How to monitor for fast changes?

Projects General Guidance
1

I have a potentiometer -- let's call it a volume knob -- that goes from 0 ohms to 10k ohms. I'm trying to come up with a good way to detect when the volume is turned up quickly. No matter where it is in its range, I'm trying to detect anytime it goes up by say 10% or more within a say 50ms. It needs to continually monitor for quick changes -- no start or stop timeframes or high/low limits to work with -- it needs to detect fast change of 10% anytime it happens within 50ms, wherever it is in the range and whenever it happens. I can't come up with a good approach that doesn't require keeping a running log of readings, which seems a bit brute-forcey... Any ideas?

2

ArduinoTom:
no start or stop timeframes or high/low limits to work with

Hi, what does that part mean?

Why not just take a reading every 50ms and compare each new reading to the last reading?

What if the analog reading is 10 or under? Any change will be at least 10%.

Paul

3

50 millisecs isn't fast as far as the Arduino is concerned.

Use the technque in the Blink Without Delay example sketch to use millis() to time things so you take a reading every 50msecs. Save the value in a variable. Have another variable that holds the previous reading. Compare the new with the old. Replace the old with the new. Repeat as needed.

...R

4

You should probably be thinking in terms of absolute rate of change (volts per second or whatever) rather than rate of percentage change. In other words, think of the "10%" or whatever as being 10% of the full scale value, not of the current value.

5

Thanks for the replies. Let me clarify the 10%. The circuit creates a 0 to 5v signal that varies over time -- sometimes slowly and sometimes quickly. By 10%, I meant anytime it goes up by more than .5 volts within 50ms. I'll need to do a little trial and error to pick those values in reality (so maybe .4v within 85ms, or whatever) jut that's the idea.

When I said no timeframe or high/low limits, I just meant there was no particular time to start the stop watch, nor any place within the 0-5v range when it reaches a trigger point.

I was hoping to avoid storing a log of values and comparing each new one to the lowest within the last 50ms, but that's the only thing I've come up with so far too...

6

I don't see any need to record multiple measurements. Can't you just record the new measurement against the previous measurement at regular intervals, calculate the implied rate-of-change and act accordingly?