RE: nrf51822, custom board, crystal design

Beau R. — Mon, 01 Feb 2016 01:57:27 GMT

Thank you. My board design is not too far from the reference designs. I do have a little less ground plane due to having some other sensors on board. I saw that my 16Mhz crystal seemed spot on which is comforting.

Strangely enough, I was able to resolve this issue by disabling the internal DC/DC converter. One new addition from my last board is the fact that we have a switching regulator on board. Lesson learned, don't enable the dc/dc mode if you are regulating your power.

RE: nrf51822, custom board, crystal design

RK — Sun, 31 Jan 2016 23:34:51 GMT

yes you can use either internal RC (with periodic calibration) or synthesized LF from HF instead of the LF crystal. Last time I asked, the calibrated RC was the recommended one of those.
Capacitors seem right. The formula in the reference sheet is Cl = C1.C2/(C1+C2) where C1 and C2 are C + Cpin and Cpin is 4pF. Rearrange that and you get C = 2.Cl - Cpin. In your case Cl is 8pF and 9pF respectively so that would give you capacitances of 12pF and 14pF, which is just about what you have. Unless you have a lot of stray capacitance from the traces connecting them, you should be in the ballpark. How close to the reference layout did you stay?

You could test the frequency outputs of your crystals with a scope if you have one (i don't) or by getting a timer running off HF or LF clock to toggle a GPIO pin in a simple 'blinky' type app and use a logic analyser. That should tell you if you are way off.

How about antenna? Crystals are tricky, antennas are beasts. If this is affecting you on either crystal source, I'd go take a look there too.