Every time I see Peter Nachtwey pop up on plctalk I just feel like to total IDIOT. I’ve passed my university calculus classes but don’t think I really learned enough to apply it; probably a me-problem.
How much “control theory” do most of you guys use vs. how much is strictly electrical/programming knowledge?
Control theory is more than calc 1/2/3, it is also, and primarily, differential equations.
I can easily state that 80% of my knowledge and learned skills are NOT programming.
To design a machine controls I have had to learn chemical engineering, fluid dynamics, hydraulics, and stuff in many other fields.
If I see a free seminar in the area (especially if it says Free Lunch) and I have the time I sign up because I might learn something I can use in the future.
That is actually how Peter got where he is. He was working on the control for a hydraulic application that no one could answer his questions, so he learned more about hydraulics than any other person has known. And got good at it.
Yeah, that is what I’ve been figuring the longer I hang around controls people. I am getting pretty good at jiggling bits in the PLC, but I am not satisfied with being merely passable for my whole career. I don’t aspire to be the greatest controls engineer in history, but I do aspire to be good.
I am halfway retired and still dont understand half or what Peter or drbitboy talk about when they go off the deep end, they both have helped me out a lot when needed, I also dont think you need to know everything, the key is being smart enough to know when you need help and have the ability to either figure it out or get help.
I would say my math is my weakest link but my friends are my strongest link… so as they say, “use what the good Lord gave yah”
My math isn’t the best either, I never took calc in school but have done some extra learning on the side in what little downtime I get. I love the posts where Peter, Max, and the DMD get into their debates. That is where I have learned the most to be honest. I don’t understand what they are talking about, so I sit down and do the math that they post to get a better understanding.
For me, like GIT, the issue is knowing where to find honest and true answers at. I detest all of the AI slop that I have been seeing lately. For me, those are the worst posts. “ChatGPT told me this should work, but it doesn’t. Please tell me how to fix it.”
You think this thread has the potential to turn into one of those Peter/drbitboy/MaxK threads? Where Brian tries to answer the question in the thread, Max just wants to trigger Peter and Peter wants to remind Max of his credentials.
What say you @drbitboy?
https://www.plctalk.net/forums/threads/the-error-squared-pi-controller-implementation.148330/
Already have my popcorn on standby!
Haha, I knew that if Max posted, Peter wouldn’t be far along.
I find the PID posts interesting, because Peter likes (i) to point out that, given a system model it is possible to implement a reasonable a priori tuning of the system+PID, and (ii) to lament that the designers never provide a system model. That means the hapless OP with the system to tune has to do system identification, and once that is done (via Cohen-Coon or similar, or even my favorite, though much-maligned, Ziegler-Nichols) we have a system model at which point item (i) above is possible. Unless there is a lot of 
time. And then anyone can tweak from there.
Most systems, if not integrating and nor deadtime-dominated, are first- or second-order plus deadtime (FOPDT or SOPDT), in which case the behavior can be modeled with a linear first- or second-order differential equation, the solution for which is an exponential with a negative time constant and possibly i in the exponent; and the Laplace Transform transforms the equation to a quadratic algebra problem. The i brings in sine and/or cosine, which explains the oscillation.
Anyway, what I am getting at is that. although it is nice, it is not necessary to fully understand the calculus in depth, because the problem was solved for most systems over half a century ago and usually boils down to a, if not easy at least straightforward, problem.
I think it’s more important to understand the three terms of the discrete PID and the system characteristic, so when there is a problem, one can, after getting some experience under the belt, build a model of what is happening and make a decent first guess which term(s) to tweak and in what direction.
I really like this answer, and I think that is ultimately why I like those posts so much. I mean, besides the bickering and one upmanship that ultimately follows lol. I am currently working with one of our electrician apprentices and teaching him about controls. He wanted to get straight into PLCs and HMIs, big surprise right? Well, I started him from the very beginning, and even quoted one of your signature lines: “i) Take care of the bits, and the bytes will take care of themselves.” (I did credit you by showing him one of your posts, lol)
We have covered bits, bytes, logic gates, Boolean Algebra, and math in different bases. Now, about two months into our time together, we have started looking at what ladder logic is, and he asked me “How can you just look at the code and figure out what the machine is doing?” Another drbitboy quote, “The PLC cares not a whit what you want it to do, but will unapologetically and inexorably do what you tell it to.” I simply look at what the machine is being told to do. It also helps that I have a mental model of exactly what the machine is supposed to do, so I can compare what it is being told to do with what it is supposed to do.
That is all everything boils down to, how well can you model something in the controller, and what paths do you need to take to get it there.
Here is a PID one for the books, still going to work on this again someday
https://www.plctalk.net/forums/threads/pid-trainer-issue-on-setup.135535/
I solved that one, the problem was the non linearity of the PV sensor (probably noise also, I could not get one of those IR range sensors to work at all; I was able to point a cellphone at the ball & beam and get good results though)
Hey, I still got game in calculus: Reddit - The heart of the internet
My Calculus is non-existent, but somehow I managed. 
My senior year of high school, I was going to take trig and calculus, but my folks moved me to a small town that didn’t offer calculus, but required me to take ag and re-take art to graduate. After graduation, I had a choice between going to college and eating, so my college was limited to a few courses after I had a good factory job, but I didn’t get far, since I was raising kids and working a rotating shift.
I have tried to follow some of the advanced PID threads, but I glaze over after a few paragraphs.