New Zoyi multimeter+oscilloscopes - ZT-70xS, up to 50MHz/250MSps (nov 2022)

[bobuhito]

@nikbry
I believe the 10Hz spec is when using "AC coupling" mode, so it doesn't apply to my measurements.  "DC coupling", which I used in the video, literally goes all the way to DC (0Hz).

[csuhi17]

 I think it's a physical limitation, low memory and slow sampling rate.  It happens very often with Fnirsi, it also happened rarely with my Owon.


I played with it a bit, and it seems that retriggering in a 100ms time base is approx. 4 seconds.
The guy messed it up in the video by placing the trigger completely on the left side.
IF he left it in the middle, it would have been seen that the trigger misses the first pulse and catches the second.
The fact that at the end he tried to zoom out or move the stopped recording to the right and did not find the first pulse was also misleading.
I can't tell for sure which part of the off-screen wave is real and which is junk.
Because we cannot see which is the end and the beginning of the memory.

I also rarely experienced this trigger error with Owon.
With Fnirsi, very often...

[apollo11fan]

Would you repeat the experiment, but with and without Normal acquistion sampling enabled?

[csuhi17]

Yes.
There is a small bug at 50 and 100ms.
With a single trigger, the wave is not completely saved, or we only see more on the display than what is in the memory maybe?
20ms is already good.

and a bug where if you capture a signal with single trigger, I press the hold/save button to capture a signal again, it doesn't always stay in single trigger mode.
switches to AUTO, if I press MENU and navigate to where the trigger settings are, it switches to Single again.
This part may be a bit incomprehensible.

[tutochkin]

bobuhito, I recently did an analysis of the fnirsi 2c23 error correction in the new 2c53t, and compared it with the work of the zoyi zt-703. I did not observe any problems with the trigger operation in the zoyi zt-703.

https://youtu.be/1E0RQBMMkVQ?si=Vl533En3Ao4EsyMk&t=111

[csuhi17]

bobuhito, I recently did an analysis of the fnirsi 2c23 error correction in the new 2c53t, and compared it with the work of the zoyi zt-703. I did not observe any problems with the trigger operation in the zoyi zt-703.

https://youtu.be/1E0RQBMMkVQ?si=Vl533En3Ao4EsyMk&t=111

I think I managed to understand Bobuhito's problem.
it is as if measurement points are no longer saved ~500ms after triggering with the 50 and 100ms time bases.
If you set single triggering and then switch the output of your signal generator to it, you can see that the continuous signal is only visible for 500ms after the trigger. then a straight line.

If you give it a continuous signal and then press the single trigger, you will see some kind of remnants from the memory after 500ms too.

Among my previous posts, this is exactly what you can see in the picture called 100ms.
Bobuhito could no longer see a second pulse when he set the delay between the two pulses to 500ms.
Some kind of bug.

Another annoying thing.
If I stop it with a single trigger or simply with the Hold button, then if I zoom in on the signal, the green arrow marking the trigger and the real trigger point move relative to each other, it does not follow.

[bobuhito]

@csuhi17
Yes, thanks for explaining it better.
I thought the video showing the missing pulse would speak for itself, but everyone's instinct naturally is to think this is somehow normal operation or user error...now, I hope Zoyi/Zotek can fix this bug soon with new firmware!

[csuhi17]

I was confused that the trigger was mentioned. 
with fast signals, I noticed that occasionally it does not trigger on the first edge.

This error could have been clearly demonstrated with a continuous signal.
 1. I set the single trigger. 
2. I touch the measuring tip to the measuring point.  Then you can see where the signal breaks.

[xKertx]

V1.42 Software update details from zotektools.com:

• Added FFT mathematical function, replacing the original format key with an FFT switch.
• Optimized low-frequency AUTO speed and accuracy.
• Enhanced the scrolling scan mode.
• Improved the cursor display issue in scrolling mode.

[matefon]

Hello there,
I just got my ZT703s a few days ago. I really enjoy it, however a few downsides: the default probe is not so good, so I ordered better ones, but the standard BNC is too big to attach, it just "stucks" before fully clicking in (it can be used this way, but I wonder if I can "wear it in" by using pliers gently). The other problem is I can't use my computer to analyze the waveforms (properly).
So I am currently working on a Python script (with the help of ChatGPT) to scan BMP files and logic-analyze the signal.
That's when I noticed, that the BMP format is "corrupted": I can open it on Win10, but if I try to open it with anything else (online image converters, ImageMagick, OpenCV, PIL) I get the "Unsupported BMP header type (56)",  "ImproperImageHeader" or on freeconvert.com: "length and filesize do not match" error.
My unit was shipped with firmware 1.3.9 and I updated to 1.4.2. Unfortunately, I don't have BMP exports from 1.3.9, so I can't test if the update caused this.
Using anything to convert it fails - except GIMP, which I just tried before posting.

The original reason to post was I couldn't convert however I tried, but after testing with GIMP, I post to let you know that this bug exists.
Also, if someone could try to convert a BMP from an older firmware with an online converter (e.g. https://www.freeconvert.com/), we could narrow down the possibilities that could cause this behaviour.

Have a nice day,
Matt

(attached capture of a PS/2 keyboard 'A' key press - 0x1C)

[xKertx]

I also have had problems with the BMP screenshots.
I used Photoshop to convert BMP to png or jpg, and it works.

[Atlan]

https://www.eevblog.com/forum/testgear/fnirsi-1013d-100mhz-tablet-oscilloscope/msg5278324/?topicseen#msg5278324

I found easy working solution for fix BMP format
Just change only one byte - 56 to 40 here (i use binary hack for .bmp, change one byte at 0x0e offset )
After that Firefox, Telegram and Python successufllly recognized screenshots.

[xfint34]

56    BITMAPV3INFOHEADER

Not officially documented, but this documentation was posted on Adobe's forums, by an employee of Adobe with a statement that the standard was at one point in the past included in official MS documentation[12]

https://en.wikipedia.org/wiki/BMP_file_format#DIB_header_(bitmap_information_header)

It is using a more obscure bitmap format. Searching BITMAPV3INFOHEADER shows quite a few results of problems with various image software not supporting it.

I wonder if they use the same code since this Zoyi meter and Finrisi have the same obscure bitmap format.

[apollo11fan]

matefon: you can downgrade your firmware. In a previous post, someone uploaded many versions of the firmware.

Have you tried opening the .BMP using the program paint.net?
https://www.getpaint.net/download.html

[matefon]

Thanks for mentioning, I didn't know if I can downgrade.
I think the easiest automated solution would be what Atlan suggested (changing a byte).
Now I have problems with the analyzing code, ChatGPT can get something right, then it brokes edge detection, then it fixes edge detection but brokes visuals... I don't have experience in NumPY and OpenCV but maybe I will need to write my own code for this.
I also plan to use a Raspberry Pi Pico as a logic analyzer, and use the ZT703s to determine the signal voltage so I can use level shifters or voltage dividers on slower circuits.

P.s.: I went to a local electronic store today to get a BNC extender, so I can fit the probes, they didn't have it but we came up with a solution to get a BNC-RCA converter and an RCA-BNC converter together
Maybe temporary solution, maybe not, I will see how it affects signal reception and how fragile does it make the on-device connector, as it is sticking out and I'm afraid I will broke it accidentally while using.

[matefon]

Also as an IT engineering student, I'd like to try debugging the firmware and add new features if I can. Does someone know about an open-source firmware for this? Or a way I can contribute? I suppose no, but worths to ask.

[Atlan]

You can write x sin(x) interpolation for 1013D oscilloscope

[xKertx]

the default probe is not so good, so I ordered better ones, but the standard BNC is too big to attach, it just "stucks" before fully clicking in

I think the real reason is that the plastic plate that is behind the bnc-connectors is too thick. Baybe if we remove it and sand the backside of it, to get it little thinner, maybe that could help.
I might try this next time i open my unit.

[matefon]

I think the real reason is that the plastic plate that is behind the bnc-connectors is too thick.

That is possible. Every manufactured component has a margin of error, and in my (and your) case it happens to be that the probe is sligthly larger, and the plastic is a little bit thicker.
I don't want to physically modify my unit just yet, so I'll be using my 'BNC extender' for now.

In a video somewhere I came across that the battery inside is not actually a 18650, but a larger one, so it puts stress on the battery holder string. Is it true?

[csuhi17]

You can see it yourself if you take it apart.
It is not difficult to disassemble, just 4 screws and carefully lift the bottom.
There is a plastic tab on both sides in the middle, which I managed to destroy, but it does not affect anything.
Mine came without a battery, I put an unprotected flat-top battery in it, I had to solder a button on it to make contact.

[xKertx]

In a video somewhere I came across that the battery inside is not actually a 18650, but a larger one, so it puts stress on the battery holder string. Is it true?

As with every mechanical part, the 18650 cells also have tolerances, so the size may vary. And it also depends if the cell has a protection circuit or not.
Also, the battery holder might be little too short (i think it is designed for unprotected cells).
In my unit the original cell was little tight and i swapped it with a Panasonic 3400mAh protected cell (because i don't want any random chinese cells in my house) that was also little tight. So no difference here.
I just shortened the spring for about one turn and it's perfect now.

[matefon]

So the battery size problem varies between units. I know I could take it apart but I thought I would ask if this is a common problem.
Thanks for the answers.

[xfint34]

In a video somewhere I came across that the battery inside is not actually a 18650, but a larger one, so it puts stress on the battery holder string. Is it true?

Mine has a protected 18650. You can tell because the length of unprotected 18650 is 65mm. Protected is a few mm longer. The cell holder in this device is definitely for unprotected cells.

The holder in mine is deformed because they jammed the unprotected cell in there. It's stuck real good. I'm afraid of trying to remove it. That's a problem for future me.

[apollo11fan]

I think you meant "protected cell" which would be taller/longer. Mine came with a protected cell. Don't you think it's a good idea to use the same? In one Youtube video, the person cuts the spring...

[matefon]

I just disassembled mine, and oh lord, there was tension. I cut off 2.5 turns from the spring. There is still some tension, but if I cut more I can't put in another 18650 cell, as the spare cells I have are slightly smaller (probably not protected). This way I can change the battery and still have less tension in the plastic battery holder.