I am using an Epson-Toyocom C-002RX  CL=12.5pF 32.768kHz (ESR 30kOhm typ.; 50kOmh max )Xtal with the CC2511 and CC2510 chips.

I have power-on startup problems on a number of boards, which i think maybe related to the fact that the ESR of the crystal may be falling below the recommended range as described in the data sheet for the CC251x, even though the Data sheet does not state a minimum, and the Xtal on the Dev board for the CC2510 has a max ESR value less than the typical value for the Xtal I am using. !?!

Also, the data sheet (section 9.6) says the oscillator is amplitude regulated to 400mV p-to-p, however I have only ever seen a swing of about 20mV unless I keep an oscilloscope probe against  pin17 of the device, in which case it slowly climbs up to approx. 200mV p-to-p over a period of 1~2 seconds. Keeping the probe attached to pin17 does not increase the probability of startup, nor does tweeking the load capacitors values to try to emulate a scope probe being attached.

If anyone has some insight into the problem, I would love to hear from you.

Thanks.

Karel

In addition, I have removed the 32K clocks from two CC2510 development boards and tried them on the boards I have which exhibit the startup problems, and get the same results, both in terms of startup failure probabilities and in terms of voltage swing.

If anyone at Chipcon/TI could give a qualitative explanation of how too large an ESR is to large, I would be most grateful.