Measuring the frequency of a short pulse train

[riccardo]

Hello,

I am looking at ways of measuring resonance of a coil/cap in parallel so that it can then be pulsed at that frequency (around 100kHz). Preferably I would like to do this with a relatively small 8-bit AVR like the ATMega 328P used in the UNO. Basically I would put a single pulse into a coil, then the ringing waveform is buffered to give a set of sqare pulses for the micro to use.

I've seen a few different methods for making freqency counters but have difficulty with them. This is due to a few reasons:

• The frequency is too high for direct pulse timing like the Arduino PulseIn command. (the accuracy is very poor)
• I dont have a continuous pulse train to sample continuously
• The pulse train can have a varying number of pulses from the ringing coil (e.g. 5 to 20)

Could anyone offer any suggestion on how to do this?

 

[Nigel Goodwin]

You simply need to rewrite the frequency counter program slightly - I'm presuming there are Arduino frequency counter programs just as there are PIC ones (50MHz counter when using a 4MHz clock).

A frequency counter normally continuously reads the frequency - you simply need to modify the code so it only reads it once, or twice, or three times, whatever you want. I would suggest that you modify it by adding a button to 'reset' the counter, which waits to be pressed after a reading has been taken - I would also add a routine before the reading is taken, that waits until a pulse has been detected before it starts the reading.

So it sits there waiting for a pulse, once it detects one it starts counting the incoming pulses (using a hardware counter - again presuming the Arduino has a suitable one), and stops counting after 1 second - then displays the reading, and waits for the button to be pressed to start again.

 

[Pommie]

If the coil and cap are in series then could you vary the frequency and see at what frequency the current is maximum? I may be wrong on this but I'm sure someone will correct me. Over 40 years since I studied resonant circuits!


Mike.
Whoops, just seen they're in parallel but think the power should still be maximum at resonance.

 

[dougy83]

You can use a timer input capture module. Basically this allows you to read a 16-bit counter value each time the input changes state (or e.g. has a positive edge). The different between two positive edges is the period of the waveform, although you can wait for as many edges as you want to get a more accurate measurement. Just subtract one [unsigned short] count from another to get the period. For 100kHz, the period will be 160 counts for 16MHz crystal of the Uno.