SDS800X HD Actual Use Cases

[mawyatt]

OK, this thread is only for actual use cases for the Siglent SDS800X HD, the Rigol DHO800X use cases can start another independent thread so as to keep the "Mines better than yours" pissing contest away!!!

Also along with the only for actual use cases above, this is not about finding bugs, perceived bugs, manual confusion/discrepancies, user wants/wishes, my brand X can do too (better than yours), and so on, it's about actual use cases of the SDS800X HD where one actually uses the DSO for measuring/displaying something useful

Hopefully some potential DSO users will find the expected following use cases valuable, and not have to sift thru all the BS flying around when it comes to these new 12 bit low cost DSOs!

We'll start off with something that just came up.

Using a precision CMOS pulse generator to create a precise 0 to 5 volt low frequency pulse we noted some peculiarity on the pulse rising edge. To view this we needed to use the Zoom feature to expand the horizontal and the vertical offset to look into the waveform details of interest at 10mv/div while using a 5 volt offset. This is a fairly large range for the offset with a 10mv/div vertical scale factor. The waveform was bouncing around some due to the high and low frequency noise present, so we employed the math function to average the Zoom waveform not the usual Channel waveform (nice ability), to reveal the average of the Zoom and vertically expanded waveform. The pulse output is delivered directly to the DSO BNC without any cabling, and also deliverd to a KS34465A DMM with a ~350mm twisted pair. Power is from a cheap 12VDC SMPS Wall Wart.

First (#27) is the waveform as mentioned connected above, second (#28) is with leads to DMM reversed!!! Note the "ringing waveform" introduced by just flipping over the Dual Banana connector on the DMM!!!

Next we powered the pulse generator with a lab power supply, and repeated above. #29 is with DMM connected normal and #30 is reversed!! Note the lower level of "ringing" compared to Wall Wart Power Supply as expected, however some remnants of the "ringing" still appear!!!

Why this behavior, this is left as an exercise for the interested readers

Anyway, hope this thread proves out useful with others contributing actual use cases, and folks being able to witness meaningful examples without all the BS. Would be nice to include the new Rigol (and others) but fear this will just turn into another pissing contest as we've all seen in other threads

Edit: Here we go again with this site messing up all the images

Best,

[shabaz]

With high-res, new ways are needed to see information that older 8-bit 'scopes didn't need.
Your 'scope traces reveal a particular need and your solution, which was the need to do "something" about the noise, because especially when zoomed-in, a single capture in some cases can be almost meaningless, and the solution in your case was the averaging. It's not something one normally realizes unless they actually try it for real, which goes to show that real-life examples are extremely useful.

Does the 'scope support other possible solutions? Averaging is for sure a good solution, but there may be cases where 'something else' is preferred, to not get rid of peaks and troughs for instance, but to still see the trend without as much noise. I'm thinking are there color grading or other options present in the zoom view, that would also be interesting to see for this use-case.

I don't have this 'scope (or the Rigol), but am curious what capabilities these instruments contain.

[mawyatt]

Good point about using the advanced "Display" capabilities for revealing more useful information contained within the waveforms, something we're sure others are much more proficient with than us.

We took the above case using the Display Modes "Infinite Persistence" with "Dots" capability. Here you can see the discrete 12 bit quantization levels with the vertical scaled to 5mv/div with 4.98V Offset

First #33 is with Color Grade OFF, second #32 is with ON, last (#34) is with Persistence changed to 30 second Color Grade ON.

Best,

[Antonio90]

I'm quite interested as to why the polarity reversal on the DMM causes the ringing.
Could it be the input capacitance of the DMM, charged by the previous pulse? Or is that a completely stupid gamble?

Edit: stupid gamble. I don't know.

[mawyatt]

Yes, it's the highly asymmetrical input capacitance of the DMM +- inputs when measured to ground reference

The DMM - input has significantly more capacitance to ground than the + input, so when the pulse generator is reverse connected its' output "sees" a much larger effective capacitance thru the DSO input ground.

Best,

[Antonio90]

Yes, it's the highly asymmetrical input capacitance of the DMM +- inputs when measured to ground reference

The DMM - input has significantly more capacitance to ground than the + input, so when the pulse generator is reverse connected its' output "sees" a much larger effective capacitance thru the DSO input ground.

Best,

I didn't know that at all. Thank you very much!

[mawyatt]

If you have a bench DMM try measuring the +- input capacitance to ground and report over here:

https://www.eevblog.com/forum/testgear/dmm-input-capacitance/

Best,

[Njk]

BTW that use case shows that a floating scope can't be always substituted with an isolating transformer.

[mawyatt]

Here's another interesting use case with the SDS800X HD. If one wishes to see a low frequency plot of capacitance vs frequency then an expensive LCR meter is usually required. There were various threads about using the built-in Bode Function to plot Impedance and Admittance using a DSO, see these for details.

https://www.eevblog.com/forum/testgear/capacitive-impedance-plots-with-sds2104x-plus-bode-function/msg4335745/#msg4335745

https://www.eevblog.com/forum/testgear/admittance-measurements-with-dso-awg-with-bode-function/msg4491952/#msg4491952

Following the analysis in the above threads, the Bode Plot is defined as:

Bode Function Display = Vo/Vi where Vi is the signal created by a coupled signal generator, and Vo is the voltage across the DUT.

If we add a series impedance Zr with the DUT (Z), then (neglecting DSO sensing currents) the DUT current is:

I = (Vi-Vo)/Zr, where Vo is the output or voltage across DUT, and:

I= Vo/Z, and equating I:

Vo/Z = (Vi-Vo)/Zr

Z = Zr(Vo)/(Vi-Vo)

If Vo is << Vi (meaning Zr >> Z), then:

Z ~ Zr(Vo/Vi)

Normally we make Zr a Reference Resistor, but what if we make Zr a Reference Capacitor, what then?

Z ~ Zcr(Vo/Vi), where Zcr is reference capacitance impedance and what if we are interested in Z being just capacitance, thus:

Zc ~ Zcr(Vo/Vi), where Zc = 1/jwC and Zcr = 1/jwCr

1/jwC ~ (1/jwCr)Vo/Vi, and:

C ~ Cr*/(Vo/Vi)

So if we can "fool" the Bode Plot and swap the Ch1 (Vi) with Ch2 (Vo), then:

C ~ Cr*Bode Plot!!

Of course we can just set the Bode Plot Input and Output to Ch2 and Ch1 respective and not need to reverse the probes and "fool" the Bode Plot!!

Ok, now that we've got thru all that, does it work!!

Yep, sure does

Here's an example of a Precision Polystyrene 1uF Reference Capacitor (1.01594uF) and a 100uF Electrolytic Capacitor.

The DUT Capacitance is directly displayed in dB to the Reference Capacitor, so the resulting DUT Capacitance is the product of the display and Reference Capacitor. Note how the effective DUT capacitance begins to drop off with increasing frequency. Results agree well with Lab Bench LCR Meter (TH2830).

DUT Capacitance result is Cr * 10^(Reading/20)

Not a replacement for a plot displaying LCR meter, but in a pinch might help.

Anyway, hope some folks find this useful.

Best,

[mawyatt]

Added a few more plots.

#37 is a 470uF Electrolytic that measures 472.6uF (TH2832) @ 100Hz, with C Reference (1.01594uF) is Cr*10^(53.28/20) or 468.7uF @ 100Hz.

#39 is 1000uF Electrolytic that measures 997.8uF and plot shows Cr*10^(59.84/20) or 997.4uF @ 100Hz.

#41 is 47uF Electrolytic that measures 44.8uF and plot shows Cr*10^(33.12/20) or 46.0uF at 100Hz.

Edit: Technique also works with inductors. Since the inductive impedance is direct rather than inversely proportional to value there's no need to "invert" the Bode Plot. Follow the above analysis and insert inductive impedance rather than capacitive to show how this works. Of course one needs a Reference Quality Inductor which is usually rare, so we just used a small SMD 2.4mH as Lref, see #43. Note the Reference Impedance must be >> than the DUT Impedance for the "assumption" mentioned above to work, so a larger value inductor is required for Lref, rather than the smaller value Cref when plotting capacitance.

Anyway, details left for the interested readers and here's an example with a 100uH inductor using a 2.4mH reference. Inductor measured 100.5uH with T2830 at 10KHz and plot shows 96.0uH. Better result should be expected with quality reference!!


Best

[mawyatt]

Here's a few uses with the Bode plot showing filters.

These are interesting RC filters that exhibit a "greater than unity gain" yet have no inductors, nor transformers, transmission lines, switches, non-linear elements, or active devices, just a couple resistors and a couple capacitors.

One is a Low Pass and the other a High Pass, and both show ~0.66dBV gain. Probably not much use but interesting anyway, analysis left up to the reader

Included some LTspice Sims for reference.

BTW a single JFET also exhibits a Passive Voltage Gain without any additional components, nor power supply!!

Gotta love this little SDS800, puts a grin on our face every time we use it

Will add more later if folks are interested.

Best

[joeqsmith]

I wonder if you can't make PDN measurements with it?   I tried using the NanoVNA and get ok-ish results.  Do they support anything like this native?

[mawyatt]

Would expect so, as it's already been shown how to make Impedance and Admittance measurements utilizing the Bode Function.

One could use a current probe for the current measurement for V/I or I/V plots, but also a simple small sense resistor works. The 12 bit resolution, low noise, and large dynamic range (range scaling) come into play here.

If Siglent makes a firmware update that allows the math channels to be part of the Bode Function, then this opens up a lot for direct use cases without need for additional support probes/components/equipment.

As we noted on numerous occasions, contrary to some naysayers, this Bode Function is extremely useful in the hands of knowledgable users

Best,

[KungFuJosh]

this Bode Function is extremely useful in the hands of knowledgable users

...and me too. 😉

[mawyatt]

Here's a couple plots of the classic passive Twin T Filter, just a few resistors and capacitors for a ~47dB notch

Fun stuff indeed 

Best,

[tszaboo]

Would expect so, as it's already been shown how to make Impedance and Admittance measurements utilizing the Bode Function.

Best,

I've been thinking about a PDN measurement for a while with this scope. I think a low output resistance preamp would be needed for it. And a DC block. And a preamplifier with specific input impedances. And some way to calibrate the whole system. And some script to de-embed the results from the measurements. It's not going to be easy.
It's also kind of difficult to use the results. They would be limited to ~200MHz. Systems that define PDN impedances are typically high speed digital, where wider band measurements would be needed. Of course you can connect it to the Arduino and show that it's a bad design, but what is the design requirement for a microcontroller like that?

[joeqsmith]

I had made up some standards down to 100uohms.   The NanoVNA just doesn't have good enough performance to resolve it.  Still, $50.   

I think I documented everything if you go to my dropbox and select the NanoVNA directory, you will find the software users manual.   

One problem with the NanoVNA when making these measurements is you really have to slow things down.   It seems like a sweep would take me several minutes to run.   The other problem is it was limited to about 50kHz.  You could run it lower with different firmware but the performance was too poor to be useful.   

If you are looking for actual use cases as your thread suggests, and want to try it out, I'm interested to see this measurement  with the Siglent.   You may have to write some software to do it, but then again, wouldn't be surprised if it didn't already have it.

***
I did make a video demonstrating the basics and goes over some of the problems.   Also shows using the software to tame a PDN.  The three hurdles with that whole setup were breaking the ground loop, blocking the DC and standards to verify performance.   

[mawyatt]

Would expect so, as it's already been shown how to make Impedance and Admittance measurements utilizing the Bode Function.

Best,

I've been thinking about a PDN measurement for a while with this scope. I think a low output resistance preamp would be needed for it. And a DC block. And a preamplifier with specific input impedances. And some way to calibrate the whole system. And some script to de-embed the results from the measurements. It's not going to be easy.
It's also kind of difficult to use the results. They would be limited to ~200MHz. Systems that define PDN impedances are typically high speed digital, where wider band measurements would be needed. Of course you can connect it to the Arduino and show that it's a bad design, but what is the design requirement for a microcontroller like that?

At present the upper end is limited to 120MHz, maybe Siglent can extend this if the AWG is capable (SDG6000X for example).

Here's a couple links showing the Bode Function involving Injection Locking, which shows how well the selectivity of the Synchronous Sampling/Detection or whatever algorithm Siglent has utilized behaves. Impressive IMO and could benefit other uses where large interference is present.

https://www.eevblog.com/forum/testgear/siglent-sds800x-hd-12-bit-dsos-coming/msg5441336/#msg5441336

https://www.eevblog.com/forum/projects/injection-locked-peltz-oscillator-with-bode-analysis/

Best recommendation we can provide is get one and play around, you can always return it, however we'll wager you'll keep it and it will put a grin on your face as well

We are a very well seasoned semi-retired research scientist/engineer with a fairly extensive background. Grew up with nothing put HP/AG/KS and Tek equipment over the years, so it takes quite an instrument to impress us, the SDS800X HD has succeeded admirably!! We have a DHO814 as well, it's impressive and quite a good little DSO, but the SDS is just better is almost every way and much more "Pro Like".

Best,   

[joeqsmith]

Plan to buy a new low end one at some point just to try out.  The 800 has a smaller display than my tablet.  It's difficult anymore to see these small screens without glasses.   Doesn't appear the 800 supports an external monitor as well.   Does look like a nice scope otherwise and it's certainly affordable.

[mawyatt]

I had made up some standards down to 100uohms.   The NanoVNA just doesn't have good enough performance to resolve it.  Still, $50.   

I think I documented everything if you go to my dropbox and select the NanoVNA directory, you will find the software users manual.   

One problem with the NanoVNA when making these measurements is you really have to slow things down.   It seems like a sweep would take me several minutes to run.   The other problem is it was limited to about 50kHz.  You could run it lower with different firmware but the performance was too poor to be useful.   

If you are looking for actual use cases as your thread suggests, and want to try it out, I'm interested to see this measurement  with the Siglent.   You may have to write some software to do it, but then again, wouldn't be surprised if it didn't already have it.

***
I did make a video demonstrating the basics and goes over some of the problems.   Also shows using the software to tame a PDN.  The three hurdles with that whole setup were breaking the ground loop, blocking the DC and standards to verify performance.   

That looks like an interesting project, however software isn't our bag. Also seems this will likely get quite involved, so we'll let others more capable than ourselves tackle this.

Maybe you might consider acquiring a SD800X HD, doubt you'll be dissapoined

Best,

[mawyatt]

Plan to buy a new low end one at some point just to try out.  The 800 has a smaller display than my tablet.  It's difficult anymore to see these small screens without glasses.   Doesn't appear the 800 supports an external monitor as well.   Does look like a nice scope otherwise and it's certainly affordable.

We are 76 and bad eyes (cataracts and such), but can read the DSO reasonably well. The remote Web server works really well and isn't too slow and we use this often.

Best,

[joeqsmith]

Maybe tszaboo will make an attempt at the PDN project. 

Another subject that had recently came up again was measuring SRF of an inductor.  I show measuring it with the cheap VNA but then also with a scope.  Again a pretty good actual use case for this particular scope.  Even if you can't drive a sharp edge, you can certainly square it up easy enough.  Your only dealing with a single component and much easier to setup and measure.   

https://www.eevblog.com/forum/beginners/measuring-the-self-resonant-frequency-of-an-inductor/msg5112405/#msg5112405

I'll see if I can find a video review showing that remote interface.   

[mawyatt]

Another use for this DSO and the Bode Function is for regulator PSRR measurements. This is a measure of the regulators ability to attenuate unwanted signals/fluctuations from the input to the output.

We have a couple popular +-5 volt linear regulators, the 78L05 and 79L05 respectively.

The setup is rather simple with a series R (10 ohms) for the input power source and shunt C (0.1uF) for stability connected to the regulator input. Also a larger coupling C (100uF) from the regulator input to the Bode Signal source AWG output, this C provides DC isolation for the AWG. One could also use a coupling transformer common for use in closed loop measurements, but pay attention to the DC thru current as it can saturate the transformer core.

The regulator output is loaded with a shunt R (1K) and shunt C (0.1uF) for stability.

Bode connections are CH1 to the Regulator Input and CH2 to the Regulator Output. This will produce a dB ratio of output to input voltage and thus a negative dB representation. PSRR is generally specified as a + dB term, so one can swap the Bode inputs either physically or thru the Bode Configuration if necessary, however we prefer the negative connotation as this indicates an output less than input more of a Gain type display.

Anyway, here's couple plots showing the 78L05 (+5V) #2 and 79L05 (-5V) #3 PSRR. Note the inferior rejection of the negative 79L05 wrt the positive 78L05, both setups were identical except for polarity. The "response peaks" are partially attributed to setup (Protoboard), altho the 79L05 still appears worse in that respect.

For those interested the PSRR is influenced by input-output voltage differential and load current, so one could create a "family" of curves if desired.

Edit: Added a 78L15 #4 and 79L12 #5.

Best

[mawyatt]

Now that we've got the Regulator PSRR, how about taking a look at the Regulator Output Impedance Z as a function of Frequency. We can apply the Bode function for this as well!!

Configure the Regulator setup with decoupling C on Input and Output, use a large value (100uF) for the Input and small for the Output (0.1uF) so we can view the Regulator Output Z as a function of Frequency.

Attached the AWG output to the load Resistor (1K) normal Ground end and DSO CH1 to same end, then DSO CH2 to Regulator Output which connects to the load R.

If we call the AWG output as V2 and the Regulator Output as V1, then:

Z (Regulator Output Impedance) = V1/I, where I is the R load current.

I = (V2-V1)/R

Z = V1/((V2-V1)/R), or R(V1/(V2-V1))

If V1 is << V2, implying R is >> Z, then,

Z ~ R(V1/V2)

Since Bode is in dB, and Y axis is 20Log(V1/V2)

Then Z is R* 10^(Bode Display)/20

Does it work??

Yep sure does

Here's a plot #6 of a 78L05 output Z using a 1K load. Just add 20Log R (60dB) to Y axis, and then it's in dB ohms.

So at 100Hz the 78L05 Output Z is 13.8 milliohms, and 69.5 milliohms at 100KHz. Note the inductive nature (+ Slope) of the output Z beginning at ~10KHz.

Plot #7 is the 79L05 negative Regulator, 11.5 milliohms at 100Hz and 4 ohms at 100KHz 

Anyway, this seems like another useful capability of these new "Entry Level" DSOs!!

Best,

[Performa01]

Since Bode is in dB, and Y axis is 20Log(V1/V2)

It is worth mentioning that Bode Plot doesn't have to be in dB. It can also use Vpp and Vrms.