curve fitting problem

[KDGREAT.]

hi everybody,

i want to investigate thermocouple behaviour to a temperature  transient change.am using USB 6008 and gate & trigger VI with curve fiiting VI. Whenever I run the program, I get the following error message- Analysis:  The system of equations cannot be solved because the input matrix is singular  with the Curve fitting VI being highlighted.

Anybody with clue on what am not doing right would be highly appreciated.

Pls find the attached file.

Thanks.

KDGREAT

[altenbach]

Can you include some typical data?

Are the starting values reasonable?

[KDGREAT.]

Thanks Altabanch.

My initial conditions: a=13,b=1,c=25.

The data is a thermocouple signal (response to a sudden temperature change). I also need to be able to capture the signal using GATE and TRIGGER VI.

Thanks.

[altenbach]

---

@KDGREAT. wrote:
My initial conditions: a=13,b=1,c=25.
The data is a thermocouple signal (response to a sudden temperature change). I also need to be able to capture the signal using GATE and TRIGGER VI.

---

I don't really care where the data comes from, but why can't you attach some typical data so we can try to reproduce the error??

If you graph some simulated data using your initial conditions, does it look anything like your data?

In the VI you attached, the initial conditions are 1,1,1 and not 13,1,25

[KDGREAT.]

Altabanch,

sorry for the wrong initial guess on the vi ,have not actual input it when pasted it but the initial conditions still remains the same : 13,1,25.

The data is almost like a DC signal with some little noise. The closest to it by simulation is by select a DC on simulate signal palette with addition of uniform white noise. That's why I said its a thermocouple signal.

The model I chose is a(1-exp(-x/b))c   . a,b,c are constants.

Thanks for your attention.

[altenbach]

Well, if it is only a DC singnal, it is a flat line. How do you possibly expect to get 3 parameters out of it???

What is a typical range for X?

[KDGREAT.]

Well at a steady state, 3 parameters will be difficult to get, but I want to capture the response of the signal to a sudden change in temperature, say from 25-45 deg c . That's where the problem lies.

Thanks.

[altenbach]

Well, then you should record the temperature response to the sudden change in temperature and stop collecting once it has stabilized at the new temperature. Now fit the entire trace. 

[KDGREAT.]

Thanks once again. pls how can I go about that. Hope am not disturbing much. Am a beginner.

Thanks.

[G2S]

Hi,

I am also having the same issue as the OP. I have attached the VI for review but I cannot get the Curve Fitting to find the non-linear best fit from the data ouput of the trigger and gate VI ? ?

G2S