Taking the derivative of measured data?

[RyanKlimowski]

Hey I've been trying to take the derivative of some measured data from a digital rotary motion sensor but I'm having a few problems. It would seem since there are so many data points there are alot of areas on my Distance graph are flat for a very brief period of time making the derivative very small. Which I'm assuming is why the graph looks like this.

If any one would like to look at my VI file and help me smooth out my speed graph that would great.

Thanks

Ryan

[Ben]

It looks like you are doing software timed data acquistion so the variation you are seing maybe more a measurement of the Jitter than the singal itself.

Just my 2 cents,

Ben

[johnsold]

The derivative of a step is an impulse. So the result looks reasonable to me (without looking at the VI).

One good way to get a derivative of noisy data is to use the Savitsky-Golay filter method. Depending on whihc version of LV you have, the details vary because NI has made changes in the implementations of the S_G filters.

Lynn

[Lysandros]

---

@Ben wrote:

It looks like you are doing software timed data acquistion so the variation you are seing maybe more a measurement of the Jitter than the singal itself.

Just my 2 cents,

Ben

---

Ben, I was searching for derivatives on labview in the forum and bumped on your comment. When you say "software timed acquisition" do you mean that the A/D converter is triggered by the computer clock in order to acquired data? Is it possible to use hardware timed acquisition for a usb device like the NI-6211? 

Thank you