Practical DIY projects for Micro Controllers

[Sercrets]

So I have looked online, on reddit, quora, etc, and we all know the cliche answers, such as "Oh make an alarm system for your room", or "Build a simple snake game", etc. and whilst those ideas may be cool, they don't really amuse someone who is older than a 12 year old.
So what I'm really asking is what are some practical applications for things to make. Obviously it all depends which microcontroller you use, so some examples for super cheap 30 cent controllers to 3 dollar controllers, to above and beyond would be nice. I am new to microcontrollers, however I own several arduinos, and am looking into purchasing other micro controllers. So to sum it all up, what are actual useful/fun projects to make that do more than amuse a mere child and are actually worth making.

Thanks in advance, Cheers

Sercrets.

[Grandchuck]

Books about this are available.  Have you looked at these?  https://create.arduino.cc/projecthub/projects/tags/microcontroller?page=1

[jpanhalt]

My first project 20+ years ago was a semi-automated retriever for the winch line used to launch model sailplanes to thermalling altitude.

Not many 12-year olds in that hobby.  Of course, you may not be interested either.  Several other projects followed related to my hobby farm. 

My current project is something that might have more general interest.  I have to cook to eat, but I enjoy doing it too.  My project is a cooking thermometer that predicts done time reasonably accurately, once the roast temperature starts to rise.  Next phase will be to add color, like this:  http://up.csail.mit.edu/science-of-cooking/home-screen.html

Bon appetit.

[Sercrets]

My first project 20+ years ago was a semi-automated retriever for the winch line used to launch model sailplanes to thermalling altitude.

Funny enough I was actually interested in getting into model sailplanes, but never got into it due to the winch issue. Do you have any more info on how you did it? You recent project looks very interesting. Any tips for how to select which microcontroller to buy for a project? No sense spending more on something too powerful/underpowered for a project... Which performance aspects should I look out for etc? Thanks in advance; cheers.

Books about this are available.  Have you looked at these?  https://create.arduino.cc/projecthub/projects/tags/microcontroller?page=1

Briefly I have, however most of it never really interested me; however I see new things have been posted on making a boat-loader out of an uno to program another microcontroller.

[jpanhalt]

Sure, here are some pictures:(attached)

That basically is a very large spinning reel that required a weekend of machining and a large aluminum skillet.  The MCU part was a model servo latch coordinated with the winch pedal so retrieve and launch couldn't happen at the same time.  That would be bad.

EDIT: I should add that backlash on the winch and retriever speed were my first attempts at high current PWM.  They were 100 to 200 A at 12V.  The MCU played a role in that too.  The last image is the comptroller for the retriever.  The one for the winch was more robust.

EDIT2:  Just noticed the retriever half open is an old photo.  My first attempt for the latch was a solenoid.  That was not reliable.  The final version used a mini servo driven by a 12F509 (as I recall).  That allowed the two positions (open/closed) to adjusted accurately and reproducibly.  There was a sensor on the latch to determine whether it was closed (i.e., launch position). A backup sensor (IR proximity) was used for safety.  It really would have been a mess to launch with the retriever in retrieve position.  I think the proper Australian term is  "a gutsa."

[Sercrets]

Interesting, I shall have to look into gliders again. Any more suggestions from anyone is greatly appreciated

[Circlotron]

Replace the mechanical timer of an old washing machine with something entirely electronic. There are several inputs e.g. water level, door open, out of balance, as well as various user controls. Outputs are water solenoids, motor power, motor direction, buzzer, indicator LEDs. The sky is the limit.

[RJSV]

   Hi there:
   I will be needing a 2nd 'Base Station', having 4 small motors, each probably needing bi-directional drive, standard 'H' bridge.  Toy sized Dc PM motors draw around 800 mA or less, on 3 V to 5 V 'H' bridge outputs.

   I can do the enclosure, and any mechanical tasks needed.  I'd rather write the microcode, but here's the thing:
   Years of disability makes for 'Chronic Fatigue' which is outcome that many Lyme Disease folks mention
That means I can write machine language CODE, no problem, especially real-time primatives, (like 1 millisecond interval structured into code loops.)
   Problems arise, more with development tools and the myrid details needed to support a decent Editor / Compiler, and functions dealing with pulses; moderate accuracy is ok, and not much below 100 uSec.
   The 'bank', of 4 Dc motors, is 'tracked' by a 16 byte
( X four channels)  table containing status bytes, etc. per each of the individual motor - output channel coming from the BASE Station.
   Of course, the expanded function version needs 5 of those.  I'd suggest mem allocation, just a simple 16 bytes which makes for a complete status indexing table below 2048 bytes.

  Plenty of that table might end up blank; The main need / convenience, is to have a 'error' count, for (later) tracking and diagnosing, problems with the serial NETWORK, sourced by the BASE Station.
Likely need to upload that sort of TESTING / Switch Characterizing data, (it's all informal during development).

    Oh, and WHAT IT IS:
   The project, is a NETWORK Switch, and...no; It doesn't run at 2.4 GHZ,  not even close:
   The thing runs at speeds around  1/20'th HERTZ !
That's 0.05 hz. as a mechanical device.

   Code is no problem, but we might do best, if I just flesh out the firmware, leaving more (routine) aspects like ROM burning etc. TBD.
Thanks, there!

[RJSV]

   Photo of first proto BASE STATION #1:
   Manual push-buttons for bi-directional motor drive, keeping pulses as short as possible (0.200 Seconds, minimum 'finger' delay.  Single motor output channel, only.

   Second version, possibly Arduino, for issue impulses at 200 mSec each, for advancing switch actuation travel.
   Test fixture also for (stable) so-named 'Pull-Away' timing characterization.

   You would see that 'mock-up' assembly, for each of 5 'channels', each driven by running CODE, in Arduino, or multiple simple 8-bit controllers.

   I'd love to do this control center VIA a vintage COMMODORE64.
YEA, YEA
 
   

[John B]

The real problem is that you have many solutions in search of a problem. What do you need to solve around your house, or maybe in your car that you could do so using electronics? A few things around my house have been automatic lighting and ventilation.

[james_s]

My first microcontroller project 22 years ago was a clock based on some B-7971 alphanumeric nixie tubes I had acquired. After that I started building a lot of different stuff. That clock is still working, sitting on the mantle above the fireplace behind me as I write this.

[brucehoult]

You seem to be approaching this from the wrong angle.

You need to ask yourself "what problems do I need to solve?" And then consider whether the solution should be mechanical, electro-mechanical, electronic, digital, or what. Not all problems need a microcontroller, or even electronics, but many things that 40 years ago would be best done with a few 7400 chips, a 555 or two, maybe an op-amp are now better done with a microcontroller plus at most some opto-isolation or power transistor for the electronic part.

No one can tell you what specs you need. The input and output pins/interfaces and amount of RAM and ROM/flash is determined by your application. CPU power-wise you can get a long long way with a 16 MHz 8 bit 2k RAM 32k ROM ATMega328 like in the Uno, but at the same time a 600-960 MHz, 32 bit, FPU, 1 MB RAM, 2 MB flash Teensy 4.0 is only $20, so unless you're planning mass-production why limit yourself?

Boards running Linux and with 64 to 512 MB of RAM can cost even less -- but with fewer I/O pins.

(one reason: the ATMega will be much easier if you want to write every single line of code to set the hardware up yourself)

[rstofer]

While not necessarily useful projects in themselves, they do provide a toolbox of solutions:

Arduino Starter Kit (UNO included)

https://www.amazon.com/ELEGOO-Project-Tutorial-Controller-Projects/dp/B01D8KOZF4

Raspberry PI Starter Kit (PI not included)

https://www.amazon.com/Freenove-Raspberry-Processing-Tutorials-Components/dp/B06W54L7B5

In both cases, the tutorials are excellent and code is provided.  PI development is done on the target machine so installation is a breeze.  The Arduino code will build on whatever platform is hosting the Arduino IDE.

There are MANY other kits on Amazon.  I picked these because I have actually bought them for myself.  They are truly excellent.  You can also search Amazon for robot kits, either Arduino or Raspberry Pi.  The Pi versions are very powerful with capability for things like computer vision, object detection, networking, etc.  Code by user, generally, but there is a lot of stuff on the Internet.

In any case, these kits help build up tools that might be useful for other applications.

[eugene]

So what I'm really asking is what are some practical applications for things to make.

There are no practical applications when you can buy something ready-made for less money.

What projects have you completed already?

[Old Printer]

Life long model airplane builder here, and flew sailplanes competitively for many years. Winch & hi-start launching are the “pure” forms of the sport, but they require a substantial amount of reasonably mowed flat land. Standard winch layout needs about 1000 feet, and has to be into the wind. If the wind direction changes by more than 45 degrees the whole setup has to be changed. We used to fly on sod farms. Technology of the last 15 years or so includes lipoly batteries and brushless motors which equal very light and powerful power systems, and coupled with gearbox’s and folding props make for a built in launching system with little weight and drag penalties. This opens up many smaller flying sites and a whole lot less work than dragging around a winch/retriever/deepcycle battery setup. Been there, done that, don’t miss it. There is an entire class of sailplane competition devoted to this type of flying that includes an altimeter based motor control that cuts off the power at 500 feet. The prop folds back flat against the fuselage nose section and all but disappears. Most sailplanes require weight to be added to the nose because of the long tail moment, and ballast is routinely added if flying in more than a moderate breeze. Combined this pretty much negates the added weight of a modern power system. Add in the convenience factor and you will see very few winches or hi-starts in use any more. Modern airfoils and low drag design and construction dominate the current sailplane class. Light weight, high drag “floaters” from the 1970’s have been relegated to the nostalgia class to be flown in dead calm conditions, and even they can accommodate a modern power system without much if any performance penalty. Sailplanes are all about performance, and catching a good thermal and watching your plane climb to the limit of sight on mother nature's energy is a thrill you won’t forget. Sorry if this is off topic, it’s an old passion of mine and I get carried away. 😀

[james_s]

I started flying models in 2014 after being interested in it since I was a small child. Unfortunately for a long time the cost and other commitments kept the barrier of entry too high, but with the advent of cheap foamies, powerful LiPo-brushless power systems that could match glow engines in performance, dirt cheap electronics and effective simulators that all changed. I was able to self teach and crash hundreds of simulated planes before I ever tried flying a real one and I easily transitioned to physical models. I know a lot of guys moan about how the cheap ready built planes and electric power have ruined the hobby, but the way I see it, these developments enabled me and thousands of others to participate in a hobby that was on life support as the grumpy old timers were dying off. 

[Kleinstein]

I target problem you like to get a solution for is a good target, but it may be too complicated for the start.

For the first experience it helps to start with something really simple, that is something like a simple counter,  bliking LEDs, a simple signal generator or thermometer.  It usually takes some practice to get used to a µC and it helps to have simple projects one actually finishes. Depending on the µC used, one may not need much hardware and could use a bread-board for this. I would just taking the code and plan from the internet much practice, though it may help for the very first setup to get the IDE set up and get a starting point.

[Old Printer]

. I know a lot of guys moan about how the cheap ready built planes and electric power have ruined the hobby, but the way I see it, these developments enabled me and thousands of others to participate in a hobby that was on life support as the grumpy old timers were dying off.

Dead on James. I like to build as much as fly, but many don’t have the inclination or time. If it were not for ready built models that hobby would have disappeared many years ago.

[Infraviolet]

To make a project interesting you really do have to have a specific use case lined up for it, coming up with those ideas of which bothers in your life can be solved by electronic projects is quite hard. If you can't come up with a specific use case, come up with an idea or a really cool overall system, then you have lots of little projects making parts for it, each mini project can give you chance to work with different kinds of functionality. Maybe make a robotic arm, you get to have parts on handling motor control, absolute and incremental encoder reading, feedback loops, power supply regulation, maybe make an optoelectronic sensor near the gripper or some sort of proximity sensing between the sections of the arm to prevent self-collision (when you find contrived circustances which your kinematics programming couldn't prevent). If you have separate arduino compatible ICs running separate sections of the arm you get to do comunication protocols between them...

[BradC]

So what I'm really asking is what are some practical applications for things to make.

What itch do you have that needs scratching? I have a 34 year old Volvo with a speedometer that is +/- 10%. I built an auxiliary speedometer that is accurate to 0.01%, that self calibrates using GPS and has a display resolution of 1Kph. It had a display resolution of 0.1Kph but that turned out to be a bad idea.

Prior to that, the first problem was I wanted to use a Parallax Propeller uC and that didn't have a compiler for Linux, so I had to write that before I built the speedometer.

Like everyone else has said, pick a problem and solve it.

[westfw]

I would really like a gadget that I can connect to an arbitrarily sized small (< 20W?) solar panel, then stick it someplace and have it report (in a couple of days) how much power is available with that positioning.  This should support various panel voltages (at least ~5V and ~12V) and do some sort of creative loading of the panel so that it has to actually deliver power.
There are just all these spots where I'd like to stick a solar this or that, that are NOT close to "full sun", but might get enough sun to power the things I'm interested in (with a suitably oversized panel, perhaps.  They're now cheap enough that one doesn't have to arrange for the minimum possible sized panel.)
It'd be a nice, politically correct, project.  And actually useful.

[RJSV]

Yes west.. I feel same way, I have several of the 2.7 W trickle chargers for automotive 12 V.
Those plus a couple of outlets, that are for charging a USB cell phone, whatever'.
Those 'aux' 12 V outlets also have a typical 'cigarette' lighter outlet.

[brucehoult]

I would really like a gadget that I can connect to an arbitrarily sized small (< 20W?) solar panel, then stick it someplace and have it report (in a couple of days) how much power is available with that positioning.  This should support various panel voltages (at least ~5V and ~12V) and do some sort of creative loading of the panel so that it has to actually deliver power.
There are just all these spots where I'd like to stick a solar this or that, that are NOT close to "full sun", but might get enough sun to power the things I'm interested in (with a suitably oversized panel, perhaps.  They're now cheap enough that one doesn't have to arrange for the minimum possible sized panel.)
It'd be a nice, politically correct, project.  And actually useful.

Yeah, that would be a cool thing.

It might want to sweep the loading to find the point for maximum power. Is it (for a given panel) always at the same voltage, regardless of light level? I don't know.

I've never been in a living situation where it has made sense to get solar, but now I'm renting a farmhouse with a large yard and no neighbours maybe it would be interesting.

I'm on a low user plan which means I pay NZ$0.34 per day instead of $2.02 per day, but as a result the electricity is $0.3817 per kWh instead of $0.3231. So I'm paying a very steep US$0.26/kWh and it's would be perhaps well worth trying to generate a little myself.

Particularly nice would be if I can run my 4 kW cooling portable air conditioner off solar on hot summer days :-) Without it I'm using around 10 kWh a day, but with it I can be using up to around 30 to 32 kWh maximum in early-mid February. Though that was sooo hot with nights 22 or 23 C that I was using it 24/7. More normal is 15-20 kWh total in the day if I'm running it from say 10-11 AM to 6-7 PM. So 5-10 kWh for the AC unit itself.

As it's on a 3 pin plug, it would be easy to separate it from the house power.