[SOLVED] large scale production of RF cellular band and its production spread

[Andrew77]

Hello,
I'm engaged in a pre-test for a large scale (hundreds) production a 4 stages rf power amplifier, targeted around 30W for cellular bands and a gain of around 45 dB.

By using COG & NP0 capacitors&resistors with 1% of tolerance, and ceramics (AVX) capacitors for the higher power side, what could be a correct estimation of variations for the overall gain?
Could I say that the average gain, for example referring to one single band,
is 45 dB +/- 1dB?
Could production spread can infer to a only +/- 1 dB of variation?
thanks

 

[FvM]

Firstly the capacitors have tolerance specifications which can be quite different, e.g. 5 % downto 1 % or even lower.
Secondly there's a sensitivity function that describes how a specific parameter, e.g. gain depends on the variation of each component value. It depends on the circuit and can be determined either in a simulation, a rough hand calculation, or empirically.

So without detailed circuit information, the question can't be answered. In a regular design process, component properties like tolerances would be considered from the start. ("Design for manufacturing")

 

[vfone]

30W at cellular band (perhaps up to 2.2GHz) would be a challenge to meet the +/- 1dB variation, using standard components.
Today semiconductors have very predictable behavior, but standard passive components and cheap PCB layout at high frequencies and high power could vary more than predictions.