What does PID stand for?

[luckyflyer]

In the Scorpion software I'm curious what doe PID stand for?

[luckyflyer]

Answered my own question, Wikipedia https://en.m.wikipedia.org/wiki/PID_controller

[jhamel]

Pelvic Inflammatory Disease.





In this context, it stands for Proportional, Integral, and Derivative.

Folks facile with physics and calculus can write books and dissertations on this subject. I'll try to explain it in basic terms more easily understood by the rest of us at the risk of offending the more educated folks. At is applies to us, P, I, and D are parameters of stabilization. We generally adjust them using the P-gain, I-gain, and D-gain. P, I, and D ultimately affect the electronic stabilization feedback loops used in our FBL controllers in all three axis (rudder, elevator, and aileron). We generally get to adjust the P-gain and I-gain, but D-gain may be either adjustable or fixed depending on the FBL unit. In general, the P-gain is the main gain. This P-gain determines the speed with which the helicopter will correct for outside disturbance like wind. It works in the present and it has a very short term memory. This means that as the wind acts upon the helicopter, the P-gain of the FBL unit does what it can to bring the helicopter back to the position right before the wind acted on it, but because the P-gain has short term memory, is does not remember where exactly the pre-wind position was. When the P-gain is set too low, the speed of correction [for wind, etc.] is slow, so there is quite bit of drift. If set too high, you will notice a low frequency (slow compared to I-gain) hunting around the original position before the outside force (wind, etc.) acted on the helicopter. For flybar heads, the flybar mechanically controls the P-gain and you adjust the "P-gain" stabilization by adjusting the length of the flybar rod, the size and weight of the paddles, and the COG of the flybar by adding additional weights along the flybar rod. For pre-Heading Hold tail gyros, the P-gain was the main adjustable determinant for "rate-mode" rudder stabilization. Compared to the more accurate and meticulous I-gain, the P-gain works in the present and "lives for the moment". The P-gain does not learn from the past (I-gain), and does not care about the future (D-gain).

The addition of the I-gain parameter to the Tail gyros and later full FBL controllers was the real game changer. This is the so called Heading Hold parameter. Integral has longer term memory than Proportional, so it remembers better what the original position before the outside force (wind, etc.) acted on the helicopter. Integral is more "accurate" than Proportional, but it is slower than Proportional. So you still need an optimal P-gain for speed or correction. The I-gain compensates/corrects the drift (called error) left behind by the P-gain. When the I-gain is set too low, you will notice a drift (head and/or tail). When the I-gain is set too high, you will notice high frequency (higher frequency than high P-gain) oscillations (head and/or tail). The I-gain remembers and learns from the past, but it applies this past knowledge to the present slower than P-gain. So I-gain is "book smart" and "meticulous/anal-retentive" whereas P-gain is more about "action", but it does not remember very well its destination. The P-gain gets it in the "ball park" fast and the I-gain "locks it in".

The D-gain "dampens" the P and I gain. Not all the FBL units let you adjust this parameter. Think of a gust of wind suddenly stop or you letting go the transmitter sticks back to neutral. The D-gain determines how fast and how well the helicopter will come back to pre-disturbance condition. If set too low, the reaction will be slow. For instance, you are doing fast pirouettes, and you let go of the rudder stick. If the D-gain is too low, the tail will continue to pirouette for another 1/16 to 1/8 of a turn before it stops. If the D-gain is set too high, when you let go of the rudder stick, the tail will aggressively oscillate as it stops. Same happens with elevator and aileron.

Again, my explanation is not meant to be a comprehensive lecture on calculus. It is not my intention to minimize the depth of such a complex subject or insult anyone's intelligence. I mean it to be just a simple introduction to the PID concept and how it applies to RC helicopter modelers.

[luckyflyer]

Very good explanation, are you possibly a teacher?

[jhamel]

Nah. The closest I come to teaching is giving short presentations in morning reports.

[Airagitator]

Good answer.

Please check / clarify two things:

P-oscillation is faster than I-oscaillation I think. I is more like swinging. Alos, with to low I it feels not so "locked in".

Lower D-gain results in sharp behaviour, fast stops, tail-kick-back and higher values make stops sluggish and lead to aggresive oscillation in conjunction with high P.

Thats what I have noticed with brain and neuron.

[toadiscoil]

This is probably the best "for dummies" (like me) explanation of PID I have ever seen.

My question is about the main head. In the tail it is fairly obvious. The Heli puros and you check how locked in it stops. Start with P gain get it tonwag and get it down. Then play with I gain until stops are precise. But what about the head? Will that be up and down?

Also when doing a pitch pump and the tail blowing out is that a tail or a head gain adjustment?

Sorry for the dumb questions this season I want to learn more about tuning. I know I could play with it and find out but I don't have much time so I want to maximize the time I spend tuning by knowing the theory first.

Originally Posted by curmudgeon

Pelvic Inflammatory Disease.





In this context, it stands for Proportional, Integral, and Derivative.

Folks facile with physics and calculus can write books and dissertations on this subject. I'll try to explain it in basic terms more easily understood by the rest of us at the risk of offending the more educated folks. At is applies to us, P, I, and D are parameters of stabilization. We generally adjust them using the P-gain, I-gain, and D-gain. P, I, and D ultimately affect the electronic stabilization feedback loops used in our FBL controllers in all three axis (rudder, elevator, and aileron). We generally get to adjust the P-gain and I-gain, but D-gain may be either adjustable or fixed depending on the FBL unit. In general, the P-gain is the main gain. This P-gain determines the speed with which the helicopter will correct for outside disturbance like wind. It works in the present and it has a very short term memory. This means that as the wind acts upon the helicopter, the P-gain of the FBL unit does what it can to bring the helicopter back to the position right before the wind acted on it, but because the P-gain has short term memory, is does not remember where exactly the pre-wind position was. When the P-gain is set too low, the speed of correction [for wind, etc.] is slow, so there is quite bit of drift. If set too high, you will notice a low frequency (slow compared to I-gain) hunting around the original position before the outside force (wind, etc.) acted on the helicopter. For flybar heads, the flybar mechanically controls the P-gain and you adjust the "P-gain" stabilization by adjusting the length of the flybar rod, the size and weight of the paddles, and the COG of the flybar by adding additional weights along the flybar rod. For pre-Heading Hold tail gyros, the P-gain was the main adjustable determinant for "rate-mode" rudder stabilization. Compared to the more accurate and meticulous I-gain, the P-gain works in the present and "lives for the moment". The P-gain does not learn from the past (I-gain), and does not care about the future (D-gain).

The addition of the I-gain parameter to the Tail gyros and later full FBL controllers was the real game changer. This is the so called Heading Hold parameter. Integral has longer term memory than Proportional, so it remembers better what the original position before the outside force (wind, etc.) acted on the helicopter. Integral is more "accurate" than Proportional, but it is slower than Proportional. So you still need an optimal P-gain for speed or correction. The I-gain compensates/corrects the drift (called error) left behind by the P-gain. When the I-gain is set too low, you will notice a drift (head and/or tail). When the I-gain is set too high, you will notice high frequency (higher frequency than high P-gain) oscillations (head and/or tail). The I-gain remembers and learns from the past, but it applies this past knowledge to the present slower than P-gain. So I-gain is "book smart" and "meticulous/anal-retentive" whereas P-gain is more about "action", but it does not remember very well its destination. The P-gain gets it in the "ball park" fast and the I-gain "locks it in".

The D-gain "dampens" the P and I gain. Not all the FBL units let you adjust this parameter. Think of a gust of wind suddenly stop or you letting go the transmitter sticks back to neutral. The D-gain determines how fast and how well the helicopter will come back to pre-disturbance condition. If set too low, the reaction will be slow. For instance, you are doing fast pirouettes, and you let go of the rudder stick. If the D-gain is too low, the tail will continue to pirouette for another 1/16 to 1/8 of a turn before it stops. If the D-gain is set too high, when you let go of the rudder stick, the tail will aggressively oscillate as it stops. Same happens with elevator and aileron.

Again, my explanation is not meant to be a comprehensive lecture on calculus. It is not my intention to minimize the depth of such a complex subject or insult anyone's intelligence. I mean it to be just a simple introduction to the PID concept and how it applies to RC helicopter modelers.

[Relisys190]

proportional–integral–derivative

[Relisys190]

P accounts for present values of the error. For example, if the error is large and positive, the control output will also be large and positive.


I accounts for past values of the error. For example, if the current output is not sufficiently strong, the integral of the error will accumulate over time, and the controller will respond by applying a stronger action.

D accounts for possible future trends of the error, based on its current rate of change.[2]. For example, continuing the P example above, when the large positive control output succeeds in bringing the error closer to zero, it also puts the process on a path to large negative error in the near future; in this case, the derivative turns negative and the D module reduces the strength of the action to prevent this overshot.