I have an idea for a project that would involve both digital and analog inputs resulting in a PWM output. What board or boards should I get?
More detail:
I would be setting up a transmission in a Dodge pickup to go into torque converter lockup when a certain speed would be reached, also coming out of lockup for shifts, then going back in when the gear is secured. I’d have one hall effect sensor input for speed and one analog input for line pressure (to know the shift points), these would determine when the torque converter clutch solenoid would activate and deactivate. The solenoid is a 12v circuit, I believe the sensors are 5v.
any arduino board would meet the spec. Draw a circuit schematic so that all the info is there to make a choice. The arduinos all have analog pins, digital pins and PWM. you need to think about power supply. protection in an electrically noisy environment and robustness. Plus any other requirements you might have for your specific project
I figured twisted pair wiring would be necessary for the speed sensor. However, is the board itself sensitive to noise also? It would be mounted in the cab, so I’m not too concerned about robustness.
This would be my first go at more complex controls, so I know very little about the particulars.
OF course it is. All depends on what the "noise" is and it's voltage. Noise is also unwanted current pulses on your DC power wiring. You can limit the susceptibility by shielding and bypass and feedthru capacitors. Noise can be conducted or induced by capacitance between adjacent conductors or by RF radiation.
You will not know until you have something developed to test with.
Get a few nano or esp32 boards (if you want wifi). Get sensors together. Get a breadboard. If you can blah an oscilloscope it is useful but a multimeter at a pinch.
Before all that draw a schematic
Get one component on the breadboard with the uC (microcontroller) and get it working with the most basic code. Only ever do 1 small bit at a time. Always name sketches, variables etc well and descriptively. Open new sketches any time you make a significant change so that you have an iterative approach. Once a thing is working then work to encapsulate the code (look this up) it will help when you come to combine it.
Bugs will always exist between your last saved working iteration and your current active code. If you do it altogether bugs can hide anywhere.
If you have a scope you can test your signal wires for noise and use capacitors etc to clean it up. Otherwise you will just have trial and error and have to research best practice.
Breadboard circuits only have to represent real world so LEDs can be used in place of relays etc to represent your logic. Then you move to perfboard and soldered connections for further bench top testing and finally in the field testing. If everything works then you develop a pcb using an online manufacturer and then test it
A long time ago a single board computer based on a Z80 was built. A 7805 was used to power the CPU, EPROM, RAM and a CTC. It worked for 10 years without any hick up. There's hope! Decoupling caps near every device Vcc does well.