Hi,
current configuration:
1150 output into balun, then DC block, 9 dB pi-pad, then input matching ckt of RFMD RFPA0133 (half watt ISM band PA).
The PA is unconditionally stable.
With the PA set to low or mid output power:
The CC1150 produces the Tx burst at the correct level at the correct frequency, and when idle, is quiet.
With the PA set to high power:
At low programmed output levels (> 20 dBm), the 1150 produces the desired output, plus a simultaneous free running tone roughly in the 480-500 MHz range, about 12 dB higher than the desired one.
At higher settings for the 1150, the desired tone is produced during the Tx burst, but reverts to free running VCO during idle as above. This free running tone disappears during the Tx bursts, reappears during the idle state.
Loading via my finger over the Xtal does not change anything.
Loading via my finger over the 1150 IC shifts the free running frequency up by ~10 MHz (the locked desired frequency does not change).
Adding the matched pad between the 1150 and PA proved to be enough isolation to eliminate this behavior for 1 or 2 units. It remains on hte other. Balance of workmanship for solder joints appears very good, and board design follows reference, uses 0402's, ground- surround on top layer and all grounds tied by vias with 050" spacings on average.Power supplies noise greatest during SPI writes and does not exceed 150 mV (closer to 50 mV when SPI inactive).As for SW, we are doing the auto calibrates and SRES after the part is turned on.
Seems like load pulling / knocking out the lock when the PA input impedance (and drain current) changes rapidly (as when RF signal becomes present or goes away).
Any recommendations greatly appreciated Regards, Stuart
