Display time when signal deriv equals zero

[Gzimmer]

Say I have a simulated sine wave, I'd like to have an indicator show the time at which the wave's derivative equals zero. How would I go about this?'

EDIT: For simplicity, say I only care about the first time the signal equals zero.

[RTSLVU]

Maybe I am over simplifying it but if you input your derivative in place of the x control the indicator will change to the time x=0

[johnsold]

Double = zero is a poor practice.  Why? Because of the way numbers are represented in binary in the computer the result of a calculation which is nominally zero may actually appear as a very tiny value.  It is better to compare to a range of values near zero.  You can use "Machine epsilon" (a predefined constant equal to the smallest non-zero value available on your computer), or some larger tolerance chosen on the basis of knowledge of your process.  For example if you have an 8-bit A/D converter with a 10 V full scale range, the smallest signal you can measure is 10/256 = 0.03906...  In such a system anything less than about 0.02 is probably zero.

Use In Range and Coerce? rather than Equals Zero?

Edit: The other issue is that a numerical derivative may never be zero.  If a data sample does not lie exactly at the peak the derivative values will change sign on either side of the peak but may never be close to zero.

Lynn

[RTSLVU]

I understand your point and in cases like you mentioned I would probably end up doing something like that.

But I have ran into more cases where doubles will equal exactly zero and need to act accordingly.

First thing off the top of my head, I work for a power supply manufacturer and often need to test transient response of our PSU's by step loading 0 to full load and full load to 0. When I step from full load to 0 I not only send the (double) value 0 to the electronic load but also use a (x)=0 to turn the load's input relay off fully disconnecting it from the UUT.  My vi's also have the opposite where (x)<>0 engages the loads input relay.  

Notice I said (x)<>0 because the load has an internal crowbar circuit for short circuit testing, I set the load value to -1 and the load applies a short circuit. 

[johnsold]

If you are sending a zero from a control or something like that you are probably OK. You do need to watch controls which have the increment set to a fractional value.  Increments which are not exact binary fractions can cause round-off errors at zero.  Calculations which should result in zero are more likely to cause problems.

Lynn