this could sound a repeated answer but to me it seems to be not so trivial. Or at least, it is so due to my ignorance.
I need to use an RF6886 power amplifier to interface it with a 50 ohm antenna. Are present the schematics of my frequency range (400-450MHz) and it is stated that it works pretty good with these 50 ohm.
With the PCB stackup used, 4 layers PCB, the RF traces are surrounded by GND planes on the LAyer 1 and the Layer 2 is a complete GND plane. Other signals/GND are on remaining planes. With this structure, the 50 ohm is achieved by a net of width around 0.65 mm (or less, now it is not importante the exact value).
Here comes the problem: the reference schematic (in here, page 10) reports the component used, but I feel unconfidend on the physical parameters of the nets before the antenna connector, therefore everything that connects the components of the matching network. Moreover, the output uses all of 6 pins of one side. But unifying them in a large net, means having a much lower impedance (just as every other combination of this output).
I could simply use the reference layout, but I didn't find it yet.
How can I extrapolate the allowed net characteristics (Width), while the only thing known is the maximum VSWR? The length should not be a problem (I suppose), being more more less than the wavelength.
Not to mention the input: it is a sort of CLC T-network, but I'm totally unaware of the input impedance of the chip, so there is the same issue. I'm entering to it with a net of 50 ohms, but it is the only assumption that seems valid. I'm sure I am missing something.
A possible answer, for the OUTPUT only: In page 9 of that datasheet talks about load impedances of the reference schematic: assuming that it is the load saw just at the output pins, so before the matching network, I can extrapolate a net dimensions. But how about all the intermedium connections of the output network? (so…as just mentioned before)
I hope to be clear. Thanks and sorry if it is too silly.
Best Answer
You could simulate your layout to estimate the VSWR the chip will be seeing.
You'll need to consider the reflections from the layout, layout variability, and the antenna and its connector to determine the worst case VSWR your design will produce.
Note that the 5:1 maximum VSWR specified in the datasheet represents a 67% reflection. It should be fairly straightforward to achieve better than this spec at 450 MHz.
This one I'm stumped on. The datasheet doesn't specify any input impedance or VSWR that I can see. The equivalent circuit they show implies a low-impedance input. You may want to call your sales rep to solve this one.