Crystal Oscillator tester
About a year ago, while working on a project dealing in cc2500 wireless communication, It was a pretty simple system, one transmitter and one receiver. Everything was right but it didn’t work. After fighting with it for a full day, I found that the crystal oscillator was faulty. After that day, every time I buy new crystals, I put ’em in my bin only after testing it.
Testing crystal oscillator
I used to test on a board I made for LED binary clock, I replaced the crystal with a female burg strip so I can insert crystals, board has a mega328 which runs the ASCII Table sketch. Inserted the crystal, hit reset and see the ASCII table on serial monitor, if it runs okay, quartz crystal goes in storage bin. This was cumbersome, as I needed a whole PC to see the crystal was okay.
The other day, I got 20 crystals out of which just one worked, during testing, for a second I doubted that my circuit may have gone faulty but it was alright. I took them back to vendor and good thing he replaced the 19. So the problem was I buy the crystals, test them at my workshop and return them if they don’t work. I made a GLCDTerm few weeks ago which was perfect for this, I could hookup a 9v battery to it and power my test circuit from, and see the serial data right on the GLCD. So I head down to get my replacements and check every one right there at the vendors place.
Rohin - Mar 6, 2015
I think this depends on the error margin of the crystal you are using. I had similar experience with CC2541. The datasheet demands a crystal with 20ppm error for clock frequency stability (typically you get SMD versions of crystals that are small in size for this range - expensive by almost 3-4 times). The ordinary crystals such as this (http://in.element14.com/iqd-frequency-products/lfxtal003240/crystal-16mhz/dp/9713700?MER=MER-BPR-4PPM-PD-IQD-CO&) have 30 - 50ppm error. I think this could be the reason why you had to discard multiple crystals!