Siglent SDS1x04X-E BodePlot II (SFRA) features and testings

[rf-loop]

In some special cases perhaps also this kind of principle is possible?  (as my nickname tell I'm more RF.... )
(of course it need well known ref Mic etc...  and very carefully istalled so it do not mess Microphone under test and it get as perfectly same as Mic under test get.)

Thanks for pointing this out. It is of course viable, but now requires an calibrated measuring microphone, which makes the setup even more expensive. Just like reference sound pressure sources, measuring microphones do not have a perfectly flat response, but are calibrated, i.e. their response is precisely known and the corresponding correction can be applied to the measurement results.

EDIT: I think your point is that the sound pressure source needs not be calibrated, but it still needs to be high quality to come close to a point shaped source without partial oscillation of the membrane to give a uniform sound field.

Very interesting. It appears that you're suggesting that the test should use a live capsule on the microphone. I often test the mic prior to installing a capsule. So the mic would have a ~72pF dummy load instead of the capsule, and then I insert a signal directly to the mic circuit.

Shouldn't this be significantly simpler to test the circuit that way?

Thanks,
Josh

First your question about my images editing.
Because these are technical pictures, example instrument display copy (never use jpg, jpg is for photographs)  and need simple things - like txt, some highlights or color balance etc and of course it have quite nice advanced batch conversion/adjust tools. Fast and easy but still can do many things and it is not mammoth like some photographers Photoshop or GIMP etc.
I have used over 20 years Irfan Skiljan's nice Irfanview.  For this kind of simple technical images purposes it have today lot of tools. It is today, and have been long long time,  lot of more than just viewer what is was over 20 years ago when IrfanView was launched.. So this name is bit fun today, but name have history. In beginning it wast just more like simple viewer, at this time world was bit different.

It is not suitable for real The Photographs editing. Camera Photographs are other things.
(today with Irfan you still can do really meany things.. (Irfan + plugins package))



But then return to the actual subject.

So you are testing microphone without microphone.
I think this do not belong to this thread at all. It belongs under topic something like "how to measure microphone frequency response"

If you are testing microphone amplifier circuit freq. response alone - then you need tell you are testing this kind of electronic circuit (explanation).  Microphone is receiving acoustic waves in air. (because you talk this kind of microphone. Of course, the medium in which mechanical vibrations are transmitted can be something other than air.) Normally when we talk microphone (music, speak etc sounds) frequency response we talk this frequency response how it move acoustic waves in air to electric signal. Other question is after then how these electronic circuits handle this microphone "capsule" signal. These circuists are not "microphone". and if you talk alone these amplifier / buffer circuits response you do not talk microphone frequency response. So these kind of things very easy mess whole discussion and soon, as also here now, we are like out from BodePlot II things.

Microphone real freq response measurement is imho really different case. It need sound. For other than crap result it is quite complex case and need - hard knowledge how to do it and also instruments for do it. There also BodePlot II can use.

If you  want only test this microphone amplifier without microphone itself (capsule) you can of course do it also using BodePlot II. But now it looks your problem is not at all BodePlot II things. You need perhaps help how to do this test setup between your circuit and  SFRA system. Please open new topic for it and give deep detailed information about your this DUT.  Schematics and specifications what ever are available. Perhaps then someone(s) can try help how to do it. Includind also perhaps some tips how to then finally do BP II settings. But first you need some help for do hardware test setup for this.  First you need good setup.  BPII settings are then in final case.
 
For your self learn how to use BPII itself:

I recommend you do some very very simple circuit, just example RC filter.
Then run BPII with it and exercise and try understand all settings what they are and how they affect what you get on display. Also including understanding level scales. Adjusting display when BPII is running
 
Sidenote:(Siglent need improve this vertical autoscale so that user can select it or use last user adjusted vertical and remember it if stop and adjust example horizontal or other parameters. Now it is frustrating to adjust these scales every time again after start sweep because it always goes to autoscale and least I have not find how to deny vertical autoscale for level and phase draw.)

I highly recommend to firtst exercise with known circuits if it is not clear how to use SFRA (BodePlot II). For base level exercise/learn  how to use BPII , imho, it is best to use known circuit what result we already roughly know and simplest circuit is example 1st or 2nd order RC low pass (or something like it, band pass, high pass etc). If you want simplest possible DUT and frequency can be "what ever" because it is only for learning. You can use setup (if you are sure you can set SDG for tracking mode with equal outputs so that CH2 is tracking CH1 with 0 phase, 0 level and 0 freq difference.)

Connect CH1 to scope CH1 (as DUTin) (coaxial, 50ohm termination)

Connect probe 1. BNC end  to SDG CH2 Output BNC.
Connect probe 2. to scope channel (example CH2) as DUTout.
Connect probes GND wires together. Connect probe tips together. Set both probes for 1x.
These probes are now your DUT.  This DUT is low pass filter, corner somewhere well under 10MHz

If you now want look lot of lower freq things.
Just keep this setup but take example 100ohm resistor and 100nF capacitor. Connect capacitor other end to probes GND joint. Open probes tips joint. Connect capacitor other end to probe 2. tip. Connect also resistor other end to this joint.   Then connect resistor other end to probe 1. tip.
Start exercise using BPII default settings and set sweep logarithmic and start from example 1kHz to 10MHz.
Of course you can do some other filter or cascaded filters.
Here can find nice tools for calculate these and also get calculated responses. So you can see how close practice and theory goes when you look your BodePlot II results. (of course calculus are made using ideal components what do not exist at  all in real world, never. Even DC voltage exist only in theory books but not in real world. Never.)

But I hope we can see this "microphone what is not microphone" testing things goes to under some other new topic or other appropriate thread.

[KungFuJosh]

I'll try those simple tests and eventually start a new thread if I decide to keep the SDG. This seems to be more hassle than it may be worth.

BTW- That was somebody else that asked about your graphics. My day job is web and graphic design. I'm well versed in the adobe suite, and I have irfanview too.

Thanks,
Josh

[rf-loop]

I'll try those simple tests and eventually start a new thread if I decide to keep the SDG. This seems to be more hassle than it may be worth.

BTW- That was somebody else that asked about your graphics. My day job is web and graphic design. I'm well versed in the adobe suite, and I have irfanview too.

Thanks,
Josh

Sorry, it was @Electro Fan who ask this.

Btw, I do not understand what "hassle" there is. SFRA works just as SFRA works. Even if you buy big brand SFRA machine there is very important to think and understand what things machine do not for you.  User must arrange a suitable test connection setup between DUT and Analyzer. User is responsible about test signals are right and DUT and test instrument see suitable and good quality signals.  Some times it need also thinking, knowledge and experience. First before all, need fully understand how test instrument works and what are all things there. I want easy, immediately all for me just as in movies and stories. It do not work like this.
I have been in situations where need first design test setup. In some cases it need lot of work.  Finally after long long work and design etc... then power up test instrument and push three - five buttons and turn some knobs and after some minutes there is nice result on screen. If someone look when start instrument and how easy all goes to final nice graph on display... then he may think why company pay to you this salary. Just because he did not know what was 99.99% of this work what final and result he simply see, one dirty graph on the screen..

It is because every time you measure something (connect something, probe, wire or what ever) to DUT you change DUT. More or less. Yes in many simple low freq audiuo things this must not care so much except perhaps if go to some state of art things. Some times it is major problem and sometimes it may need deep knowledge and hard design etc to know how to arrange probing and how to know how much it affect DUT and how much it bend result what we hope is as accurate truth as possible.   Never overlook how much need do work for good reliable test setup, its is perhaps most important (and some times frustrating) thing when do some testing.
Turning instrument knobs is most easy peace of this cake. And even in this, doing instrument settings, many even experienced peoples are messing because they do not know how instrument is designed to work. So, also knowledge need include - always: "know your tools and instruments". And this is what need exercise and study and it IS hard work and it need do or if not  then just better stay with toys and playing some simple things. There is no free lounges. Home works need do or if not then some  day all feels "difficult" and messy. Look some videos around internet and you know what I mean...but it is not appropriate to talk about it here.

[tautech]

This ^^^ and so much all this !

Practice with simple pass filters and get to understand BPII. It does take time but this investment in time and X-E plus AWG is much less than other equipment that has similar capability.

[mrprecision]

I have tested the Bode Plot Function of my scope (SDS1104X-E) and have been very enthusiastic about it so far. I connected the scope to the generator (DG2042X) via Ethernet, this worked without any problems right away. However, there seems to be a problem with the Bode Plot function.
In the simple example I tested a passive filter of the 1st order with 100Hz at -3dB. The sweep for the Bode Plot function was chosen from 10Hz to 1kHz with 10 points per decade. "Channel Gain Auto" was selected under the channel settings.
Then the Bode Plot measurement was started. The first run is as expected. However, the first point in the Bode Plot is incorrectly determined in all further runs. In my view, this is a mistake. However, if you select "Channel Gain Hold" under the channel settings, this incorrect measurement does not occur. It also appears to occur only at low start frequencies.

Video
http://www.schaeck.eu/download/eev/siglent_bode_2_error.mp4

[Pitrsek]

Any clue as to why is the low limit set to 10Hz?

[tautech]

Any clue as to why is the low limit set to 10Hz?

Offer a valid use case for under 10 Hz and we'll send this request to Siglent.

[n3mmr]

Any clue as to why is the low limit set to 10Hz?

Offer a valid use case for under 10 Hz and we'll send this request to Siglent.

For really-high-quality audio power amplifiers, a bode plot of the range 0.1 Hz to, say, 100 Hz will enable a designer to accurately predict the audibility of artefacts coming from attempts to limit the vlf range by steep slope high pass filtering of subsonics.

[mrprecision]

I have tested the Bode Plot Function of my scope (SDS1104X-E) and have been very enthusiastic about it so far. I connected the scope to the generator (DG2042X) via Ethernet, this worked without any problems right away. However, there seems to be a problem with the Bode Plot function.
In the simple example I tested a passive filter of the 1st order with 100Hz at -3dB. The sweep for the Bode Plot function was chosen from 10Hz to 1kHz with 10 points per decade. "Channel Gain Auto" was selected under the channel settings.
Then the Bode Plot measurement was started. The first run is as expected. However, the first point in the Bode Plot is incorrectly determined in all further runs. In my view, this is a mistake. However, if you select "Channel Gain Hold" under the channel settings, this incorrect measurement does not occur. It also appears to occur only at low start frequencies.

Video
http://www.schaeck.eu/download/eev/siglent_bode_2_error.mp4

Back to my Question: But why is the already existing function only working for the first run?

[Performa01]

I have tested the Bode Plot Function of my scope (SDS1104X-E) and have been very enthusiastic about it so far. I connected the scope to the generator (DG2042X) via Ethernet, this worked without any problems right away. However, there seems to be a problem with the Bode Plot function.
In the simple example I tested a passive filter of the 1st order with 100Hz at -3dB. The sweep for the Bode Plot function was chosen from 10Hz to 1kHz with 10 points per decade. "Channel Gain Auto" was selected under the channel settings.
Then the Bode Plot measurement was started. The first run is as expected. However, the first point in the Bode Plot is incorrectly determined in all further runs. In my view, this is a mistake. However, if you select "Channel Gain Hold" under the channel settings, this incorrect measurement does not occur. It also appears to occur only at low start frequencies.

Video
http://www.schaeck.eu/download/eev/siglent_bode_2_error.mp4

Back to my Question: But why is the already existing function only working for the first run?

Yes, there's a timing problem whenever the vertical gain of the measurement channel exceeds 1V/div.

There is a base range up to 100mV/div that doesn't require an input attenuator.
For the vertical gain settings up to 1V/div the first 20dB attenuator is activated.
For the vertical gain settings up to 10V/div the scond 20dB attenuator is activated on top of that.

These attenuators are relay operated and quite obviously the 2nd attenuator causes additional time delays that are not allowed for in the Bode plot firmware.

Knowing this, the workaround is simple. Just lower the stimulus level by 3dB, so that the vertical gain of the measurement channel (Ch.2) never exceeds 1V/div. With this setting you can have glitch-free measurements over more than 100dB dynamic range (the first attenuator doesn't pose a problem):


SDS1104X-E_Bode_LP100Hz_10-10MHz+3dBV

In the screenshot above, the 100Hz lowpass filter is measured up to 10MHz. As can be seen, my filter works as expected up to some 2MHz, then there is a resonance with a maximum attenuation of about 100dB and after that, the attenuation is decreasing again. The phase plot also clearly documents this.

Numerous passes have been made and the data at 10Hz have always been correct. Of course I've also tried lower spans, including the setting that you've used (10Hz to 1kHz). It all works fine.

[Performa01]

Any clue as to why is the low limit set to 10Hz?

The Bode Plotter uses AC coupled input channels, which simplifies operation significantly. Even though the corner frequency is below 3Hz, the phase shift is significant even above that frequency and slight differences between channels (component tolerances) cause additional phase errors in the measurement channels with regard to the reference channel. This is the reason why 10Hz has been chosen.

In DC coupled mode, any DC offset can easily ruin the dynamic range. The DC offset might be too high to be compensated by the internal offset DAC, and even if it is within range, this would cause additional effort.

Knowing all this, I've still requested a lower frequency limit of at least 0.1Hz a long time ago, but it seems that Siglent doesn't see much benefit in spending the effort to implement this, since the practical use cases are limited indeed.

One Argument is that the Bode Plot (that is not fast anyway) would get painfully slow at frequencies that low: at 100mHz, a single measurement takes at least 10 seconds, filtering/averaging not even factored in. This means, if Siglent ever implements that, you need to be aware that this option is for tests running over night only

[mrprecision]

These attenuators are relay operated and quite obviously the 2nd attenuator causes additional time delays that are not allowed for in the Bode plot firmware

I understand the problem with the 2nd attenuator, you think that could be fixed by Siglent. Why are no additional time delays are allowed for the bode plot firmware?

[rf-loop]

These attenuators are relay operated and quite obviously the 2nd attenuator causes additional time delays that are not allowed for in the Bode plot firmware

I understand the problem with the 2nd attenuator, you think that could be fixed by Siglent. Why are no additional time delays are allowed for the bode plot firmware?

I belive there is wrong word (allowed). Perhaps it is something like "additional time delays are not been adquately addressed"  Lets hope some future FW fix this.

[rf-loop]

One Argument is that the Bode Plot (that is not fast anyway) would get painfully slow at frequencies that low: at 100mHz, a single measurement takes at least 10 seconds, filtering/averaging not even factored in. This means, if Siglent ever implements that, you need to be aware that this option is for tests running over night only

^ This.

@mrprecision
SDS BodePlot works using frequency selective "receiver" principle (receiver center freq sweeps together with generator sweep, bit like SA with tracking generator. Even more, BP also adjust RBW during sweep related to current freq. It adjust receiver center during sweep and also it adjust Resolution Band Width automatically proportional to frequency) and naturally this part is very slow when go to very low frequencies specially with sub Hz selectivity because...
Try play under 10Hz and down to under 1Hz with good resolution using FFT...  we just meet limits of this kind of systems. 0.5kEur oscilloscope can not be everything. Still its SFRA performance is far over any others in this <1kEur price group (scope + generator) and even in bit higher class. With these limits we need live.
And of course hope they find solution for this attenuator delay time better or what ever is this reason. It have been quite long time this problem has been found.

[mrprecision]

It seems to be a timing problem with the 2nd attenuator.

Secound run with 6Vpp:


Secound run with 0.5Vpp

[Pitrsek]

The phase difference between channels should be taken care off during calibration, that's not a reason at all. Drop in the injection voltage could be compensated on the generator side.
Actually I believe that they really should include a DC coupling mode, especially if this thing is coming to scopes with 50R inputs(most of the scopes that I worked with had 50R input DC coupled).
And with 50R inputs, there is full new world of S parameters and impedance measurements... 
Yes, measuring low frequency takes time, that's give fact. I'm not asking for making it a default settings, I'm asking for a possibility to go lower, if the user wants to. 
Use case is rather simple, try measure open loop gain of an opamp... For a lot of slower opamps, the dominant pole is rather low.

[Performa01]

The phase difference between channels should be taken care off during calibration, that's not a reason at all. Drop in the injection voltage could be compensated on the generator side.

Calibration in a DSO takes care of the runtime differences between channels, which in turn lead to a continually increasing phase shift proportional to the input frequency. A DSO cannot compensate for the phase errors caused by slight differences in the corner frequencies of the AC coupled input path.

What you’re describing, sounds much more like the user calibration of a VNA test setup. In contrast, Bode Plotters don’t do a calibration/normalization run.

Actually I believe that they really should include a DC coupling mode, especially if this thing is coming to scopes with 50R inputs(most of the scopes that I worked with had 50R input DC coupled).

Yes indeed, all Siglent DSOs beyond the SDS1000X-E entry level have switchable 50 ohm inputs, starting with the SDS2000X-E. And yes, these inputs are either DC or AC coupled, regardless of the selected input impedance.

And with 50R inputs, there is full new world of S parameters and impedance measurements... 

Once again you are talking about a VNA here – and apart from that I don’t think I have ever seen s-parameters used below 10Hz

Seriously though, you don’t suggest that a cheap DSO should be able to replace a VNA with s-parameter test set, do you? Or – to get back to the <10Hz discussion – a Dynamic Signal Analyzer?

The fact that the Bode Plot in Siglent DSOs is more than just a me-too feature, hence can be used for some serious work in certain areas, cannot hide the fact that it is just a software module in an entry level (or midrange, in case of the SDS5000X) DSO, hence it's got to have some limitations.

Yes, measuring low frequency takes time, that's give fact. I'm not asking for making it a default settings, I'm asking for a possibility to go lower, if the user wants to.

You need not convince me. As stated earlier, I’ve already requested this quite a while ago, because I also know use cases for this, like very slow high precision control loops.

The problem is that Siglent has lots of projects and limited resources, so they need to prioritize things. Extending the range of the Bode Plot certainly isn’t high on their list.

In fact I constantly need to promote and defend the Bode Plotter, especially against consultants coming from the higher end industries, because the Bode Plot is generally seen as a hobbyist feature, or for one man workshops at best. I fight for it, simply because I was once young and on a budget myself and want hobbyists to get something useful for the most common use cases without having to reach out for some half-broken ancient boat anchor.
 
Under these preconditions, I’m glad that we still got the Bode Plot II and I’m positive it will see further improvement - just not as a high priority.

Use case is rather simple, try measure open loop gain of an opamp... For a lot of slower opamps, the dominant pole is rather low.

While this sounds like a use case, I don’t think many will do that.

The characterization of chips in general, like operational amplifiers, is the task of the semiconductor manufacturer (who certainly won’t use an SDS1004X-E for this) and they publish the results in the data sheets. So why not just use that?

The only other scenario I could think of would be the incoming components inspection in the industries using that chip, but once again I cannot imagine them using a cheap DSO for the task. And they will most likely not care for the corner frequency in the open loop response, but just measure the frequency response at the intended gain instead.

In general, measuring the open loop response of an OpAmp is far from being straight forward and consequently not common at all.

[Pitrsek]

We've started to mix two thing together,  let me separate that.
10Hz - I consider the limit to be bit to high, I'm not telling it's useless. I think it could be improved. Open loop gain was an example because that's what I'm doing now. But in general any slow loop would benefit from the expanded bandwidth - sometime the dominant pole is really low. The open loop measurement is actually rather simple and straight forward. If you have inverting configuration, measure voltage at virtual ground (pre-amp is needed) and divide by opamp output voltage. If it's no-ninverting configuration, differential amplifier is needed to extract opamp input voltage.
You'd be surprised how many vendors do not bother to state the details of the open loop gain... I think it's a bit of a chicken and egg problem.  It's not done regularly because you do not have the needed equipment, and it kinda works most of the time(until it does not...). Yes, it's a bit of niche at this time.
It's a same as loop gain measurement - look at the amount of literature and info that's available now. It's not like the theory has changed or that the issues were not there 20yr ago, it's just that tools were not as widespread as are now. Now every scope mfgr now has an appnote about loop gain measurement.
In my previous jobs, before we bought impedance/loop gain analyser, there was not much general interest in impedance profile of discrete parts, or loop measurement. Then we bought analyser, and everybody saw how useful it is. The company bought another one to keep up with demand from engineers... Once you make it accessible, then everyone is like "how could i lived without it"?

I'm not suggesting that you should be able to replace your VNA with a budget scope, I'm suggesting that with some extra effort, the bode plotter could be made much more useful. And most of the required stuff is already there!
Personally I consider impedance profile measurements to be extremely useful - again, some manufacturers do not provide models/data, sometime the bias effect are missing, or you'd like to know output impedance of power supply, or a filter. There are a lot of good reasons why you'd like to do impedance sweeps.  Yes, you have a hard limits in bandwidth and dynamic range and you'll never have the same performance as dedicated equipment, but for quite some measurements it could well be enough. 

Even if it's a software module, i believe it's extremely useful. And if calibration routines are added, so you can make impedance sweeps, it could be pretty awesome. I'm well aware that full calibration routines are not trivial, but it's no rocket science either(especially the normalisation).

TLDR - it's great feature, I'm glad it's there, it could be improved to make it quite a bit more useful.

 

[Pitrsek]

Actually since they do sell VNA, they ought to have a decent grasp about the calibration routine. Probably different team, on different floor, but the know how is there. Just saying, It could be a nice intro in the impedance measurements for users on budget. And it would be useful...
You'd like to have a bigger dynamic range and higher frequency limit? Here, there's our VNA.

And this would conclude my 2cents of unsolicited free advice 

[mrprecision]

I had contact toSiglent, they will fix the bodeplot 2 bug with the next firmware version.

[mrprecision]

I have tried out to measure the impedance vs. frequency. I was surprised how good this worked out. After the first tries I have build a connection box for better results. I have measured than a 1uF Wima capacitor. This turned out very good and I have compared it with the results of a Rode&Schwarz network analyzer. For privat use the results are very good up to 20MHz. If I set the probes correct than the y-axis scaling is also correct in dB. You only have to calculate Z=10^(dBvalue/20) and you get the value in ohms. Someone tried this also out?

See the results:

[khutch004]

Any clue as to why is the low limit set to 10Hz?

Offer a valid use case for under 10 Hz and we'll send this request to Siglent.

For really-high-quality audio power amplifiers, a bode plot of the range 0.1 Hz to, say, 100 Hz will enable a designer to accurately predict the audibility of artefacts coming from attempts to limit the vlf range by steep slope high pass filtering of subsonics.

People building bass guitar amps and high pass filters for them are interested in evaluating the very low frequency response. They often use an HPF to prevent the cone excursions from exceeding safe limits if subsonic information appears at the amplifier inputs. Bass cabs often walk a fine line between being able to generate a high output level at 30 to 60 Hz and blowing up when presented with low frequency noise and often some help from the amp or a filter in front of the amp is needed.

A better question to ask is the 10Hz low limit arbitrary or is it driven some important cost or reliability concern?

If it was just an arbitrary limit then why NOT extend it as far as possible?

People building filters for some scientific applications will want to test at sub 10Hz frequencies. There may not be terribly many such people but again if the limit is arbitrary then it should be removed.

[mrprecision]

I have tried out to measure the impedance vs. frequency. I was surprised how good this worked out. After the first tries I have build a connection box for better results. I have measured than a 1uF Wima capacitor. This turned out very good and I have compared it with the results of a Rode&Schwarz network analyzer. For privat use the results are very good up to 20MHz. If I set the probes correct than the y-axis scaling is also correct in dB. You only have to calculate Z=10^(dBvalue/20) and you get the value in ohms. Someone tried this also out?

Because some users asked me, here are pictures of my connection boy for the impedance measurement with the Bode Plot II function. Has anyone tried to measure the impedance vs. frequency with the Bode Plot II function?

[rf-loop]

Nice work.

---
About new SDS2000X Plus Bode Plot.
SDS2kX+ have 50MHz internal signal generator. Of course with it can not do setups what use SDG with 2 output settings.

I have not SDS2000X Plus and here in China all who knows are now spring holidays.

Who can check if this model can use internal signal generatorfor BP use only without FG license installed. Even with scopes with external SAG USB generator do not need SAG  license if use this only for BodePlot.
If use SDS2kX+ internal signal gen only for BP (BP controls it automatically) do it work without FG licence.

[mrprecision]

About new SDS2000X Plus Bode Plot.
SDS2kX+ have 50MHz internal signal generator. Of course with it can not do setups what use SDG with 2 output settings.

I have not SDS2000X Plus and here in China all who knows are now spring holidays.

Who can check if this model can use internal signal generatorfor BP use only without FG license installed. Even with scopes with external SAG USB generator do not need SAG  license if use this only for BodePlot.
If use SDS2kX+ internal signal gen only for BP (BP controls it automatically) do it work without FG licence.

I use the Scope SDS1104X-E and the generator SDG2122X and I don't need any licence for the BodePlot and FunktionGenerator. I connect them over ethernet.