Professional frequency measurement Upgrade

Today I came across some very interesting videos about professional frequency measurement error ( Agilent 53131A Frequency Counter Oven Upgrade) and it gave me the idea to better the precision of the frequency counter I am working on.

I was quite happy until now with the accuracy of about 100 ppm measured on the Arduino compatible frequency counter, but I realize that it is probable that we can get 1 order of precision better just (with a redesign…) by using a much better crystal oscillator. Until now I opted for a regular 16 Mhz crystal and if one considers what is reported on this discussion about the Arduino clock frequency accuracy it appears that the 100 ppm error that I measured is very coherent with the error due to the clock. So probably we could get down to 10 ppm or less using a more accurate clock.

Then I found that this Abracon LLC ASVTX-09-16.000MHZ-T TCXO might be a not too expensive solution. It has a frequency accuracy of +/- 2.5 ppm.

But I need some more information on using this TCXO instead of a standard parallel crystal.

What is your opinion about this? Do you know about a successful clocking of the Atmega328p using a 16 Mhz oscillator with an error under 10 ppm ?

3 Comments

  1. online learning

    Simply desire to say your article is as surprising. The clarity in your post is simply excellent and i could assume you are an expert on this subject. Well with your permission allow me to grab your feed to keep up to date with forthcoming post. Thanks a million and please carry on the rewarding work.

    1. mm
      admin

      Thank you very much. I finally opted for the best crystal available on the market +/- 10 ppm frequenncy accuracy and +/- 10 ppm change with temparerature. Aging 2 ppm. So I hope the frequency counter will reach an error under 30 ppm with this simple change. Using a Tcxo may be considered for un upgraded version of the board in the future (it needs a little re-design).

      1. mm
        admin

        Hello, this is to let you know that indeed the next version will use a TCXO with a frequency stability over temperature of only 0.5 ppm and this will be coherent with the objective of a precision of a couple of parts per million. That is about the maximum we can get from an Atmega328p based board.

Leave a Reply

Your email address will not be published. Required fields are marked *