50 ohm matching question

Neewbie....

What is the proper way to connect/match components in an RF front end assuming all have 50 ohm IO? Microstrip traces? LC matching networks, no matching required?

Thanks

Rich

Which method to select depend on your application, components cost money, traces cost PCB space. If matching is needed depend on your requirements. If no matching results in good enough matching is further matching improvement above specification, which you not get paid for.

It depends. If you are at 88MHz, you will not want to try to use MicroStrips. If you
are at 14GHz you will not want to use L's,C's, R's unless they are like 0201 size....

You have to know what is the Impedance of one component them match it to the
next. You should read a book a good, cheap, and easy to understand book is
"RF Circuit Design" By Bowick. You can also Google RF Matching.

RF Match


In the photo the Red Dot is the Components Impedance and every time the white line
changes direction that is due to a new Element/Matching Component.

I am at 1GHz. I am putting together an RF front end, LNA and SAW, to a receiver IC. The antenna, LNA, and SAW filter all have 50 ohm IO so perhaps I do not need matching. I looking into connecting the components with microstrip, according to a online calculator the trace should be 140 mil thick...........

Rich

Copper thickness is usually 1/2oz(0.7mil) or 1oz(1.4mil). What substrate are you using?

I am using 1oz Cu and FR-4 0.062"

Thanks

Rich

the material and thickness you are using is standard for non-RF circuits. For the frequency you are operating at, you should consider using Rogers material. At 1 GHz, the Rogers 4003 is decent. At even higher frequency, above 6GHz or so, you should start using Alumina or Quartz.

The high frequency portion is small, SMA, lna, saw, and then directly into rx ic. Perhaps fr4 would suffice?

I don't think FR 4 would suffice for microwave application. At microwave frequency, the dielectric constant variance is high ranging from 3.9-4.5, this is a guess. Also, the loss tangent is quite high, on the order of 0.035. Rogers material has low dielectric constant variance and lower loss tangent.

OK, good to know, thank you very much for your help. I will recalculate the microstrip with Rogers 4003.

Rich

1GHz is low enough to use discrete components (0603).
Now, even though all your parts are 50 ohms you still should put some "Hooks" in
your layout for unforeseen issues.

Between every device Antn,SAW, & LNA. you should put pads for a series and shunt
component (an "L-Matching Network". If everything is perfect (Simulation ≡ Design)
then you will just need to put a 0 ohm resistor in the Series and everything is great.
But if the Simulation and the Physical design don't match you can tune up the design.

It really depends what line length you need.

For 0.5mm FR4 thickness, the 50 ohm line width is 1mm and the insertion loss is ~0.05dB per 10mm line length.

That is really a good idea. It is always best to have options.

Thanks to all for the advice.