New 2ch pocket DSO+SG - Zeeweii DSO2512G

[nathi]

Hi everyone my name is Nkosinathi(Emmanuel), this is my first time posting posting in this forum.  Have been watching this topic (DSO2512G) since I want a handheld oscilloscope for repairing SMPS, amplifier and TV'S but yesterday I saw on AliExpress that fnirsi have released a new 2 in 1 handheld oscilloscope which look nice on paper. Hopefully I will see a review of this oscilloscope in the forum so I can compare.

ZAR 2,704.16  40%OFF | FNIRSI DPOX180H Handheld Phosphor Digital Oscilloscope 180MHz -3DB 50000wfms/s Dual Channel 20Mhz Signal Generator ZOOM XY FFT
https://a.aliexpress.com/_mLSTvk0

Please forgive me English is not my 1st language incase some confusion.

Kindly regards

Nathi

[Aldo22]

Hi nathi

The two devices are not quite in the same league price-wise.
The DSO2512G costs about $80 and the new FNIRSI about $140.

The DPOX180H thread is here:
https://www.eevblog.com/forum/testgear/new-toy()-scope-fnirsi-dpox180h-claimed-180mhz500msps-(may-2023)/msg4868756/

[nathi]

Hi aldo22

You right and thanks for replying and the link for new fnirsi handheld oscilloscope.

[tatel]

Beware that FNIRSI 1013D and 1014D are reported as having just 100mV/div input sensitivity at 1x

https://www.eevblog.com/forum/testgear/fnirsi-1013d-100mhz-tablet-oscilloscope/msg4577548/#msg4577548

So I would wait until a detailed review of the new FNIRSI appears

[Aldo22]

Beware that FNIRSI 1013D and 1014D are reported as having just 100mV/div input sensitivity at 1x

https://www.eevblog.com/forum/testgear/fnirsi-1013d-100mhz-tablet-oscilloscope/msg4577548/#msg4577548

So I would wait until a detailed review of the new FNIRSI appears

I don't know if this is true for the DPOX180H, I guess it's not that bad, but what I noticed is that the DSO2512G measures RMS voltage quite accurately even at low voltages.
At least it is quite consistent with the AN8008.
I was a bit surprised by this.

[jebem]

Following David's recommendation, as soon as I got my ZEEWEII DSO2512G, I took it apart.
I had a brief look to confirm that is was working before have it dismantled.

It came with Hardware v9.14 and Firmware v1.2.8C.
As I ordered it from the official ZEEWEII  shop in AliExpress, for about 97 EUR (!), somehow I was expecting to get the latest v9.15 H/W and v1.2.9C.
But no, I got one from the old batch.

As expected from the bandwidth tests published here by the fellow members that took the time to do it (Thank you!), the analog stage seems to be designed like the other low cost scopes. A few opamps (but no VGA/PGA in sight), and resistive ladder attenuators selected by analog switches (74HC4051) and the like.

That said, the design includes some nice details, like several dedicate regulated power supply lines for digital and analog circuits.
The PCB build quality looks good with no obvious soldering defects, no last minute piggyback installed components, perfect PCB finish.

The only issue I found is related to the chassis design and to the metal shield covering the analog section.
Once you undo the four screws at the back cover and have it removed, the PCB seems to be accessible for removal as well.
Please be aware, as it is not enough to undo the visible screws fixing the PCB to the front panel.
The metal shield must be removed by unsoldering the two joints, to reveal and give access to one last screw that also need to be undone.

Identifying those Chinese code markings on the SMD components is time consuming.
This is what I could find out so far.

[Aldo22]

Hello jebem!

Wow, great job! 

It came with Hardware v9.14 and Firmware v1.2.8C.
As I ordered it from the official ZEEWEII  shop in AliExpress, for about 97 EUR (!), somehow I was expecting to get the latest v9.15 H/W and v1.2.9C.
But no, I got one from the old batch.

Strange that you got the previous version (the same as I have), but probably one has no claim to a particular version.

And € 97, that's with VAT (IVA) included.
For other destination countries (e.g. CH) Aliexpress shows the prices without VAT and the regulations are different anyway.
It's a bit difficult to compare in an international forum, that's why I always quote the prices without taxes if possible.

Thank you!

Btw. How do you identify the components with the scratched markings?

EDIT: I think you are right with the AG1KLPQ48.

[tatel]

@jebem: thank you very much

Now, do we have any reasons to think this is truly a 500 MSa/s device?

[jebem]

Btw. How do you identify the components with the scratched markings?

EDIT: I think you are right with the AG1KLPQ48.

In my unit there are only two ICs without markings.
The pinout of both agrees with the identification codes I have chosen.
Is it correct?
I guess we will never know.

[Aldo22]

Hi jebem

In my unit there are only two ICs without markings.
The pinout of both agrees with the identification codes I have chosen.
Is it correct?
I guess we will never know.

All I can say is that the AG1KLPQ48 is most likely correctly identified, as what I can still read on mine matches that mark. AGM is correct anyway.

I didn't remove the metal shield on mine, so I can't say much about the other chip.
But yes, VDD is on pins 28 and 46 of the MXT2088 afaics.

Thank you.

[jebem]

Btw. How do you identify the components with the scratched markings?

EDIT: I think you are right with the AG1KLPQ48.

Hi Aldo!

Thanks for your comments.

In my unit there are only two ICs without markings.
The pinout and connections of both ICs agrees with the identification codes I have chosen.
The AG1KLPQ48 FPGA id was not scrapped in some previous pictures/videos, I believe. That's why I took note of it back then.
Also, from what I have read in Chinese websites, the Lattice FPGA was replaced by the newer and smaller and more efficient AG1K series.

See here (sometimes the website may not respond):
https://m.elecfans.com/d/comp-2786412-136355882485081825280.html
http://www.agm-micro.com/products.aspx?lang=&id=22&p=14

As for the ADC, the pinout matches the MXT2088 that in turn is identical to the AD9288.
Why did I chose the MXT2088  over the AD9288?
Well, this is a Chinese product made with Chinese components after all.
I don't think they need to import the AD9288 when they can produce the MXT2088  at home at a lower cost.

See here:
http://www.mxtronics.com/
http://www.mxtronics.com/n107/n124/n181/n184/c692/attr/2630.pdf

[jebem]

@jebem: thank you very much

Now, do we have any reasons to think this is truly a 500 MSa/s device?

Well, it is highly improbable as it all seems that the ADC in use (released with the code marking removed) looks identical in the pinout to a MXT2088 (100MSPS per channel).
Unless the manufacturer finally decide to disclosure the mysterious ADC identification and specs, that is.

Apparently they like to hide things. On my unit they hid the FPGA identification as well.

And as seen in some pictures of older H/W releases that did not use shielded analog circuits, they hid the identification of two 16-pin IC's, as if they were some sophisticated VGA or similar.
However, on my newer unit using a soldered shielding, these two ICs were shipped with their ID markings, so we know those are just 74HC4051 to control the resistive attenuator as seen in so many designs at this price range.

This is my initial evaluation, but I can be wrong.
When I have the time, I will dig deeper into the reverse engineering of the circuit, but I would be surprised to learn that a 97 EUR (VAT and S&H included) can compete in design against devices costing 3 times more. That makes no sense to me.

[Aldo22]

In my unit there are only two ICs without markings.
The pinout and connections of both ICs agrees with the identification codes I have chosen.
The AG1KLPQ48 FPGA id was not scrapped in some previous pictures/videos, I believe. That's why I took note of it back then.

My chip wasn't scratched that thoroughly. If you know what could be written there, you can verify it. (Attachment).

I don't think they need to import the AD9288 when they can produce the MXT2088  at home at a lower cost.

It is certainly not wrong to assume that the cheapest possible (chinese) components were used here.

@jebem: thank you very much

Now, do we have any reasons to think this is truly a 500 MSa/s device?

Well, it is highly improbable

To be honest, the question of the sampling rate is not that important to me.
I know now what it can do and where it has weaknesses
For me it is simply the cheapest 2-channel scope on the market and in that respect it is unrivaled, regardless of whether it can do everything they claim.

I find it quite useful and it hardly causes any trouble when using it. And it's quickly put away in the closet when you no longer need it.
For me it is okay.

Thank you!

[jebem]

In my unit there are only two ICs without markings.
The pinout and connections of both ICs agrees with the identification codes I have chosen.
The AG1KLPQ48 FPGA id was not scrapped in some previous pictures/videos, I believe. That's why I took note of it back then.

My chip wasn't scratched that thoroughly. If you know what could be written there, you can verify it. (Attachment).

Nice picture!
Did you used a microscope for that?
So we know for sure what is the FPGA in use.

My dumbphone camera is barely adequate for this task.

I am using a magnifying glass with LED illumination to have a closer look on the unmarked ICs on my unit, but I can't see any leftover marks on it.
I guess they dug deeper to get all traces of ink off the chip surface.
Passing the nail over the surface, a depression is felt when entering the area that was thinned.

[Aldo22]

Nice picture!
Did you used a microscope for that?
So we know for sure what is the FPGA in use.

My dumbphone camera is barely adequate for this task.

No, I've just illuminated it with a LED flashlight at a shallow angle and then photographed with my not-so-new and rather inexpensive Redmi Note 7.
Then cropped in Gimp and increased the contrast.
But yes, you can read it better this way than from the naked eye.

Cheers!

[Aldo22]

Hi Jebem!

Can you tell me what this part is good for? (attachment)
What can you adjust with it?

Thank you!

[jebem]

Hi Jebem!

Can you tell me what this part is good for? (attachment)
What can you adjust with it?

Thank you!

That sems to be a trimmer capacitor, one per vertical channel.

In my scope unit, they locked the rotary cursor by dropping a bit of transparent silicone or similar product.
That is odd, as (good quality) trimmer caps do not need to be locked in this way, as the rotary cursor does not move around easily even when subjected to mechanical vibration.

Usually the trimmer caps are used as a RF compensation devices to adjust the response of a RC or LC network. In this application it would be a RC network.


Clearly the analog section is by far the more complex circuit design, followed by the power supply.
In the analog section I still have to insert all the passive components and get courage and spare time to collect the schematic diagram.

[Aldo22]

Hi Jebem

Thanks for the answer.
I figured it was an adjustable capacitor, I just didn't know what for.
I guess it's best not to touch it?

Btw. Why are you dissecting the device so thoroughly?
Is it just for fun, or do you have a specific purpose?

Thank you!

[jebem]

I figured it was an adjustable capacitor, I just didn't know what for.
I guess it's best not to touch it?

Yes, I would not readjust it, as it will modify the response of one of the attenuators, I don't know which one it is because the manufacturer do not publish the schematic diagram

Btw. Why are you dissecting the device so thoroughly?
Is it just for fun, or do you have a specific purpose?

Basically it is just out of curiosity and have some fun in the process.
I believe that we all learn from each other, and it is always interesting to find out what design solutions were chosen to achieve the same goal. Different people, different solutions.
 

[tatel]

To be honest, the question of the sampling rate is not that important to me.
I know now what it can do and where it has weaknesses
For me it is simply the cheapest 2-channel scope on the market and in that respect it is unrivaled, regardless of whether it can do everything they claim.

I find it quite useful and it hardly causes any trouble when using it. And it's quickly put away in the closet when you no longer need it.
For me it is okay.

Thank you!

Well those weaknesses arent't clear to me yet. Yeah: it could be said it's a 18-20 MHZ BW device, it isn't able to do a poor man differential measurement and does an useless FFT instead. I has RMS measurement which Owon HDS242 doesn't. That's quite clear now.

But what about:

input sensitivity: Is it 100 mv/div at 1x like FNIRSI 1013D, or is it better? Is it true or is it just a digital zoom like FNIRSI's 50 mV/div?

noise: I see reports about Owon HDS family having quite a bit of noise which could become worse with time. How much noise does DSO2512G have?

Temperature drift: Again HDS200 family seems to have quite a bit of temperature drift. Is it the same on DSO2512G?

Firmware bugs/updates: Did you detected any bugs? Can the firmware be updated?

Please note I'm not bashing the device. I just think it would be worth to do some more testing.

[Aldo22]

Yeah: it could be said it's a 18-20 MHZ BW device,

How do you define that? How do you come up with it?
Do you own such a device now?
I don't see anything special happening at 20MHz and I have such a device.
Why do you say something like that?

Please note I'm not bashing the device.

Yes, you are!
Just don't buy it if you hate it. Why bother?

[jebem]

I had a more closer look into the scope circuit.
Attached are some drafts covering mainly the analog details.

I think I had enough of it for now and I am closing my unit before it finally breaks for good with so much manipulation and probing. Specially worrying is the display flat cable connector that may break lose from the PCB if I continue picking on it.

From the Vertical Analog Input channel circuit it is clear now that signal amplification is not used, as the OpAmps are configured basically with unity gain.

However the analog input signal is attenuated when required, though.
There are two main attenuator sets, one set per analog channel.

Using Channel 1 as example:

The first attenuator is located at the analog input and it is controller by the mechanical relay K1.
The input signal is loaded by a total of 1010KR (R34+R33+R32).
For small signals where the scope sensitivity is set at the higher values, the R27/C10 RC network passes the signal to the buffer OpAmp U16 without attenuation.
For stronger signals where the scope sensitivity is set at lower sensitivities, there is a threshold point on one of the sensitivity selection where the mechanical relay is switched over (we can hear it), so the input signal is now taken from R32 where the signal is attenuated about 33 times.

The second attenuator is a ladder of 6 levels selected by the analog switch U18.
This ladder attenuator uses low resistance values (total load on the U16 buffer OpAmp output of just 60R). This approach would benefit a better response on high frequencies.

In order to protect the first buffer OpAmp (U16 on CH1 and U17 on CH2), when the input signal amplitude is excessive for the selected sensitivity, it will be clipped by two pairs of fast switching diodes (D6 on CH1 and D7 on CH2).

Concerning the "poor" bandwidth response above 40MHz and up to the claimed 120MHz, it is a consequence of the design choice used on all of these low cost scopes. That is to be expected, nothing new here.

Among several limiting factors, I would point out:

- There are several parasitic capacitive loads in parallel with the signal, like the clipping diodes (D6/D7), the analog switches ICs (U18/U19/U23/24);

- The first buffer OpAmp (U16/U17),  despite running in unity gain, doesn't seen capable of a "flat" response from DC up to at very least 120MHz.

Alleged this IC is a "8091", with no brand marks. Even being a Chinese clone with identical performance to a genuine Analog Devices, a AD8091 is specified as 110MHz -3dB bandwidth in optimal conditions.

Also they chose to use a heavy 60R load on the output of these buffer OpAmps. The AD specs referes to a minimum of 100R loads, well above the used 60R; this likely would result in increased signal distortion depending on the ability of the IC response across all bandwidth.
 
A better compatible pinout OpAmp with at least 300MHz -3dB bandwidth could be used here, like the AD8009, although with increased costs.

All in all, I like this design for what it have costed.

[tatel]

Yeah: it could be said it's a 18-20 MHZ BW device,

How do you define that? How do you come up with it?
Do you own such a device now?
I don't see anything special happening at 20MHz and I have such a device.
Why do you say something like that?

Please note I'm not bashing the device.

Yes, you are!
Just don't buy it if you hate it. Why bother?

@jebem: thank you very much for the information you provide.

@Aldo22: Now you are throwing a temper tantrum. Please grow up and spare us any drama. I have some questions for you:

1) It has been said this device has a 100 MSa/s ADC. If so, I can't understand how it could be good up to 120 MHz bandwidth. I'm sure you can, probably mentioning something fancy like the ancient aliens that gave civilization to us or something like that. Could you demonstrate DSO2512G has a faster than 100 Msa/s ADC instead? If you can, please do so. I would really like to hear that, even if a real 20 MHz bandwidth already seems to me good enough for the price.

2) I define DSO2512G as a 18-20 MHz device because:

a) I do own a DSO154Pro also from Zeeweii, which I think is quite a fine device for the price, It is advertised as a 18 MHz device, one channel, with a €40 price tag. So I think DSO2512G is like a siamese couple of DSO154Pro, perhaps a little more, with a €100 price tag. To me, that seems to add up, I mean, similar bandwidth, etc, but two channels, not one, for double the price.

b) It has been reported DSO2512G has about 36-40 MHz bandwidth. Please note this quote is from you:

So in 2 channel mode not overclocked: 48Mhz -> 4.8Mhz -> 16MHz
2-channel mode overclocked: 60Mhz -> 6Mhz -> 20MHz

And then we have this quote from gitm:

A general comment about the DSO2512G.  It's bandwidth is not 100/120 MHz.  That's bullshit. It's actual bandwidth is about 35/37 MHz and that's with decent probes, 250 Mhz 1X/10X from Probemaster.

I'm sure you remember how yor reply included a boxing emoji and how you took a couple jabs after that. I can't help but think a 100 Msa/s ADC adds up with about 36-40 MHz bandwidth. I asked if that was with just one channel active so it would be about 18-20 MHz with both channels active. No answer to that, so I can only assume that's the case.

Your assertion that perhaps this device would have been designed to work with only small signals to have higher bandwidth is ludicrous. To me that means that I should use a 100x probe when working with, say, 24 V signals. No way. I'm going to say its a 20 MHz device instead. I say 20 MHz since it's a dual-channel device, and I want to know what it's able to do with both channels active and real-world signals. Those half-truth assertions can remain for marketing people and other thieves of that kind.

3) I'm sure you don't see anything special at 20 MHz but I have noted you don't want to see anything that could mean you gullibly bought it thinking Zeeweii marketing hype is true. You are the only guy here gringing anytime somebody says that advertised bandwidth has to be an outright lie. If you continue that way, I --and probably others-- will have to think your opinion isn't valuable. But, hey, that's not my problem, please feel free to do as you want. Ancient centro-american natives thought that to make his penis bleed would make the gods happy so droughts could be avoided. I'm fine with that too --as long as it's not my penis the one to be wounded-- but know what? Droughts can't be avoided that way. Perhaps the gods are to be blamed for that?

4) No, I'm not bashing this device. I even could buy it knowing its shortcomings, wich aren't that bad for the price. Particularly after knowing other devices in that price range aren't perfect either. Do you need to gringe if I say this is not a 120 MHz device? Well, that's your problem.

5) No, I don't feel any hate for any device. It would be ludicrous to hate an unanimated thing. I do hate marketing thieves, however.

6) Perhaps I'll buy this device, perhaps not. But anyway I'm not going to recommend it as a 120 MHz device. Can you understand that?

Last but not least, I don't really care if you reply to this post or what could that reply say. I already  wasted too much time with your "arguments". Particularlly if they are just temper trantrums.

[Aldo22]

Hi tatel!

2) I define DSO2512G as a 18-20 MHz device because:

a) I do own a DSO154Pro also from Zeeweii, which I think is quite a fine device for the price, It is advertised as a 18 MHz device, one channel, with a €40 price tag. So I think DSO2512G is like a siamese couple of DSO154Pro, perhaps a little more, with a €100 price tag. To me, that seems to add up, I mean, similar bandwidth, etc, but two channels, not one, for double the price.

No, that's wrong.
The similar 1-Channel version is the is the DSO1511G http://www.zeeweii.com/productinfo/dso1511g.html
The DSO154Pro is far less powerful.
Look at the square wave and you'll see if you understand a little bit about it.

b) It has been reported DSO2512G has about 36-40 MHz bandwidth. Please note this quote is from you:

So in 2 channel mode not overclocked: 48Mhz -> 4.8Mhz -> 16MHz
2-channel mode overclocked: 60Mhz -> 6Mhz -> 20MHz

Yes. It is quite normal that in 2-channel mode there is less bandwidth/sampling rate than in 1-channel mode.
The specification is always the highest possible bandwidth.
What is the problem?

I call it a 120MHz bandwidth device because it can measure a 120MHz signal with less than -3dB of amplitude drop and there is no significant amplitude drop until then.

I have shown this a thousand times, but if you don't want to believe it, I can't help you.
Maybe there are other definitions of bandwidth that include jitter etc.  but I can't measure/quantify that, and I don't really care.
I can use the DSO2512G to measure the frequency of my FM test transmitter at, say, 118 MHz and that is enough for me.

It probably has different characteristics at higher input voltage or at -200 °C, but it works for my 3-5 volt stuff with the 10x probes.
That's all I need, and that's probably what most buyers of this unit need.

Cheers.

[tatel]

The similar 1-Channel version is the is the DSO1511G http://www.zeeweii.com/productinfo/dso1511g.html
The DSO154Pro is far less powerful.
Look at the square wave and you'll see if you understand a little bit about it.

Sorry I couldn't be bothered anymore.

The specification is always the highest possible bandwidth.
What is the problem?

Problem is we are speaking about real bandwidth, not marketing hype. You can do whatever you want; I'll speak about what the device can do with all channels active. Otherwise I would feel like a marketing guy trying to trick customers into buying by, well, basically lying through the teeth. I think advertising a device as dual-channel while giving the bandwith figure it can reach with just one channel active is nevertheless dishonest.

I call it a 120MHz bandwidth device because it can measure a 120MHz signal with less than -3dB of amplitude drop and there is no significant amplitude drop until then.
I have shown this a thousand times, but if you don't want to believe it, I can't help you.

Please excuse me for being blunt, but enough is enough. Yes you did sent many many times your measurements, made with a €4 "signal generator" from some random guy somewhere. Characteristics of that "signal generator" are still unknown even to you, the owner. You have been suggested to ask the seller, but obviously you didn't. We still have to know what it's output amplitude really is, let alone knowing if that amplitude is really flat along the bandwidth.

Not to mention these fancy placements for terminators/attenuators/whatever that invalidated most of your measuremements. Just saying.

I can't trust your results, sorry. Yet there are people with real FGs/AWGs that reported different results. Results that seem to add up.

So yes, at this point you really can't help not just me, but any prospective buyer. Sadly, quite the contrary, I would say.

that's probably what most buyers of this unit need

I for one don't need (want) to look at any 118 MHz signal with this device. I mean, should I want to do any semi-serious project, even at hobbyst level. To play is a different matter, one would do that with some toy. Just for fun.

Unless new evidence appears, looking at any 118 MHz signal with DSO2512G, I wouldn't have any hope for any real good results. Which is exactly what I ask my devices to do, no matter the price range. I could happily use it to look at a 18 MHz signal, however.

If you want to cheat yourself by thinking DSO2512G can do it fine, I have no problems with that, the same way I don't have any problems with the way mesoamerican natives avoided droughts.

But please don't mislead any prospective buyer into thinking it. I find your assertions quite reckless, to say the least. Again,for the last time: I couldn't agree with radiolistener's assertion DSO2512G doesn't hold any value, but I agree even less with your assertion this is a "120 MHz" device.

Perhaps DSO2512G's ADC identification is wrong, perhaps not. But either way you recently agreed it's a 100 MSa/s ADC. Yet you insist DSO2512G is a 120 MHz device (60 MHz we should say, I think, but anyway). You have been asked to explain how a 100 MSa/s device could go up to 120 MHz bandwidth  You obviously choosed not to do so. Why? Please note this is a rethoric question.

Now, this is the last time I reply any post from you. It's a waste of time, like getting into an argument with wolverine, that famously got an adamantium-clad skull 

Cheers