyk.htmlsi=5bR81SNJ__rC_MGY
