Smoothing

[altenbach]

---

MrLearners wrote:

This is the screenshot of the graph block diagram.

---

So where do you still have problems? The filtered data you show in the small snippet looks fine and there is no inversion. In fact simulating your graph using the unfiltered data ptbypt in a loop gives about the same result. (see attached)

[altenbach]

In fact, all you need is a chart as follows:

 

[MrLearners]

Hi Altenbach,

I am not sure why I am having this problem. You see in this image, the value is not even close. Is it normal, or perhaps this filter dosen't work for me?

[altenbach]

What do you get if you run the VI that I have attached above?

[MrLearners]

Hi Altenbach,

I used your method in my code and run it. The image is what I get, in the graph there is only 6 pts. In the data I attached previously, there is 52 pts. The filtered data and unfiltered data differ alot.

[altenbach]

You only show 6 points, not 52. You still did not tell me if you get a reasonable result by running my example with the default data that is already in it. I am trying to figure out if the algorithm or defaults have changed.

Would it be possible to attach the actual data (52 pts) that you think does not work well.

[MrLearners]

Hi Altenbach,

[AKA_TG]

Im not a filter expert but I have gotten very satisfactory results in the past using the median pt by pt filter

(using a rank of 2 or 3.)

This usually maintains the original shape of the curve without too much data loss.

[altenbach]

What is a "pt value"?

A point-by-point approach typically always lags a little bit, because it only has one-sided information and cannot predict points that haven't been measured yet.

[MrLearners]

Hi Altenbach,

Here is more data,