EEVblog Electronics Community Forum

Products => Test Equipment => Topic started by: Anand on June 08, 2015, 06:49:18 pm

Title: Rigol DS1054z FFT bug(s)
Post by: Anand on June 08, 2015, 06:49:18 pm
Ok, since we're on a kill streak, here's some more Rigol DS1054z bugs.
Input signal: 1kHz calibration square wave.
DC input, Trigger level: 1.5V, 500us/div
With trigger level = 2.8V, bug's gone.

Switch to 2ms/div - FFT with lots of aliasing.
Switch to high-res with 120k memory depth - aliasing's gone.
High-res / 120k points is the only mode where the FFT for this signal is correct, from 10ms - 200us.
Other modes show aliasing in some timebases and Anti-aliasing option doesn't do shit.
Title: Re: Rigol DS1054z FFT bug(s)
Post by: alsetalokin4017 on June 08, 2015, 08:14:00 pm
Yep, mine does all that too. I was able to reproduce all 4 of your screenshots. Here's a couple that show the error (and "correct") at higher resolution:

(using Blackman window in the FFT)



Title: Re: Rigol DS1054z FFT bug(s)
Post by: Anand on June 09, 2015, 07:27:13 pm
God, FFT on this scope f**king sucks d*ck.
Rectangle "window" is unusable. :rant:
Title: Re: Rigol DS1054z FFT bug(s)
Post by: W7NGA on June 09, 2015, 07:52:36 pm
Your FFT display is exactly what I would expect.

I am not seeing an issue here.
Title: Re: Rigol DS1054z FFT bug(s)
Post by: Anand on June 09, 2015, 09:00:42 pm
Your FFT display is exactly what I would expect.

I am not seeing an issue here.
Expect for $400? Expect from a chinese product? Would you care to elaborate?
Because I expect a FFT of a square wave without aliasing and without some notch filter with the cutoff frequency inside my f*cking spectrum.
Something like this:
(http://upload.wikimedia.org/wikipedia/commons/thumb/b/b5/Spectrum_square_oscillation.jpg/350px-Spectrum_square_oscillation.jpg)
Title: Re: Rigol DS1054z FFT bug(s)
Post by: W7NGA on June 09, 2015, 09:42:07 pm
It's sampling theory within a constrained system. Your perfect square wave does not exist.
From the FFT array perspective, here is your 'square' wave. Not square at all .. plenty of even-order energy.

This signal looks much the same on my $6000 Agilent DSO. Except, they do the memory management for you.
I personally prefer letting the user make memory assessments and expecting the user to understand sampling and FFT constraints.
Of course, this also allows one to shoot themselves in the foot if they aren't astute ...  |O

It's not that difficult to do meaningful work if you work within the constraints of your system.

Engineering 101

$380 well spent (see second photo)
Title: Re: Rigol DS1054z FFT bug(s)
Post by: alsetalokin4017 on June 09, 2015, 11:40:22 pm
There is no doubt that the square wave from the calibrator isn't "perfect". But why is the display of the FFT so different, depending on whether "Normal" or "High Precision" is selected in "Acquire" mode? And... since as you point out there is "plenty of even-order energy"... should that not then be showing up _better_ when "high precision" is selected?  What exactly does "high precision" in this scope do, anyway?

Yes, I do agree that for the price it's still a bargain. But we novices are misled enough already, without an oscilloscope taking us by the nose and leading us down some garden-path full of aliasing.
Title: Re: Rigol DS1054z FFT bug(s)
Post by: W7NGA on June 10, 2015, 02:35:51 am
Interesting discussion, and I don't mean to be an apologist for all things Rigol.
In fact, there is some merit to the original posters complaint and I shouldn't dismiss it.

Firstly, the Rigol calibration square wave is better than I thought and there is almost zero energy beyond the expected odd-harmonics and the fundamental.
I show this in the first image, from my conventional spectrum analyzer. The signal is outside the working bandwidth but still provides a good relative indication of harmonic content.
The perfect square wave has discontinuities whereas the real world square wave does not. However, there is such a thing as a distorted (non perfect) square wave with additional FFT terms.

The second image is from the Agilent 7014B where the user has no control of the sampling parameters. Here, Agilent does a better job of memory management.
Rigol has an Auto memory depth mode that still allows sampling artifacts to leak thru. This would be confusing to the beginner and a source of potential measurement error/confusion.

The FFT can be a gnarly beast to tame. Clearly, there are critical relationships between array size, sample-rate, memory depth, truncations and windowing, etc. where the user has to have a firm
understanding of their sampling system. I am a little surprised that Rigol doesn't shield the user better in Auto mode but I don't find it limiting. I will concede that it can be confusing however.

dan W7NGA
Title: Re: Rigol DS1054z FFT bug(s)
Post by: miguelvp on June 10, 2015, 02:41:08 am
There is no doubt that the square wave from the calibrator isn't "perfect". But why is the display of the FFT so different, depending on whether "Normal" or "High Precision" is selected in "Acquire" mode? And... since as you point out there is "plenty of even-order energy"... should that not then be showing up _better_ when "high precision" is selected?  What exactly does "high precision" in this scope do, anyway?

Yes, I do agree that for the price it's still a bargain. But we novices are misled enough already, without an oscilloscope taking us by the nose and leading us down some garden-path full of aliasing.

I already replied to your other thread but I'll put it in here as well:


Check w2aew's videos regarding this:

If you don't feel like going through the videos, on Normal mode it's throwing a lot of samples away depending on the sample rate you select, on high res it doesn't.

https://www.youtube.com/watch?v=oRf-IpG6XAw (https://www.youtube.com/watch?v=oRf-IpG6XAw)

https://www.youtube.com/watch?v=EqRvUyPhbu8 (https://www.youtube.com/watch?v=EqRvUyPhbu8)
Title: Re: Rigol DS1054z FFT bug(s)
Post by: alsetalokin4017 on June 10, 2015, 04:51:38 pm
Well, it turns out that the manual recommends "Average" mode with AC-coupled input for the FFT. And indeed it works better that way.

Title: Re: Rigol DS1054z FFT bug(s)
Post by: Anand on June 10, 2015, 05:34:20 pm
Well, it turns out that the manual recommends "Average" mode with AC-coupled input for the FFT. And indeed it works better that way.
I've actually read that today, too!  ;D
 Unfortunately, no, it doesn't do anything against aliasing (pic 2):
Pic 1: timebase 1ms, FFT clean, OK.
Pic 2: timebase 2ms, FFT aliased.
Pic 3: timebase 5ms, FFT clean, OK.

Title: Re: Rigol DS1054z FFT bug(s)
Post by: W7NGA on June 10, 2015, 05:36:07 pm
This caught me at a bad time since I am leaving for S.E. Asia for 7 weeks and just put most of the test equipment into storage.
I worked for Tektronix and have a dozen (or more) scopes here at the lab. I have a Tektronix TDS2014B, a terrible scope when
compared to the Rigol, and as I recall it cost 4x what the DS1054Z cost but it is similar in performance.

I will retrieve the TDS2014B later in the day and run through these same tests. I am interested in a FFT comparison between it and the Rigol.

Will report back later ...

dan W7NGA
Title: Re: Rigol DS1054z FFT bug(s)
Post by: Anand on June 10, 2015, 05:41:10 pm
This caught me at a bad time since I am leaving for S.E. Asia for 7 weeks and just put most of the test equipment into storage.
I worked for Tektronix and have a dozen (or more) scopes here at the lab. I have a Tektronix TDS2014B, a terrible scope when
compared to the Rigol, and as I recall it cost 4x what the DS1054Z cost but it is similar in performance.

I will retrieve the TDS2014B later in the day and run through these same tests. I am interested in a FFT comparison between it and the Rigol.

Will report back later ...

dan W7NGA
Yes! Thank you, I look forward to the results. :D
Title: Re: Rigol DS1054z FFT bug(s)
Post by: Anand on June 10, 2015, 05:49:25 pm
God, look at this crap (the FFT, not the whole scope - the scope's fine, for the most part):
Same signal (1kHz square wave), Average mode acq, 1024x, 24Meg mem depth.
I stopped (single shot) the acquisition.

Pic 1: 200us/div - FFT clean, ok.
Pic 2: I go to 500us/div - FFT clean, ok.
Pic 3: I go BACK to 200us/div - FFT's f***ed. The same signal, the same settings.
Pic 4: 1ms/div - now has aliasing, versus image from the above post witch was clean.
Pic 5: 2ms/div - even more aliasing vesus images form the previous post.
Title: Re: Rigol DS1054z FFT bug(s)
Post by: alsetalokin4017 on June 10, 2015, 07:56:26 pm
It seems to me that the FFT is actually working pretty well, now that I've tried it on some other, higher frequency signals, and looking at it at higher resolution (that is, not displaying out to the 40th harmonic or so!)

These FFTs change very little whether any of the 4 acquisition modes are used, Auto vs. other memory depths, nor if AC or DC coupling is selected.
Title: Re: Rigol DS1054z FFT bug(s)
Post by: W7NGA on June 11, 2015, 02:15:48 pm
I retrieved my Tektronix TDS 2014B (100 Mhz, 1 Gs/s) out of storage and ran the FFT tests on the Rigol 1 khz calibrator.
After using this thing for a few minutes, and remembering what I paid for it, I have to tell you that the Rigol DS1054Z is the best hobbyist bargain ever!

What a jewel!

As shown in the photo, even with Auto memory management, there is no getting around the undersampling artifacts.
Using my 'real' spectrum analyzer and narrowing the RBW I can see even-ordered harmonics but they are greater than -50dB down from the fundamental.

The TDS 2014B cleans up as the sampling rate is increased, as does the Rigol.

Enjoy your DS1054Z .. as I said, I am thrilled and will be purchasing a few more for the classes I teach.

Off to Vietnam ... cheers

dan W7NGA