Actually made it easier to understand how the PID works on my aluminum smelter I'm building thanks for the explanation!

your language is very simple to conversation so I really understand about this lesson, thanks so much

Great video. Looking forward to the follow up video.

I liked this a lot. Funny and educational at the same time.

@AccidentalScience

Жыл бұрын

Thank you, I'll make more. What would you like to be the next topic?

You've done a good job!

@AccidentalScience

Жыл бұрын

Thank you sir.

Thanks for the nice video, but I think the details may be a bit off: the circuit you show with the resistor and capacitor in the feedback in parallel make a "lag compensator"/"integrating with limited low frequency gain"/"proportional with high-frequency roll-off" response rather than a "PI" response. Putting a resistor in series with the capacitor allows achieving a PI response which is proportional for high frequencies and integrating for low frequencies. Google for "op-amp PI regulator" to find some example schematics. Limiting low frequency gain is something that can still be useful for PI to improve overload recovery characteristics.

@AccidentalScience

Жыл бұрын

Thank you for commenting and providing feedbak, also sorry for the late reply but every once in a while a comment happens to fall down in my list of notifications, and I noticed your comment right today. I am not sure at what resistor you are referring to, but I suppose it is the one that is in parallel to the capacitor that I indicated as "derivative capacitor" (which is an improper terminology but I did that to make things simpler) shown at 7:37 mark. The following assumes that, but please let me know if I mistaken the point of your comment. If you put there a resistor in series with the capacitor *and remove* the parallel resistor you end up to completely lose the feedback proportionality! You can't have a practical circuit that has a response as you mentioned, with no control over the proportionality (at full frequency spectrum), but again if I mistaken something, let me know. The formula is this: u(t) = K(e(t) + 1/Ti int(0->t, e(t) δt) + Td δe(t)/δt) Where K stands for the proportional gain, e the error, Ti and Td are the integral and derivative terms, respectively, and t of course is time. As you can see from the formula the derivative can be seen as kinda prediction of the future status of the controlled actuator, that you can practically achieve anticipating the feedback signal (not the feedback loop of the gain OPAMP) through a capacitance in parallel to the resistor that provides the feedback signal to the summation node. Adjusting the ratio between these two *paths* allow to determine the best derivative point. Of course it is not the only way to achieve this, but it is just a simple example, hopefully simple to understand. I found a good book in Control System Design by Karl Johan Åström (I had to find Sweedish character to write this! :) ) that I really recommend. Greetings from the Alps.

@FJL4215

Жыл бұрын

@@AccidentalScience Thanks! I agree that formula is the expected behavior. I was thinking of the "proportional resistor" and "integrating capacitor" which are in parallel but I would expect to be in series. For example, take a small static error. This should give a control signal which ramps up towards infinity as the time goes to infinity in the formula. In the circuit, the proportional resistor limits the final value. Or in other words, an ideal PI regulator should have infinite DC gain but thr circuit has a limit set by the "proportional" resistor. This limit can be useful, but real proportional action is also very useful to have.

@AccidentalScience

Жыл бұрын

Ah gotcha. I misunderstood the capacitor you intended. So you meant the resistor, and capacitor in the feedback loop of the error amplifier (the OPAMP that has the summation node). Yes, the parallel resistor limits the gain indeed. At 10:10 mark you can see a more realistic representation of the circuit. You'll notice that the parallel resistor has been replaced with a potentiometer (Kp) that adjusts the gain over the full spectrum. Usually this is adjusted as large as possible, though in certain applications it is required to reduce a bit the gain in a trade-off with precision and stability (or sometime just to avoid noise). Also it is useful for the calibration stage, where at first the Kp is reduced to keep gain low and making other adjustments with Ki and Kd. The capacitor is connected in series with another potentiometer (Ki) that allows to adjust the integral term. Finally a further potentiometer (Kd) allows to adjust the derivative term. In real applications the effect of this last C+R series is approx. double the effect of the fixed resistor that provides the ratio with the resistor that carries the set point signal, so the ratio between the input set point and the actuator's feedback is determined (i.e. determining the speed of a motor in relation to the input set point voltage). Also consider that where you see a resistor you shouldn't assume that it necessarily reduces the gain, it could be large enough that the gain still is large, and again could be functional for the calibration stage. Note that even with a digital PID you still need to have control over the gain, that is the proportional term, maybe just for the autocalibration stage where with a low gain a step response provides the information required to adjust the other terms. I am not super expert in digital PID so take this just as a rough indication.

Love your videos! Gotta leave a comment for you at least!

@AccidentalScience

Жыл бұрын

Thank you Brian.

India, Dongargarh (For algorithm 😁)

@AccidentalScience

Жыл бұрын

Nice town.