excel maths geek qu...
 

MegaSack DRAW - This year's winner is user - rgwb
We will be in touch

[Closed] excel maths geek question

8 Posts
4 Users
0 Reactions
117 Views
Posts: 0
Free Member
Topic starter
 

i've been given a parabolic mirror, which i have scanned with a CMM.

i now have XY point data in Excel for a cross section of the mirror, does anyone know how i can find the quadratic constants to best fit my data?

(Y = aX^2 + bX + c): i want to know a b and c...

i was hoping i could use goal seek / solver, or even the quadratic trendline functions, but they've all given up.

please use short words in your replies, i'm a bit thick, thankyou.


 
Posted : 28/01/2010 2:27 pm
Posts: 6980
Free Member
 

solve it on paper?


 
Posted : 28/01/2010 2:37 pm
Posts: 0
Free Member
Topic starter
 

that would take the rest of time...

(for the accuracy i'm aiming for, i've got close to a million combinations of a b and c, and over 200 data points - are you suggesting i sit down and try 200million calculations?)


 
Posted : 28/01/2010 2:47 pm
Posts: 6332
Free Member
 

plot it on an XY graph then right-click any plotted point and select "Add Trendline" then select Polynomial and select the tick boxes for "Display Equation" and indeed "Display R2"

That will give you the best fitting values of a, b, and c

That should do it - why do you say it didn't work?


 
Posted : 28/01/2010 2:51 pm
Posts: 0
Free Member
Topic starter
 

er, i tried that, and the trend line i'm given looks nothing like my data.


 
Posted : 28/01/2010 3:08 pm
Posts: 6332
Free Member
 

send me the data to reggiegasket [at] yahoo.co.uk and I'll have a look at it.


 
Posted : 28/01/2010 3:14 pm
Posts: 0
Free Member
 

Thinks Reggie is either very motivated by helping people or has way too much time on his hands!


 
Posted : 28/01/2010 3:17 pm
Posts: 6332
Free Member
 

it's a slow afternoon.... 8)

I analyse data for a living though, so it shouldn't burn much time....


 
Posted : 28/01/2010 3:42 pm
Posts: 6332
Free Member
 

check your mail Adam


 
Posted : 28/01/2010 5:11 pm