PID - the D

[macmet]

Hello Everyone,

I have a general question regarding PIDs that I hope is easy to answer. I've done a search of this forum and I'm not sure I've seen the answer, but if I have, I didn't understand it or missed it.

In PID controllers, the D is the derivative term. I saw somewhere it described as the rate of the change made to the output value based on the error. (i.e. the cruise control on a car going up a slight hill vs. a steep hill).

My question is, if D is a rate, why is the units used for Kd given in time? AB, Bailey, Honeywell, Siemens and Yokogawa all use a unit of time for the Kd value according to my search. How is that time applied?

I apologize if I overlooked this question elsewhere on here, but I searched for it without luck.

 

[IRstuff]

see:

http://en.wikipedia.org/wiki/File:PID_en.svg

TTFN
faq731-376

 

[macmet]

Thanks IRStuff.

I've seen that article. I have it open on my browser actually. I'm not quite how that applies a constant time value.

What is the time based on? I'm not a controls guy, but I would like to understand the practical applications of those values. The PI terms seem easier to understand.

I work at a Boiler/HX company, how is the D term applied to this type of product line?

 

[BrianE22]

Say you're controlling temperature in degrees. Derivative would give you degrees/sec. So Kd would have to be in units of seconds so the product, Kd x degrees/sec. gives you degrees.

 

[IRstuff]

Really?

K_D*de(t)/dt The derivative of the output error as a function of time results in "err/time"

TTFN
faq731-376

 

[macmet]

I was thinking the reason was as Brian noted. I was hoping to receive confirmation. Does that mean PIDs are always applied to functions with time?

IRStuff, I'm not sure if you're commenting on my lack of knowledge on this subject. If you are, it is new to me. I am trying to learn more about it.

 

[DRWeig]

Get this book, it's very inexpensive and the author has extreme credentials. No, it's not me.

http://www.straightlinecontrol.com/

Here is an illustration and description that speaks to your question directly. It's in the bottom 1/3 of the article.

http://www.jashaw.com/pid/description.html

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[xnuke]

IRStuff,

Most control engineers that are setting up PIDs will use the derivative term on the process variable instead of the error, so BrianE22's units are correct for that type of application. This is especially the case if the PID has to track changing setpoints, as often occurs with temperature ramp-and-soak requirements.

Using the derivative term on the process variable instead of the error prevents large transients in controller output that the derivative term may cause due to rapid changes in the setpoint (and therefore, error). These large transients can be avoided if the PID has a derivative term filter that is set up properly, but since many control textbooks don't even talk about the filter, many engineers may not know how to properly apply it and leave it turned off.

xnuke
"Live and act within the limit of your knowledge and keep expanding it to the limit of your life." Ayn Rand, Atlas Shrugged.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[IRstuff]

I don't think I contradicted Brian22's units anywhere.

M, My response was to your comment regarding "constant time value." Perhaps you can expand on what you meant by that.

TTFN
faq731-376