PlainDAQ - open source DAQ module for Raspberry Pi Pico

[Berni]

No problem.

Don't worry not all kinds of compression are hugely complicated. The RLE (Run Length Encoding) is incredibly simple even, you simply rewrite "ABCCCCCCCCCD" as "AB9xCD" shortening repeating characters by counting them and saying "Insert X number of these here". But as you can figure this only works well with very repeating data. For wavelet transform it is best to check a good youtube video explanation on what it is and how it works. It basically decomposes the waveform into wavey pulses called wavelets (Much like the Fourier transform decomposes into sine waves). Simple signals can be represented by summing together just a few of these wavelet shapes, so only those need to actually be written down, thus massively reducing the amount of data. Decompression is then simply adding those wavelets up again to get the waveform back. This is pretty much how JPEG and a lot of other lossy image compression works.

For the case of this DAQ it probably makes more sense to instead just capture high speed ADC data to RAM and then transfer it as a single capture. That way you avoid the USB bottleneck all together while a lot of DAQ applications are fine with just fixed length captures rather than a continuous stream.

[palpurul]

For the case of this DAQ it probably makes more sense to instead just capture high speed ADC data to RAM and then transfer it as a single capture. That way you avoid the USB bottleneck all together while a lot of DAQ applications are fine with just fixed length captures rather than a continuous stream.

I'll start with no compression and try to find out the maximum data length that I can stream to PC I guess.
I looked around for large SPI rams (PSRAMs and SRAMs) and those are the things that I've found.
https://www.mouser.com.tr/c/semiconductors/memory-ics/dram/?type=PSRAM%20%28Pseudo%20SRAM%29
They are a bit expensive for this application I definitely want to try them out!

Maybe PlainDAQ comes with a memory depth option, and the option includes one of these PSRAMs as extra

[palpurul]

The RLE (Run Length Encoding) is incredibly simple even, you simply rewrite "ABCCCCCCCCCD" as "AB9xCD" shortening repeating characters by counting them and saying "Insert X number of these here". But as you can figure this only works well with very repeating data.

That could work great with a logic analyzer. 

[Berni]

The RLE (Run Length Encoding) is incredibly simple even, you simply rewrite "ABCCCCCCCCCD" as "AB9xCD" shortening repeating characters by counting them and saying "Insert X number of these here". But as you can figure this only works well with very repeating data.

That could work great with a logic analyzer. 

Yep this is indeed what they use on a lot of logic analyzers to squeeze way way more data into the sample memory. When nothing is moving it is using pretty much 0 memory.

[palpurul]

GUI Update

Movable/Interactive Vertical cursors for viewing real-time statistical data.

While waiting for the new PCBs, I had time to work on the GUI for PlainDAQ.

Here is a nice gif to show how it works:


It works by pressing and hold down the middle mouse button. When the cursors are active the plot shows the time difference between the cursors, RMS and peak-to-peak value of the samples in between the cursors in the upper left corner as you can see.

What do you think I should add to the statistical measurement?
Here is a list of the measurement capabilities that I like to add:
✔RMS
✔Average
✔Std. Dev.
✔Peak-to-peak
✔Min.
✔Max.

If you want to try it, it's in my Github

[PlainName]

It works by pressing and hold down the middle mouse button.

On Windows the middle mouse button isn't too common. Although I am not a user of this project I can say that my logitech mouse isn't set up for middle button (and when it is it's a curled finger operation, so a bit of a faff). Perhaps a configurable option, or ctl+right/left might be appreciated by some users.

[palpurul]

It works by pressing and hold down the middle mouse button.

On Windows the middle mouse button isn't too common. Although I am not a user of this project I can say that my logitech mouse isn't set up for middle button (and when it is it's a curled finger operation, so a bit of a faff). Perhaps a configurable option, or ctl+right/left might be appreciated by some users.

Thanks for the feedback.
Maybe I just put a button that turns on/off the cursor and let the user place wherever they want by picking a cursor with left button and dragging it with holding the left button.
Does that sound more intiutive?

[PlainName]

It sounds a little unusual, to be honest! But, on the other hand, it should be easily discoverable, which arbitrary clicks and presses tend not to be. Any reason you couldn't have both? The button for users who don't use that feature much (or are new to the software) and middle-click for power-users?

[palpurul]

It sounds a little unusual, to be honest! But, on the other hand, it should be easily discoverable, which arbitrary clicks and presses tend not to be.

I'll think about it.
I thought it would be very intuitive, but it seems like it's only intuitive for me

Any reason you couldn't have both? The button for users who don't use that feature much (or are new to the software) and middle-click for power-users?

I think it can be done, but I just want to keep it simple.
I'll make one more iteration show both versions. 

[palpurul]

I've been working on a new revision of PlainDAQ and I had an empty space left on my PCB. I decided to fill it up with some visuals related to sampling theory.

Here is what I've drawn: The thick line supposed to be the signal that's being measured and the dotted line supposed to be the reconstructed waveform. To be honest not sure if it's terrrible accurate depiction.  (Well I'm pretty sure it's not accurate now  )


Let us hope that it will be printed well on the silkscreen layer 
I drew this on inkscape and imported it to KiCAD via image converter.

[palpurul]

I’ve been working on a number of different aspects of PlainDAQ lately. In particular, I’ve been cost optimizing the PCB, picking components that are easy to source, and working on an intuitive graphical user interface.

First of all please join the poll on what module to use, I need your feedback.
You can also vote in my Twiteer poll: https://twitter.com/AlperenAkkuncu/status/1521929542218264577

Recent Changes to the PCB

• I moved all the terminals to the same side so they are easier to use. (In the previous version, analog inputs and outputs were on opposite sides, which would have been inconvenient when using both inputs and outputs.)
• The front-end is a bit more complicated, but it's a lot cheaper and has the same performance.
• The form factor has changed as well. PlainDAQ was previously 75 x 60 mm, and it's now an 80x60 mm rectangle, so it has a slightly larger footprint.
• I added a new visual that displays the basic input and output capabilities of the board along with a cool diagram.
• I added a Qwiic connector.
• I changed the terminal to a more affordable type. The new terminals also accept Dupont cables which is going to make it easier for you to interface with PlainDAQ.

Help Me Decide Between Wi-Fi & Bluetooth…
I need change the wireless module, but I need your help to decide between…

• An ESP-WROOM-02 Wi-Fi module (first render below)
• An HM-BT4502(A) BLE module (second render below)

New PlainDAQ with ESP-WROOM-02 (Wi-Fi)

New PlainDAQ with HM-BT4502(A) (BLE)


Please join the poll on what module to use, I need your feedback.

I also like to show the visuals that I added

• Designed a simple logo
• Added a simple visual depicting aliasing
• Added informative text about PlainDAQ

Here it is, tell me what you think 

[Berni]

Neat looking board there.

Oh and your sinx/x interpolation appears to be aliasing due to a too low sample rate. (Then again otherwise seeing the dotted line would be difficult otherwise since it would line up almost perfectly)

[palpurul]

PROJECT LOG: Developing a reconstruction filter for the DAC in PlainDAQ, to get nice looking sinewaves.
Compared Sallen-key(SK) and Multiple feedback(MFB) topologies.
MFB performed much better in the stop-band.

LTSPICE files:
https://github.com/AlperenAkkuncu/PlainDAQ/tree/main/Development/SPICE_simulation/Active_filter_comparison

I had to choose between sallen-key or Multiple feedback topology, so I compared them.

Here are the circuits that I compared:


Here are the results:


Multiple feedback performed much better in attenuating the stop band.
I added them to the schematic in my KiCAD project. I also added a switch to switch between filtered and non-filtered DAC output to see the difference between, it's going to be a fun little experiment. 
With the help of the switch I will be able to generate square waves with faster rise-time by using the non-filtered DAC output which is a plus.


We also release a new update on CrowdSupply. It's got the new PCB and sneak peak of the GUI.
Subscribe if you want to get the future updates.

[palpurul]

Hello,
It's been a quite long sine my last post. There are quite a few changes in PlainDAQ

First of all please check the update that we released on CrowdSupply and subscribe from the same page if you want to get updated: update
It's got the revision of PlainDAQ and the sneak peak of the GUI that'll be used with PlainDAQ.

In the meantime I prepared some visuals for PlainDAQ which shows the connections and the components used in it.

Here they are:

Brief:


Components:


Let me know what you think, what do you think about the colors? Do visuals looks descriptive?

Thanks!

[palpurul]

GUI update

Hello once again!
Finally I've had time to work on PlainDAQ's GUI. I worked on vertical cursors and I came up with two ways to control vertical cursors and I am trying to decide which way is the best.

Please see the video and let me know which way you think is the best.



In the video I am showing two ways to control cursors. The first method is by middle mouse button and second way is by using ON/OFF button and left click for cursors. This is my first video edit, I hope it looks descriptive and good. 

Subscribe to PlainDAQ if you want to get updated when we launch: https://www.crowdsupply.com/kuncu-teknoloji/plaindaq
Note: Pre-launch page will be updated soon. The pictures shown in the pre-launch page are old.

[PlainName]

The right one seems better to me, partly because the left one isn't discoverable and uses non-intuitive controls. Only today I've spent more time that I would be embarrassed to let on trying to remember how to exit the menu on a RD6002 PSU (no, none of the keys used to get into the menu, or the ones used to navigate it - you press the rotary dial). A left mouse click to remove something isn't something I'd guess, and it'd drive me mad trying to use that to make things happen.

But mostly because the right one is how scopes tend to do it - there is usually an explicit cursors on/off  button

[palpurul]

The right one seems better to me, partly because the left one isn't discoverable and uses non-intuitive controls. Only today I've spent more time that I would be embarrassed to let on trying to remember how to exit the menu on a RD6002 PSU (no, none of the keys used to get into the menu, or the ones used to navigate it - you press the rotary dial). A left mouse click to remove something isn't something I'd guess, and it'd drive me mad trying to use that to make things happen.

But mostly because the right one is how scopes tend to do it - there is usually an explicit cursors on/off  button

I agree, I think you reccomeded this method when I first post a video about it

Thank you for your contribution! 

[PlainName]

You're welcome

[PlainName]

Another thought on this: with the tickbox/button thing to activate the cursors, it would be much easier to use (and design to use) a touch screen without any change. No idea why that occurred to me when reading about electric cars!

[palpurul]

Another thought on this: with the tickbox/button thing to activate the cursors, it would be much easier to use (and design to use) a touch screen without any change. No idea why that occurred to me when reading about electric cars!

That actually makes more sense, why didn't I think of that LoL 

[palpurul]

Hello everyone!

After a stupidly long delay, I finally managed to get PCBs for our new design, so I used them to assemble two boards, and I think they look great. I hope you all enjoy the new look!

Here are the picture of the PlainDAQ, let me know what you think 👇👇👇





I know they don't look that consistent, I am still learning about photography and my phone is really crap 

Check out our new page: https://www.crowdsupply.com/kuncu-teknoloji/plaindaq

This is the update that we released about the new boards: https://www.crowdsupply.com/kuncu-teknoloji/plaindaq/updates/our-new-design-is-assembled

I have made the following changes in this design:
✅ Obviously the new form factor
✅ Added a Wi-Fi Module
✅ Switched to MCP4911 DAC because of availability.

Some people suggested me to remove the Wi-Fi module because new raspberry pi pico W already includes and therefore I am thinking about adding new features, what do you think I should add?

PS: I've also designed I2C controllable filters, but more on that later (you can see it on the update page.

🔵 Lastly Subscribe if you want to get notified when we go live: SUB

[Marco]

I don't see much in the way of input protection (no larger SMD resistors, not much space behind the connectors). Can those inputs take mains voltage?

[palpurul]

I don't see much in the way of input protection (no larger SMD resistors, not much space behind the connectors). Can those inputs take mains voltage?

Input protection is something on my list, but I haven't implemented anything in terms of input protection.

It cannot take mains voltage that's too large, maybe you're thinking that it's like a multimeter, but it's more like an oscilloscope.

Input range in this revision is:
+/-4V, +/-2V, +/-1V

However, I am in the process of collecting feedback I can move the ranges to:
+/-8V, +/-4V, +/-2V
+/-12V, +/-6V, +/-3V
+/-16V, +/-8V, +/-4V

[Marco]

A 220k resistor with a parallel capacitor is cheap.

[palpurul]

A 220k resistor with a parallel capacitor is cheap.

THis might be something that I don't know, could you explain to me how that protects the inputs?

That's the input section of PlainDAQ. It's terminated to ground with a 1 Meg resistor and a small series resistor. I know that series resistor can protect to some level, but I'm sure it's not going to protect against mains voltage