Why crystal 16MHz doesn't show 16MHz frequency when measured? What are the causes?

Discussion in 'The Projects Forum' started by nazree, Jul 25, 2008.

  1. nazree

    Thread Starter Member

    Jun 23, 2008
    13
    0
    Hello fellow friends and engineers.

    I am not an expert in electronics. I am facing a problem that affects a lot of my time to troubleshoot but still can't figure out the causes of this problem..

    Would appreciate if there is any expert in here..can you share your experience or your views on this?

    Why crystal SMT (surface mount) 16MHz doesn't show 16MHz frequency when measured? What are the typical causes? I tried measuring using oscilloscope both pins, but the value is not 16MHz, but in kHz and sometimes smaller...and the value keeps changing.

    When I measured the voltage across the oscilloscope, there is no voltage at all. Does this mean the crystal is damaged?

    I have tried to change to a new crystal, but still same result.

    For your information, they crystal is used to provide 16MHz to the microcontroller 8051 (pins XTAL1 and XTAL2 of the 8051).

    I used 22pF as the parallel capacitors across the crystal.

    If anyone here can give your views, please let me know. Your helps and assistance is highly appreciated.

    Thanks.

    Nazree

    PS: I will attach the picture of the diagram later...
     
  2. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    It is entirely possible that the level is too low to see, or perhaps it isn't oscillating at all. Do you have an RF meter, spectrum analyzer, or high freq oscope to verify something is actually there?
     
  3. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,692
    901
    Here is another suggestion. Write a simple count program. For example, pulse a pin every 100 counts. Then measure the frequency on the pin.

    That would allow the crystal to be in its normal operating environment and avoid such artifacts and overloading as have been mentioned. John
     
  4. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    Is this crystal feeding a microcontroller? I am willing to bet you are loading the crystal down with your probes so that it cannot oscillate. Check your probe capacitance!

    Steve
     
  5. Tahmid

    Active Member

    Jul 2, 2008
    344
    25
    Hi,

    The crystal oscillator will not show proper frequency if you just connect it to a frequency meter / oscilloscope. You need a proper circuit along with it, that could be made with a transistor and a couple of resistors / capacitors. I use 4Mhz xtal oscillator for my pic micro 16f84a and the crystal is working exactly the way it should. I test it with my transistorized circuit.

    Thanks.
     
  6. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    The reason, I believe, is what I mentioned previously. You're loading a really high impedance circuit with a lower impedance that what it can drive.

    Proper Circuit --> High Impedance Buffer

    You can easily use a voltage follower (opamp ckt) for this. You don't need any caps or resistors, just a properly powered opamp.

    If this is a micro, you can infer the clock by setting PWM and comparing calculated versus measured. The higher the period, the more the precision of the estimation.

    Stephen
     
  7. nazree

    Thread Starter Member

    Jun 23, 2008
    13
    0
    Hello guys.

    Firstly I am impressed with your inputs and opinions. This is among the best electronic forum. Many thanks guys.

    Secondly sorry for the late reply. I haven't logged in for several days.

    Thirdly, I have solved my problem. After changing the loading capacitors (22pf) to new ones and crystal to a new one, the crystal was still not working.

    But when I downloaded the code into microcontroller 8051, it worked ! But i am still confused, because the code has been downloaded so many times into the microcontroller. My question is, do we need to download the code once again into the 8051, if we change the crystal? I don't think so...

    But is there any other possible reasons, why do we need to download the code again? Any opinion?

    Thanks in advance guys. You are really expert..

    Nazree
     
  8. nanovate

    Distinguished Member

    May 7, 2007
    665
    1
    If your crystal was bad to begin with you may not have properly downloaded the code the first time.
     
  9. theamber

    Active Member

    Jun 13, 2008
    318
    0
  10. nazree

    Thread Starter Member

    Jun 23, 2008
    13
    0
    Thanks guys for your kind helpful replies. Highly appreciated.

    Nazree
     
  11. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,346
  12. nazree

    Thread Starter Member

    Jun 23, 2008
    13
    0
    Thanks Bertus. That is helpful.

    Nazree
     
  13. RiJoRI

    Well-Known Member

    Aug 15, 2007
    536
    26
    I have a digital scope, and was trying to find out why my 8051 wasn't working properly. I saw a beautiful wave on the crystal input. The only problem was that the frequency was VERY low. I made the timebase smaller -- from milliseconds to micro- or nano- seconds (don't remember which) -- and the proper frequency wave showed up.

    If the scope is reading the wave too slowly, you will get an artificial wave -- the numbers look reasonable, but they are not realistic. This matter is usually discussed under Nyquist frequency.

    --Rich
     
  14. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Some crystals require a series resistor between the microcontroller and them to work properly. Check its datasheet.
     
Loading...