Couple of Questions Regarding SCP Instruction and Sensor Calibration

Hey guys, I've got a pH/conductivity transmitter and a 1762-IF4 A/D for my micrologix 1200. I'm using an SCP command with implied decimal places to turn the value into a scale for 0 to 1400 for pH (corresponding to 0.00 to 14.00 pH) and the conductivity is scaled for 0 to 100 for mS/cm (corresponding to 0.0 to 10.0 mS/cm ).

Now, to calibrate these two, it comes down to basically displaying the current sensor value on the screen and changing the offset or trimmer pots to set the sensor to the buffer solution currently being measured.

So I've got a couple questions, firstly can I have the SCP command run every loop through? I know that I've set the card to update without filtering for 130mS update so it probably won't be having a new value every loop but Im just concerned if weather the SCP command will take up significant processing time? I'm guessing not..

Now as for actually taking measurements, I'd imagine I'm going to want to take some averaged measurements over time.. So this brings up the question of whether there is a way to see if the measurement has updated....Here is an example, lets say I have measurement values of 5,3,6,2. Now the actual average of these should be (5+3+6+2)/4 = 4. However, if I just simply added it up every cycle I could end up with (5+5+3+3+6+6+2)/7=4.28 which would not be correct... So I'm wondering if there is a way to only update the average with new measurements?

Thanks!

You could use a compare to see if the value has changed and move the new value to a new register use this register for your display

I thought of that but for averaging how will I know if the measurement remains the same? I dont think it will count towards the average as faras I can tell?

Like bce123 said, compare SCP output to a storage register, if the values are NEQ, take action, then move new value to storage value. That way you only take samples for averaging when the value changes.

Let's say that N7:0 is the value returned from SCP each scan. N7:1 is the stored value. We'll use N7:10-19 for the samples to be averaged. We'll store the sum in F8:0 in case it's too large for a 16-bit register. N7:2 will be the average of 10 samples.

This method will give you a rolling average of the last 10 samples that will only be updated when a new value comes in, rather than every scan. This will only return an accurate average once all 10 sample registers (or you can adjust for however many samples you need) are full. If that is an issue, then you can add an ADD block to increment an integer every time a sample is added until the queue is full, then divide the sum by the sample count, rather than MUL by .1, i.e.: