Why can't I measure output of xtal on scope?

Discussion in 'Embedded Systems and Microcontrollers' started by lowprofile, Mar 5, 2012.

  1. lowprofile

    Thread Starter Member

    Oct 31, 2011

    As part of my microcontrollers class, there is a VERY basic assignment: Hook up a crystal resonator as shown in the datasheet.

    I have done so, and I believe this is well-known enough that I shouldn't have to include a schematic (though I can make and upload one if it's necessary).

    It works great, I have used the 20MHz crystal for many projects already in this configuration. The problem is in trying to use a scope to read that frequency. As soon as I apply the probe, oscillation stops, the PIC does not move to the next instruction. I have a LED "traffic light" that constantly cycles, but probing either the clock input or clock output on the PIC (16F877A), the oscillation stops.

    I am only trying to do this because it's stated as a step in my school workbook. However, I note that in the book, the author is actually using a 20MHz OSCILLATOR (4-pin package) rather than the crystal. She states that they are basically the same.

    I suspect that if I had the Oscillator and not just the crystal that I would be able to measure this frequency.

    I also suspect that the impedance of the probe is loading the resonator circuit, but I don't understand this.

    My questions are:
    Why does attempting to measure this frequency cause the oscillations to stop?
    Would this happen if I used an oscillator instead?

  2. bertus


    Apr 5, 2008

    It can happpen that the load capacity of the probe of the oscilloscope is to large.
    If that happens, the oscillator may stop.
    basicaly this circuit is used for the oscillator at a MCU:


    The attached PDF will give you more ifo.

    lowprofile likes this.
  3. t06afre

    AAC Fanatic!

    May 11, 2009
    You could try to use 74HC04 or 74HCT04 as buffer between your scope probe and the oscillator output.
    lowprofile likes this.
  4. lowprofile

    Thread Starter Member

    Oct 31, 2011
    Thanks guys! Excellent posts!

    I'm going to experiment a little and try a buffer as well as temporarily removing the existing parallel capacitors (to be "replaced" by the capacitance of the probe, temporarily) and I intend to report back, in the hopes that someone else who needs to know will find this.
  5. wmodavis

    Distinguished Member

    Oct 23, 2010
    An important thing to learn in the world of instrumentation (ex. hooking a scope, voltmeter etc. up to a circuit) is that: the measuring device will always alter the circuit operation to some extent such that the measuring device will display a reading that is not the same as when the measuring device is not connected to the circuit.

    Your job (if you decide to take up the task) is to minimize those effects so that your readings are real close to what actually exists without the measuring device connected to the circuit. It takes some knowledge of the circuit being measured and the measuring device characteristics and ohm's law.

    Good luck!
  6. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    My 2 cents is to not to measure these points directly. Anything you add changes the circuit, and if you remove it after the measurement you have a different circuit resulting. There exist some very expensive powered probes (once called FET probes) that add the smallest load to the crystal.

    You can read the frequency off an external oscillator as it has a low impedance output that is isolated from the crystal.