DreamSourceLab DSTouch / DS4T1012 / DS4T252

[geochaben]

I was too hasty. I see you are right from the DSView manual. Any hope they will add DSTouch in the future?

[RAPo]

I don´t think so. I emailed them, and they said DSViiew was only for the usb-scopes and the LA's.

[cynric]

Did you guys test the protocol decoders yet? I'm a little irritated on how they want to decode SPI from 2 inputs - is it just a clock synchronous single lane de-serializer? Or does it calculate the sampling points from a manually-entered clock rate and hooks up to a chip select and a data line? Two data lines and nothing else? Lots of options are thinkable, but none would strike me as ideal.. And the manual doesn't go into detail at all.

Beside from that I really like what I'm seeing. How about pre- and post-trigger options in single-shot or normal mode? Is that possible over the whole buffer depth? Like setting the trigger on the very end of the acquisition period and then zooming in and scrolling through the whole data set.
Is it then possible to add a protocol decoding layer afterwards?

I've had a DSLogic U3Pro16 for years, am really fond of it and actually prefer it over the Saleae at work, so the high-speed DSTouch might be a good company for that. Probably much better than the Fnirsi DPOX180H I got last year (which I do like, but has so many not thought-through short-comings)..

Thanks guys!

[jebem]

Hello,
A little follow-up after receiving my DSOTouch DS4T252 (50Mhz bandwidth).

I have tried to make an unboxing+review here (in french because my english speaking is.....  but there are subtitles   ) :


I'm not a professional "Youtuber" and I don't have any good camera except my old Android phone
I hope it will help some of you who, like me, are looking for a portable, affordable, fast and silent DSO with a FPGA
I have put all the information in the video description

Also, I have been in touch with their support, they are really nice and reply fast 

So I highly recommend this little DSO 


Edit: A funny checking I have done some minutes ago 

2 great tools together measuring each parts of the other one

Ref. Model DS4T252

Thank you for your honest reviews!

I was looking for a low cost handheld oscilloscope able to acquire signals in the order of 1mV, a Holdoff Trigger, and using the required hardware components to implement a "proper" design.
I believe this is the one. So I have just ordered it from the DreamSourceLab shop in AliExpress.

These guys at DreamSourceLab go to the point when advertising their oscilloscopes - No smoke and lights, just straight technical information.

Also I put here (Thank you, Aldo) a review from the well known Kerry Wong covering this model. He seems to be non biased when reviewing products.
Here for this DS4T252 model:

[jebem]

Brief comparison between the main hardware components used on the DSTouch  DS4T1012 (150MHz ) and on the DSTouch DS4T252 (50MHz).

The cheaper DS4T252 50MHz model shares many components with the DS4T1012 150MHz model like the FPGA, MCU, Flash memories and Battery management.
However, it uses a much less capable ADC/sampling clock generator, and a smaller Waveform sampling storage.

So far couldn't identify the Vertical channels active components due to low resolution pictures taken from Chinese videos.
Here the cheaper DS4T252 feature a 1mV/div, while the DS4T1012 keeps on the safe side with just 10mV/div.
That said, the Kerry Wong review on the DS4T252 shows a lot of noise, around 300 to 500uV when selecting the 1mV/div scale.
Well, we can't expect a low cost device to perform like a high end one.

Component_description	|	DS4T1012_(PCB:_TOUCH1G-V1011-P0715)	|	DS4T252_(PCB: DST252-V911-P0524)
PCB upper side	|		|
MCU	|	STM32H750 ARM Cortex M7	|	STM32H750 ARM Cortex M7
Waveform sampling storage	|	9DK17 D9PTK MT41K128M16JT-125 2Gbit DRAM	|	W9864G6KH winbound 64Mbit
Battery Manager	|	BQ25601	|	BQ25601
High Speed Data Processing	|	XC6SLX16 xilinx SPARTAN-6 FPGA	|	XC6SLX16 xilinx SPARTAN-6 FPGA
Vertical channels ADC	|	HAD1511 1000MSPS (one-channel), 500MSPS (two-channels)	|	MXT2088 200MSPS (one-channel), 100MSPS (two-channels)
ADC Sampling Clock Generator	|	ADF4360-7 24-pin IC, Integrated Synthesizer and VCO	|	8-pin IC?? Close to the ADC, but can not see details due to low resolution pictures
PCB under side	|		|
128Mbit Flash Memory	|	W25Q128JV winbound	|	W25Q128JV winbound
64Mbit Flash Memory	|	W25Q64FV winbound	|	W25Q64FV winbound

References
DS4T252 (50MHz) Analysis and teardown (Chinese language): https://www.bilibili.com/video/BV1in9nYtE2q/
DS4T1012 (150MHz) Teardown with analysis by Kerry Wong: http://www.kerrywong.com/2023/09/29/teardown-of-a-dreamsourcelab-dstouch-ds4t1012-oscilloscope/#google_vignette

[jebem]

Just received my new toy DSTouch DS4T252.

Very compact unit.
First impressions are all positive.

Very good carrying case.
Complete set of cables: USB-C to USB-A male, MCX Signal Generator output to BNC male, and two P6060 X1/X10 Oscilloscope probes.

Firmware V1.0.

[jebem]

Absolutely random, non controlled, not properly impedance adapted lines, tests on my new DreamSourceLab DSTouch DS4T252 unit.


1. Battery

It seems that it arrived with a fully charged battery, if the batt. indicator is to be believed.
After playing with the oscilloscope settings and features for around 1.5 hours, the batt. indicator was around 75%.
So it seems the claimed battery life of 3-6 hours is achievable, even if keep the meter busy to avoid it to enter into idle sleep mode.

Charging current was 5VDC at around 1.5A (with battery at around 75% charge state).
 

2. High level signal handling

Used a less than ideal Juntek JDS-2900 Signal Generator - Very noisy and amplitude imprecision at low amplitude levels and a bit jittery as well.

Trigger is almost perfect, and the Trigger Holdoff feature is very efficient handling those noisy/complex waveforms.

2.1. Single channel:

• A 10 kHz square wave is captured perfectly.
• A 20 MHz square wave is still recognizable as such with a bit of goodwill. Above that, it becomes a sinusoide.
• A 20 MHz sinusoidal wave is captured perfectly.
• A 50 MHz sinusoidal wave is captured easily, although with a small attenuation - It goes above that without issues despite increasing attenuation.

2.2. Dual channel:

• Up to 47 MHz sinusoidal wave is still acceptable, although with some artifacts, jitter, and strange voltage reading - Above that, measurements are all erratic.

3. Low level signal handling

• 1mV/div internal noise. As seen on some YT video reviews, it indeed shows around 700uV of internal noise.
  Shorting the BNC input didn't reduce it.
  I went away from the noisy workbench to a balcony with no electrical cables and turned off the router. Same result.
  This noise is being picked from internal EMF sources, and the oscilloscope shows erratic frequency readings changing with time base as well.
• 10 kHz square wave 3mVpp, registered as 5.9mVpp at 1mV/div and 20uS/div.
  A lot of noise coming from the Signal Generator.
  To trigger this waveform in a stable way, the Trigger Holdoff was set around 60% of the sliding bar.
• Same 10 kHz square wave 3mVpp, registered as 4.3mVpp at 1mV/div and 20nS/div.

[0xdeadbeef]

What I find pretty much underwhelming here is the total lack of any trigger options beyond the bare minimum.
According to the manual, there is only a rising edge or falling edge trigger. There is note even a "both edges" trigger, letting aside all the other most basic trigger options (Pulse Width, Nth Edge, Timeout etc.) which would have been really, really easy to implement.
Admittedly, an edge trigger is the one you will need most but they really didn't put any kind of effort here to do more than the absolute minimum.

[ptluis]

Your Juntek is really noisy. I post two photos of my unit using a Feeltech 6800 and Uni-T UTG932+ (+ means 30MHz model hacked to 60MHz)

The time to fully charge DS4T252 is 2h38, total capacity 3137mAh

The ground noise on my two DS4T252 units are between 566uV and 600uV

[jebem]

Your Juntek is really noisy. I post two photos of my unit using a Feeltech 6800 and Uni-T UTG932+ (+ means 30MHz model hacked to 60MHz)

The time to fully charge DS4T252 is 2h38, total capacity 3137mAh

The ground noise on my two DS4T252 units are between 566uV and 600uV

Yap, the Juntek JDS-2900 was a low cost buy just to have something able to generate large signals at 50MHz here in my mom's home where I spend more time than on my home 500km away.

Thanks for checking the low level signal registering using better Signal Generators.
It seems that most of the noise seen on your pictures is related to the oscilloscope vertical amplifier internal noise.
Did your scope triggered in auto mode those small signals, or did you had to enable trigger holdoff?

[jebem]

What I find pretty much underwhelming here is the total lack of any trigger options beyond the bare minimum.
According to the manual, there is only a rising edge or falling edge trigger. There is note even a "both edges" trigger, letting aside all the other most basic trigger options (Pulse Width, Nth Edge, Timeout etc.) which would have been really, really easy to implement.
Admittedly, an edge trigger is the one you will need most but they really didn't put any kind of effort here to do more than the absolute minimum.

Indeed.
Apparently the trigger options are very limited on these low cost oscilloscopes.

However this model features Trigger holdoff (up to 1 second), something that most of these  cheap models miss. It is really useful to get a stable registration while keep sampling and displaying on complex waveforms like AM modulated signals of the ones I posted above with a mix of square wave with lost of added noise.

[ptluis]

Your Juntek is really noisy. I post two photos of my unit using a Feeltech 6800 and Uni-T UTG932+ (+ means 30MHz model hacked to 60MHz)

The time to fully charge DS4T252 is 2h38, total capacity 3137mAh

The ground noise on my two DS4T252 units are between 566uV and 600uV

Yap, the Juntek JDS-2900 was a low cost buy just to have something able to generate large signals at 50MHz here in my mom's home where I spend more time than on my home 500km away.

Thanks for checking the low level signal registering using better Signal Generators.
It seems that most of the noise seen on your pictures is related to the oscilloscope vertical amplifier internal noise.
Did your scope triggered in auto mode those small signals, or did you had to enable trigger holdoff?

Did you try running your Juntek from battery? you will reduce noise dramatically. My Fy6800 was running from AC but my UTG932 is running from battery.

Regarding triggering, my 2 scopes DS4T252 were able to trigger in Auto mode these small signals, after triggering the signal I just had to adjust vertical resolution from 2mV to 1mV manually, in fact the trigger is able to catch really slow signals of 1Hz and adjusting the horizontal in roll mode.

The only time I had to use Holdoff was in AM signal 10MHz sine carrier with 300Hz sine mod

When trigger is not able to show a stable signal in Auto triggering, I can easily get the signal to show manually. But won't happen to often.


I'm really surprised how good this device is, measurements are accurate within specs and between different units I didn't notice differences using the same cabling set, maybe 2-3 digits nothing to be worry about, DC signal are accurate. Really good. FFT is usefull with accuracy hardly seen on pocket devices. I'm able to use FFT on mV square signals and it is able to identify correctly all harmonics.

I've found this device casually, never heard of them although their around for a few years. I glad I found them.

Also they told me that a 4Ch portable scope is in development, no release date yet, I'm looking forward to seeing it.

[jebem]

Your Juntek is really noisy. I post two photos of my unit using a Feeltech 6800 and Uni-T UTG932+ (+ means 30MHz model hacked to 60MHz)

The time to fully charge DS4T252 is 2h38, total capacity 3137mAh

The ground noise on my two DS4T252 units are between 566uV and 600uV

Yap, the Juntek JDS-2900 was a low cost buy just to have something able to generate large signals at 50MHz here in my mom's home where I spend more time than on my home 500km away.

Thanks for checking the low level signal registering using better Signal Generators.
It seems that most of the noise seen on your pictures is related to the oscilloscope vertical amplifier internal noise.
Did your scope triggered in auto mode those small signals, or did you had to enable trigger holdoff?

Did you try running your Juntek from battery? you will reduce noise dramatically. My Fy6800 was running from AC but my UTG932 is running from battery.

Didn't tried on batteries, as I only use it for large amplitude signals. I have no doubt that its AC SMPS adapter is a major source of noise.
The test on small signal was intended to see how the DS4T252 would handle a complex signal like that, and it passed in flying colors.

For lower signals I use a TinySA Ultra, and sometimes I get nostalgic and use my old HEATHKIT IG-102S RF signal generator to troubleshoot old radios as a hobby.

EDIT:
I have just recall that I have a FY6900-40 at home and it is also noisy at low levels, despite having a slightly better power supply.
The issue is the same, SMPS are noisy by nature and it must be contained, something that low cost devices don't have.
Below a picture of it.
Later on, I added ferrite beads to the power supply DC flat cable but it had negligible effect.
Then I got the components to power it from rechargeable batteries with battery management kit, but again, these instruments are not really good at supplying low level signals, so I forgot about it all and got the TinySA Ultra.

[dcm]

I am looking for advice for amateur use in a home workshop: audio circuits, automotive diagnostics, and testing power supplies, e.g., ripple and noise measurements.
I am considering buying the DS4T252 or the Owon HDS2102S. I’m stuck — which one would you choose?

[ptluis]

Two more photos

AWG UTG932+
Battery operated

50 cm coaxial cable RG316
No 50 ohm termination

AM signal
10MHz sinewave Carrier (Amplitude of 3mVpp and 10mVpp)
300Hz sinewave mod at 100%

Stable signal archived using holdoff feature

[ptluis]

Some DC measurements comparing with DM40C

Voltage reference chip is AD584LH

The DS4T252 has an option called "amplitude fine tuning" to manually calibrate DC values, but I didn't do it because I don't feel the need to. For me the displayed values are accurate enough.

[jebem]

I am looking for advice for amateur use in a home workshop: audio circuits, automotive diagnostics, and testing power supplies, e.g., ripple and noise measurements.
I am considering buying the DS4T252 or the Owon HDS2102S. I’m stuck — which one would you choose?

I believe any one of those two models will serve you well as handheld oscilloscopes to be used in a workbench at home.

However the Owon HDS2102S form design may also fit well in the field/harsh environments due to its more rugged case, and it includes a multimeter.

Not that the DS4T252 is fragile, but its case style seems to be less practical to be used in the field IMHO.

[jebem]

Some DC measurements comparing with DM40C

Voltage reference chip is AD584LH

The DS4T252 has an option called "amplitude fine tuning" to manually calibrate DC values, but I didn't do it because I don't feel the need to. For me the displayed values are accurate enough.

Impressive. I was not expecting that level of precision on these low cost scopes.

So far, the only thing I don't like in the DSTouch DS4T252, it is the case construction.
It seems that it is caseless - They just sandwiched the assemblies and then applied vulcanized rubber around it to form a case.
It seems to be virtually impossible to open this case to peek around inside without deforming or even destroying the rubber case.
After looking to some videos where it has been dismantled, I am not keen to do the same thing on my unit 

[jebem]

One DS4T252 internal EMF noise source found (probably not unique but surely it is the larger one).

Colour TFT displays with TouchScreen are a well known major source of EMF interference.

(I have several advanced DSP  receivers and only the ones equipped with a colour TFT display are picking up strong interference when using the local telescopic antenna on LW, MW and even HF lower bands. In practice this is not a real issue as we tend to use external antennas enough far away from the display.
One of them uses a colour touchscreen (TS) display and is the worst offender. This is because they do use TS polls instead of interrupts to detect a pressure).

As I suspected, the colour display with TS used in this DS4T252 oscilloscope is not different, and it indeed generates a lot of EMF interference.

To verify this issue, I used another low sensitivity oscilloscope (a Zeeweii DSO3D12  on 10mV/div) with a short loop probe at X1, and move the probe around the DS4T252 device.
The largest noise is clearing coming from the front left lower corner, decreasing  fast when moving the probe away from it.

This TS display noise seems to being picked up by the vertical amplifiers and visible when selecting the maximum sensitivity of 1mV/div.
I suspect that the vertical amplifiers may not be shielded or are insufficiently shielded at the PCB side facing the Display.

[ptluis]

Some DC measurements comparing with DM40C

Voltage reference chip is AD584LH

The DS4T252 has an option called "amplitude fine tuning" to manually calibrate DC values, but I didn't do it because I don't feel the need to. For me the displayed values are accurate enough.

Impressive. I was not expecting that level of precision on these low cost scopes.

So far, the only thing I don't like in the DSTouch DS4T252, it is the case construction.
It seems that it is caseless - They just sandwiched the assemblies and then applied vulcanized rubber around it to form a case.
It seems to be virtually impossible to open this case to peek around inside without deforming or even destroying the rubber case.
After looking to some videos where it has been dismantled, I am not keen to do the same thing on my unit 

Yes that's what I don't like either, to disassemble it we need to unglue the protective rubber, but let's hope we don't need to replace the battery for a long time.

This scope also runs directly from the charger but I've noticed that from time to time it is topping up the battery,

I did some consumption tests when the scope is running from the charger (although I don't recommend because of the increased noise it adds to signals, better to run from battery


The following values may vary depending on what the scope is doing, I tested with a simple sinewave:

Backlight 1/3
1CH, RUN, 540mA, 2.703W
1CH, STOP, 395.4mA, 1.977W

2CH, RUN, 592.6mA, 2.963W
2CH, STOP, 385.2mA, 1.926W


When both channels are enabled the increased consumption is around 260mW (52mA!)

Battery life: 2CH ON, RUN, triggering sweep 1KHz - 1MHz sine wave, after 5 hours the battery indicator was still green, then I had to turn off everything. So with some care it's possible to reach 6 hours or more.

It is able to trigger CAN signals although it cannot decode (yet) for obvious reasons. "Yet" means, maybe they add it.

Post your bugs or dream features you would like to see added to the next firmware release.

And that's all folks!

[ptluis]

One DS4T252 internal EMF noise source found (probably not unique but surely it is the larger one).

Colour TFT displays with TouchScreen are a well known major source of EMF interference.

(I have several advanced DSP  receivers and only the ones equipped with a colour TFT display are picking up strong interference when using the local telescopic antenna on LW, MW and even HF lower bands. In practice this is not a real issue as we tend to use external antennas enough far away from the display.
One of them uses a colour touchscreen (TS) display and is the worst offender. This is because they do use TS polls instead of interrupts to detect a pressure).

As I suspected, the colour display with TS used in this DS4T252 oscilloscope is not different, and it indeed generates a lot of EMF interference.

To verify this issue, I used another low sensitivity oscilloscope (a Zeeweii DSO3D12  on 10mV/div) with a short loop probe at X1, and move the probe around the DS4T252 device.
The largest noise is clearing coming from the front left lower corner, decreasing  fast when moving the probe away from it.

This TS display noise seems to being picked up by the vertical amplifiers and visible when selecting the maximum sensitivity of 1mV/div.
I suspect that the vertical amplifiers may not be shielded or are insufficiently shielded at the PCB side facing the Display.

Interesting that I couldn't replicate the noise test the same way you did it, can't see any noise on both scopes. Only if I set vertical to 1mV/div on the scope with the probe I'm able to see the "noise" at 1.5mV level

Then i decided to try a different approach instead of a shorted loop, i used a ferrite inductance as an antenna to "amplify" the noises and I'm able to detect more information from the lower middle to the left, on the back you can get a stronger noise detection.

With the little trick of using a "ferrite inductance" you can even "see" smps control chip working signals without having to touch the chip

[lapinc]

The DS4T252 has an option called "amplitude fine tuning" to manually calibrate DC values, but I didn't do it because I don't feel the need to. For me the displayed values are accurate enough.

Somehow "Amplitude Tuning" does not work on my unit - it is present in the menu (along with the "Auto Calibration"), but, unlike "Auto Calibration", nothing happens when I choose it.

[ptluis]

The DS4T252 has an option called "amplitude fine tuning" to manually calibrate DC values, but I didn't do it because I don't feel the need to. For me the displayed values are accurate enough.

Somehow "Amplitude Tuning" does not work on my unit - it is present in the menu (along with the "Auto Calibration"), but, unlike "Auto Calibration", nothing happens when I choose it.

Not on mine either. i haven't tested it because auto calibration  was done correctly. Anyway I'm making a list to send to dreamlabs containing small bugs on UI, improvements and features that could be added to the scope. This bug was added.
 

[jebem]

Interesting that I couldn't replicate the noise test the same way you did it, can't see any noise on both scopes. Only if I set vertical to 1mV/div on the scope with the probe I'm able to see the "noise" at 1.5mV level

Then i decided to try a different approach instead of a shorted loop, i used a ferrite inductance as an antenna to "amplify" the noises and I'm able to detect more information from the lower middle to the left, on the back you can get a stronger noise detection.

With the little trick of using a "ferrite inductance" you can even "see" smps control chip working signals without having to touch the chip

My probe is just a 100MHz oscilloscope probe set on X1 and shorted by the ground cable to form a small loop, and the oscilloscope was set to X1 as well, with 10mV/div.

Try moving the probe *slowly* around the border of the left lower corner of the DUT as seen in my pictures, where the antenna is the ground cable.
A little deviation and the signal indeed reduces dramatically and fade away on a 10mV/div scope.

My DUT unit is generating in excess of 20mVpp on a single point that I guess is closer to the Touch Screen IC from the front side. This TS chip is installed on the flat cable. This noise amplitude and frequency readings changes with the selected time base on the probing oscilloscope.

If you can't detect it, probably you have a different version of DS4T252 hardware assembly.

My DS4T252 unit measures in excess of 700uVpp of internal noise on 1mV/div, while yours shows quite lower levels of noise.
I wonder if the manufacturer didn't install the shield covers at the front side of the PCB assembly facing the display on some production batches from where my unit came from.
That would partially explain why your units are quieter than mine.

Anyway, I am not planing to open my unit at this time to check the shielding.

[ptluis]

Interesting that I couldn't replicate the noise test the same way you did it, can't see any noise on both scopes. Only if I set vertical to 1mV/div on the scope with the probe I'm able to see the "noise" at 1.5mV level

Then i decided to try a different approach instead of a shorted loop, i used a ferrite inductance as an antenna to "amplify" the noises and I'm able to detect more information from the lower middle to the left, on the back you can get a stronger noise detection.

With the little trick of using a "ferrite inductance" you can even "see" smps control chip working signals without having to touch the chip

My probe is just a 100MHz oscilloscope probe set on X1 and shorted by the ground cable to form a small loop, and the oscilloscope was set to X1 as well, with 10mV/div.

Try moving the probe *slowly* around the border of the left lower corner of the DUT as seen in my pictures, where the antenna is the ground cable.
A little deviation and the signal indeed reduces dramatically and fade away on a 10mV/div scope.

My DUT unit is generating in excess of 20mVpp on a single point that I guess is closer to the Touch Screen IC from the front side. This TS chip is installed on the flat cable. This noise amplitude and frequency readings changes with the selected time base on the probing oscilloscope.

If you can't detect it, probably you have a different version of DS4T252 hardware assembly.

My DS4T252 unit measures in excess of 700uVpp of internal noise on 1mV/div, while yours shows quite lower levels of noise.
I wonder if the manufacturer didn't install the shield covers at the front side of the PCB assembly facing the display on some production batches from where my unit came from.
That would partially explain why your units are quieter than mine.

Anyway, I am not planing to open my unit at this time to check the shielding.

I'll repeat the test again tonight and see if I can catch anything like you did.
when did you bought your scope? I bought both directly from the manufacturer, the first one 15 days ago, and the second 7 days ago. Overall I'm satisfied with them, the 2M memory are really handy. My understanding is that they are still working on them, improving the firmware and correcting bugs.