LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Has Temperature Stabilized?

This is probably pretty basic but I am very rusty on calculus.

 

I have a thermal chamber. I want to set a temperature and then know when the temperature has been within about one degree of the setpoint for a given number of minutes. Say thirty minutes.

 

I know the easy way is by comparing the setpoint to the temperature and to use a timer. But there must be a more elegant way. Using the easy method will mean that if someone opens the door and the temperature changes by two degrees for a short amount of time means waiting another thirty minutes.

=====================
LabVIEW 2012


0 Kudos
Message 1 of 7
(3,869 Views)

Interesting!

 

Try this.  Maintain a history of the chamber setpoint and the chamber temperature.  Subtract the setpoint from the chamber temps to get "Thermal Impulse Moments" (a measure of the work the chamber must do to control the temperature) scale these values by e * time (to "Weight the most recient values higher)  the thermal transfer to the DUT (or whatever you are soaking will follow a natural log function so we are justified to weight the recient impulse over the earlier by some multiple of e related to the thermal impedance of the item to be soaked)  sum the result and divide by time.  Determine imperically the threshold that demonstrates the item is at temp.  


"Should be" isn't "Is" -Jay
Message 2 of 7
(3,864 Views)

Steve,

 

Are you saying it is OK for someone to open the door and that this will not change the decision about stabilization?

 

Two things: 1. What is the character of the system response? Does it approach the setpoint smoothly and always from the same side (like a first order system) or can it overshoot and oscillate about the setpoint (higher order system)?  2. How close is the one degree limit to the resolution and accuracy of the system?  Does the same temperature sensor provide information to the temperature controller and to your monitoring system or are the sensors separate?  How much noise is in the temperature data?

 

If you have a first order system (or a higher order system tuned so that it does not overshoot) and a single sensor, then you might be able to fit the temperature/time data to a suitable exponential curve (1 - exp(-t/tau)) and use that to predict when the stabilization has occurred.  For systems with overshoot or mutliple sensors I think you might have a much more complicated decision process.

 

Lynn

Message 3 of 7
(3,863 Views)

@johnsold wrote:

Steve,

 

 you might be able to fit the temperature/time data to a suitable exponential curve (1 - exp(-t/tau)) and use that to predict when the stabilization has occurred. ..

 

Lynn



ditto


"Should be" isn't "Is" -Jay
0 Kudos
Message 4 of 7
(3,860 Views)

johnsold, it is possible that the chamber has been at a stable temperature for several hours. Someone may open the door to change something. What we have been doing is to wait 30 minutes from the time that the setpoint and temperature are nearly equal. There is a button to skip the 30 minute soak time. I would like the system to know when the temperature is stable - it may take less time depending on the temperature the chamber was at when the door was opened.

 

#1 The chamber is an industrial thermal chamber. It does not overshoot or oscillate.

 

#2 I am not sure of the resolution but the accuracy is pretty good. It is possible to set it to a temperature and it will stabalize to within 1/2 degree Celsius. Maybe better.

 

No multiple sensors or overshoot.

 

You and Jeff gave me some stuff to think about. Thanks.

=====================
LabVIEW 2012


0 Kudos
Message 5 of 7
(3,854 Views)

My Q&D-try to find a simple solution.

Soaktime.PNG

 

Felix

0 Kudos
Message 6 of 7
(3,812 Views)

I know this is not exactly what you were asking for but I do thermal testing.

 

We use UL standard UL1778 Section 49 to determine temperature stabilization.


"Temperature change shall be less than 2 degrees C over 3 consectutive 15 minute invervals."

 

Here is an action engine I created to determine when the devices I test have reached thermal stabilization. You feed it an array containing the last temperature readings. In my case I meaaure the temperatures every minute and display them on a chart them every 15 miutes I send the temperatures to this action engine to test for stabilization.

 

========================
=== Engineer Ambiguously ===
========================
Message 7 of 7
(3,808 Views)