VOLTCRAFT DOV series

[Aldo22]

I connected mine via a USB hub.
Maybe that will help?

Very briliant idea, Aldo22, thank you very much!
Old USB 2.0 hub did it!
 

Great! 
But maybe it's a mechanical problem after all.
My dongle works fine when it's plugged directly into the scope.

[Aldo22]

Can someone help me:
I have got WiFi/bluetooth adapter, that mentioned earlier Aldo22, EP-N8568,
...
Or do I have a faulty adapter?

in both firmware V1.0.1.8.2 and V1.0.1.0.17 there is a wlan module /system/vendor/realtek/wlan.ko and looking inside the binary it seems to refer to two device alias:

• vid=7392 pid=A611
• vid=0BDA pid=B720

would you both evava and Aldo22 or someone other has a working wifi check if your wlan-bt device is one of such referred vid:pid ? (putting the dongle on linux computer with lsusb command -- for example 7392:A611 or 0BDA:B720 could be some other if it is compiled inside the kernel itself and not as module... and discerning the right device...
if you only have windows you should find a way yourself to see vid and pid)

I actually did that here:
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6344390/#msg6344390

[evava]

I connected mine via a USB hub.
Maybe that will help?

Very briliant idea, Aldo22, thank you very much!
Old USB 2.0 hub did it!
 

Great! 
But maybe it's a mechanical problem after all.
My dongle works fine when it's plugged directly into the scope.

WebControl - that’s something! 

[evava]

Can someone help me:
I have got WiFi/bluetooth adapter, that mentioned earlier Aldo22, EP-N8568,
...
Or do I have a faulty adapter?

in both firmware V1.0.1.8.2 and V1.0.1.0.17 there is a wlan module /system/vendor/realtek/wlan.ko and looking inside the binary it seems to refer to two device alias:

• vid=7392 pid=A611
• vid=0BDA pid=B720

would you both evava and Aldo22 or someone other has a working wifi check if your wlan-bt device is one of such referred vid:pid ? (putting the dongle on linux computer with lsusb command -- for example 7392:A611 or 0BDA:B720 could be some other if it is compiled inside the kernel itself and not as module... and discerning the right device...
if you only have windows you should find a way yourself to see vid and pid)

I actually did that here:
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6344390/#msg6344390

And I here:
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6344392/#msg6344392

[TomC]

So maby there is hardware missing to support the segmented memory fully.
Do we have an ADS800 owner that can test if it's working there ?
Well at least it's working basically.

I hadn't tried this until now. It seems to me that it essentially works the way the Owon manual describes it. To try it I set up my AWG to sweep from 1KHz to 25kHz with a sweep time of 20ms (see attachment 1). With the scope set up to 1.25GSa/s and 10M depth (about 8ms of signal capture), a single trigger may or may not capture the transition from 25kHz to 1kHz (see attachment 2). According to the manual, with segmentation I should be able to capture this transition due to the improved memory utilization. So I set up segmentation and in this case it allowed a maximum of 60 segments. The Owon manual also states:

(4) Once entered, if the number of segments is 100, the instrument will
acquire 100 frames of triggered waveforms. After acquisition, these 100
frames will be displayed on the screen.

Attachments 3 and 4 seem to show just this. The seg Acquired phase (frames of triggered waveforms), and the Draw phase (frames will be displayed on the screen). Now, if you manage to press Stop during the Draw phase, you'll get, in this case, the 60 segments in the Waveform Player (see attachment 5). And using the Multipurpose knob to scroll through the segments the transition from 25kHz to 1kHz can be found in multiple segments (see attachments 6 and 7). What I don't see is the timestamp mentioned in the manual. Not sure that it's necessary.

[goaty]

Now, if you manage to press Stop during the Draw phase, you'll get, in this case, the 60 segments in the Waveform Player (see attachment 5).

So basically it´s same as on DOV. It does not stop on it´s own ? "if you manage to press Stop".

Also, ususally the use-case is a bit different.
You have short burst of data with long pause in between.
Then you waste lots of memory with "nothing". With segmentation you do not record the empty data between bursts, but only the bursts (and some before and after).

(In your case, it´s more a problem of a trigger that´s not available - frequency step trigger - than that of segmentation. If you use one full 50M record buffer it´s the same as using 50 1M segments)

If you have no SPI data or serial, I used a AM-modulated signal. Just some random signal that is modulated with square pulses (short pulse, looooong pause) from 0% to 100% amplitude.

Just saw this: if it has recorded 60 segments at 10Mpts, this scope must have 600M memory !?
-> Probably the FPGA-Memory is 50Mpts, but the "Android Tablet" memory is of course much bigger. So normal segmentation means you have a dead time of transferring the FPGA memory to android.
Thats why they have the fast segementation as well, which allows only 1kSamples.
I need to check the dead time, or has it been done in this thread already ?

[TomC]

So basically it´s same as on DOV. It does not stop on it´s own ? "if you manage to press Stop".

Also, ususally the use-case is a bit different.
You have short burst of data with long pause in between.
Then you waste lots of memory with "nothing". With segmentation you do not record the empty data between bursts, but only the bursts (and some before and after).

(In your case, it´s more a problem of a trigger that´s not available - frequency step trigger - than that of segmentation. If you use one full 50M record buffer it´s the same as using 50 1M segments)

If you have no SPI data or serial, I used a AM-modulated signal. Just some random signal that is modulated with square pulses (short pulse, looooong pause) from 0% to 100% amplitude.

Just saw this: if it has recorded 60 segments at 10Mpts, this scope must have 600M memory !?
-> Probably the FPGA-Memory is 50Mpts, but the "Android Tablet" memory is of course much bigger. So normal segmentation means you have a dead time of transferring the FPGA memory to android.
Thats why they have the fast segementation as well, which allows only 1kSamples.
I need to check the dead time, or has it been done in this thread already ?

The example I used was just a quick way to try to verify the description in the manual, not really a good example of where you would use segmentation to solve real life problems. I've been investigating for a while what you describe on your last edit. And yes, the only explanation I see is a two-tier memory scheme. Got ChatGPT involved and the general architecture is shown on attachment 1. The detailed explanation is on attachment 2. The ChatGPT explanation seems quite plausible to me, so I'll leave it open for comments.

[goaty]

I played a bit with the trigger output.
For 50MPts, dead time is 20ms.

In segmentation, it´s about 130µs.

 And 85µs in fast segmentation (1kPts).


Funny: In normal acquisition, the dead time is less when using 10kPts (82µs) compared to 1kPts (117µs).


I must admit that I'm not quite understanding the trigger output.
Neither low not high pulse corresponds with the calculated time of
Horizontal scale *18/ sampling frequency.
Ah whatever. What was I trying to do?

[TomC]

Funny: In normal acquisition, the dead time is less when using 10kPts (82µs) compared to 1kPts (117µs).

At 10kPts there are 6 cycles within 2ms, at 1kPts there are 8 cycles within 2ms. I'm not sure how to interpret this, but it seems overall that 1kPts is acquiring faster.

[TomC]

Funny: In normal acquisition, the dead time is less when using 10kPts (82µs) compared to 1kPts (117µs).

At 10kPts there are 6 cycles within 2ms, at 1kPts there are 8 cycles within 2ms. I'm not sure how to interpret this, but it seems overall that 1kPts is acquiring faster.

This what I think is happening. Just to get the approximate pulse length on the 10kPts trigger out signal I'm looking at the first 2ms, there are about six 83µs neg pulses and five 300µs positive pulses. With that in mind, let's remember that the center of the acquisition period more or less coincides with the positive excursion of the trigger. So in this case the active period is more or less 83 + 83 = 166µs, and the dead period more or less 300 - 83 = 217µs.

For the 1kPts, the signal is pretty close to a square wave. The negative pulses are around 118µs and the positive pulses are around 132µs. So in this case the active period is more or less 118 + 118 = 236µs, and the dead period more or less 132 - 118 = 14µs.

[Aldo22]

The behavior of “High Res” is a bit hard to predict.
Whether it’s “effective” depends on s/div and memory depth.
Attached are a few examples:

Mode: Sample, 1 ms/div, 1 M
Mode: HR, 1 ms/div, 1 M
Mode: Sample, 200 μs/div, 100 k
Mode: HR, 200 μs/div, 100 k
Mode: HR, 200 μs/div, 1 M

[TomC]

The behavior of “High Res” is a bit hard to predict.

According to the manual:

High Res: Reduce and improve the signal-to-noise ratio on aperiodic (single-shot) waveforms.

Don't know for sure If this will make a difference. I'll try to test it later.

[Aldo22]

According to the manual:

High Res: Reduce and improve the signal-to-noise ratio on aperiodic (single-shot) waveforms.

Don't know for sure If this will make a difference. I'll try to test it later.

I think this is meant to contrast with “Average.”
“Average” only works with periodic waveforms and a stable trigger.
HR works in all cases, but apparently not with all settings.

[goaty]

HR works in all cases, but apparently not with all settings.

Usually what it does is combine multiple samples to increase the ENOB. At least that´s how RTB is doing.
Basically it´s averaging per sample, not per trace as normal averaging mode does.
So I guess the aquisition depth (1k, 10k, ...) is limited for HR as well as the horizontal scale of course.

[Aldo22]

HR works in all cases, but apparently not with all settings.

Usually what it does is combine multiple samples to increase the ENOB. At least that´s how RTB is doing.
Basically it´s averaging per sample, not per trace as normal averaging mode does.
So I guess the aquisition depth (1k, 10k, ...) is limited for HR as well as the horizontal scale of course.

Based purely on observation, there appears to be a limit at 25 MSa/s.
Up to that point, High Res seems to be effective.
Above that (e.g., from 62.5 MSa/s onwards), there is hardly any visible difference.

Can anyone confirm this, or am I overlooking something?

[TomC]

Based purely on observation, there appears to be a limit at 25 MSa/s.
Up to that point, High Res seems to be effective.
Above that (e.g., from 62.5 MSa/s onwards), there is hardly any visible difference.

Can anyone confirm this, or am I overlooking something?

I'm observing the same thing (see the screenshots). Based on the info I gathered from the following links, I think there is a limit sampling rate where the DOV/ADS scopes can no longer oversample the signal enough to support High Res. The limit seems to be anything beyond 62.5MSas.

https://www.electronicdesign.com/technologies/test-measurement/article/21801317/why-you-should-care-about-oscilloscope-acquisition-modes

https://www.google.com/search?q=On+an+ADS824A%2C+explain+the+High+Res+Acquisition+option&oq=On+an+ADS824A%2C+explain+the+High+Res+Acquisition+option&gs_lcrp=EgRlZGdlKgYIABBFGDkyBggAEEUYOdIBCTE3Njg1ajBqN6gCALACAA&sourceid=chrome&source=chrome.ob&ie=UTF-8

[Aldo22]

Just one small thing I noticed: "Single Shot" sets "Average" mode to "Sample".
But "Stop" does not.

[Arek]

I buy this dongle and it does not work, don't buy it.

[Frederik]

I am looking into the possibility of adding the missing AFG components to my VOLTCRAFT/OWON scope. Microchip has posted a picture of the ADS 914 PCB and I tried to make a part list of what’s missing. The components that I can’t identify are marked with a question mark in the list. I hope some of your guys can help to identify.
 
If there is interest we could define a AFG component kit together. A better picture of the ADS 9xxx PCB would help.

Link to Microchop’s picture
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6277830/#msg6277830

Link to the partlist.xlsx
https://docs.google.com/spreadsheets/d/1R1Aayle2ksnMFIUz9xdALPenyD_pSGie/edit?usp=drivesdk&ouid=106656474749086766187&rtpof=true&sd=true

[evava]

I buy this dongle and it does not work, don't buy it.

I bought this on Allegro, and it works for me, but only through USB 2.0 hub.
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6334940/#msg6334940

and Bluetooth works fine, with bluetooth speaker, e.g. usable for videos from eevblog/YouTube (like Aldo22 showed earlier)

[sdroamt0]

https://it.banggood.com/OWON-ADS804A-ADS814A-ADS824A-Digital-Oscilloscope-70MHz-4Channel-1_25GSa-or-s-12-bit-Vertical-Resolution-High-Res-Waveform-Capture-50M-Memory-Depth-p-2032175.html?cur_warehouse=CN&ID=6332318

what do you think, what are the chances that these are the original owon versions with the new mainboard and then without ripple?

[goaty]

The Edimax I ordered from Pollin works, but only Wi-Fi. Bluetooth is not working.
It works directly in front USB.
VID is 7392
PID is A611

There must be a tiny difference in the chip or whatever that causes Bluetooth not to work.
I cannot switch it on in the settings.

[SL75]

https://it.banggood.com/OWON-ADS804A-ADS814A-ADS824A-Digital-Oscilloscope-70MHz-4Channel-1_25GSa-or-s-12-bit-Vertical-Resolution-High-Res-Waveform-Capture-50M-Memory-Depth-p-2032175.html?cur_warehouse=CN&ID=6332318

what do you think, what are the chances that these are the original owon versions with the new mainboard and then without ripple?

Unfortunately, no one can really say that unless they just bought such a device and the seller only sells devices from the same production batch. 
By the way, the new circuit board isn’t completely “ripple-free.” Owon didn’t get rid of the 7.6 kHz ripple caused by the ripple current drawn by the EXOR gates. Even though the 7.6 kHz disturbance is much smaller, you can still reduce it significantly through circuit modifications, as has been shown several times here.

[TomC]

I bought this on Allegro, and it works for me, but only through USB 2.0 hub.
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6334940/#msg6334940

I have one just like it I bought from Amazon. I normally use it with a hub, so I can also have a mouse & keyboard and a flash drive, but it also works directly connected to the scope's USB port. The attachment is a scan of the box it came in.

[Aldo22]

I bought this on Allegro, and it works for me, but only through USB 2.0 hub.

I have one just like it I bought from Amazon. I normally use it with a hub, so I can also have a mouse & keyboard and a flash drive, but it also works directly connected to the scope's USB port. The attachment is a scan of the box it came in.

I don't think that's a distinguishing factor.
evava seems to have a mechanical problem with his dongle.
He also mentioned issues with the PC ports.
It's probably not a widespread problem with this dongle.