floating point math zero

[jyang72211]

Attached my code.  If you run the code with the default input, you will see a few NaN.  

[johnsold]

The calculation of IaLN_RMS results in 4.999999999999996, not 5. Set the dispaly format to show 16 significant digits.  Thus the calculation of Ia_THD is the result of taking the square root of -35.527E-15 = NaN.

As others have pointed out this type of calculation can result in errors.

Lynn

[jyang72211]

Is there a way to limit all the numbers through out my program to 3 decimal points in my calculation?

[altenbach · Accepted solution]

You will still potentially run into problems because decimal fractions don't necessarily have exact binary fractions. One way would be to make a small subVI that multiplies by 1000, rounds to the nearest integer, then divides by 1000 again.

Overall, you have way too much duplicate code. Why not operate on arrays?

[jyang72211]

Hi Alten,

How would using array cut down on duplicate code?  I am using cluster, because I am using it to make my code more documented, but it maybe backfiring on me.  I do have to go back to the code and clean it up though.  Can you give me an example of what you mean?

Yik

[altenbach]

Here's the THD code case, for example.

Your code:

Same using arrays:

[Yamaeda]

Some modifications which highlights the reduced code duplication.

/Y

[jyang72211]

Hi Yama,

I am using 2009.  Can you re-sent code in 2009?

Yik

[jyang72211]

Hi Alten,

That's nice!!!!  I wish I can give you 10 Kudos.  

Yik