CD4521B: Crystal stays High, no oscillation

Hi Daniel,

First, I should say that this is not a part developed by TI, so I have limited information to help here. The device was developed by Harris and acquired by TI. I will do my best to help with what I have available to me.

-

Based on an initial inspection, I think you are trying to operate at too high of a frequency for this device. There are a couple of other things that will help you, but I think you need a separate crystal driver to get the 4.2 MHz you are looking for.

You might look at getting an SN74LVC1GU04 to drive the crystal separately.

Even with an external driver, the CD4521B would need to be operating above 5V to work with over 4 MHz input (4MHz is listed as the max for Vcc = 5V).

I might recommend first dividing the frequency of the crystal with something like an SN74LVC1G80, then you can operate everything at 5V and the input to the CD4521B would just be half of what it would have been and you can take your output from one earlier output stage.

-

Since the crystal recommends a load of 18pF, the actual capacitor value that should be used is closer to 14 pF, although 18pF should produce oscillation, just with some frequency offset. There's a good explanation on this website (not affiliated with TI).

Additionally, R114 should be a value larger than 0 for optimal results - usually 1kohm works, but you should at least get oscillation with 0 ohm (just with more ringing). The app note I link below has some images that show how R_S impacts operation as well as a more in-depth explanation.

-

The internal circuit 'driving' the crystal is just an unbuffered inverter. The recommended circuit is a bit odd to me

It looks to me that the circuit would not oscillate in the configuration provided:

The more traditional circuit for driving a crystal is this:

Or from one of our app notes on the topic (Use of the CMOS Unbuffered Inverter in Oscillator Circuits): 

I'm not sure if the connection was forgotten / this was a mistake - or if that circuit was intended and it's just not a crystal oscillator configuration that I am familiar with.