Map function with arrays?

[Daniel Wood]

Hi guys

I have a working circuit at the moment measuring a sensor and I am using the arduino map function to calibrate my results

float map(float x)
{
return (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min;
}

Everything works fine but there are some non-linear regions to my sensor which is throwing off the results, does anyone know of a version of this method of calibration that can use arrays as inputs so I can do a 10 to 20 point calibration?

 

[be80be]

Not going to get close using map. Map going to round off anyway so if you had a table its just going to be divided into parts with map.

You'll have to write your own map and call it something like Nmap because map is built in to arduino ide
The example you posted is part of it
For the mathematically inclined, here's the whole function

long map(long x, long in_min, long in_max, long out_min, long out_max)
{
return (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min;
}

 

[misterT]

Isn't that just a linear interpolation (equation of a straight line)?

I hate how Arduino tries to make things "simple" by making people lazy, or "reluctant to learn".. no need to work out simple linear functions anymore.

Graph the input of the sensors versus the output you wan't. Then we can figure out how to do the "mapping".
I guess you need a piecewise linear function, which you can turn into lookup table (array).. and maybe use interpolation for extra accuracy.
(google "piecewise linear function" and "linear interpolation")

 

[ClydeCrashKop]

I got a lot of help doing something similar on this thread.

https://www.electro-tech-online.com/threads/converting-adc-readings-to-degrees.138402/#post-1147501

There is a link to a calculator in there.

 

[Ian Rogers]

Do you want mine????

up to 8 point calibration....

 

[Daniel Wood]

Thanks guys!

Ian, I am half way through solving this thing.. I think all I have to do is wrap the function in a for statement and it should be good to go
(moved to android basic)

Sub MultiCalibrate


  Dim x As Float
  Dim y As Float
  Dim i As Int

  For i As Int = 0 To NO_OF_POINTS
 
  If x >= INN(i) AND x <= INN(i+1)
     y = (x-INN(i))*(OUT(i)-OUT(i+1))/(INN(i)-INN(i+1))+OUT(i)
  End If
 
  Return y
End Sub

 

[Ian Rogers]

(moved to android basic)

Cool!! having fun...

If you have an array of calpoints.. You only need to map to the four points that are relevant..

Use you analogue input to determine which four cal points to use.. The lower two become your zero and the upper two your span...

Once you determine the four points you can pass that to the map function....

 

[misterT]

Your IF can be simpler

For i As Int = 1 To NO_OF_POINTS
If x < INN(i)
.. You have found the right "i".

 

[Ian Rogers]

This is how we are rolling

 

[Daniel Wood]

Thanks T for telling me the correct terminology
Well it works like a charm now. I'll leave it here for reference in case anyone needs it in the future

'Both arrays must be equal in length
OUT = Array As Int(0,67,105,209,326,419,520,631,727,838,942,1040,1150,1253,1371,1462,1569,1682,1779,1881,1987,2078,2210,2315,2409,2513,2620,2730,2835,2944,3034,3150,6430,9518,15810,0)
INN = Array As Int(2,4 ,5  ,7  ,12 ,15 ,19 ,22 ,26 ,30 ,34 ,37  ,41  ,44  ,48  ,52  ,55  ,59  ,62  ,66  ,69  ,73  ,77  ,81  ,84  ,87  ,91  ,95  ,100 ,105 ,109 ,110 ,216 ,318 ,514,0)


Sub MultiCalibrate

  Dim i As Int
  Dim x As Long

  x = PARAM(2)

  For i = 0 To INN.Length-1
    If (x <= INN(i)) OR (i >= INN.Length-2) Then
        temp1 = (((x-INN(i))*(OUT(i)-OUT(i+1))/(INN(i)-INN(i+1))+OUT(i))/1000)
        Exit
    End If
  Next

End Sub

Ian: Haha the lazy sods! Tell the boys that I envy them!.. You might be interested in the App I'm making.. I finally got that android application working!

 

[Daniel Wood]

Cool!! having fun...
Use you analogue input to determine which four cal points to use.. The lower two become your zero and the upper two your span...

I've been calibrating stuff at yours for years