Matrix user forums

this frequency meter works great, however i discovered something odd--> this only works accuratly with a tmr2 frequency of 1000 hz, for example reading 5khz input the meter reads 5Khz but if i set tmr2 interrupt to 250hz it reads 4988 hz instead,.... so for kicks i tried a 16.384Mhz crystal instead of internal osc for a lower even tmr2 interrupt freq, i tried 1000 hz again and it was spot on, then i tried 62hz tmr2 interrupt frequency, it was 30 hz low at 4970 roughly,...... i would have figured that using a low tmr2 interrupt frequency my interrupt on pin input frequency would have been more accurate than at a high frequency tmr2 interrupt since at a high interrupt speed there is less time to service the interrupt on pin macro, but that theory was proven backwards(why did this happen?)

Hi Brandon

Had a look at your program, and was getting a bit confused by the results.

Have written my own version, differences from yours are:

• uses 16f877a
  the timer enable and disable are in macros
  LCD update is in a macro
  uses timer1 as counter
  main is used to initialize and has an empty loop
  not using variables to control count etc
  has an output on A0 for calibration (1sec)

As I thought that there could be a problem with using multiple interrupts, I decided to try something different . So after a look at the datasheet I saw that it was possible to use a timer as a counter with an external input. In the Timer_Enable and Timer_Disable macros the C code is used to enable and disable the counter. Also in these macros is an output to port A for timing measurements/calibration.

The macro Frequency is used to read the counter and reset it with the C code, and then display the result.

Using the output on port A I have got the following results, freq is the timer2 int and time is measured on port A

• freq time
  10000 0.9986
  1000 0.9986
  500 0.9975
  250 0.9954

By changing the target value of tick you should be able to calibrate the unit, plus you could use it to change the frequency range if so desired.
If a higher accuracy is required then use the highest frequency possible, as this will allow finer calibration.
So for the above reading, target should be

• 10014
  1001
  501
  251

Still cannot explain the difference in reading

Gary