New 2ch pocket DSO+SG - Zeeweii DSO2512G

[tatel]

No one needs to see a sine wave.
The criterion for reliability is the square wave.

Well, whatever. But a real useful signal, say a 3 Vpp or a 5 Vpp one. Not a strange corner case, please

[Aldo22]

Hi Tatel

I don't have a signal generator that is anywhere near 5 volts at high frequencies.
Maybe 4thDoctorWhoFan can help?

Do you have an official guideline on how to accurately measure bandwidth on oscilloscopes? Also regarding the amplitude?

Cheers

[tatel]

Do you have an official guideline on how to accurately measure bandwidth on oscilloscopes? Also regarding the amplitude?

All I'm able to do can be watched on the videos about DSO154. I recorded how the sine wave behaves to see where the -3dB point is. Then I recorded the square wave to see up to where it's good.

Cheers

[Aldo22]

I don't understand what you want.
You say "So, to be able to assess what DSO2512 can actually do, what we need is a video showing how a known good 5 Vpp sinewave looks at different frecuencies."

Why 5 Vpp why not 20 Vpp?

If this is important to you, then show an official, authoritative definition of how bandwidth is measured with oscilloscopes.
Maybe we can find someone who has the necessary equipment to measure under these conditions.

For me personally, it's not that important.
If I feel like measuring 100MHz @20Vpp, I'll buy a 100x probe, then I will have 200mV at the input.
The DSO2512G supports 100x.

The question with the amplitude instability interests me more.

[tatel]

I don't understand what you want.
You say "So, to be able to assess what DSO2512 can actually do, what we need is a video showing how a known good 5 Vpp sinewave looks at different frecuencies."

Why 5 Vpp why not 20 Vpp?

If this is important to you, then show an official, authoritative definition of how bandwidth is measured with oscilloscopes.
Maybe we can find someone who has the necessary equipment to measure under these conditions.

For me personally, it's not that important.
If I feel like measuring 100MHz @20Vpp, I'll buy a 100x probe, then I will have 200mV at the input.
The DSO2512G supports 100x.

The question with the amplitude instability interests me more.

Man, 5 or 20 volts, wouldn't matter, as long it's over that hundred of milivolts range that seems to be where DSO2512 is doing funny things. It souldn't. It should be the same bandwidth no matter signal amplitude. At least up to those 40 V that seem to be what it can take at the input.

To me, if 4thDoctorWhoFan says he measured a 35-40 MHz bandwidth, just one channel activated, with an HP device(IIRC), that's enough to me. It adds up. I don't really need anything more. I'm not the guy wondering why he's seeing unstability on DSO2512 at 70 MHz. To me the answer is: it's unreliable at that bandwidth. As simple as that. Just saying.

In the same way, if 4thDoctorWhoFan says the bandwidth drops like a stone if signal amplitude is 1 Vpp or more, that's it. Why that bandwidth would be greater if input signal is in the -7 dBm range, remains to be seen. But I don't really care. Most microcontrollers are working at 3,3 V right now. Arduinos work at 5 V. Only thing I look at hundred of milivolt range is PSU ripple/noise, and that's usually under 200 KHz.

I don't know what that bizarre thing at -7 dBm range could be, but if it can't be done over 1 Vpp, then it can't be taken seriously, I would say. After seeing that screenshot, it doesn't really seem to me it's doing it so well anyway.

[4thDoctorWhoFan]

3) 4thDoctorWhoFan reported he measured 35 MHz bandwidth with just one channel activated IIRC. That means about 17,5 MHz with both channels, isn't that right? With overcloking, it goes to 40 MHz, again only one channel activated, so I guess it could do 20 MHz with both channels activated. This is with a 1Vpp signal, it isn't?

The 35Mhz bandwidth was with a 1Vrms signal which is about 2.8Vp-p.
I then dropped the signal to .1Vrms (.282Vp-p) and measured a bandwidth of about 97Mhz.

I mentioned in a previous post how I did not expect such a big difference in bandwidth depending on the input signal level.

[tatel]

3) 4thDoctorWhoFan reported he measured 35 MHz bandwidth with just one channel activated IIRC. That means about 17,5 MHz with both channels, isn't that right? With overcloking, it goes to 40 MHz, again only one channel activated, so I guess it could do 20 MHz with both channels activated. This is with a 1Vpp signal, it isn't?

The 35Mhz bandwidth was with a 1Vrms signal which is about 2.8Vp-p.
I then dropped the signal to .1Vrms (.282Vp-p) and measured a bandwidth of about 97Mhz.

I mentioned in a previous post how I did not expect such a big difference in bandwidth depending on the input signal level.

OK I stand corrected. 2.8 Vpp not 1. But bottom line remains the same.

[Aldo22]

Hi 4thDoctorWhoFan!

The 35Mhz bandwidth was with a 1Vrms signal which is about 2.8Vp-p.
I then dropped the signal to .1Vrms (.282Vp-p) and measured a bandwidth of about 97Mhz.

Do you mean 1Vrms at the source or at the input of the oscilloscope?
At the input, with a probe in 10x mode (in 1x mode you wouldn't get over a 6 or so MHz) or a Load Resistor, that would be 10 Vrms at the source?
Or am understanding that wrong?

[4thDoctorWhoFan]

Hi 4thDoctorWhoFan!

The 35Mhz bandwidth was with a 1Vrms signal which is about 2.8Vp-p.
I then dropped the signal to .1Vrms (.282Vp-p) and measured a bandwidth of about 97Mhz.

Do you mean 1Vrms at the source or at the input of the oscilloscope?
At the input, with a probe in 10x mode (in 1x mode you wouldn't get over a 6 or so MHz) or a Load Resistor, that would be 10 Vrms at the source?
Or am understanding that wrong?

I did not use a probe.
I used a direct connection with a BNC cable with an inline 50 ohm load from the signal generator to the O-scope.  You should connect your TinySA the same way but you will obviously need a SMA to BNC cable.

[Aldo22]

I did not use a probe.
I used a direct connection with a BNC cable with an inline 50 ohm load from the signal generator to the O-scope.  You should connect your TinySA the same way but you will obviously need a SMA to BNC cable.

I'm just trying to figure out where this boundary you're describing really is.
As I understand it, only the voltage at the input of the oscilloscope can be relevant, not the voltage at the probe tips or before the load resistor.
What voltage is that with the 50 ohm load?
(Maybe that's different with better signal generators, but the TinySA delivers lower voltage with load resistor).

With the test probes in 10x mode it would be 100mV at the input in that case.
In this range I see no problem with my device.

Do you understand what I mean?
I am just a Noob trying to think! 

You should connect your TinySA the same way but you will obviously need a SMA to BNC cable.

That's what I do
https://www.eevblog.com/forum/testgear/new-2ch-pocket-dsosg-sigpeak-dso2512g/?action=dlattach;attach=1766672;image

Thank you!

[4thDoctorWhoFan]

That's what I do

Then I don't understand the probe questions.  Just use a direct cable connection.  Voltage from source will be the input to O-scope.

[Aldo22]

Then I don't understand the probe questions.  Just use a direct cable connection.  Voltage from source will be the input to O-scope.

What I am tying to say: If there is a problem with 1Vrms at the input of the oscilloscope, then it would take 10Vrms at the tip of a 10x probe to see the problem, right?
I know that you are not using probes in this case, but I would have to.

[rtek1000]

This DSO seems to be manufactured by the same company of the DSO1511E.

Both the software and hardware are very similar. It seems to be the 2 version channel of that device.

I own a DSO1511E, and I would have bought a two channel version if they had it back then.

I am quite satisfied with the 1 channel version:
https://www.eevblog.com/forum/testgear/dso1511e-review/

This new one has 2 channels, X-Y mode, multiple waveform signal generator and better button layout / usability than the 1 channel version.
The only downside seems to be that it doesn't have the reference waveform function and it is bigger than it's predecessor.

It's so tempting, that maybe I will buy it too  .

I agree, I saw the photo in the first post and immediately identified the hardware of the DSO1511, I have one.

I even had to repair the bootloader, see here:
https://github.com/rtek1000/DSO1511e_Plus_Unbrick

It would be interesting to be able to add a channel on the DSO1511.

[Aldo22]

It would be interesting to be able to add a channel on the DSO1511.

I don't imagine that will be easy.
The DSO1511 is also quite a bit smaller than the DSO2512G.
Below you can see a picture of the inside of my DSO2512G.

[Aldo22]

Hi

Is this theoretically correct as shown in the attachment to determine the rise time or is there more to consider?
The horizontal cursors are at roughly 10% and 90% of the PK-PK.
I am not so much interested in the measurement setup or the accuracy, but more in the basic principles.

I thought about asking here, but I wasn't sure if I should reopen this old thread. 

Thanks!

[wasedadoc]

Not 10% and 90% of peak to peak when there is undershoot and/or overshoot. 10 and 90 of the "steady" state before under- and after overshoot.

https://rfmw.em.keysight.com/digitalphotonics/flexdca/ug/content/topics/tdr-tdt-mode/time-toolbar/rise_time.htm

[Aldo22]

Not 10% and 90% of peak to peak when there is undershoot and/or overshoot. 10 and 90 of the "steady" state before under- and after overshoot.

Thank you!
Very helpful!
So it would be better this way? (Approx.)

[wasedadoc]

Yes.

[Aldo22]

Here is an anecdote about the "toy manufacturer".

I read from a man who had a problem measuring a signal with negative polarity (DC measurement, plus to ground). The oscilloscope showed values outside the tolerance range.
He complained about this to the store and within a week, he received by email a new firmware and instructions how to calibrate the inputs himself.
Now it works precisely and he is satisfied.
So it doesn't look like they are trying to rip you off.

I then also asked if I could get this calibration option and also got a reply within 2 days:

"Hi, we don't recommend you to upgrade it, we only have improved firmware for V1.2.9C version now. Unpredictable errors may occur after other version upgrades."

It's understandable that they don't backport this option to older hardware versions, but at least I got a qualified answer.

I hope the manufacturers of real oscilloscopes are as responsive as the toy manufacturer. 

[jebem]

Today I have ordered this DSO2512G scope from the official ZEEWEII shop on AliExpress.

Thank you all for your posts and information.

I understand that the current Hardware version would be at least v9.15, as seen in some pictures published by satisfied customers.

This 9.15 version looks similar to the previous v9.14 where a proper metal shield covers the input signal processing components.

Older H/W versions still use a electromagnetic shield by using a glued metallized paper on the back cover close to the battery.

Differences between H/W versions are not limited to the shielding technique, though. PCB layout, heatsink on at least one IC (USB battery charger?), and at least one component replacement (the Lattice VLSI IC CPLD FPGA was replaced by a physically smaller AG1KLPQ48 AGM Microelectronics.

From some posts above, several questions were raised and some were not answered with 100% certain.
I have downloaded the user manual.

Concerning the bandwidth figure, the manual clearly states that it is restricted to 5 MHz when setting the scope vertical input at X1. Full bandwidth is reached only when using the Channel 1 set at X10 (and disabling Channel 2), and enabling the overclock option.

This bandwidth restriction is not exclusive of this scope. Other equipment from other brands not related to this design also specify similar specs.

I understand that a small signal imposes great effort in the analog circuit design that needs to amplify it to a minimum level before it can be passed to the ADC in order to get more than just noise numbers at the output.

Therefore, the analog section to handle small signals is usually dedicated, and that is usual to use signal relays to redirect the fragile input signals to a proper analog amplifier able to generate an identical shaped output with a higher level.

Designing a flat, low noise, wide bandwidth analog amplifier is a very difficult task.
It takes more than just reading IC datasheets and run some simulations, the secret is in the real implementation with expensive components, and requires a lot of time for experimentation.
This takes money, much more than what we are paying for this product.
Therefore the small signal analog amplifier was designed to respond kind of flat for a narrower bandwidth figure, in this case, 5 MHz. 

As for the larger input signals, they have enough amplitude to be handled directly by the ADC, although a proper design would use some intermediate network components to shape the signal.


 

[Aldo22]

Hi jebem!
I am curious if you like the DSO2512G.

Concerning the bandwidth figure, the manual clearly states that it is restricted to 5 MHz when setting the scope vertical input at X1. Full bandwidth is reached only when using the Channel 1 set at X10 (and disabling Channel 2), and enabling the overclock option.

I don't know if I misunderstand you, but the bandwidth limitation is a feature of the probes in 1x mode, not of the oscilloscope.
See e.g. here

As I understand it, the 1x/10x switch on the oscilloscope simply changes the measurement numbers (x1/x10/x100), it does not change the way the signal is displayed.

Sorry if I misunderstood you.

Have fun!

[jebem]

Hi Aldo!

I will have to wait a few weeks to get my scope at my hands...

Concerning the bandwidth limits, of course any probe will act as low pass filter, specially on 1:1 mode (X1), so they limit the upper band for sure.
Good point.

I do not know the probes quality included with these kits. But I know that not all probes are created equal.
I would expect a good 100 MHz probe to have around 15 to 20 MHz upper limit on X1, not just 5 MHz.
I realize that the 100 MHz probes are not adequate for a alleged 100 MHz scope. We should be using here an absolute minimum of 200 MHz good quality probes, not these low costs jobs.

But I might be wrong, testing in a lab would be required to test and verify.
The scope would need to be connected to a source generator using a properly terminated transmission line (coaxial cable), removing the scope probes from the circuit.
And then select the vertical sensitivity in the lowest mV range, scope on X1, and check the bandwidth.

Did someone test it already?

I would be surprised if the upper bandwidth limit would be close to the specified 100 MHz, or even 20 MHz for that matter.
That would mean that this scope small signal analog amplifier circuit is top notch with an incredible flat wide band.

The point here is that the ADC IC's are not able to deal with low level signals that will close to the noise floor of the ADC, also consider that these cheap ADC's uses just 8-bit (256 levels), so the input analog signal, if small, needs to be amplified and shaped to present an adequate voltage level, and a flat curve along the frequency band to let the ADC do its work properly.
For that purpose, a sensitive analog amplifier is used, and these are expensive to design and implement if they need to have a flat wide band of, say, 100 MHz. So these low level signal amplifiers typically have modest bandwidth values due to  cost savings.

For stronger signals it does not make sense to keep using a sensitive analog amplifier, so typically it is switched out from circuit and another network circuit is switched in to attenuate the signal to adequate levels to be accepted by the ADC IC.

Hence the analog section typically would include more than just one network filter/amplifier for all the ranges, and these network filters can be inserted or removed depending on the vertical sensitivity selector position.

Of course really cheap low cost scopes have only one basic network filter with a resistive ladder attenuator and compensator capacitors, but these equipment types will not be able to handle  small signals due to the limitations of the low cost ADC chips.

[Aldo22]

Hi Jebem!

I do not know the probes quality included with these kits. But I know that not all probes are created equal.

You get the usual P6100 6/100 MHz probes.
You can't expect more for $80 including shipping.
At least you get two probes, which is not the case with Owon handhelds for example!

I would expect a good 100 MHz probe to have around 15 to 20 MHz upper limit on X1, not just 5 MHz.

Hmm, not sure. My PP-250 have a declared BW of 10/250 MHz.

The scope would need to be connected to a source generator using a properly terminated transmission line (coaxial cable), removing the scope probes from the circuit.
And then select the vertical sensitivity in the lowest mV range, scope on X1, and check the bandwidth.

Did someone test it already?

Do you mean like in the attachment?
10mV is the lowest range. SG is a TinySA (Sine, -36dBm) with coax and 50 Ω dummy load. Overclocking mode from 100MHz on.

I don't see an amplitude problem.
But as I said before, the signal on the screen is not always as crisp as in the screenshots.
On the screen you sometimes see wider or even various adjacent lines at higher frequencies.
With persistence set to 1s it becomes smoother.

It is a cheap device, the cheapest 2-channel oscilloscope afaics, maybe a toy.
You should not look for the "fly in the ointment", but enjoy what it can do. 

[jebem]

Hi Aldo,

Thanks for the reply and information.

From your screenshots, the small signal response looks fine to me. Small jitter in the signal is to be expected.
That clears out any internal constrains concerning the bandwidth upper limit on small signals.
Excellent!
I would say that we would need to spend way more to get better results.

So the 5 MHz limitation is only related to the probes included in the package.
The nominal 6 MHz LPB at X1 of similar probes is close enough to the specified 5 MHz for the scope on X1 seen in the user guide.


Looking closer to published pictures of the older PCB versions (where there is no soldered metal shield over the vertical input section), I have a rough idea on what they have designed.
Initially I assumed that a dedicated small signal amplifier was in use to handle the mV range, but after your test results I do not think they did it in that way (still it is used in more elaborated and expensive devices).

In fact it seems they did better than that for this product price range, despite the average/poor input sensitivity of 10mV/div that in fact points to the probable design choice in use.

More likely the vertical amplifier/attenuator is a single circuit handling all the vertical sensitivity levels, kind of a PGA (programmable gain amplifier), managed by the MCU (F1C100S AllWinner Tech SoC ARM926EJ-S) IC that controls one 3-to-8 decoder (SM74HC595D) IC per vertical channel. These SM74HC595D decoders seems to be acting as control lines actuating on the PGA section level selectors.

Unfortunately the Chinese plant decided to keep three crucial IC's identification a secret, by erasing the silkscreen labels of them.

Two of these IC's (16-pin) are used in the vertical PGA section (one per channel) and these make sense to be analog PGA devices.

The 3rd one being the ADC IC (48-pin) that other members have mentioned above as likely being an AD9288 or compatible device.

As others commented above, marketing the scope as running at 500 MSPS is strange to say the least, as a single AD9288 is able to run at 100 MSPS per ADC channel with a total of 200 MSPS.

200 MSPS would be the absolut minimum to reconstruct a sinusoide at 100 MHz in one vertical channel as clamed by the manufacturer.
Using both vertical channels leaves just 100 MSPS per channel and this would allow to read up to 50 MHz at most.
This is not far from what the scope manufacturer specifies.

Either the ADC IC is a custom Chinese IC able to run at 500 MSPS (based on AD9288 or not) and the claim is correct, or else if it is a genuine/cloned AD9288, then it looks highly improbable, even with overclocking, to run the IC at the claimed rates.

I will dismantle my unit after it arrives to have a more detailed look and try some sort of reverse engineering, as I usually do for all my acquired devices, specially when the manufacturer do not release service documentation.



Unrelated talk:

I have used professional scopes since my days in school in the 70's and have worked professionally in electronics as used in several lab environments including medicine investigation, telecommunications (RACAL, STORNO, FURUNO, DECCA, REDIFON, RCA, and a few other brands),  and later doing computer hardware maintenance in the 80's, so I am familiar with a few generations of scopes, most of them from Tektronix and HP, well before the DSO generation.

Then I moved to computer software and forgot my electronics background for a while, but kept my radio and calculator collector hobby.
Currently I have a couple of scopes for my private usage, one being a CENTRAD from the 70's (100% analog, CRT, transistorized with a few TTL chips), and more recently I acquired a Owon SDS1102 DSO and  a FeelElec FY6900 signal generator.

In order to check, align and repair my collection of radios, mainly classic analog and a few DSP based as well, I use a vintage HEATHKIT IG-102S signal generator (valve based) and a classic, vintage, professional, FLUKE 1920A frequency counter.

See, I would need a clean and low signal level generator, in the range of a few micro-Volt and below (0.2uV), to inject at the different receiver stages, and for that I managed to use the HEATHKIT despite it is not a professional equipment, missing a calibrated output attenuator (I use a diy ladder attenuator, not calibrated).
The modern FeelElec FY6900 looks good in the specs, but the signal output levels are more oriented to digital and analog high level signal devices.

Finally, I miss the 500 MHz Tektronix I used in my telecomm workshop in the 70's to  be able to read low level analog signals from my double-conversion heterodyne radios, where the 1st IF frequency is on the 40 to 60 MHz range.
The Owon SDS1102 is almost able to do the job (it can read the local oscillators because of the highish signal levels), but I really would need a 200 MHz wideband with something like 1mV per division to do proper readings.

Why now this tiny cheap ZEEWEII DSO2512G ?
Well, I like to collect and analyse electronic devices, specially small ones in my spare time.

[Aldo22]

Hi Jebem!

despite the average/poor input sensitivity of 10mV/div that in fact points to the probable design choice in use.

10mV/div is just fine for the lowest price range.
There are more expensive devices from e.g. Fnirsi that have 20mV/div or even 50mV/div.

I'm interested to hear what you find out!

I'm just the opposite of you, it's the first oscilloscope for me and I'm having fun with it and already learned quite a bit.
Contrary to other opinions, I think it is suitable for beginners because of the ease of use and the manageable options.
In the end, you can also learn something from its limitations.

I didn't buy it because of any bandwidth or sampling rate specifications, but because two channels can definitely do more than one and it was the cheapest two channel device.
If it can show a nice square wave on both channels at, say 4MHz, I'm happy for that price.

Even if I might buy a "real" oscilloscope later, this one has its justification.
It is small and handy, runs on battery and has a rubber cover, so it does not fall down when you put it somewhere in the engine compartment, for example. 

The case measures 18x14cm and holds the device and all accessories.
You can put it under the pillow.