ADF4382A RF output power in divide-by-2 and divide-by-4 modes

I'll see what I can find and will get back to you on your question. 

Here are the results from a measurement I took on the ADF4382A evaluation board. While the board was bolted to the spectrum analyzer, the trace losses of the barrel SMA connector and the RF trace on the eval board have not been removed. The data from 11 to 21 GHz matches fairly well with Figure 24 in the datasheet. Regarding part to part tolerance in the output power, I don't have a firm number on that. Since the part is fabricated on silicon which typically has less variance than GaAs or GaN processes, I would guess that you are talking about something in the +/-2 dB range from part to part. - Eamon

A few more comments. The data was taken at 1 GHz increments. It's important to note that operation is not guaranteed at the following frequencies

1. below 2.875 GHz

2. between 5.25 and 5.75 GHz.

3. between 10.5 and 11.5 GHz

4. above 21 GHz. 
    

   I also received a plot from a colleague that was taken on a different setup during device characterization. This goes from 5 to 21 GHz and includes performance over temperature. It aligns fairly well with the other plot. And while there are no gaps in the frequency range, the same statements above about guaranteed frequency range, apply.  - Eamon