CC1200: Queries regarding CC1200

1. Stray capacitance is PCB parasitics and internal crystal oscillator core parasitics. They do not change when you change crystal. For this reason the new crystal should should be specified for the same crystal loading capacitance as the one you are replacing.

The crystal you refer is specified for a loading capacitance of 7 pF, whereas the old one was 10 pF. Thus, if you do not change (i.e. lower the crystal loading capacitors on the PCB) the load will be higher than 7 pF and the crystal frequency will not be at 38.88 MHz, but lower.

2. An ESR of 60 ohm will work. Note that crystal startup time depends on ESR; lower ESR gives faster start up time.

3. The frequency initial tolerance, ageing, and temperature drift must be accounted for when setting the RX filter BW. The RX filter BW needs to be wide enough to fit the transmitted signal BW and allow for frequency offset between RX and TX. There are several postings on this subject on E2E. In short: RX filter BW >= Signal BW + 4 x crystal error (in ppm) x RF frequency

CC1200 has a feature called feedback to PLL (FB2PLL) which increases the effective RX filter BW without degrading the noise bandwidth and hence the sensitivity. Please refer to the CC1200 user guide and processors.wiki.ti.com/.../Category:Sub-1GHz for more details. In short: Feedback to the PLL "increases the RX filter" and thus allow for a less accurate crystal to be used. The noise bandwidth, and hence sensitivity, will not increase when enabling this feature. Setting FREQOFF_CFG to 0x30 enables feedback to PLL and increases BW from "programmed RX filter BW" to "programmed RX filter BW +/- RX filter BW/4". As an example, RX filter BW is programmed to 10 kHz, the feedback to PLL increases this filter to 15 kHz (although 10 kHz is still the noise BW).

Hi Sverre,

Thanks for your replies. Points-2 & 3 are pretty clear but please find below my comments regarding point-1.

1. This questions was in general as to what external capacitance tolerance a particular crystal can handle. Also what safe assumption of stray capacitance could be made like 2pF or 3pF or 5pF provided we have followed layout guidelines of crystal.

The reason being suppose I couldn't find a crystal with 10pF capacitance then can I safely use some crystal having 7pF load capacitance (or 13pF crystal on higher side). I want to know general band of capacitance the internal oscillator can handle safely.

Thanks,
jagdish

CC1200 was designed and characterized for a 10 pF crystal capacitive loading. 10 pF is therefore the recommended value and I find it strange that a crystal with this spec is hard to find.

We have 2-3 pF parasitic capacitance on our reference design so this is "a safe assumption of stray capacitance". Although, 10 pF is the recommended capacitive load I believe also 7 pF and 13 pF will be functional. When using 7 pF the frequency becomes more sensitive to changes in board capacitance. When using 13 pF the startup time goes up (Startup time goes with the 1/CL^2)

Noted & thanks for clarifications. Getting some frequency like 38.8MHz in 3.2X2.5mm size with -40'C to +85'C & 10pF was harder as I couldn't locate parts easily in digikey & mouser.

Thanks,
jagdish