New 2ch pocket DSO+SG - Zeeweii DSO2512G

[Aldo22]

Too high ripple on 8 and 12 MHz. Try to connect si5351 module with coax cable directly to the device and install pass through 50 Ω dummy load on device side for proper matching.

Done that, but it looks a lot uglier than with the ground spring.
What could be the problem?

EDIT: It looks somewhat better with coax but without the load resistor, but not as good as with the probe.

[tatel]

Can somebody confirm/deny this device doesn't have any antialiasing filter?

I think to make an oscilloscope without any antialiasing filter would be a very stupid thing to do. OTOH I can't see why anyone would make that statement unless it's true?

Anyway, is it good enough to look at PSU ripple/noise? Can work up to 20 MHz?

[Aldo22]

Can somebody confirm/deny this device doesn't have any antialiasing filter?

I think to make an oscilloscope without any antialiasing filter would be a very stupid thing to do. OTOH I can't see why anyone would make that statement unless it's true?

Anyway, is it good enough to look at PSU ripple/noise? Can work up to 20 MHz?

If user radiolistener says it doesn't have an antialiasing filter, I believe him.
I can't judge this as a newbie.

But I'm also not sure if that's important to me.
For me, this device is simply the cheapest handheld 2-channel (toy-!) oscilloscope on the market if you don't want to deal with used devices.
I wouldn't think of comparing this to a decent $500+ desktop oscilloscope.
Even as a layman, I recognize that there must be a big difference.

But that doesn't make the device useless to me.
For my purposes it works so far.
I learned a lot from it, better than from any book.

I do not know what the "usable" bandwidth of it is and how that is defined.
You can see from my screenshots how a 20MHz signal looks with this device.

I'm happy if it shows a decent 5 MHz square wave. I don't expect more for this price.
It has about 5x higher bandwidth than the DSO154Pro.

For me it is a good device for the first contact with the world of oscilloscopes.
Others hate it for reasons I don't fully understand.
I can understand that they ignore it or smile at it, but I don't see how someone can hate it.

[tatel]

I agree with all you said. However I want to know. Call it curiosy if you want. I think there is something worth to be learnt in this matter.

[Aldo22]

As far as PSU ripple is concerned: I don't know how to measure it exactly and what it should look like.
I can only show you a picture of what it looks like when I connect the probe to a USB charger.

[tatel]

Well, first of all, what I'm interested in right now, is aliasing. I want to know more about it and this device could be a good place to look at it. That's interesting in itself. So far I readed this: https://www.asdlib.org/onlineArticles/elabware/Scheeline_ADC/ADC_NumRep_Nyquist.html

Then we have the first derivative. If aliasing makes DSO2512 untrustworthy, even for looking at PSUs, then I wouldn't recommend it again. Analog 20 MHz scopes would still be the way to go for me. 

But I have to put my finger into that wound first.

It's said ripple/noise are usually well under 2 MHz and indeed the three /four SMPSs I have look into, showed it under 200 KHz. Of course, DC is 0 Hz. So, to me, its really difficult to believe aliasing could be making this waveforms false. However switching noise has higher frequency. I'm willing to learn and ready to be proven wrong.

See attachment.  That comes from a cheap 12 V chinese SMPS if memory serves me well. This was done with my benchtop scope, which is unlocked to 100 MHz. You can see the voltage ripple as the caps charge/discharge and the switching noise. Even Zeeweii DSO154 showed remercably similar things, but it's clearly not a 20 MHz BW device.

PSUs can pick a lot of noise almost from everywhere. So it's a tricky task. One advice is to make the first measurement with the PSU turned off so you can then make a second measurement with the PSU turned on. Then you have an idea of what really comes from the PSU.

Since you have, say 150 mV riding of top of 12V, you would use AC coupling and 1x attenuation. Some sort of coax cable with BNC on one side and attached to PSU's output on the other side. You also limit the bandwidth to 20 MHz. PSU output impedance is low so you want also a 50 ohm terminator. You have to be careful about terminator's power rating.  W=V squared/ohms. To look at 12V and above you would use a voltage divider and take that into account. Dave has a video on it. I used a 10W RF dummy load with a T adapter instead.

Of course it's also said it's better done by using some sort of differential means of measurement to get rid of that noise. However differential probes are usually HV while ripple/noise is on the mV range. So one would like to use some sort of LV differential amplifier, which is far from cheap.

You would want also check if PSU overshoot at turning on/off, etc. DC coupling for that, of course.

Generally speaking, you want to make sure that no components will be killed by any overvoltage and that things like ADCs aren't disturbed by the noise.

[radiolistener]

Is this what you mean?

yes

Done that, but it looks a lot uglier than with the ground spring.
What could be the problem?

EDIT: It looks somewhat better with coax but without the load resistor, but not as good as with the probe.

with load resistor it should show just a signal
without resistor it should show a signal + ringing in the cable

But this is true if device has high impedance input 1 MΩ.
If it has low impedance, the result will depends on impedance.

With proper pass through dummy load and direct cable connection it should show almost ideal square wave which form is limited just by device bandwidth.

Is the second image taken without pass through dummy load?
If so, it looks that the device has low impedance input. May be close to 50 Ω. This is not good.

Try to measure device input with a usual DMM. What it shows?
If you have VNA measure device impedance within 0-100 MHz.

Can somebody confirm/deny this device doesn't have any antialiasing filter?

I don't have this device and I didn't tested it, but all such kind of Chinese devices doesn't have it, because there is no bandwidth headroom, so adding anti-alias filter leads to a very small bandwidth and buyers won't buy it due to a small digits on the box. So, I think this one is not exception.

[Aldo22]

@radiolistener: Thank you.
I will test your advices later.
In the meantime, I have taken another foto.
It shows a much cleaner signal using load resistor and probe in 1x mode.
For me as a newbie, I would have thought that the signal would be nicer with the coax, but it's the other way around.
What is the difference between the probe in 1x mode and coax? The coax is only 30cm long.
Everything else including load resistor is the same.
Thanks!

[Aldo22]

Is the second image taken without pass through dummy load?

Yes! But there is no problem with the dummy load with the probes in 1x.

Try to measure device input with a usual DMM. What it shows?
If you have VNA measure device impedance within 0-100 MHz.

Sorry, I don't really understand.
Should I measure the signal generator or the DSO?
What should I measure exactly? Resistance? Capacitance?
I am Noob.

Thank you!

[radiolistener]

For me as a newbie, I would have thought that the signal would be nicer with the coax, but it's the other way around.

yes, if device has proper 1 MΩ it should show the best waveform with direct cable connection and 50 Ω pass through dummy load on the oscilloscope side.

if it is not, there is something wrong with impedance.
Check your dummy load. Measure it with DMM, what it shows?

Also measure your device input with DMM when device is powered on, what it shows?

What is the difference between the probe in 1x mode and coax? The coax is only 30cm long.

1x probe will works as low pass filter and with probe you should get more slow edges.

In that case it can be used as some kind of antialiasing filter. But the probe response is not flat like it should be for antialiasing filter.

What it shows with coax and dummy load on oscilloscope side?


The picture with dummy load and probe is not good, because it has ripple on the top and bottom of square wave, si5351 module output doesn't have such ripple. It has about 1 ns edge, so with proper matched connection you can estimate rise time of your DSO.

[radiolistener]

Should I measure the signal generator or the DSO?
What should I measure exactly? Resistance? Capacitance?

Put DMM on resistance measurement mode and measure DSO input when DSO is powered on. In such way you can see input impedance at DC, it should be close to 1 MΩ.

The same measure resistance between GND and center pin on your dummy load to check if it's ok. It should be 50.0 Ω. If it's not, it is broken.

In order to measure impedance within 0...100 MHz you're needs to use VNA (Vector Network Analyzer) such as NanoVNA.

[radiolistener]

Here is how 10 MHz si5351 with direct cable connection looks on Siglent. There is a little pulse near the edge - this is a rigning, it's position depends on the cable length. In this case cable has about 1 meter, if I remember correctly.

[Aldo22]

Check your dummy load. Measure it with DMM, what it shows?

It has 49.6 Ohm

Also measure your device input with DMM when device is powered on, what it shows?

With my An8008, it shows 0. Measurement does not work this way somehow.

What it shows with coax and dummy load on oscilloscope side?

This ist the first picture on this page. The "ugly" one.

The picture with dummy load and probe is not good, because it has ripple on the top and bottom of square wave, si5351 module output doesn't have such ripple. It has about 1 ns edge, so with proper matched connection you can estimate rise time of your DSO.

I will see what it shows with TinySA.
But we are at 10MHz here. You said my toy is only good for 5 MHz, so this is maybe the limit of the toy-DSO.

Thank you!

[radiolistener]

It has 49.6 Ohm

not the best, but acceptable VSWR=1.008

With my An8008, it shows 0. Measurement does not work this way somehow.

0 Ω ? Or over than 10 MΩ?

[Aldo22]

[0 Ω ? Or over than 10 MΩ?

Ah, now I see! I have to switch the DSO to DC mode, then it shows. 0.995M.

[radiolistener]

What it shows with coax and dummy load on oscilloscope side?

This ist the first picture on this page. The "ugly" one.

The waveform is unstable? This is interesting. I don't see the reason for instability. Can you show the better quality image?

Ah, now I see! I have to switch the DSO to DC mode, then it shows. 0.995M.

it's also ok. It should show nice square waveform with coax cable and dummy load.

Try to take it with DC coupling.

[Aldo22]

The waveform is unstable? This is interesting. I don't see the reason for instability. Can you show the better quality image?

Yes, it's strange because with the probes, it looks OK for me.
It's at 10MHz, so the ripples with the probes are probably due to the limitation of the toy-DSO.

I thought that the coax is probably not the best, but it's so short...
RG-58/U 50 Ohm RoHS

Is this image better?
 

[Aldo22]

Btw. The problem exists not only at 10MHz, it looks also ugly with the coax at 2.02 MHz, for example.
The signal looks also better with the 1x probe at 2.02MHz.
Both are with load resistor.

EDIT: I see that the voltage ist totally different with the coax and the 1x probe.

[tatel]

@aldo22

[4thDoctorWhoFan]

I just received my ZEEWEII DSO2512G from AliExpress.

Is it possible to update the firmware?
If so, where would I find the latest firmware?

Thanks!

[Aldo22]

@aldo22

Thanks, I think I've watched this before, but I'll check it again tomorrow.

[Aldo22]

Is it possible to update the firmware?
If so, where would I find the latest firmware?

In theory yes.
But so far, there are only updates for the 1-channel version: https://jumpshare.com/b/LjQzEy3seDkQSP0D5kAg
If ever, updates for the 2-channel version should appear here afaik: https://jumpshare.com/b/dJ4idH7EgKKI2ykVFNf2

There is also a youtube channel: https://www.youtube.com/@zeeweii5631/videos and a website: http://www.zeeweii.com/

[radiolistener]

It's at 10MHz, so the ripples with the probes are probably due to the limitation of the toy-DSO.

The ripple appears here due to a bad signal matching. This is because you're using probe, that's expected result.

The main issue here is that it cannot show signal with proper 50 Ω matching when you use direct coax cable connection with 50 Ω dummy load on the DSO connector. According to your DMM measurements at DC, impedances are ok. It's still unclear how it works at AC. But at a glance all should be ok. But it shows strange unstable waveform. That's not normal.

But I don't have idea what is the reason for that.
May be some software or trigger implementation bug which happens due to a very sharp edges.
It should not work in that way.

EDIT: I see that the voltage ist totally different with the coax and the 1x probe.

In both cases voltage is incorrect. si5351 module has about 1.5 Vpp.

Probably there is something wrong with impedance at high frequency.

Try to change frequency, the amplitude should be the same at any frequency. If it has significant change, it means that something is wrong with impedance matching.

[Aldo22]

@radiolistener: Thank you for your patience!

May be some software or trigger implementation bug which happens due to a very sharp edges.

Hmm, but then why does it look much better without a load resistor, only connected directly with coax?
Please see the attached images.
2.02Mhz looks almost the same as with the probes and 9.9MHz has some jitter but it looks much smoother than with load resistor.
What doesn't work at all is the combination of coax and load resistor.
The load resistor on the probe also looks good.

In both cases voltage is incorrect. si5351 module has about 1.5 Vpp.

Probably there is something wrong with impedance at high frequency.

Try to change frequency, the amplitude should be the same at any frequency. If it has significant change, it means that something is wrong with impedance matching.

To be honest, this is the first time I'm experimenting with this load resistor.
I tried it with my old signal generator (XR2206 clone), because I can fine-tune it amplitude wise.
How does this work? Do I have to convert the volt measurement by hand?
On my DSO, I can only choose between 1x,10x,100x but not any other calculations.

I have mesaured:

Signal generator settings are: 1KHz sine, 1V PKPK
DMM connected directly to the output of the SG, before the probe/resistor.

Without Load Resistor / Probe 1x:
DSO: PKPK: 995mv / RMS: 320mV
DMM: RMS 330mV

With Load Resistor / Probe 1x:
DSO: PKPK: 46.6mv / RMS: 15.2mV
DMM: RMS 121.9mV

No signal with load resistor in 10x mode.

I don't understand how to interpret the values using the load resistor.
Or do I have a fundamental misunderstanding?

Thank you!

[4thDoctorWhoFan]

Is it possible to update the firmware?
If so, where would I find the latest firmware?

In theory yes.
But so far, there are only updates for the 1-channel version: https://jumpshare.com/b/LjQzEy3seDkQSP0D5kAg
If ever, updates for the 2-channel version should appear here afaik: https://jumpshare.com/b/dJ4idH7EgKKI2ykVFNf2

There is also a youtube channel: https://www.youtube.com/@zeeweii5631/videos and a website: http://www.zeeweii.com/

Thanks for the info. 

FYI, I measured a bandwidth of approximately 35Mhz on my scope in normal mode.
When in overclocking mode, the bandwidth increased to about 40Mhz then the signal dropped off.  However, around 80Mhz the signal level increased and I was able to get to 120Mhz measured & displayed on the internal frequency counter.