PID contol, I value maximum?

[kurabiye]

Hello, I am trying to tune my PID system for different temperature values. I am using 2 power supplies and a temperature monitor for temperature contol, however in Lakeshore 330 manual I have seen a practical method as follows : Now my question is what should be my maximum I value to be able to proceed right when I am using Labviews PID function and not using Lakeshore 330?

3.3.5.1 Setting Gain (Proportional)
Enter a gain (Proportional) value from 000 to 999. To enter a gain value, press the P key. The lower Control
window display shows the current P setting (the default is 50) with the units place blinking. Use the numeric
keypad to enter a new setting. Press Enter to accept the new Proportional setting or Escape to return the
normal display and retain the old setting.
To experimentally determine proper gain, use the following procedure. Set Autotune to Manual, then turn off
both Reset (I) and Rate (D). Enter a nominal gain of 50. Verify that the heater turns on; if not, increase the
gain setting until the heater turns on, then let the system stabilize. It stabilizes at some point below the setpoint
(typically 2 to 3 K below). Increase the gain by factors of two until the system temperature begins to oscillate.
Adjust the gain for small sustained oscillations. Measure the period of these oscillations to determine the
correct setting for reset. Reduce the gain by a factor of two to three until the temperature again stabilizes. Be
sure to allow time at each setting for the system to stabilize (if it will). Some systems and cryogenic sensors
with low sensitivity require the maximum gain.
3.3.5.2 Setting Reset (Integral)
Adjusts the reset time constant of reset (Integral) in the control function. Reset time in seconds = (999/Value
Entered). For example, if the Reset setting is 20 the reset time in seconds is about 50 (999/20=50). Enter a
value from 1 to 999. A reset of zero makes the controller proportional only. To enter a reset value, press the I
key. The lower Control window display shows the current I setting (the default is 20) with the units place
blinking. Use the numeric keypad to enter a new setting. Press Enter to accept the new setting or Escape to
return the normal display and retain the old setting.

[becidea]

PID değerlerini sistemin kendisi ile belirlersiniz. Yazdıklarınızı -manuelde olanları- takip ederek size en uygun PID değerlerini belirlediğiniz sıcaklık değerinde elde etmeniz en güvenli ve sağlıklı yoldur. Sonra bu değerleri LS-330 un labview driver ı ile yapabilirsiniz. Labview' de PID için çok örnek var lakin yorulmadan LS 330 un kendi driver ı gerçekten güzel. Bende LS 336 var ve driverlerı yetiyor. More than enough güzel bir tanım olurdu

Selamlar.

MG

[kurabiye]

*** I am not using 330, just posted that as an example. Now my question is how can I change I maximum (999) according to my system?

MG Cevap icin tesekkurler ama benim sorum biraz daha farkli. Ben 330 sistemini degil, sadece power supply ile temperature okuyabilecegim bir multimeter kullaniyorum. 330 un manualini ornek olmasi acisindan post ettim. Cozemeye calistigim sey, 330 da maximum I olarak aldigi 999 degerini ben kendi sistemimde hangi numeric ile degistirecegim?

Thanks,

Tesekkurler

[becidea]

Hi,

For controlling PID as you wanted, you will need PID toolkint in order to control temperature of heater via power supply.

Check this: http://forums.ni.com/t5/LabVIEW/PID-temperature-control-through-voltage-Amperes-and-thermocouple/m-p...

Best

MG

[kurabiye]

Hi, 

  Yes I do have a PID control VI which works fine, It controls my heaters and temperature sensor. It has P, I and D as inputs so when I am tuning them, I manually calculate them and give them as inputs (there is also autotune option but It does not work perfect in some cases thats why I am not following that path). So for tuning, I need to calculate I from the gain (P) and period of the oscillations which I already have. The only unknown here is the maximum I value to be able to calculate I appropriately. What should be the maximum value of I ? 

Thanks

[nathand]

What method of tuning are you using?  Why do you need a maximum I value?  Note that if you're using the LabVIEW PID, the I value is time and a larger value means a longer (slower) integral response.  You can use as large a value as you want but as that time grows it will have less of an effect.

[kurabiye]

The reason for asking the maximum value is calculating the appropriate I as it is done in Lakeshore 330 Manual (the first post that I have sent has the detailed description for that ==> I=max reset value/period). 

There is also another method as attached on the pdf, which tries different PID Methods and tunes that way but that one seemed longer to me than the method used in Lakeshore 330 manual.

[nathand]

---

@kurabiye wrote:

The reason for asking the maximum value is calculating the appropriate I as it is done in Lakeshore 330 Manual (the first post that I have sent has the detailed description for that ==> I=max reset value/period).

---

You've misunderstood.  The instructions for the Lakeshore says that the reset time in seconds is 999/Value Entered.  So, you calculate the reset time based on the period of oscillation and then compute the appropriate value to enter.  For the LabVIEW PID function, the I value is simply the reset time (in MINUTES, not seconds); there is no need for a "maximum" value and division.  I'd guess that the Lakeshore is set up that way because it can only handle integer math, unlike LabVIEW which can of course do full floating point calculations.

[kurabiye]

-Nathand, thanks for the helpful information. so I will use period of the oscillations in min for the I value, right? I am attaching a jpeg file for the PID settings of 6,0,0 respectively so if we assume that period is 40sec than I would be 40/60=0.66min ?

-Is there any limitation for P and D ? 

[nathand]

The integral, or reset, time is not usually equal to the period of oscillation.  All of the tuning methods I've seen divide the oscillation period by some factor to calculate a reasonable integral time value.  Otherwise you have the right idea, although I'm a little surprised that in pure proportional control your system overshoots so much and then settles into stable oscillations - for some reason that looks odd to me.

What do you mean by "limitation"?  You can enter any value you want, but if you want good performance you of course need to use values that are reasonable for your system.