CC1101 PATABLE Output power for 0x77

I checked the design memo for the PA on CC1101 and PATABLE[7:4]=7 is not mapped to any internal control signals so 0x77 is basically a illegal value. If anything the maximum output power will be around 0 dBm.

Hi,

I`ve been trying to create an adaptive power scheme for a cc1101 application. I`ve checked the Design note refering to Programming PATABLE for desired output power, and from the typical output tables listed for 868MHz operation (and also 433MHz) , it seems pretty random what is the output power vs patable value relation.  

For example the LSB  nibble increasing results  either an increase or a decrease in output power, and 0x6X are invalid values.

Is there some explanation on what is the internal structure for these registers?

I would like to have some way of knowing how to cause some increase/decrease in output power given the current patable value to adjust the RSSI on the other end (which is being exchanged at regular intervals between nodes). I am using the 868 EM and plan to move to a 433Mhz custom pcb promptly, thank you.

The PA table on CC1101 is a bit tricky to understand. When designing it the intention was to add a wrapper translating a easy to understand table that customers uses to the internal, slightly random, control bits in the PA. This wrapper was never made...

PATABLE[7:6] select a PA output stage. The function of the LSB bits is dependent on the stage used. If you look at PA settings with the same setting for PATABLE[7:6] setting in the app note you should be able to see if you have to go up or down on the LSBs to adjust power.

That said the preferred solution would be to find the output powers you want to use from the app note and build up a table where you select a given PA setting if you want to increase or decrease the PA power by a given amount.

Both answers were incredibly useful. However it seems that the best option is to preselect a set of values and iterate through them. Also, on the lowest power output stage PA_POWER[7:6]=00, it seems that PA_POWER[5:4] behave kind of like in PA_POWER[7:6]=01, creating 4 groups of values that are discontinuous. This is based on the typical output tables, where values of 0 to 64, when plotted, show 4 different groups of output power. Can you verify that?

Yes, for PA_POWER[7:6]=00 the output power is devided into 4 groups where the output power in one group is partially overlapping the output power in the next group.