How to Prevent Sestos PID Controller from Overshooting Temperature?

AI Thread Summary
The discussion focuses on controlling the overshoot of a Sestos PID temperature controller, specifically the D1S-VR-220 model, when trying to maintain a temperature of 37.3 degrees Celsius. Users highlight the initial overshoot to 38.5-38.8 degrees before stabilizing at the desired temperature and inquire about possible adjustments to prevent this. Key factors influencing temperature control include thermal time lag, rates of temperature rise and drop, and the set-point tolerance. Suggestions for tuning the controller include reducing the hysteresis time and adjusting PID parameters, although these changes may slow response times and affect disturbance correction. Ultimately, some level of overshoot may be unavoidable due to the complexities of thermal systems.
Aviv rotman
Messages
5
Reaction score
1
TL;DR Summary
Help setting SESTOS controller parameters
Hello, I use Sestos PID temperature controller D1S-VR-220,
http://www.sestos-hk.com/english/download/d1s-en.pdf

I need to maintain 37.3 degrees Celsius. Uses a 130W heating element. The problem is with the initial operation of the controller. In initial operation the temperature rises to 38.5-38.8 and only then begins to drop to the desired temperature. my question, Is it possible to prevent or restrict the controller from going over 37.3 degrees? Is this the method of work and there is nothing to do?

Thank you very much for answering. Help setting SESTOS controller parameters
 
Engineering news on Phys.org
Before setting PID parameters, you need to establish:

1) What is the thermal time lag? That's the time from turning the heating element on until the temperature sensor first detects a change in temperature.

2) What is the rate of temperature rise in degrees per (minute / second) when the heating element stays energized?

3) What is the rate of temperature drop when the heating element is deenergized?

4) What is the temperature tolerance when at operating temperature?

In order to control to a temperature, the temperature change during the thermal time lag must be less than the tolerance. Ideally less than than 1/10 the tolerance, but less than 1/3 should work. That applies to both the rate of temperature increase and the rate of temperature decrease.

And don't forget disturbance inputs.
 
There is a time-honored way to tune controllers in the field.

https://en.wikipedia.org/wiki/Ziegler–Nichols_method

Performance is not guaranteed to meet any of the criteria you named. You'll have to find that out by experiment.
 
  • Like
Likes DaveE
The parameters I changed is the one marked in red. The truth is that everything works perfectly, the only thing that bothers me is that the temperature exceeds what was set in the initial operation I understand that the controller does this for "learning" purposes. Is there a particular parameter I can change that will not happen? Thank you very much I really appreciate.
 
010.jpg
 
Aviv rotman said:
Is it possible to prevent or restrict the controller from going over 37.3 degrees? Is there a particular parameter I can change that will not happen?
The set-point. :devil:
 
  • Like
Likes anorlunda
Aviv rotman said:
Is there a particular parameter I can change that will not happen?
Probably not. There are usually at least two parameters. If you lower all of the gains, eventually you'll get to a point where it won't overshoot. However, the response will be very slow and won't correct for disturbances as well. Most engineers choose to tolerate a little overshoot as part of their performance trade-offs. It's a complicated subject, IMO, that can't be easily answered with a generally applicable cookbook. Although @anorlunda showed you a good version in post #3.

Yes, it can be complicated. Especially for thermal systems since they are often not linear. There's usually a time delay (i.e. extra phase shift) built in because heat has to diffuse between the heater-oven-sensor locations. That tends to make overshoot greater than you would expect from the textbook cases.
 
Thank you very much.👍
 
  • Like
Likes berkeman
You haven't shown us what the time/temperature curve looks like, especially for the initial overshoot (turn-on transit). However, a look at the default settings suggests:

Try reducing the Hysteresis Time from the current 120 seconds to perhaps 90, or even 45, seconds, and see what happens.

Cheers,
Tom
 
  • #10
I'll do that and check. I will update you if there is an improvement.
Thank you I really appreciate💙
 

Similar threads

Back
Top