Spectrum/THD using FFT on Siglent scope

[jph118]

Hey everyone,

Trying to use my Siglent SDS1200CML 100mHz scope for some faux-spectrum analyzing and THD reading via FFT, but not getting very good results compared to what I see others doing on YouTube. I suppose I’ve hit the limitations of my scope, but any tips?

This is primarily for P/P output balancing, SVT-style as per Dumble’s recommendation that I’ve read on this forum. I’ve used the other methods, PI plates, power tube cathodes, etc, but would like to visually see the harmonics at output if possible.

Thanks!

[pompeiisneaks]

the FFT in most newer scopes is software based, and not super 'real time' means they can be a bit slow to update, and you often need something quite consistent over a period to get good results. I did a video of it once on my rigol and it seemed meh at best to me.

~Phil

[Synchu]

What Phil said.
I have Siglent 1102CNL. Spent quite a bit of time trying to make it work, but it didn't. Why they are stating it as a feature is beyond me

Niki

[jph118]

What Phil said.
I have Siglent 1102CNL. Spent quite a bit of time trying to make it work, but it didn't. Why they are stating it as a feature is beyond me

Glad it’s not just me! I think our scopes are basically the same, too.

[bepone]

Trying to use my Siglent SDS1200CML 100mHz scope for some faux-spectrum analyzing and THD reading via FFT, ..
but would like to visually see the harmonics at output if possible.

i dont know what are the options on this scope but FFT (auto set function existing)?
usually you can do FFT live or over the pause button depending of the scope functions and memory/record length.
bandwith need to be limited , freq. span , frequency min set 100Hz if you are monitoring 1kHz, and fmax 10kHz,
RBW resolution bandwith put as much as possible to fasten the scope up, but not too high, in 1kHz signal starting point for RBW is 20-50 Hz..

[pompeiisneaks]

My general understanding though, is that basically due to FFT being a very complex mathematical equation, software isn't fast enough to update it well, so it is slow and laggy. The hardware based ones have an ASIC designed to do the algorithm extremely quickly and are therefore way way better if you're attempting to do real time analysis instead of shaky snapshots of semi relevant data.

I.e. I did almost everything you mention from what I recall and definitely could see some of the harmonics, but they'd sometimes be hard to tell signal from noise etc.

~Phil

[jph118]

I.e. I did almost everything you mention from what I recall and definitely could see some of the harmonics, but they'd sometimes be hard to tell signal from noise etc.

Yes, that’s about where I’m at… I can see what I think is a 2nd harmonic, but can’t discern it as accurately as I’d like.

Either way, using the SVT method with cathodes and a DMM sounds pretty good anyhow.

[jph118]

FYI, was able to rig my Mac for this using iSpectrum free software, the line out of my dummy load, and Avid audio interface… maybe not as accurate as the better scopes and RTAs out there, but gives me a great display of the harmonic spectrum with the ability to observe adjustments in real time, I.e. watching 2nd harmonic increase/decrease with PI adjustment, etc.

[trobbins]

Your digital scope has 8-bit vertical resolution - which is the main reason that any FFT your scope can calculate will have quite poor dB dynamic range (max to min dB display range). Pretty much any soundcard will be at least 16-bit, and typically 24bit and maybe up to 32-bit vertical resolution, which means any FFT derived plot (eg. a spectrum showing and calculating harmonic levels) you see in a forum will typically show >100 to 150dB dynamic range, and so harmonics are easily discerned.

[Chappy]

A few years ago I researched all of the lower cost scopes. I settled on a Keysight DSOX1102G as it seems to produce acceptable FFT plots while staying within my budget. It has a built in function generator, and also generates Bode frequency/phase plots. Perfect for quickly testing audio devices. Unfortunately it is significantly more expensive than the model that the OP has but I thought I would share it's capabilities here for those that are interested in efficient repeated testing of audio devices. Attached are a few tests that I did on guitar pedals. I also plan on using it for analysing and testing amplifiers.

[bepone]

i was thinking or keysight or rs but at the end picked rs- rohde schwartz nice graph

[jph118]

A few years ago I researched all of the lower cost scopes. I settled on a Keysight DSOX1102G as it seems to produce acceptable FFT plots while staying within my budget. It has a built in function generator, and also generates Bode frequency/phase plots. Perfect for quickly testing audio devices. Unfortunately it is significantly more expensive than the model that the OP has but I thought I would share it's capabilities here for those that are interested in efficient repeated testing of audio devices. Attached are a few tests that I did on guitar pedals. I also plan on using it for analysing and testing amplifiers.

Thanks for this, I may consider it for my next purchase… those shots looks great!!

[Chappy]

The first image is a screen capture direct to a usb stick which you can plug into the front of the scope. The second, which doesnt look as nice was taken with a handheld camera. Note that Keysight has different versions of this scope. Make sure you get the one with the internal function generator otherwise you dont get the bode plot functionality.

[trobbins]

The first plot sort of indicates a 90dB max signal to noise floor capability? Can the x and y axis be set up for typical division sizes, or do you have to rely on cursor settings? Can harmonic levels also be displayed in % of fundamental? Have they updated firmware/capabilities over the years? With respect to the frequency response plot, what are the low and high frequency limits (ie. <20Hz, and >20kHz), and can you toggle to a plot of phase response?

[Chappy]

The first plot sort of indicates a 90dB max signal to noise floor capability? Can the x and y axis be set up for typical division sizes, or do you have to rely on cursor settings? Can harmonic levels also be displayed in % of fundamental? Have they updated firmware/capabilities over the years? With respect to the frequency response plot, what are the low and high frequency limits (ie. <20Hz, and >20kHz), and can you toggle to a plot of phase response?

First Q:
FFT X-axis can be set from from 0Hz to multiples of 1,2,5,10 out to 100GHZ (odd since the scope is limited to <100MHz)
FFt Y-axis can be set from to 1,2,5,10,20,50,100dB. Can also be set to mV but I prefer the dB scale in order to see the peak of the fundamental along with the harmonics. After selecting the scale you can then offset the Y axis to raise or lower the whole plot. The small white FFT arrow on the top left moves up and down with the offset and I believe that it is the 0db reference with all peaks below being negative values. When switched to mV that cursor moves to the bottom left corner and thus I assume it represents 0mV.

Second Q:
Harmonics cant be displayed as a percentage of the fundamental however with manipulation of the X1/Y1 X2/Y2 cursors you can set the X1/Y1 on the fundamental and then manually scroll the X2 to the harmonics. On the bottom of the screen numerical values are displayed for absolute values of Y1 and Y2 and the delta between Y1 and Y2. The small dash yellow lines represent the X1 and Y1 cursor lines. Long dash yellow lines represent the X2 Y2 lines.

Third Q:
I have updated the firmware once, over a year ago. I need to check my account with Keysight to see if there have been any other updates since then. I have yet to encounter any bugs or serious faults with my limited use of this device over the last few years. So far I have mainly been using old analog Tek scopes: 465, 2445, 7603.

Fourth Q:
Freq response plot goes from 10Hz up to 20Mhz

Fifth Q:
you can toggle between amplitude and phase plots or display both

Sidenote:
the internal function generator goes from 100mHz up to 20Mhz. Multiple waveforms are available.


Attached are a few FFT plots showing a 1KHz sinewave with its harmonics sourced from the built in wave generator.

Hope that helps and answers all of your questions.

Steve