PID heating control with SLC500

[mac_g3 0]

NT4.RSSHello All. I'm French ! ... and my English is not better !! My process is : -SLC500 controller, with one 1746-NT4 thermocouple/mv card, one 1746-IB16 16 input 24VDC, one 1746-OB16 16 output static 24VDC. The thermocouple sensor is connected to the channel 0 of the 1746-NT4, and a solid state relay is connected to the 0 output of the 1746-OB16. The 220VAC power circuit of the solid sate relay is connected to a small steel heating element, the thermocouple measure the temperature of the heating resistor. When i try to run the process .. the temperature control appears to run good ... and the accuracy ... is not better ( underflow or overflow the setting point that i want ). When i observe the process .. i can modify the setting of the PID control ... and .. after long time ... the regulation is .... approximative good . I search a solution to increase the accuracy of the regulation ( like the "autotune" function on certain temperature controller in small box ). The PLC programm ... is a note from the AB programming manual ....., i send in next file my PLC programm, built with RSLogix 500. Anyone have idea how to make a best PLC programm to obtain a better regulation ? Or anyone have a best PLC program ... and send to me ? Any helps is greattly appreciated. Thanks and greetings from France, Stéphane.

[vettedrivr 0]

Stephane, You are attempting to control and analog system with a digital output, this is never going to give you good regulation. When the output comes on and turns the heating coil comes on at full power, the temperature in your process increases until you reach the setpoint. When the setpoint is reached, the coil turns off and you get over shoot as the coil does not cool off immediately. Then as the process cools and the controller deadband comes into play, you will have to reach a certain temperature under the setpoint before the heating coil comes back on. The process continues to lose heat until the coil can catch back up. If you can replace the solid state relay with and variable voltage transformer that increases or decreases voltage with a small motor and drive the motor with and analog output card you will get much better regulation. The problem... motor driven variable voltage transformers are very expensive.

[Peter Nachtwey 15]

Yes, and unnecessary if update SSR update time is fast enough. Mac_g3. Got an Apple? You should visit www.controlguru.com. It will tell you how to calculate the PID gains after you know more about your system. download hotrod.zip and the sliding mode control program from http://www.plcs.net/downloads/index.php?di...;directory=Misc Hotrod.zip is an example program for SLC50x where it simulates a SOPDT temperature system using both analog or timed proportional outputs

[Snerkel 0]

As you are using an OB16 you can make the output proportional using a cyclic timer. To make a cyclic timer use a standard TON and put the done bit in series with the coil. ---¦/¦------------------(T4:0) T4:0/DN Use the T4:0.ACC to compare against the PID output, so using a 0.01 timebase one second equals 100. Scale the PID output to 0-100% Then use a compare T4:0.ACC < PID value -----------------------(Output) ¦ PID value = 100 ----------¦ Edited 19 Jul 2006 by Snerkel

[AutomationWizard 0]

It sounds like you are using a time proportional PID configuration to create discrete pulses of duty cycle determined by the PID instruction. This is a very good technique. You can probably improve your control with better tuneing of the PID. For more information on PID tuneing from someone who is an expert, see the web site at http://www.controlguru.com/ . You can also change the cycle time of your pulses. Shorter cycle time will likely lead to better control as long as it doesn't get faster than the switching time of your output circuit. There is PID tuneing software called RSTune available from Rockwell, but it isn't a guaranteed fix. You might also want to look into something called pulse width modulation.

[TimWilborne 108]

And we follow the link and we find Peter Nachtwey. Yep, he's the stinky stuff. We already know that

[Sleepy Wombat 27]

Hey, there was my sliding mode excel spread sheet i uploaded there ages ago also...extrapolated from the equations provided by Peter...

[TimWilborne 108]

I did a seach for "Wombat Code" but it didn't return any results

[Sleepy Wombat 27]

The Wombat has left a deposit in the download section Sliding Mode Control Excel Spreadsheet

[TimWilborne 108]

Thanks for the Wombat Code Sleepy, I follow you now