Siglent SDS2000X Plus

[Martin72]

If there are two, switch the probes between them.
You could also "short-circuit" the two inputs and then look again.
If the 100Mhz "peak" has then almost disappeared, you are catching an interference at 100Mhz.

[tautech]

Why EasyScopeX and not the webserver ?

[MathWizard]

My scope is connected w/USB, so I'm guessing the webbrowser won't work.

I thought I was on full memory depth as well. I was on 2M, and if I go to 200M, the 100MHz signal is gone, So it's just something in the math it does, and/or with the FM radio signals right next to it.

I'll have to watch out for that when changing timebase and whatever the mem depth does.


After calibration, my CH2 noise floor seems to be jumping from around what CH4 was in the around -80's dBV, to down in the -90's

W/ just CH1/3 on, Ch1 seems pretty stable around -90's, but CH3 is jumping from -100's to -90's. Is that normal ?  The FFT has always seemed to vary every few screen updates, I always figured it's related to the PSU and the mains voltage level.

[Performa01]

Due to the architecture of modern high-speed ADCs, it is not at all uncommon to have spurious signals from subharmonics of the sample frequency.

If we want to avoid spurs, it’s a good idea to take advantage on the deep memory and the long FFT that the SDS2000 provides. Setting the max. mem depth to 20 Mpts instead of just 2, thus increasing the acquisition sample rate from 400 MSa/s to 2 GSa/s and the FFT sample rate to 1 GSa/s, we get a fairly clean result, see attached screenshot.


SDS2354X_Plus_Spurs_500MHz

This measure lowers the noise floor (because the noise above 200 MHz is no longer included) and shifts the lowest ADC-clock related spurious frequency up to 500 MHz.

[tautech]

My scope is connected w/USB, so I'm guessing the webbrowser won't work.

The webserver is somewhat easier to use but it’s best if it’s connected via a LAN switch where you only need specify the scopes IP and enter its IP into a web browser.

You can use either and the webserver doesn’t need additional SW or drivers for connectivity.

[MathWizard]

In case you miss it, I added

After calibration, my CH2 noise floor seems to be jumping from around what CH4 was in the around -80's dBV, to down in the -90's. Every few screen updates, the whole thing jumps.


W/ just CH1/3 on, 200mV/div ,Ch1 seems pretty stable around -90's, but CH3 is jumping from -100's to -90's. Is that normal ?  The FFT has always seemed to vary every few screen updates, I always figured it's related to the PSU and the mains voltage level.

[MathWizard]

I still can't get the Bode plot working. I have my SDG hooked up over USB and it tests OK, I set both SDG CH's to 50 Ohm, and both on the scope to 50, and 1x, before selecting CH coupling, and on the scope going into BP.

My DUT is certainly not 50 ohm's anything, it's a differential mixer, I worked out the output impedance, it's not 50 of course. But I'm not entering it's transfer function into my scope.

You can see my settings below, I have a BNC M2M from CH1 to CH1, and a BNC to alligator clip from AWG CH2 to an input cap of the mixer. And the output is right on the output collector, with another BNC to alligator clip, from scope CH3. That cable might be a bit dodgy, I'll swap them and see if I can find another one.

I had the scope set to AC coupling, so no output cap on the collector, shouldn't matter.

But shouldn't that be working, without me using external terminators, or feed through terminators ? I'm getting more used to the names of all this stuff, but just trying to buy some good cables today on Digikey, IDK what else I really need for just the Bode Plot.

I have an RF textbook that's ok, there's lot's in there on coax and transmission lines, plenty to do there, but I expected the Bode plot to work, IDK what I'm over looking still.


Also on a legit site like Digikey, what's the difference between a $3 1GHz 50 terminator, and a $20 one ?? Maybe some other numbers, but will one wear out way quicker I wonder ? Same with BNC RG-58 cables, are the cheapest models on Digikey, just going to break in a few months or 1-2 years of normal use ?

[Peter_O]

I have a BNC M2M from CH1 to CH1, and a BNC to alligator clip from AWG CH2 to an input cap of the mixer.

I'd suspect the scope to control only one AWG channel and would connect AWG CH1 to SDS CH1 as well as to DUT input.

[MathWizard]

I've done it like this, w/o the 50ohm FTT, I guess this picture is for my 1204X-E which doesn't have internal 50R to GND.

[2N3055]

I still can't get the Bode plot working. I have my SDG hooked up over USB and it tests OK, I set both SDG CH's to 50 Ohm, and both on the scope to 50, and 1x, before selecting CH coupling, and on the scope going into BP.

My DUT is certainly not 50 ohm's anything, it's a differential mixer, I worked out the output impedance, it's not 50 of course. But I'm not entering it's transfer function into my scope.

You can see my settings below, I have a BNC M2M from CH1 to CH1, and a BNC to alligator clip from AWG CH2 to an input cap of the mixer. And the output is right on the output collector, with another BNC to alligator clip, from scope CH3. That cable might be a bit dodgy, I'll swap them and see if I can find another one.

I had the scope set to AC coupling, so no output cap on the collector, shouldn't matter.

But shouldn't that be working, without me using external terminators, or feed through terminators ? I'm getting more used to the names of all this stuff, but just trying to buy some good cables today on Digikey, IDK what else I really need for just the Bode Plot.

I have an RF textbook that's ok, there's lot's in there on coax and transmission lines, plenty to do there, but I expected the Bode plot to work, IDK what I'm over looking still.


Also on a legit site like Digikey, what's the difference between a $3 1GHz 50 terminator, and a $20 one ?? Maybe some other numbers, but will one wear out way quicker I wonder ? Same with BNC RG-58 cables, are the cheapest models on Digikey, just going to break in a few months or 1-2 years of normal use ?

Like Peter said, scope is controlling CH1 of AWG. Scope CH1 and DUT input are connected in parallel and to CH1 of AWG. CH2 of AWG is not used.

You could, though, go into AWG settings and set channels locked. In which case AWG CH2 would faithfully follow AWG CH1 as scope puts it through it's paces..

[MathWizard]

Yes on my AWG, CH1 is copied onto CH2, at that point only CH1 can be set, and CH2 does the same.

[2N3055]

As sanity check we go to basics.

Connect a BNC to BNC coax from AWG CH1 to CH1 on scope and Another one from AWG CH2 to CH3 of scope.
Run a sweep that should show 1:1 transfer function. Small discrepancies at higher frequencies are possible.

Once you verify that works well, then we know it is connections to DUT that are problem..

[Performa01]

I still can't get the Bode plot working. I have my SDG hooked up over USB and it tests OK, I set both SDG CH's to 50 Ohm, and both on the scope to 50, and 1x, before selecting CH coupling, and on the scope going into BP.

My DUT is certainly not 50 ohm's anything, it's a differential mixer, I worked out the output impedance, it's not 50 of course. But I'm not entering it's transfer function into my scope.

Did you work out the input impedance as well? I assume it is not 50 ohm either.

The first thing to do is making sure that the port impedances match. For low frequencies and if the input is high impedance, then you should set the AWG output to high-Z, even though this will not affect anything but the displayed output level, and even more importantly, also set the reference channel on the DSO to high-Z.

For frequencies above some 10 MHz, proper port matching should be aimed at. Like in the real world of professional HF, appropriate transformers / matching networks have to be added to the inputs and outputs of the (mixer) module. From my experience, depending on the architecture a mixer is not guaranteed to work properly if its port impedances are not properly matched.

The output port cannot be loaded with 50 ohms and a length of coax cable (which will add capacitance) if it is not 50 ohms itself. If a proper match cannot be obtained, then a low capacitance (preferably active) probe is to be used. Yet I think it is better and also much cheaper to design the output for proper 50 ohms impedance by means of transformers or matching networks in conjunction with diplexers to ensure wideband impedance matching.

The next issue is that any mixer has two input ports – what is connected to the 2nd port? If it is just DC, i.e. the mixer is used as a voltage controlled attenuator, then you can do the Bode Plot just like with any passive DUT.

If the mixer is actually used as a mixer, with some oscillator signal into the 2nd port, then the output frequency will be different than the input frequency and Bode Plot cannot work because it uses a frequency selective detector.

This is where the dual output AWG comes in handy, because you can set a constant frequency shift between the two channels, making sure that the expected output frequency of the mixer is the same as the reference frequency used for Bode Plot.

[MathWizard]

Yeah I just started thinking about 50R attached to the circuit, and of course even if my scope really has a DC blocking cap, 50R is killing my signal. I've been calculating some impedance matching stuff lately, and 50 ohms always "scares me" because that would normally be a lot of current if across Vcc. But it went over my head IRL.

1 cable is not very good either.


I'll try a passive filter, see if that works.


For the mixer I have input impedance eqn's too, at DC it's near 1750R. I've done max power transfer and quality factor and stuff, but yeah, I guess I'm forgetting what that means IRL.


But if the scope is looking at the input and output, can't it figure out the gain and phase without caring about the impedance attached ?

[tautech]

While not the 2kX Plus, there is much we can learn about Bode plot here:
https://www.eevblog.com/forum/testgear/siglent-sds1x04x-e-bodeplot-ii-(sfra)-features-and-testing-(coming)/

[KungFuJosh]

Is 0.01V enough to test the DUT?

[MathWizard]

Well impedance mismatch aside, now I have something that looks reasonable. When I turn on the BP operation button, the scope only turns on AWG CH1, and NOT CH2 aswell. It seems if I turn CH2 on before starting the BP, it stays on, and I get something reasonable.

On the AWG, all I've been doing is going to Ultility/CH Copy Coupling/ and turning on Tracking, channel coupling is all default 0's. So shouldn't the scope turn on both channels ?


And no 10mV was not enough for that loading. Right now the scope is on 1M w/ the BNC coax cables, and now too much voltage, at 0.8Vrms input.

[rf-loop]

Well impedance mismatch aside, now I have something that looks reasonable. When I turn on the BP operation button, the scope only turns on AWG CH1, and NOT CH2 aswell. It seems if I turn CH2 on before starting the BP, it stays on, and I get something reasonable.

On the AWG, all I've been doing is going to Ultility/CH Copy Coupling/ and turning on Tracking, channel coupling is all default 0's. So shouldn't the scope turn on both channels ?


And no 10mV was not enough for that loading. Right now the scope is on 1M w/ the BNC coax cables, and now too much voltage, at 0.8Vrms input.

BodePlot control only AWG Ch1

And AWG, even when AWG Ch2 Track AWG Ch1,  AWG Ch2 on/off do not follow AWG Ch1 on/off.

Bode Plot  do not know anything about AWG Ch2. It do not even know it is there. And in some cases it is nice it do not at all touch to Ch2.

(also it do not know if AWG Ch2 have frequency or level offset to AWG Ch1  and user settable offset is very important - mandatory -  in some BodePlot applications. Example with frequency offset  BodePlot do not know that the DUT changes the frequency, example mixer... DUT can even be receiver... DUT in is antenna in... and DUT out is example 1st IF.
In some cases it is also nice that other AWG channel have constant amplitude and other follow BodePlot variable Level sweep.
There is so many possibilities.
This Siglent BodePlot with external AWG really gives many possibilities when the user learns to bend it to his own needs. Of course, it should also (sometimes when other things don't override this) be further developed to be better and more developed - like so many other things.

[MathWizard]

So am I supposed to use a splitter, like a T, and would I put that at the AWG CH1 or scope CH1,  shouldn't really matter if they are good parts ? and then to DUT input ?

So what is that siglent picture I posted of ?? Is that incorrect for the SDS2000X+ ?

I'll try this with my 1204X-E, if that doesn't work, 

[mawyatt]

This SDS2000X+ ability to Offset the Bode Function wrt Phase, Frequency and/or Amplitude utilizing an external Dual Channel AWG (like SDG2042X or 6022X) is so powerful in the hands of Knowledgable users.

The SDS2000X+ Bode implementation is a Masterpiece feature and is so useful and powerful!! A tribute to the core developers who obviously knew what they were doing with extensive "hands on" experience and what one might expect from the likes of Keysight, R&S, Tektronix and LeCroy in their full featured and expensive instruments.

The SDS2000X+ Bode function is somewhat slow, but very powerful and feature rich. Suspect the HD version with 12bit ADCs and faster processing is even better, and hopefully so for the newer upcoming SDS800 and 1000HD!!

And Siglent keeps improving these DSO/MSO features

Best, 

[MathWizard]

So is this upper picture incorrect for the 2000x+ ??

[mawyatt]

Here's an example that may help clarify how to setup and use the Frequency Offset function with the Bode Plot.

We are using a very simple "Mixer" with a 2N3904 NPN transistor. This is biased with 5VDC thru a 100 ohm resistor to the collector, the emitter is ground. From the collector a 10K resistor is connected with the other side to a shunt to ground capacitor, this forms a simple low pass filter which becomes the "Output" signal. The 2N3904 base is connected to a 1K resistor to a squarewave generator (SDG6022X) with 2VPP at 1MHz called LO or Local Oscillator. Another resistor (10K) is connected to the base thru a series capacitor to CH2 of the AWG (SDG2042X) called the DUT "Input" Signal.

This configuration forms a simple mixer where the transistor is gated ON and OFF by the squarewave (LO) and thus binary multiplying or "mixing" the sine wave Input from the AWG Ch2. This should translate the frequency of the Input signal by the LO Frequency.

The AWG (SDG2042X) is setup with Channel Coupling Frequency Deviation of 1MHz and Ch1 is connected to C1 of the DSO, C2 is connected with a scope probe (10X) to the Output mentioned above. Bode Decade Sweep is set to 100-100KHz with 31 steps/Dec at 2V amplitude.

What should happen is the AWG frequency of Ch2 should be frequency offset by the Deviation setting (1MHz) and this is the Input Signal to the Mixer DUT, so as the Bode Function sweeps the AWG from 100 to 100KHz from AWG Ch1 as seen by DSO C1. AWG Ch2 is frequency Offset by 1MHz is the DUT Input, and the Bode C2 senses the DUT Output as being swept from 100 to 100KHz where the actual Input Signal is swept from 1MHz +100 to 1MHz +100KHz.

The DUT Mixer with an LO of 1MHz, then displays a Low Pass Filter output "from 1MHz" as shown below.

As also mentioned one can do Amplitude Offsets and Phase Offsets with similar setup!!

Very nice feature set built within the superb implementation of the SDS2000X+ Bode Function when operating a like AWG!!

Anyway, hope this helps understand how the SDS2000X+ Frequency Deviation Offset can be applied with highly non-linear devices like "Mixers" are employed as DUT devices, and produce frequency translation.

Best,

[Martin72]

BTW,

An application note about the Bodeplot II:

https://www.batronix.com/files/Siglent/Oszilloskope/SDS1000X-E/SDS1000X-E-BodePlotII.pdf

[mawyatt]

Here's a little expansion of the Bode Frequency Offset mentioned above. We've added two more channels to the Bode Function, so now using all 4 channel capability of the SDS2000X+.

The setup is the same 2N3904 based "Mixer" mentioned above, and we've added 2 additional outputs. Another Low Pass Filter @ ~500Hz (other @ 1.6KHz) and a Bandpass Filter @ 10KHz implemented with a LC, so 2 LP filters and BP Filter as outputs.

The LO and Offset are 1MHz. What should happen in the input signal is swept from 100 to 100KHz from the AWG Ch1, AWG Ch2 is Frequency Offset by 1MHz, so it's output sweeps from 1MHz + 100Hz to 1MHz + 100KHz as setup with the AWG (SDG2042X) Channel Coupling Feature with the CH2-CH1 Frequency Deviation setting of 1MHz which "Offsets" Ch2 by the Frequency Deviation setting, which causes Ch2 to sweep 1MHz above Ch1.

With the LO (SDG6022X) set to 1MHz, this causes the 2N3904 to switch (Mix) at 1MHz and frequency translate the Input from AWG Ch2 which is being swept from 1MHz + 100Hz to 1MHz + 100KHz per the Bode Settings (100Hz to 100KHz). The 2N3904 collector output is translated back down by the "Mixing Action" to 100Hz to 100KHz and sensed by the DSO channels, where DSO C1 is directly sensing the AWG CH1 setting (not frequency translated), C2 senses with 10X probe the 1.6KHz LPF, C3 senses the 500Hz LPF and C4 senses the BPF output.

Quite a powerful and useful implementation of the Bode Function, as this example demonstrates

Best,

[mawyatt]

Here's the setup.

Best,