Cant get FILO array to work to filter AI input

Cant get FILO array to work to filter AI input

Hi, Probably some stupid mistake, I have a wildly fluctuating AI input. So Im trying to filter it by populating a Array[20] at first scan with the current AI values. Then I try to pop the oldest AI value from the array with LFU and push the newest AI value into the array with LFL. Then I use AVE to average the array and use that value. The AVE funct work, but my LFU/LFL just keeps loading Array[0] with the current value.....AND not pulling the old values out of the end of the array.

Yah, been there, done that: you want FIFO via FFL/FFU, not FILO via LFL/LFU. FILO is First-In-Last-Out so it is also LIFO Last-In-First-Out i.e Newest-In-Newest-Out i.e the LFU instruction will pop the newest value, rather than the oldest value.


Also, don't trigger the FFU until the Control.Position is equal to Control.Length (that may mean Control.DN is 1 i.e. [XIC Control.DN] is an alternative to EQU Control.Positon Control.Length).


Finally, the AVE should divide by the Control.Position; alternatively the code could scale the AVE (sum/20) value by 20/Control.Position, but beware of divide by zero.

You should be able to filter the signal per channel. Have you tried that? Also, try the RLIM instruction, it will probably do what you need. Lastly, don't forget that RA has a lot of code already written for a lot of applications. Take a look at their P_AIn AOI from PlantPax, specifically for Filtered PV. I don't use PlantPax as it's bloated, but I use some of their stuff in my own code, this filter is one of them. It works well. It looks like you are making things harder than they need to be.

Still having trouble with FFL and FFU, the position always stays 0 like its constant and if I manually edit it, it just resets it from POS=20 to POS=0

A simple filter would be a whole lot simpler.

But, why is it fluctuating?

See txt...

I considered a last 2 input average, to delete spikes, but unfurtunately I am seeing like a 10-20 second periodic undulations. They put in a TMF meter without enough up and down straight pipe distance.


Anyways I had no luck on FIFO stuff, so I ended up just using a counter to count through and populate an array[20] then (AVE) average that. So I could dampen down the undulation.

Something RLIM could have easily handled.

FIFO is best for this, LIFO won't work as stated.

Why are you using a timer bit to drive both your load and unload instructions? That will not work. You can use a timer to load the FIFO with FFL, but the unload should only be done each time the FIFO control block's .DN bit is set, as that unloads the oldest entry.


Using just a timer bit, you are loading, and immediately unloading the stack.