Wednesday, January 8, 2025

New test equipment: Siglent DSO and AWG

I recently decided to step up my test equipment game by getting a digital storage oscilloscope (DSO) and an arbitrary waveform generator (AWG). These will come in handy for fixing stuff, or when making electronic devices. 


The AWG (left) and the DSO (right)

After some research, I chose instruments made by Siglent for four main reasons. First, they are both very capable instruments for their price. The nearest competitor is Rigol, but Siglent beats them in almost all specs and operations, for just a few $ more. 

Second, these devices are made in China. The very stable genius/convicted felon/insurrectionist in clown makeup says he's going to impose tariffs on everything from China. That means the prices of these instruments may rise dramatically in the coming months, so get them now or maybe never.

The DSO is the SDS804X which is a 12 bit (8.4 bit ENOB, SFDR 35 dBc), 4 channel, 2 Gsps, 70 MHz BW scope. Siglent also makes 100 MHz BW and 200 MHz BW versions of this scope with greater memory depth. In fact, all three scopes use the exact same hardware, the only differences are the model number label on the front panel. The higher BW and memory depth models simply enable those features in firmware. 

This brings me to the third reason for buying these devices instead of others. Someone figured out (or leaked from the factory) a keygen that enables you to turn on the wider BW and memory depth so you can turn your 70 MHz scope in to the 200 MHz BW scope. Of course, making the modification voids the warranty, so it's a good idea to make sure everything works as it is supposed to before upgrading. If you had a warranty covered problem with the scope, maybe it could be downgraded back to 70 MHz BW before trying to get service, but who knows if there's a log file somewhere showing the mod and unmod.

The fourth reason for buying two Siglent products is that they can talk to each other via USB or network. That can be useful for certain types of measurements. More on that below.

The scope comes with 4x switchable 1x/10x, "70 MHz" probes. However, the 70 MHz probes are identical to the 200 MHz probes, minus some accessories that come with the 200 MHz probes. So even after the "upgrade" the probes will be fine.

I bought the DSO new, via amazon, and the AWG, used, via ebay. 


The DSO: SDS804x


When I started working as an engineer back in the early 80s, a scope with 1/4 this one's capabilities would have weighed about 80 lbs, and would come on a wheeled cart, and would have cost about 2 year's pay.

What sort of capabilities? The ability to do math on the captured waveforms is a big plus. That includes, among many other built in functions, the ability to do Bode plots (aka frequency response curves), and Fast Fourier Transform that converts time domain data captured by the scope to frequency domain - i.e. spectral - plots. Rise and fall time measurements, signal frequency, etc., with full statistics are just a couple button pushes away. Captured data can be stored on a thumb drive and then imported into Matlab or similar software for more heavy-duty number crunching.

Dual trace analog CRT based scopes could not monitor two points in a circuit simultaneously because the CRT could only provide one electron beam. A two channel scope typically alternated between the two channels, sweeping one trace then the other one the CRT. There was something called "chop" mode that would rapidly switch the sweep back and forth between the two channels to "sort of" monitor two points in a circuit at the same time, but you had to be very careful about interpreting what the scope was showing you in that mode.

OTOH, a single sweep on an analog scope showed the waveform in real time. There were also DSOs that used CRT displays, so they could capture a waveform and display multiple traces on the screen from memory. They suffered from the same limitations that today's DSO do, except these days the memory and A to D converters are much cheaper so today's DSOs use much more of both.

DSOs typically have multiple A to D converters and can sweep multiple channels simultaneously, so you can monitor multiple points in a circuit in "real-time" (within the limits of the sample rate, of course).

Like dual trace CRT based scopes, DSO's require constant vigilance to avoid being fooled by  what the scope is showing you. The Siglent scope always displays the sample rate and memory depth on the screen. That's not just for convenience. The actual BW of a measurement is 1/10th of the sample rate. Though the spec says it's a 200 MHz BW scope, that is the maximum BW it can achieve when it's using 1 channel operating a 2 GSPS and using the full 100 Mpts memory. The sample rate, and BW drop as a function of the memory depth, sweep setting, and the number of channels in use. For example, if I am using one channel and set the scope to 100 ms/div, it runs at 100 MSPS and uses the full 100 Mpts in memory. That's 10 MHz BW, not 200! If I turn on a second channel the memory is split between the two and the sample rate drops to 50 MSPS with only 5 MHz BW in each channel. If you're looking for a transient that lasts a few picoseconds, you're not going to see it.

There's a very good explanation of DSOs vs analog oscilloscopes here.

As the article linked above points out, DSOs can do a lot of math that the old analog scopes couldn't. Rise time, fall time, period, frequency, skew between channels, etc. can all be displayed, and since the numbers are calculated from the waveform in memory, you can set up multiple measurements and display them at the same time. It also presents statistics on the measurements including min/max, mean, standard deviation, etc.

The scope also does dozens of things geared toward factory production of electronic gear that I'll never make use of. The 243 page user manual (and several others) is here.


The AWG: SDG1032x


The AWG is a two channel, 14 bit, function/arbitrary waveform generator that works up to 30 MHz. Another keygen enables it to be bumped up to 60 MHz because it uses the exact same hardware as the SDG1064x AWG. I had some trouble getting the keygen to work on my AWG, but I'll keep trying. In the meantime 30MHz is fine.

The AWG has dozens of waveforms built in, and can use data generated by free software called EasyWave and entered via network to create almost any waveform you might want/need. It also has multiple modulations built in, and the two channels outputs can be combined or synced. The manuals are located here.



Sine wave specs for the AWG. Distortion at audio frequencies isn't spectacular, but adequate for many tests.





So what can you do with this stuff?


As an example, I recently bought an Advance Paris A12 stereo amplifier. Among it's many features are two line level subwoofer outputs with switchable LPF frequencies of 75 and 150 Hz. I have inquired about the slope of the LPF and have been unable to get a response anywhere. So how about measuring it?

If I connect the AWG, the DSO, and the amp together like this:


The test setup. Channel 1 of both the AWG and the DSO are for the reference signal. Channel 2 is for the device under test (DUT-  the amplifier). There are BNC connectors on the AWG and DSO, and phono connectors on the amp.


I can set the DSO to do a Bode plot that will cause the AWG to sweep the signal frequency while it monitors the output from the amp and plots the gain and phase difference between the input and output signals. The AWG is set so that channel 1 and 2 are synced meaning their frequency and phase are locked together- as one changes, the other will change equally, under control of the DSO. The DSO will compare the signal coming through the amp to the signal coming directly from the AWG and plot the differences in amplitude and phase vs. the frequency. 

There's a good explanation of Bode plots here.- I'm going to have to do some studying- it's been a long time since I dealt with this sort of thing.


To verify my settings, I connected the AWG channels 1 and 2 directly to the DSO channels 1 and 2 and ran the sweep. The image below shows essentially zero difference in amplitude and phase as it should. The minor differences seen may be due to differences in the two cables I used- one was a BNC to BNC cable, the other a scope probe, just poked into the AWG output jack, and/or the limited resolution of the scope.


Bode plot of the AWG connected straight through to the DSO, frequency swept from 10 Hz to 5 kHz.

The LPF in the amp can be switched from 75 Hz to 150 Hz, so I ran two tests, one for each LPF setting, sweeping sine waves from the AWG over a frequency range from 10 Hz to 2000 Hz, measuring 20 points per decade. I set the AWG to 100 mVrms out so I wouldn't overload the amplifier, and disconnected the speakers and subwoofer before running any of the tests. I adjusted the volume control on the amp to get the test signal close to the 0 dB line on the Bode plot.

I connected the amplifier using two BNC to RCA (phono) cables with the LPF switch set to 75 Hz and swept again. This time I turned on two cursors (X1, X2) on the amplitude trace as well as 3 dB point, gain and phase margin measurements (P1, 2, and 3).


The 75 Hz LPF test. The DSO says the 3 dB point is 80 Hz (P1, lower left), and cursors say it's at 81 Hz. It probably varies a little depending on where you start. P2 shows gain margin. P3 is supposed to show phase margin- not sure what happened there.




I switched the LPF to 150 Hz and ran the test again.

The 150 Hz LPF filter test. I've turned on cursors and a couple measurements. The cursors show the 3 dB point of the filter is at 166 Hz and the scope's measure of the 3 dB point is 167 Hz. The scope and I probably chose slightly different starting points for the measurement.


The DSO can save the data from these plots as .CSV files like this one from the 75 Hz test:

Instrument Name,SDS824X HD
Serial Number,SDS08A0X806378
Software Version,3.8.12.1.1.3.8
Awg Type,USB
DUT Input Source,CH1
DUT Output Source1,CH2
DUT Output Source2,None
DUT Output Source3,None
DUT Channel Gain,Auto
Sweep Type,Simple
Awg Amplitude,0.1V
Awg Offset,0V
Awg Amplitude reference level,1V
Awg Load,HighZ
Awg Amplitude Unit,Vrms
Sweep Mode,Logarithmic
Start Frequency,10Hz
Stop Frequency,2000Hz
Sweep Line,50
Sweep Log(dec),20
Amplitude Mode,Vout/Vin
Amplitude Axis Type,Logarithmic
Amplitude Axis Range,-92dB,-28dB
Phase Unit,Degree
Phase Axis Range,-120Deg,80Deg
Bode Data
Number of Points,48
Frequency(Hz),CH2 Amplitude(dB),CH2 Phase(Deg)
10,-0.316206946,176.898533
11.2201845,-0.301349145,174.827017
12.5892541,-0.295593542,172.720797
14.1253754,-0.285307544,170.482712
15.8489319,-0.282871402,168.078595
17.7827941,-0.278952734,165.495564
19.9526231,-0.276636798,162.72877
22.3872114,-0.283139497,159.650485
25.1188643,-0.286777113,156.27685
28.1838293,-0.302849299,152.508918
31.6227766,-0.325629345,148.256579
35.4813389,-0.37922284,143.414019
39.8107171,-0.448551886,137.938013
44.6683592,-0.57699543,131.685234
50.1187234,-0.785198164,124.551014
56.2341325,-1.1070296,116.547249
63.0957344,-1.58892266,107.680373
70.7945784,-2.27801731,98.1979734
79.4328235,-3.19725386,88.408361
89.1250938,-4.36847629,78.6493903
100,-5.76128114,69.3966066
112.201845,-7.33886852,60.8239298
125.892541,-9.07129381,53.1396052
141.253754,-10.8753204,46.4291769
158.489319,-12.7822801,40.5519963
177.827941,-14.7305102,35.324053
199.526231,-16.6876153,30.6849254
223.872114,-18.6739893,26.8189145
251.188643,-20.6655185,23.2595056
281.838293,-22.6610596,20.0543472
316.227766,-24.6691554,17.3480928
354.813389,-26.6556163,14.7279327
398.107171,-28.6618495,12.4576434
446.683592,-30.6646106,10.2932372
501.187234,-32.6885167,8.5660686
562.341325,-34.6574611,7.25368276
630.957344,-36.5675565,5.67887063
707.945784,-38.7709119,2.96040434
794.328235,-40.6839971,0.931052533
891.250938,-43.4931255,-4.32564563
1000,-44.6084714,-1.41852283
1122.01845,-46.6796759,-4.34549142
1258.92541,-48.7276961,-5.72869398
1412.53754,-50.9672037,-6.92954931
1584.89319,-52.9628139,-10.5570804
1778.27941,-54.8710069,-11.4217083
1995.26231,-57.0067772,-13.761997
2000,-57.1240924,-13.1864203


From the plots and the data, we can see that the 75 Hz filter hits the -3 dB point at about 80 Hz, and the slope after that is about 12 dB per octave (or 20 dB per decade), making this a 2nd order LPF. The 150 Hz plot and data show the 3dB point to be about 160 Hz, with the same 12 dB per octave slope.

Why does knowing this matter? If you were going to drive a subwoofer that didn't have it's own crossover you'd want to know how much of the higher frequency energy is going to be presented to the sub's amplifier/driver. It would also be very useful if you were building an active crossover to drive a bi- or tri- amped set of speakers. Or if you were restoring old audio gear and wanted to see if the tone controls were behaving right, or wanted to see the limits of their effect. It has many other uses for tuning higher frequency circuits, too.


FFT


Another one of the interesting things the DSO can do is called Fast Fourier Transform (FFT). That is a process whereby time domain data is converted to frequency domain data and plotted. It essentially turns the scope into a crude spectrum analyzer.

I hooked the scope to my CD player and played a steady 1 kHz tone from a test CD, then ran FFT and plotted the result.


Spectrum of a 1 kHz test tone played from a test CD on my VRDS-20 CD player. The first 8 harmonics are listed. Note the sidebands on either side of the 1 kHz spike. Those are 120 Hz offset from the 1 kHz tone, indicating some ripple from the power supply is getting into the output of the CD player.

This is no substitute for a real audio analyzer (or spectrum analyzer) as the DSO noise level is high and ENOB is only 8.4, but you can see power supply ripple in the signal and the relative levels of some of the harmonics. The results also depend on the selected "window" used to make the measurement. Some window types are more accurate for amplitude and others more accurate for frequency.

I will be exploring other measurements and options in the future and my test cable collection expands, and I study and practice a bit more. A DSO can become a hobby all by itself!



No comments:

Post a Comment

Leave comments or a questions here and I'll try to post a response as soon as I can.