New handheld touch DSO, Fnirsi 2C53P, claimed 50MHz/250MSps (April 2024)

[tunk]

- dual channel
- 4.3" 480*272 touch display
- 50MHz, 250MSa/s
- 10mv/div - 10v/div
- 10ns/div - 10s/div
- storage depth up to 64kByte
- signal generator: sine 10MHz, other 3MHz, 0.1-3.0V, duty cycle 0-100%
- digital multimeter
- 4000mAh lithium battery, type-C USB
- size 145x98x35mm

Manual: https://img.wqdres.com/res/0/20240314/43cc9b1cc9ee47d5a7759b015dcd415e.pdf

[tunk]

Starting to show up at aliexpress for $105 (a bit too much?):
https://www.aliexpress.com/item/1005006907937828.html

[csuhi17]

It looks a bit better, I'm not saying it's great, but if it's really going to be around $100, worth it.
And of course if the parameters are real.
We know how Fnirsi tends to lie and talk big.
64kByte is only 8k points.

Which is not visible from the picture.
It has persistence mode 1s and infinite.
math function, basic operations between two channels.

I would like to know what they thought about the data saving function in the multimeter mode.

The display could have been a little bigger, at least 6", then it would be better.

[BrokenYugo]

That they bothered including 20MHz bandwidth filters is nice. Compared to something like the DSO 2512G this seems like a better deal if the firmware is functional.

[csuhi17]

The above link is already live, allowing orders to be sent to Hungary.
For the math function, what can be "-1, -2, |1|, |2|" what can it be used for?

[pcprogrammer]

The above link is already live, allowing orders to be sent to Hungary.
For the math function, what can be "-1, -2, |1|, |2|" what can it be used for?

It supports maybe a third trace with the outcome of:
ch1 + ch2
ch1 - ch2
ch1 * ch2
ch1 / ch2
inverse of ch1
inverse of ch2

The |1| I'm not sure about. Maybe absolute value or average?

The specs for the scope might be correct though. Using overclocked ADC's it can do 250MSa/s and the 64Kbyte as total sample memory might be accurate depending on the FPGA used in the thing.

Fun toy, yes, but worth the price it is selling for, don't think so. For me at the moment 123,27 euro. 

[csuhi17]

It will depend on its performance, it is possible that it will be a match for the owon HDS42s if it meets the expectations.
250Ms/s is not much, but it does not say whether it is halved when using two channels.
For some, even that will be a lot.
Just don't have another 1013d/1014d series.
I look forward to the tests with great interest.

I would cry if the |1| would mean
It would turn an AC sine signal into a full-wave rectified sine.

[csuhi17]

I found a test on the Internet, only a few pictures and text.
I managed to interpret it in such a way that the first series may have some physical problem, and whoever bought it just now will most likely have to return it. 

https://www.mydigit.cn/forum.php?mod=viewthread&tid=451205

[Aldo22]

Do you really need so many "counts" with a DC accuracy of 0.5%+3?
The good old cheap AN870 has 0.05%+3, so the "counts" make more sense.

It's always a bit the same with these many new scopemeters:
They are neither particularly good (handheld) scopes, nor particularly good meters.
And they are not really cheap either.

[tunk]

Firmware update, May 10th 2024 (app+FPGA): https://www.fnirsi.com.cn/download/digit

[csuhi17]

I'll upload some more pictures later.
My first impression of the device before I even turned it on.
Good packaging, feels good in the hand, odorless.

After switching on, the image is good, the display is clearly visible.

After I started some basic tests.
My feelings:  Uhh, ah, 

[pcprogrammer]

I see an Anlogic FPGA, a MTX2088 ADC, but can't see the processor. Is it hidden under the shield?

With only the one ADC chip it is at max 250MSa/s on only one channel. 125MSa/s at best when both channels are enabled.

[csuhi17]

I forgot.

[pcprogrammer]

Still no processor. 

Could they have done it all in the FPGA I wonder. The eagle 4 series has more capabilities, but still.

Edit: I guess they did. See: https://www.cnx-software.com/2018/09/04/licheetang-anlogic-eg4s20-fpga-board-targets-risc-v-development/
and https://github.com/Edragon/Datasheet/blob/master/EG4S20_DataSheet_V1.4.pdf

[csuhi17]

I can't prove it, because the usual method doesn't work for this waste.
The single trigger does not work with a fast time base.


if I pause it, I can't zoom in on it more than a certain time base.

There is a problem with the low frequency, I think it is physical and not software.

I haven't gotten to the DMM and signal generator part yet, I need to take a break to calm down.

maybe
It is possible that the processor is at the bottom of the board, I did not remove it.


There is 12.8 MB of space for images and the update, which is also deleted after the update.

a few pictures.

Square signal 1Hz 10Hz 1kHz 5Mhz and 10Mhz
At 10Hz you can see what the single trigger looks like, this was the best picture out of my 10 attempts. Some had no sign at all.

[csuhi17]

A little comparison with the Owon HDS272.
I am not uploading a picture of Fnirsi, in the case of two channels or with BW limit turned on, because the signal has not changed at all.

I forgot to mention that I used a BNC-BNC cable with a 50 Ohm termination on the scope side.
To be safe, I also checked with Rigol, the signal was a nice square with 50% noise.

[Aldo22]

@csuhi17
Do I understand correctly that it cannot measure the frequencies at 5MHz and 10MHz?

Edit: For comparison 20MHz square from DSO2512G

[csuhi17]

In the pictures above, I focused on the shape of the wave.

There is a problem with the other frequencies, it seems.

I compared the two devices with the square wave so that I could see the difference between the "250MSa/s" Fnirsi and the 250MSa/s Owon, otherwise I can't even estimate the real sampling speed.

The Single trigger does not work, it captures signals randomly.
Sometimes the square signal resembled something like an EKG.

The four images below are the same square wave. 500mVpp 5Mhz
A little problem between 80nS and 200nS time base

[csuhi17]

1mh 4
3rd month 13

A 1 MHz signal needs 4 samples on the display to measure the frequency correctly.
13 for 3Mhz.
My other instruments can tell the frequency from one period.
It measures the 5 MHz frequency as 425.24 kHz, 8 periods are not enough for it. 

This made me cry. the math operation has "-1" that's ok. there is inversion on the other scopes as well.
But the |1| well. for what?

There is a change in amplitude for the two time bases, I don't understand why. 40ns-20ns

The highest frequency I could measure with it is 10Mhz, but it takes 44 periods to calculate it.
This is worse than the 1013D-1014D series.

[Atlan]

You are unteachable.  What did you expect from fnirsi?

[csuhi17]

You are unteachable.  What did you expect from fnirsi?

I was hoping that they wouldn't lie about this weak device.  My brother needed a small basic scope, he tried everything I have and it didn't work for him.  A simple electrician, he needs a few basic things.  I thought that if it knows what they wrote about it, especially since they mentioned the separate memory in it, then it will be fine.  But this is the last one.  I'll write to them, and I'll see what they answer... Unfortunately, I don't have time to deal with it now, the controller just arrived for the engine.

When I apply 500mVpp color to the input and set an offset for it, it appears for a moment that the signal jumps.
But he came right back after that.
It's like being in AC coupling all the time.
If the Sine signal is slow, its amplitude is apparently smaller.
528mVpp at 1Khz
75mVpp at 1Hz
7.5mVpp at 100mHz

This is much worse than the 1013d-1014d series.
Everyone stay away from Fnirsi oscilloscopes.
I hoped that after DPOX180H the quality would only go up, but no. 

[csuhi17]

The bottom looks like this.
From up to down
EA3036C 3Y6DEM
??
U74HC4053G-P16-R
GS8092

[pcprogrammer]

It looks like it is using a separate processor, but which one that is again the question. 

By the looks of it the 64 pin chip is connected to the display and via the two resistor packs to the FPGA.  So maybe they moved away from the Allwinner controllers to something cheaper or newer.

[csuhi17]

The low frequency distortion was because the coupling was somehow stuck in AC mode.
I switched it to DC to no avail.
OK after switching several times.

1Vpp sine wave:
Good up to ~32.5Mhz with 1 channel
750mVpp at 40Mhz and 340mVpp at 50Mhz.
good up to ~18Mhz with two channels
756mVpp at 35Mhz and 370mVpp at 50Mhz

The single trigger only works with 2uS and above, not with a faster time base.
Fnirsi also confirmed this...

when you pause it and set it to a faster time base, the wave does not change on the display, you cannot pull the wave apart. Fnirsi has not yet given an answer to this.

I played with it a bit, but I couldn't find the 64K memory, compared to the HDS272s, which only has 8k...
It can be similar to 1013d and 1014d.

The part of the multimeter is semi-usable, I did not find a "relative" function in it, not for any measurement.
Similar in accuracy to Owon DMMs.
I like the trend graph. It lasts 1 minute and 25 seconds.

Its signal generator can be used, but at higher frequencies it mixes in a little DC offset.

There are still many problems with this, I hope Fnrisi will not approach it like the previous one, releasing two updates and then trashing it.
 
However, some of the controls are very annoying.

[Aldo22]

I played with it a bit, but I couldn't find the 64K memory, compared to the HDS272s, which only has 8k...

"kByte" is not the same unit as "kpts".

Confusing, but at least they named it correctly this time.
Others write 64k, so you can only guess.