Rigol Oscilloscopes With(out) Bode Plot Feature?

[David Hess]

Keysight 3000T also has it, together with variable signal/frequency.. As far as I know Siglent 1000X-E has it too, it also have phase, variable level, and also something no other can which is multiple channels simultaneously... I presume Siglent 5000X will have the same (when it's available, if not already).

My first impulse (1) was to call shenanigans but a review of the Keysight 3000T user manual shows that it does support the needed FFT magnitude and phase function as well as FFT after differentiation so this DSO just moved to the top of my list.  The manual even mentions aligning the trigger so they had this application in mind.  All other modern DSOs may bow down before it.

Based on the Siglent 1000X-E manual, the Siglent does not return FFT phase results.  Get thee behind me, Satan!

(1) Get it?  Impulse?

[tautech]

Based on the Siglent 1000X-E manual, the Siglent does not return FFT phase results.

Correct X-E doesn't for FFT but they do for FRA/Bode plot and up to 3 phase results are available as 2N3055 mentioned.

From a year old post using earlier FW than now available,
https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg1436912/#msg1436912

Bandpass passive filter, ch4 full response, ch2 high pass

[2N3055]

I could see this being interesting to look at several stages of active filter at the same time on single correlated screen.
Or a test of an active crossover and plot 3 frequency bands simultaneously so you can see overlap.

Also, since Siglent uses external SG, in theory you could go quite high in frequency with SDG5000 SDG6000  and SDS5000X....
Another interesting thing about that scope.

Keysight 3000T FRA app is very nice and works real good. It maxes out at 20MHz because of siggen.

EDIT: yes Tautech I meant SDG6000, not SDG5000 . Even more stupid mistake since I have it on the desk 

[tautech]

I could see this being interesting to look at several stages of active filter at the same time on single correlated screen.
Or a test of an active crossover and plot 3 frequency bands simultaneously so you can see overlap.

Also, since Siglent uses external SG, in theory you could go quite high in frequency with SDG5000 and SDS5000X....
Another interesting thing about that scope.

Keysight 3000T FRA app is very nice and works real good. It maxes out at 20MHz because of siggen.

Of course you mean SDG6000X ^ 

Performa01 has the 500 MHz version and of course a pre-production SDS5104X.................

[alsetalokin4017]

...

[2N3055]

I could see this being interesting to look at several stages of active filter at the same time on single correlated screen.
Or a test of an active crossover and plot 3 frequency bands simultaneously so you can see overlap.

Also, since Siglent uses external SG, in theory you could go quite high in frequency with SDG5000 and SDS5000X....
Another interesting thing about that scope.

Keysight 3000T FRA app is very nice and works real good. It maxes out at 20MHz because of siggen.

Of course you mean SDG6000X ^ 

Performa01 has the 500 MHz version and of course a pre-production SDS5104X.................

Well I have the SDG6000 but no SDS5000X... Which, as the time goes by, I itch more and more to compare it to my MSOX3000T to see how it fares.
Unlike specification comparisons, I would like to make direct comparison how it is when doing real world tasks, like measuring, decoding, how good is MSO integration etc etc.
From what I tried quickly, it felt real nice and quick... But I would like to try side to side comparison.

For instance, I have a NFC devkit on my desk now. Keysight 3000T has NFC triggers, but if you capture whole packet it starts undersampling severely. I have to enable peak detect mode to see propper amplitude of signal. There is even a trick that Keysight devised to route trig out (that works at full sample rate) back to second input channel as a decoded signal, and then you use that to decode with Manchester decode.

That kind of thing happens all the time. It would be much better if it had more memory. I use Picoscope quite often because of that.
Having both good responsive scope and long mem would be nice...

[IDEngineer]

Correct X-E doesn't for FFT but they do for FRA/Bode plot and up to 3 phase results are available as 2N3055 mentioned.

Very impressive. I don't need a scope in the range of the 1000X-E series but for this feature, alone, I might buy one. I already have a Siglent SDG sitting right here on the bench, it might appreciate another member of the family.

I understand the difference between SNA's (SA's) and VNA's but if the SA in question is not "real time" (i.e. swept oscillator + mixer) then it's almost assuredly doing (inverse) FFT and the phase data is right there, waiting to be displayed.

EDIT: I presume the firmware is shared across all models of the 1000X-E line? Or do you have to dig through the specs of each model to confirm that each specific combination of bandwidth and channel count have the Bode Plot feature? The 1104X-E is many tens of percent cheaper than its 200MHz 1204X-E sibling and 100MHz should be plenty for a "Bode Plot Instrument" on my bench.

[2N3055]

Keysight 3000T also has it, together with variable signal/frequency.. As far as I know Siglent 1000X-E has it too, it also have phase, variable level, and also something no other can which is multiple channels simultaneously... I presume Siglent 5000X will have the same (when it's available, if not already).

My first impulse (1) was to call shenanigans but a review of the Keysight 3000T user manual shows that it does support the needed FFT magnitude and phase function as well as FFT after differentiation so this DSO just moved to the top of my list.  The manual even mentions aligning the trigger so they had this application in mind.  All other modern DSOs may bow down before it.

Based on the Siglent 1000X-E manual, the Siglent does not return FFT phase results.  Get thee behind me, Satan!

(1) Get it?  Impulse?

Well you are correct 3000T has FFT gain/phase. I believe we had a discussion about that once. In meantime I got the 3000T that got on top of my list too..
Not perfect though, I have to be careful when it starts undersampling, sampling rate drops really quickly because of small mem. But, when poking around unknown signals , I just switch on peak detect. That way I don't miss that there is something there. But, my Picoscope with 512 MPoints gets fair amount of use when 3000T memory won't do...

But we are discussing FRA function, and in that regard, Siglent seems to have good idea how it should look. I didn't test it in detail myself, so no idea how good is code quality, but very good concept.

(1) I had to integrate that charge a bit before I detected the pun...

[2N3055]

Correct X-E doesn't for FFT but they do for FRA/Bode plot and up to 3 phase results are available as 2N3055 mentioned.

I understand the difference between SNA's (SA's) and VNA's but if the SA in question is not "real time" (i.e. swept oscillator + mixer) then it's almost assuredly doing (inverse) FFT and the phase data is right there, waiting to be displayed.

Well, it doesn't have two inputs to synchronously sample both input and output to be able to get accurate phase/gain.

[IDEngineer]

BTW, there's a sort of appnote on Siglent's website illustrating the use of the Bode plot feature. Only one channel for each of amplitude and phase, but still.

http://old.siglentamerica.com/USA_website_2014/Documents/Other/Bode-plot-application.pdf

[IDEngineer]

Well, it doesn't have two inputs to synchronously sample both input and output to be able to get accurate phase/gain.

True. I was thinking of SA's I've used that do have two inputs to do the comparisons (old ham radio back in the day). I actually have a small DDS-based PC-hosted unit for working on antennas and even THAT has two inputs.

In the back of my mind it seems the plot could show phase relative to the start of the sweep, perhaps, and that could be done with a single input. After all, even real time SA's are displaying amplitude and they are only processing a single input, so the displayed amplitude is relative to some baseline - not a real time comparison against another input. For many applications I think a relative phase display would be sufficient.

EDIT: I wonder if this could be done with the math/equation features on a regular scope. My Rigol DS4000 understands phase, for example... I need to dig into what it exposes in math equations. Perhaps it's possible to hack your way there, if you're patient for the math to be processed. Not sure it supports a log X axis but if the SG generates a log sweep you'd end up with the same display format.

[2N3055]

SA has absolute level calibration. Phase is always relative relationship. A timing relationship between two edges of signal. Sweep start means nothing in that regard.

Picoscopes have FRA app written that uses scope as sampler.
I still have to test it in detail at low frequencies with 4262 16 bit scope. With 3406D it works real nice.

[IDEngineer]

Phase is always relative relationship. A timing relationship between two edges of signal.

If it's a swept sine, the zero crossings would give you that reference. You could resync no less frequently than every 180 degrees and accuracy would be dependent upon purity of the original sine and whatever distortion the circuit under test imposed upon its output. The effect of distortion could also affect a true two-input VNA, though for different reasons since it has a realtime reference input.

I'm just saying a modest Bode Plot feature that includes phase could be achieved in this manner. Wouldn't duplicate the results of a Keysight VNA, but wouldn't duplicate its price either.

I'm still impressed that two pieces of equipment from Siglent, totaling under $1K, can give you an amplitude+phase Bode Plot out to tens of megahertz.

[David Hess]

Based on the Siglent 1000X-E manual, the Siglent does not return FFT phase results.

Correct X-E doesn't for FFT but they do for FRA/Bode plot and up to 3 phase results are available as 2N3055 mentioned.

The FFT method works with single shot acquisitions and in real time albeit with lower signal to noise so it is more useful for slow systems and for real time adjustments.

Swept measurements I can do semi-automatically almost as fast as these DSOs can do it automatically so this feature as not as useful for me.

[David Hess]

Keysight 3000T also has it, together with variable signal/frequency.. As far as I know Siglent 1000X-E has it too, it also have phase, variable level, and also something no other can which is multiple channels simultaneously... I presume Siglent 5000X will have the same (when it's available, if not already).

My first impulse (1) was to call shenanigans but a review of the Keysight 3000T user manual shows that it does support the needed FFT magnitude and phase function as well as FFT after differentiation so this DSO just moved to the top of my list.  The manual even mentions aligning the trigger so they had this application in mind.  All other modern DSOs may bow down before it.

Based on the Siglent 1000X-E manual, the Siglent does not return FFT phase results.  Get thee behind me, Satan!

(1) Get it?  Impulse?

I went through the Keysight 3000T user manual in detail and am not quite so impressed now.  It seems like either Keysight deliberately crippled its functionality or whoever decided to add support for FFT phase results did not understand their application and just included them as a "checkbox" item.

But it is a start.

[vr2whf]

https://chipsnwafers.electronicsforu.com/2020/01/18/rigol-expands-its-portfolio-with-customisable-digital-oscilloscopes/

"The BodePlot function has now been integrated into all MSO5000 series oscilloscopes. Both the MSO8000 and MSO5000 models have been expanded with a 12-bit high-resolution mode. "

It seems not an official source.  Anyone can verify it?

[EEVblog]

Jump many years later and Rigol rolls out their new mso5000 series scopes. Hard to resist for the base price, can be loaded up to $5k+ for features (we all know ways around this though), and lacks this basic bode plot feature! Rigol what gives?

More importantly, their new scopes don't have hires boxcar averaging!
I've kind convinced myself (based on zero evidence) that it must have been a goof in their new ASIC and they didn't want to respin it, so they just left out the feature 

[EEVblog]

I'm still impressed that two pieces of equipment from Siglent, totaling under $1K, can give you an amplitude+phase Bode Plot out to tens of megahertz.

The analog discovery does this for 1/4 the price and at higher resolution.

[Sighound36]

https://chipsnwafers.electronicsforu.com/2020/01/18/rigol-expands-its-portfolio-with-customisable-digital-oscilloscopes/

"The BodePlot function has now been integrated into all MSO5000 series oscilloscopes. Both the MSO8000 and MSO5000 models have been expanded with a 12-bit high-resolution mode. "

It seems not an official source.  Anyone can verify it?

I would suggest that the 12 bit high resolution mode is correct for the 5000 & 8000 but as far as the bode plot function goes it's located could be a little beyond me 

[vr2whf]

https://chipsnwafers.electronicsforu.com/2020/01/18/rigol-expands-its-portfolio-with-customisable-digital-oscilloscopes/

"The BodePlot function has now been integrated into all MSO5000 series oscilloscopes. Both the MSO8000 and MSO5000 models have been expanded with a 12-bit high-resolution mode. "

It seems not an official source.  Anyone can verify it?

I would suggest that the 12 bit high resolution mode is correct for the 5000 & 8000 but as far as the bode plot function goes it's located could be a little beyond me 

The news came out 4 days ago. Not sure if Rigol has planned a firmware upgrade to introduce the Bode Plot function to MSO5000.

[Sighound36]

Hello vr2whf

I know Martin72 did make a point of mentioning it to the German Distributor a while back, I have both a 5000 and 8000 series and they have high rez and precision modes, but currently no bode plot facility which in fairness they really should have from the start point of the models inception, however we live in hope 

[IDEngineer]

I'm still impressed that two pieces of equipment from Siglent, totaling under $1K, can give you an amplitude+phase Bode Plot out to tens of megahertz.

The analog discovery does this for 1/4 the price and at higher resolution.

Good point! Might justify Yet Another Equipment Acquisition. "No, really dear, I NEED this." 

[Martin72]

More importantly, their new scopes don't have hires boxcar averaging!
I've kind convinced myself (based on zero evidence) that it must have been a goof in their new ASIC and they didn't want to respin it, so they just left out the feature 

Today (after firmware updates) they got a hires mode or do you mean something different ?

I know Martin72 did make a point of mentioning it to the German Distributor a while back

Yep, statement was, bode plot will come…
But also the next update will come in may (2019)...

[thm_w]

More importantly, their new scopes don't have hires boxcar averaging!
I've kind convinced myself (based on zero evidence) that it must have been a goof in their new ASIC and they didn't want to respin it, so they just left out the feature 

They added it in the last FW, it just doesn't work well at all... eg its say 8->9-bit, not very useful compared to 8->12-bit.
But I 100% agree its probably some ASIC screw up, now they have to figure out how to re-implement it in the FPGA or in software.

Having it in the ASIC HW would have been a great feature. Sampling at 8-10GS/s you'd get ~4x oversampling or more with basically no BW/CPU penalty.
Hell the STM32 ADC's have the same feature of HW oversampling for some years.

[EEVblog]

They added it in the last FW, it just doesn't work well at all... eg its say 8->9-bit, not very useful compared to 8->12-bit.
But I 100% agree its probably some ASIC screw up, now they have to figure out how to re-implement it in the FPGA or in software.

Having it in the ASIC HW would have been a great feature. Sampling at 8-10GS/s you'd get ~4x oversampling or more with basically no BW/CPU penalty.
Hell the STM32 ADC's have the same feature of HW oversampling for some years.

When you spend years developing an ASIC to be one of the "big boys" and compete with the likes of Keysight that have done it in their ASIC for eons, you'd think it'd be a basic requirement.
But then Rigol release several model scopes using the new ASIC and the Hires feature is mysteriously absent, only to now surface more than a year later in a firmware update...