PID Controllers

The output of a proportional only controller with zero error will be 50%.

For a PID controller with zero error, the output will depend on the startup conditions, but the smallest continuous error ( a very small fraction of a percent) will cause it to integrate to the top or bottom of its output.

Most of the time you are RIGHT.

E.g. 50% of Valve Opening is the bias value to keep a flow as close as possible to Set Point of Controlled Loop. It gives 0% - 50% - 100% range of Process Variable Regulation.

Why questions in this place most of the time are not clear and responding writers try make guesses? Is this a provocation?

What? you ask a unimportant question.

the output depend on your input and control condition. dont know input, dont know output.

but we can say that for zero erro, proportional controller will has infinite large gain.

pls pay attention to your concept

You should also check your answers...

Try to help the people, don't insult them.

His question is legitimate. A few more details would help but it is OK.

Regards,

When the PID is regulating, there is never a zero error condition. Error is what drives the system. It is always seeking the setpoint and noise will cause it to jitter about that setpoint even ever so slightly. You might say that with "zero error" the output would be stable or OFF since the PID would not be seeking but that is probably only a theoretical state or one that exists due to a determined deadband within which the PID output would be kept stable/non-changing.

The short answer, that is also a circular reference, is: It is what it is.

Yep, that's the kind of thing it is when it is that kind of thing.

Omkar 991, I believe that you are asking "If I have a P contolller (no I and no D) what is the output with zero error" - if this is the case then the output will be 0 because P factor x error (in this case 0) results in zero and not 50% as indicate in a previous response.

Exactly. With 'output bias' it could be anything the original poster chooses. A proportiona-only controller will never achieve zero error. To get there requires various amounts of integral action.

Mr or Mrs Guest:

Answer #1 was Perfectly Practical, and short - most people love this way.

For theoretically proper answers, the Question was unclear!

Omkar 991, I believe that you are asking "If I have a P contolller (no I and no D) what is the output with zero error" - if this is the case then the output will be 0 because P factor x error (in this case 0) results in zero and not 50% as indicate in a previous response.

Addition to my earlier response:

However if you are asking what the output of the P controller or the PID controller should be in order to achieve 0 error then see response from rcapper, ignore the responses from DaveR (incorrect) and cnpower (uninformative).

If you google "PID controllers" then you will get a wealth of information on line.

Hope this points you in the right direction.

DaveR answer is mostly correct. It depends on the type of controller you are using. Some of them are offset at 50% (or even adjustable) when the controller output is "zero".

He is also correct to indicate that a "zero" error condition is almost never seen and the integrator will drift to the top or bottom limit unless it has a DC discharge feature(zero DC gain).

Basically, PID controllers are implemented in different ways depending on the manufacturers and level of sophistication. Without specific details, we cannot give a specific answer.

It is not possible to achieve an exactly zero setpoint error.

If for instance you were heating a thin peace of metal and had a large heat source, the heater would over shoot both heating and cooling back to setpoint. The error is much reduced with a smaller heater and a larger load.

Also if the heated objects were of varying weight or thickness there would be some hunting.

Most newer controllers are smart controller and use a logic circuit to help with control and slide the proportional setting up and down to give a tighter control. Some controllers use fussy logic to help achieve better control of the setpoint.

What process are you trying to control?

With a steady load you can achieve an awfully close to zero setpoint error.

Let us know what you want to control and maybe we can give you a definitive answer.

PLEASE KINDLY I NEED YOUR MAIL

Thank you everybody for your prompt help.

I have referred few websites and few good books. But many of them in their representation of a basic P controller [the Laplace model], have shown the control signal to a plant as output of a proportional controller block which is having input as the error signal. It means that--

control signal = (Proportional constant of controller) * (error signal).

So if error becomes zero, there will be no control signal to the plant.

But from your replies I am convinced that the error can never be zero. Some finite value of error has to be there.

One more question--

What is the difference between a compensator and a contoller?

your comprehension is mistake, I think.

which book do you read as well as web site?

proportion control is simple amplifier of gain that is a certain number. its Laplace transfor is still itself ( constant).

error signal has serveral situation, I, II, III etc classes. yhou have to make clearly at first.

compenstor and controller are same things. but some people says that controller include power driver as well. whereas compensator is only a modifier or adjustor for weak signal.

besides, different loop circuit has different error signal.

conroller terms are different in each fields and books..

Some process such keeping the water level in a tank are integrating in nature and may work well with a proportional control. In this case it may be possible to have a very small error if the water flow goes to zero.

Other process need a constant drive to operate continuously such a constant (non zero) speed controller. In this case, the integrator in the PID keeps the drive at the proper level and the P&D gains take care of the changes.

The following definitions are open to discussion but should help.

In general, we call the PID a compensator when the process would work on its own but is made more precise by the addition of the PID controller.

The PID will be called a controller when it does most if not all of the work needed to operate the process.

The difference between the two can be very "fuzzy"...

This Process of Answering is the best Example of Disturbances created by the Question. Everybody tried to be a Controller. Some of them were better (Pre)Set than others.

Most of Water Level Controls, I have seen, were two-positioning type,i.e. without P or I or D elements in the loop.

It -process- also revealed how many words are used on this World to describe the same terms/elements/hardware & software terminology. I guess it is the result of Experts Wanting to eliminate Competition.

Try www.watlow.com they have over 500 spec. sheets.

To compensate - to offset control - a part of the control circuit?

compensator -

com·pen·sate (kmpn-st) v. com·pen·sat·ed, com·pen·sat·ing, com·pen·sates v.tr. 1. To offset; counterbalance. 2. To make satisfactory payment or reparation to; recompense or reimburse: Management compensated us for the time we worked. 3. To stabilize the purchasing power of (a monetary unit) by changing the gold content in order to counterbalance price variations. v.intr. To serve as or provide a substitute or counterbalance. [Latin compnsre, compnst- : com-, com- + pnsre, to weigh; see (s)pen- in Indo-European roots.] compen·sative (kmpn-stv, km-pns-tv) adj. compen·sator n. com·pensa·tory (km-pns-tôr, -tr) adj.

Compensator usually is meant to be an element in the controlling loop to compensate lag effect by lead or vice versa to make stability in control loop. E.g. (1 + sT) vs 1/(1+sT) in Laplace's notation. The controller is a block of elements that could have internal elements with ability to set P (amplification factor), I (integral/rate) time and D (derivative) time constant. It is possible that in simple systems one or more compensators could solve control needs. Mostly in electronics or pneumatics.