how can i detemine the fc?

[Jonas12@]

I have a signal coming in from a level sensor and it comes with noise.

the level sensor goes up to 94 but increases very slowly.

the noise is from 1 up to 2 levels up and down.

how can I know what cutoff frequency to used if I'm using the butterworth ptbypt filter.

also what samplings can I used?

[crossrulz]

---

Jonas12@ wrote:

the level sensor goes up to 94 but increases very slowly.

the noise is from 1 up to 2 levels up and down.

---

Units would help A LOT here in understanding what you are attempting to say.

Fc is the cutoff frequency of your filter.  In most designs Fc is the 3dB point, meaning the frequency in which the power is cut in half.  You typically want Fc to be slightly outside of your real signal's bandwidth.

[Jonas12@]

I'm taking signals from a sensor. The sensor takes the signals as inches. 

0 to 94 inches. I get noise from the signal and the meter goes up and down 1 inch or -1 inch all the time because I have little wave or movement in the tank full of water. For example I get 90.5 inches and then 90.43,90.70,90.20, 91.3 in inches.  I'm using a Butterworth filter PtbyPt to filter this noise out. My sampling rate (fs) is 20 and the cutoff frequency is (.125). I don't know why .125 and if I put a number below this number I actually cuts the noise, but if I put some number greater than .125  I get the same signal with noise. 

I'm using a sensor that sends signals to the CRIO. I wish It was a RC circuit where I could used the formula fc=(1/2pieCR), but all I can work with is the filter, 

I have attached some pictures that do the same as my program. Can you explain to me why they are using .1 to cut noise. 

and if I put 20 in the sampling how that affects the signal?

[Henrik_Volkers]

It would be usefull if you tell us your samplefrequency and add some real data.

Usually a water tank oscillates in modes .. the lowest (first) mode give the lowest frequency...

Easy filter: just mean last values over one first mode periode.

If you have certain modes due to fill or drain, a clever positioning of the sensor (in the 'knots') can help.

Can you sample faster? Maybe your signal contains line hum?

Oh, and you can rebuild your RC filtler in software, need one feedback 😉

Some basic information for digital signal processing (filter)  is the ratio of fc/fs    .. and Nyquist and ....

[Jonas12@]

my sample frequency  is 15 using the butterworth filter. 

I have a lot of noise going around. as you can see in my data. 

I want to have a more smooth signal. I'm also have a .1 (fc) 

I'm using a round to nearest to have a whole number. I don't know if I'm doing this right to wrong. 

I'm I using the sample frequency  and fc, right? 

[Jonas12@]

what is the fc/fs ratio? and the nyquist use for? 

[Jonas12@]

this is the filter I'm using to filter of cut the noise from sensor. 

this filter takes inches. 

sampling of 20 and fc of .125 

is this enough?...

[crossrulz]

---

Jonas12@ wrote: sampling of 20 and fc of .125

---

Again, WHAT ARE THE UNITS?  In engineering, units a EXTREMELY important just to make sure we are talking about the same thing.

At this point, I recommend you do some basic research on signal processing.  There are plenty of books and web pages on it.

[Jonas12@]

the filter PTByPT in labview only takes Hertz. 

I was assuming you knew.