Velleman WFS210 wireless oscilloscope for only £20 !

[Avacee]

You are correct that you don't get voltage values and that the unit sends screen positions to the PC/tablet as a viewer.

The software works on having 10 vertical divisions so for a setting of 500mV/div you get from -2.5V to +2.5V.
There's no offset so you are stuck with 0V being the midpoint.
The unit sends where the voltage reading will be within those bounds based on 3=Top of screen .. 128=Centre .. 252=Bottom of screen .. 250 different values or 25 per division.
So for 500mV/div a value of 3 = 2.5V .. 128 = 0V .. 252 = -2.5V .. and simple maths for the rest.

If you send the ChxY Position in a Settings message the unit will offset the values by this much.
This you could do yourself but by having the unit do it for you you can simply dump the byte stream into a viewer with no further processing.

So the datastream that comes back is the form:
byte 1: CH1_Sample1    (value between 3 and 252) .. if you view the stream with nothing attached you'll see a lot of 0x80 = 128 = 0V
byte 2: CH2_Sample1
byte 3: CH1_Sample2
byte 4: CH2_Sample2
byte 5: CH1_Sample3
byte 6: CH2_Sample3
....
byte 8192: CH2_Sample4096

Amen on it needing a new firmware 

Baudrate:
I've tried 460800 and it didn't throw an exception and returned data.
I haven't tried any throughput tests though to see what it actually connected at .. job for tomorrow!
I initially used 19200 as I found that as a default in the firmware code in usb_function_cdc.c -> CDCInitEP(void)

[Bicurico]

What I don't understand: why does the device reply with 4096 bytes?

If it is for two channels, that's 2048 bytes for each.
The screen is only about 660x400 pixel.
If each byte represents the hight of the curve, you would need two bytes for 400 pixel height (actually a few bit less).
That would still be 1024 byes for each channel or 1024 pixel. That just does not equate!

I would expect a resolution of i.e. 512 x 256 pixel and each trace would be composed of 512x2 = 1024 bytes for both channels.

At 660x400 pixel, 4096 data bytes just don't make sense.

Plus, can it be that they are sending the grid data, too? I know, I should look in the source code, but I have been busy and tired, so I am just experimenting things.

Regards,
Vitor

[Avacee]

The protocol has nothing to specify how wide the display is.

Is it on a 660x400 display, or a 1280x800 (my tablet) or a 1920x1080 monitor?
I can stretch the windows app over 3 monitors for 5760x1080 .. and it cuts off after it shows the sent samples :p
If the screen isn't wide enough to show all the samples then the excess are discarded.

Someone at Velleman decided the unit would send 4096 so it does.

Edit: Added screenshot to show cutoff

Edit2: Regarding "Is it sending the grid data"?
It is not. The Velleman app works by dividing the height of the grid control by 10 to get the pixels per division.
Altering the height of the app doesn't gain you more vertical divisions - it's always 10 divisions as mentioned in my last post about V/div and the unit's +/- range.
It then maintains a square grid by using the same number of pixels per division horizontally.

[Bicurico]

Hi,

Thank you a lot in helping me understand this.

In fact, I did not try changing the WFS210 software window size - of course I can stretch it.

Also, I was mentally blocked in thinking in terms of horizontal screen resolution - in case of an oscilloscope, it makes of course no sense: the more samples, the better and what you do with them is up to you. The problem was that I was using a composite video signal to test with and was trying to get exactly one scanline into my screen... 

I think I am now getting ready to actually start implementing something.

My next test will be to try some XY rendering. I will use a oscilloscope demo *.WAV file. Let's see what will happen. But for now I will hit the bed!

Regards,
Vitor

[IAmBack]

Got mine.

I managed how to make it work with my tablets. It appeared to be not straightforward to start it every time on my equipment. Now with little fiddling software starts every time. For those who may concern, I need to go to app manager, turn the app off (if it is running), next remove all data, and then start it again in normal way. Sometimes it needs to switch between normal and demo mode, bit finally starts every time.

I've noticed, that there is no difference in the plot with time base settings from 5us/div till 1us/div...
Anyone can confirm this? How it is on windows?(I'm quite reluctant to install sw and drivers on my laptop).

/edit/
Just pressed "autorange" button while 1MHz sinewave was connected to the input. This thing have switched to... 5ms/div, and displayed beautiful sine plot. What a excellent tool to demonstrate aliasing and Nyquist theorem!

[Avacee]

I've noticed, that there is no difference in the plot with time base settings from 5us/div till 1us/div...
Anyone can confirm this? How it is on windows?(I'm quite reluctant to install sw and drivers on my laptop).

That's confirmed as a bug by Velleman

https://forum.vellemanprojects.eu/t/wfs210-tested-and-this-product-has-issues-request-for-support-comment/27511/2

[sorenkir]

Following  Ian.M's advice, I have removed the DC-DC converter from the ADUM3160 based USB isolator and wired an USB cable connected to a phone charger:



This is working, the WFS210 is isolated from PC & bench ground and can be connected to test equipment and linked to the PC with the USB (without Wi-Fi):



Michel.

[CanadianAvenger]

I realize this is an old thread, but if anyone is interested the firmware is available on GIT (now).

https://github.com/Velleman/WFS210-Firmware


I picked one of these up off of a clearance shelf, and wanted to hack around with it. From this thread, it looks like there was at least some interest in the past from others doing the same. I think the hardware platform has potential to be useful for some applications, the current implementation sucks.

[djos]

I realize this is an old thread, but if anyone is interested the firmware is available on GIT (now).

https://github.com/Velleman/WFS210-Firmware


I picked one of these up off of a clearance shelf, and wanted to hack around with it. From this thread, it looks like there was at least some interest in the past from others doing the same. I think the hardware platform has potential to be useful for some applications, the current implementation sucks.

Mine died recently, so it's on the healing shelf for if I can ever be bothered digging into it.