Hi, I'm brand new to Arduino but have a fairly complicated project. I am trying to use an arduino Mega2056 to control 4 linear actuators (non feedback). We are attaching rotary encoders or potentiometers on a rack and pinion that is attached to each actuator. As each actuator extends, its respective rotary encoder or potentiometer will turn, allowing us to track the length of each linear actuator. The linear actuators are arranged in pairs parallel to one another and are intended to move in tandem with its parallel member. For our device to work properly, we need to be constantly adjusting the actuator length as a function of the IMU pitch value.
Is there any advantage to using potentiometers that would mean an analog input, as opposed to the rotary encoders for digital input as far as simplicity in simultaneously monitoring all four inputs at the same time. From my research, it seems that I would only be able to monitor one encoder while it was moving to count the state changes. If this is correct then I would be missing the other four encoder movements and tracking each arm would not work. With analog potentiometers, my understanding is that I can basically read the instantaneous value once per loop rather than having to read for the entire duration of the movement. Is this correct?
Any advice on which type to use, and how that would affect the implementation in the code?
Pots seem easier to me. A rotary encoder gives you a relative count so you have to know the starting
position. But, a pot only turns about 270 degrees whereas a rotary can make multiple rotations and count indefinitely, so it may depend on how you're converting linear position to rotation. (Servo motors use pots.)
I'm not sure if an analog read is faster than reading the encoder because I don't know the details of reading an encoder. Hopefully somebody else knows... An analog read is only one C++ instruction, but the multiplexed ADC doesn't "settle" instantly and you can read digital a digital pin faster than you can (accurately) read an analog pin. (And the number of C++ instructions doesn't tell you how many machine-instruction cycles the compiled code will take.)
But unless your Arduino is doing lots of other things, speed probably isn't a problem when you're comparing mechanical motion to processor speed.
Pots take longer to read, have a limited range, may be non-linear, may have different overall resistance, and wear out over time.
Which actuators do you use?
4 of these linear actuators :
A Feedback Rod Linear Actuator (same page) would give you already the required feedback.
Unfortunately, I have to work with what I have...
You mention IMU. Is this an Aerospace application? Is robustness and reliability a key criterion?
The problem with pots is that they wear out, their wipers bounce in high vibration apps, and they have a limited range of rotation.
If you can be sure your pinion doesn't rotate more than 270 deg then pots are ok, otherwise you will need to gear them which may be more trouble than it's worth.
Incremental rotary encoders require indexing which is usually not practical on a servo. They are good for velocity measurement though.
Absolute rotary encoders are the gold standard but they are frightfully expensive for high resolution, say more than 10 bits.
If you have a machine shop you can easily make up four sensors using rotary hall sensors from AMS. They are easy to use, exceptionally robust, non-contact (never wear out), have 12 bit resolution, have a digital SPI output so all four can be on an SPI bus, they are very fast to read, and they can rotate continuously - you just need to keep track of how many times they have rotated.
Check out the AS5145 from AMS. They have other Hall sensors too for precision measurement applications. The chips are cheap and you can get a few samples for free from the manufacturer.
You will need four small quadrature magnets, but these are not difficult to source.
These devices are all the rage at present in gymbals for high-end radio control transmitters, and are also showing up in r/c servos too.