How to generate a random number within a specified tolerance and control that tolerance.

[SPSUENGR]

I am trying to simulate data without using a DAQ device by using a random number generator and I want to be able to control the "range" of the random number generated though.  On the front pannel, the user would be able to input a nominal number, and the random number generator would generate numbers within plus or minus .003 of that nominal number.  The data would then be plotted on a graph.  Anyone know how to do this?  Or if it's even possible?

Just for information sake, the random number is supposed to look like thickness measurements, that's why I want to simulate plus and minus .003 because the data could be thinner or thicker than the nominal.  

Thanks in advance!!

[Darin.K · Accepted solution]

Two ways to approach this.  First, for a uniform distribution you could "Roll the Dice" which is the Random Number (0,1) function, multiply the result by 0.06 and then add Mean - 0.03 to center the distribution around Mean.  If you are simulating a real process, then my initial guess would be that the results would follow a normal distribution.  In this case you could use the Gaussian White Noise VI from the Signal Generation Pallete with a standard deviation of 0.015 if you wanted 0.03 to be the 2 sigma point, and then add Mean to the results.

[falkpl]

I always wondered why the the random number generator didnt have optional inputs like range and mean or max, min.  Most of the time this is how they are used.

[JÞB]

---

falkpl wrote:
I always wondered why the the random number generator didnt have optional inputs like range and mean or max, min.  Most of the time this is how they are used.

---

Hmm.................. Kudos on your "Idea exchange" post................

Oh,..... wait,,,,,,,,,,,,, you haven't idea'd it-  YET.

Link to here from the LV Idea X.  good_Idea

Message Edited by Jeff Bohrer on 04-09-2010 09:09 PM

[SPSUENGR]

Thanks Darin,  I used the Gaussian White Noise.vi and it worked beautifully.  I was able to simulate what I needed and impress my professor also.  Thanks again!

[altenbach]

You can create a random number with a gaussian probability profile using two equal distributed (0...1) random numbers and applying the Box Muller transform.

I wrote one many years ago, her it is. 🙂 (LabVIEW 8.0). 

Message Edited by altenbach on 04-12-2010 09:13 PM