EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: dimbmw on December 27, 2016, 10:48:53 pm
-
Hi
I am wondering what scope is best on the market for me if I need to record the waveform (from some analog sensor) to post-analyze it? Ideally I would like to record like about 10 seconds, and then manually choose starting time point and ending time point in the archive, and have the scope to do the FFT on the fragment.
Currently I own the cheapest siglent and it does not allow me to do it.
I need ''tis thing asap and will buy it onceI get a good advice here.
I've studied some videos, and when it comes to the FFT it appears that Rhode Swartz has the best fit, but it does not have the archive and the ability to run the ft on the previously recorded waveform. Or maybe I misunderstood the manual about it.
My frequencies are low (within 10 kHz).
Thanks in advance for sharing the info!
-
DMM with digitizing option (34411A, 34465A,34470A) could be better option for sensor signal. And also cheaper.
-
DMM with digitizing option (34411A, 34465A,34470A) could be better option for sensor signal. And also cheaper.
Thanks, but I'd prefer to see the recorded waveform on screen with a convenient user interface.
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
-
DMM with digitizing option (34411A, 34465A,34470A) could be better option for sensor signal. And also cheaper.
Thanks, but I'd prefer to see the recorded waveform on screen with a convenient user interface.
The idea is you use a computer for the interface and FFT computation, why does it need to be all in the scope? Its more expensive and less flexible to do it on a scope. Most of the higher end scopes will let you define the region to compute the FFT on, Agilent/Keysight called it time gated FFT:
https://www.testequity.com/documents/pdf/keysight/time-gated-fast-fourier-transforms-an.pdf (https://www.testequity.com/documents/pdf/keysight/time-gated-fast-fourier-transforms-an.pdf)
But if you don't need the memory depths and sampling rates of a high end scope the total cost of the solution is crazy compared to the sampling multimeters.
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
-
DMM with digitizing option (34411A, 34465A,34470A) could be better option for sensor signal. And also cheaper.
Thanks, but I'd prefer to see the recorded waveform on screen with a convenient user interface.
The idea is you use a computer for the interface and FFT computation, why does it need to be all in the scope? Its more expensive and less flexible to do it on a scope. Most of the higher end scopes will let you define the region to compute the FFT on, Agilent/Keysight called it time gated FFT:
https://www.testequity.com/documents/pdf/keysight/time-gated-fast-fourier-transforms-an.pdf (https://www.testequity.com/documents/pdf/keysight/time-gated-fast-fourier-transforms-an.pdf)
But if you don't need the memory depths and sampling rates of a high end scope the total cost of the solution is crazy compared to the sampling multimeters.
Do you think I can get a scope within $2K budget to do what I need ?
I will study some sample multimeters as well, if I'll fail to get the scope.
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
Only USB scopes can record continuously at low samplerates but you can end up with multi-gigabyte files which can be difficult to deal with. A regular bench top oscilloscope is limited to it's memory depth so you'd have to make the oscilloscope do repeated trigger - store (aka data logging).
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
Only USB scopes can record continuously at low samplerates but you can end up with multi-gigabyte files which can be difficult to deal with. A regular bench top oscilloscope is limited to it's memory depth so you'd have to make the oscilloscope do repeated trigger - store (aka data logging).
I only need to record about a couple of seconds.
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
Only USB scopes can record continuously at low samplerates but you can end up with multi-gigabyte files which can be difficult to deal with. A regular bench top oscilloscope is limited to it's memory depth so you'd have to make the oscilloscope do repeated trigger - store (aka data logging).
I only need to record about a couple of seconds.
For lots of data points and accuracy this is when sample rate AND memory depth is important.
I'm having a fiddle with a SDS1102X and a SDS2304X to see if they can do what you need, however I have to go out soon so won't be able to provide you with urgent answers.
At this time I'm looking at FFT at slow timebase settings in Zoom mode.
I'll come back with my findings later.
Edit
I'm off to get fluxed. :-DD
https://www.eevblog.com/forum/chat/auckland-beer-and-flux-xmas-get-together/ (https://www.eevblog.com/forum/chat/auckland-beer-and-flux-xmas-get-together/)
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
Only USB scopes can record continuously at low samplerates but you can end up with multi-gigabyte files which can be difficult to deal with. A regular bench top oscilloscope is limited to it's memory depth so you'd have to make the oscilloscope do repeated trigger - store (aka data logging).
I only need to record about a couple of seconds.
For lots of data points and accuracy this is when sample rate AND memory depth is important.
I'm having a fiddle with a SDS1102X and a SDS2304X to see if they can do what you need, however I have to go out soon so won't be able to provide you with urgent answers.
At this time I'm looking at FFT at slow timebase settings in Zoom mode.
I'll come back with my findings later.
Great, thank you very much. When you have time, just try to set the trigger, then touch the tip of the probe, record it, and see if you can run a FFT on a recorded waveform or (better) on it's subset.
-
Load the waveform into an audio processing program and select part of the waveform to do FFT on. My go-to program for this kind of stuff is still Cooledit.
If you want to do it on an oscilloscope then the GW Instek GDS2000E series may be an option. It can store waveforms to USB or a network drive using it's data logging function which then can be loaded into a reference trace and analysed using FFT (or other math functions). However it can't do FFT on a partial trace.
Thanks, I've just checked the GDS2000E manual. But did not find the waveform recording function. Data logging is smith different I believe - it is when smith is sampled not more often than 5 seconds.
Only USB scopes can record continuously at low samplerates but you can end up with multi-gigabyte files which can be difficult to deal with. A regular bench top oscilloscope is limited to it's memory depth so you'd have to make the oscilloscope do repeated trigger - store (aka data logging).
I only need to record about a couple of seconds.
For lots of data points and accuracy this is when sample rate AND memory depth is important.
I'm having a fiddle with a SDS1102X and a SDS2304X to see if they can do what you need, however I have to go out soon so won't be able to provide you with urgent answers.
At this time I'm looking at FFT at slow timebase settings in Zoom mode.
I'll come back with my findings later.
Great, thank you very much. When you have time, just try to set the trigger, then touch the tip of the probe, record it, and see if you can run a FFT on a recorded waveform or (better) on it's subset.
I'm thinking of trying to grab a Single shot, save it as a Reference waveform then do an FFT on it.
But later, got to go.
-
up to 10 KHz? hoping you don't need to go under 10 Hz, just use any usb audio interface. the 50-99$ won't "sound" too good but i assure you the preamp is perfectly linear (well that's probably part of why they don't sound too good, am i right?) Voltages too high/low? you have the volume knob!
if you need down to DC get a motu.
You have the bandwidth you need and at least 14 effective bits of resolution. FFT on a scope will always be limited by the fact that it uses and 8bit ADC (there is a nice thread that compares FFTs somewhere)
-
up to 10 KHz? hoping you don't need to go under 10 Hz, just use any usb audio interface. the 50-99$ won't "sound" too good but i assure you the preamp is perfectly linear (well that's probably part of why they don't sound too good, am i right?) Voltages too high/low? you have the volume knob!
if you need down to DC get a motu.
You have the bandwidth you need and at least 14 effective bits of resolution. FFT on a scope will always be limited by the fact that it uses and 8bit ADC (there is a nice thread that compares FFTs somewhere)
What software would you suggest for such an audio interface that will allow to capture and to perform an FFT on the fragment of the recorded waveform?
PS I have analog discovery but it does not seem to offer what I need
-
what kind of amplitude are you talking about. You would get a better answer if you were more specific about what you were recording.
You may be describing "audio". Almost any computer can record audio. If you removed DC blocking capacitors I bet they might even be able to record DC. But that (usually) wont give your precision.
There is a lot of new, increasingly affordable data acquisition hardware. For example, Red Pitaya.
Also, see software defined radio applications that work on your platform. Linrad is one.
-
up to 10 KHz? hoping you don't need to go under 10 Hz, just use any usb audio interface. the 50-99$ won't "sound" too good but i assure you the preamp is perfectly linear (well that's probably part of why they don't sound too good, am i right?) Voltages too high/low? you have the volume knob!
if you need down to DC get a motu.
You have the bandwidth you need and at least 14 effective bits of resolution. FFT on a scope will always be limited by the fact that it uses and 8bit ADC (there is a nice thread that compares FFTs somewhere)
What software would you suggest for such an audio interface that will allow to capture and to perform an FFT on the fragment of the recorded waveform?
PS I have analog discovery but it does not seem to offer what I need
ANY daw. or some other audio analysis program like cooledit/goldwave/audioanalysis/audioxplorer(on osx)... then matlab, octave, ...
-
ANY daw. or some other audio analysis program like cooledit/goldwave/audioanalysis/audioxplorer(on osx)... then matlab, octave, ...
Is cooledit and adobe audition the same software? does not look cheap.
-
Cool Edit is an old package, which I've a feeling was bought by Adobe and no doubt rebranded and made considerably more expensive as a result. Shame.
Try Audacity instead, it's free.
-
Yes try Audacity. It's used by many professionals who work with audio. Even Professional radio stations use it. Not sure the app will do what you need though. But it's free so well worth the effort to at least try.
http://www.audacityteam.org/ (http://www.audacityteam.org/)
-
It sure looks to me like your Analog Discovery has everything you need. Download the latest version of Waveforms. On the scope view, note that you can set the mode to record. Note that in the Config tab you can set the recording sample rate and the duration. On the far right, you can set 'Position' for the view so you can look around in any area of the captured waveform.
One of the neat new gadgets in the latest Waveforms is the Spectrum Analyzer.
I suppose I could use the AD to generate a 10 kHz modulated signal and then run it back into the Scope to see how this all plays. So, here it is... A 10 kHz carrier with a 100 Hz modulation. I sampled at 1 MHz and took 2,000,000 samples (2 seconds). I'm not sure what is happening with the FFT but it is what it is. I just wanted to verify that the scope will record.
I added a Spectrum Analyzer capture as well.
I'm pretty sure if you play around with the AD and the latest Waveforms, you will get where you want to be.
ETA: I am using version 1 of the Analog Discovery.
ETA: I replaced the images with ones that don't have Channel 2 cluttering up the display.
-
Cool Edit is an old package, which I've a feeling was bought by Adobe and no doubt rebranded and made considerably more expensive as a result. Shame.
Try Audacity instead, it's free.
I am also still using CoolEdit.
The older 32 bit versions can still be downloaded for free as well.
-
There is a nifty program called baudline (http://baudline.com/) which is useful for FFT analysis of a lot of different captured files.
"[size=0]Baudline is a time-frequency browser designed for scientific visualization of the spectral domain. Signal analysis is performed by Fourier, correlation, and raster transforms that create colorful spectrograms with vibrant detail. Conduct test and measurement experiments with the built in function generator, or play back audio files with a multitude of effects and filters. The baudline signal analyzer combines fast digital signal processing, versatile high speed displays, and continuous capture tools for hunting down and studying elusive signal characteristics.
Explore, Analyze, and Discover."[/size]
-
Why not just rig it into your sound card and get an FFT analyzer for a DAW? Under 10kHz is no problem for audio gear, and your 24 bit sampling will outperform any regular scope even in high resolution mode. Onboard sound may have a decent noise floor, but a dedicated interface would give you a substantially lower one, and you could record as much or as little as you want.
As for the software, I haven't used it but this plugin has been recommended to me by a guy who works in an acoustic lab, and the 2 channel version is free: http://www.bluecataudio.com/Products/Product_FreqAnalyst/ (http://www.bluecataudio.com/Products/Product_FreqAnalyst/)
-
Hey, look what this guy does with Adobe Audition. You can do pretty much the same with Audacity (free):
http://www.thereminworld.com/forums/t/28554/lets-design-and-build-a-mostly-digital-theremin?post=211434#211434 (http://www.thereminworld.com/forums/t/28554/lets-design-and-build-a-mostly-digital-theremin?post=211434#211434)
-
OK thanks everyone, I ve just did everything with my analog discovery 2 - did not know it has a recording mode.
Still would be happy to buy a great scope as this occasion made me unhappy about my entry level siglent - it was useless for me. I definitely made a wrong choice when I bought it, shame on me for not doing any due diligence.
So I am still interested what is the best scope on the market these days, with the recording feature, with the ability to run the FFT on the fragment of the recorded wave etc.
-
Before I gave up on the Siglent, I would read about the History functionality. I have no idea what it does but it seems to be a step in the right direction.
I am far from an expert with the RECORD feature of the Rigol DS1054Z but it seems to work. I chose to record just one frame and then ran the FFT on the stored waveform. It would be better to post a new thread requesting user experience with RECORD.
Anyway, here is the waveform and FFT:
-
OK thanks everyone, I ve just did everything with my analog discovery 2 - did not know it has a recording mode.
Still would be happy to buy a great scope as this occasion made me unhappy about my entry level siglent - it was useless for me. I definitely made a wrong choice when I bought it, shame on me for not doing any due diligence.
So I am still interested what is the best scope on the market these days, with the recording feature, with the ability to run the FFT on the fragment of the recorded wave etc.
Best? Probably one of the $150,000 Keysights! Just guessing...
Bang for the buck? Probably the Rigol DS1054Z.
Better scopes abound but they start at 3x the cost and go up from there. I would drive a stake in the ground for price or you will get replies from entry level to top-of-the-line.
-
Before I gave up on the Siglent, I would read about the History functionality. I have no idea what it does but it seems to be a step in the right direction.
I am far from an expert with the RECORD feature of the Rigol DS1054Z but it seems to work. I chose to record just one frame and then ran the FFT on the stored waveform. It would be better to post a new thread requesting user experience with RECORD.
Anyway, here is the waveform and FFT:
I do not see the Y axis on the FFT part, ho do one knows what the levels are? Is there a way to turn the X cursor on the FFT screen and see the levels?
-
What was carrier frequency, real carrier level and modulation frequency and depth % in previous Rigol example?
Here:
Siglent SDS1102X(+)
Input 50ohm, signal 5kHz sine carrier level -6dBm, AM with 1kHz sine and modulation depth 50%.
Scope run normally and stopped for looking history buffer and there was 93 last horizontal sweeps. (biggest number is last acquisition) Displayed here frame 37 (56 back from stop moment. FFT. FFT level Ref -3dBm. 10dB/div. Horizontal 2kHz/div. (manually moved center for 7kHz)
For low frequency (audio) FFT analyzing I will recommend - just sound card and some good software. (if your signal is suitable for these)
For RF analyzing. Real SA.
If use scope what have limited processing power for FFT and very limited data length for FFT there is other way. If in hobby use one have time and time is not money, as it is in profession works. I do not know how in other scopes but with Siglent, just insert USB stick and save waveform .CSV. (Or transfer using PC with EasyScopeX) Then use what ever suitable software for FFT (or what ever kind of analyze) (Warning: 14Msample .CSV file is - huge. Look your software limits)
(https://www.eevblog.com/forum/testgear/need-urgent-advice-on-which-scope-to-buy-to-record-waveforms/?action=dlattach;attach=281469;image)
if push "Math" button there can see FFT current settings. Cursors can use for FFT (selecting cursor source Math instead of CH1 or CH2 or Ref A or Ref B). In this image level is dBm, scale 10dB/div and Reference (top) level -3dBm.
-
What was carrier frequency, real carrier level and modulation frequency and depth % in previous Rigol example?
Here:
Siglent SDS1102X(+)
Input 50ohm, signal 5kHz sine carrier level -6dBm, AM with 1kHz sine and modulation depth 50%.
Scope run normally and stopped for looking history buffer and there was 93 last horizontal sweeps. (biggest number is last acquisition) Displayed here frame 37 (56 back from stop moment. FFT. FFT level Ref -3dBm. 10dB/div. Horizontal 2kHz/div. (manually moved center for 7kHz)
For low frequency (audio) FFT analyzing I will recommend - just sound card and some good software. (if your signal is suitable for these)
For RF analyzing. Real SA.
If use scope what have limited processing power for FFT and very limited data length for FFT there is other way. If in hobby use one have time and time is not money, as it is in profession works. I do not know how in other scopes but with Siglent, just insert USB stick and save waveform .CSV. (Or transfer using PC with EasyScopeX) Then use what ever suitable software for FFT (or what ever kind of analyze) (Warning: 14Msample .CSV file is - huge. Look your software limits)
(https://www.eevblog.com/forum/testgear/need-urgent-advice-on-which-scope-to-buy-to-record-waveforms/?action=dlattach;attach=281469;image)
if push "Math" button there can see FFT current settings. Cursors can use for FFT (selecting cursor source Math instead of CH1 or CH2 or Ref A or Ref B). In this image level is dBm, scale 10dB/div and Reference (top) level -3dBm.
So in your example - is the FFT run on just one frame or entire waveform?