are 50 ohm transmission lines required for 2.4GHz amplifier?

I am designing a 2.4 GHz pre-amp, is it required to have 50-ohm transmission lines (microstrip I'm guessing) at the nodes of the circuit? or will regular copper work? should I put components back to back with each other, or put 50-ohm transmission line in between? or could I just use a regular multipurpose pcb board and solder the components together?

Are you aware that "50-ohm transmission line" is also made from copper, and the 50 ohm is NOT resistance?

That magic 50 Ohm is calculated from ZL=sqrt(L'/C') of the transmission line, where L' and C' are capacitance and inductance per unit length. So the 50 Ohm line is made copper with a certain width, to get the L'/C' ratio right.

Using such an impedance controlled line matters where the line length is longer than 1/10 of a wavelength. For short lines, there is little effect, i.e. width is not critical then.

Thank you, Volker, so would it matter if the distance was really short between the components of the matching circuits, and DC components? and how about ground? does the distance of ground matter? I read about microstrip and stripline; if I was to design the length and width of the copper (conductor), then place that on a board (dielectric), and I have copper underneath the board (ground), then this would be called microstrip?

All depends what you mean by 'really short'. 2.4GHz has a wavelength of 125mm. Given Volker's comment, that gives you about 1cm before you need to worry.
How far apart are the components you are thinking about?
The wavelength of DC is infinitely long so 1/10 of that is still probably larger than your board (and a lot else besides).
You ground should be a big fat 'ground plane' where possible that has a little impedance as possible - what issues are you thinking about with that?
Susan

I was just thinking about putting the components back to back, with no transmission line, would this be ok? how about for the DC biasing also? last question, so with the big flat ground plane, board, and transmission conductor on top = microstrip: like this:https://imgur.com/a/r8jij ? If so, then I need to put into effect the board itself (as the dielectric substrate?) right?

skatefast08 said:

I was just thinking about putting the components back to back, with no transmission line, would this be ok?

Click to expand...

Sure, no length = no issue. Again: any long connection with wrong dimensions will cause issues, because it transforms the impedance seen on the other end of the line. Keeping the line short minimizes that effect.

skatefast08 said:

how about for the DC biasing also?

Click to expand...

No problem. DC on the line (wire) has no side effect on RF characteristics.

skatefast08 said:

last question, so with the big flat ground plane, board, and transmission conductor on top = microstrip: like this:https://imgur.com/a/r8jij ? If so, then I need to put into effect the board itself (as the dielectric substrate?) right?

Click to expand...

Yes, if electrical line length exceeds 1/10 of the wavelength, an effect called impedance transformation becomes visible and relevant. Then, you need to look in more detail at the line characteristics, and that determined by line width and distance to ground below and substrate material.

- - - Updated - - -

Aussie Susan said:

All depends what you mean by 'really short'. 2.4GHz has a wavelength of 125mm. Given Volker's comment, that gives you about 1cm before you need to worry.

Click to expand...

I should have been more precise: the wavelength in this calculation is the effective wavelength for this line, not the wavelength in air. For FR4 substrate, I would estimate 125mm/2 so that we need to worry about line impedance for "normal" lines longer than 6mm.

Special cases with very high or low impedance level require special attention, because then a very inductive (narrow) or capacitive (wide) line might have an effect even at short length.

where you can fit them in, we all use 50 ohm line widths with matched IC circuits.

Thank you so much! I understand now!!!

volker@muehlhaus said:

Using such an impedance controlled line matters where the line length is longer than 1/10 of a wavelength. For short lines, there is little effect, i.e. width is not critical then.

Click to expand...

do you mean the line between components is 1/10 of a wavelength, or do you mean the whole circuit length?

Any length of trace will have a specific impedance that depends on the frequency, the characteristics of the copper, the PCB it is on and nearby tracks etc..
Similarly, any component can be thought of as an impedance as well, but you can think of this as an impedance to ground or the supply rail (this is a very broad generalisation).
Therefore if you have a track that goes between just 2 components then you have a 'pi' network with an impedance to ground (say) in parallel to the input at one end, a series impedance going to the next component which is another parallel impedance with the output.
If that track then extends to another component, then you have another series impedance connected to another parallel impedance and so on.
Now, most components are designed to either provide a designed impedance (resistors, inductors etc.) or have a very high impedance (e.g. an op-amp input) which makes then effectively 'disappear' as a load on the track.
Therefore you really need to look at the track you are dealing with and consider the components that are connected to it and the impact they will have on the signal at the frequency you are dealing with.
Susan

skatefast08 said:

do you mean the line between components is 1/10 of a wavelength, or do you mean the whole circuit length?

Click to expand...

I apologize if I caused confusion. The rule is for connecting 50 ohm circuit blocks. For your LNA where you cascade many SMD that are all NOT 50 ohm blocks, it makes no sense to use 50 ohm lines. In your LNA layout, having the large 50 ohm lines at non-50 ohm points just adds parasitics, and makes the layout large, resulting in poor performance. For your LNA, use short lines, try to keep the layout as compact as possible.

volker@muehlhaus said:

I apologize if I caused confusion. The rule is for connecting 50 ohm circuit blocks. For your LNA where you cascade many SMD that are all NOT 50 ohm blocks, it makes no sense to use 50 ohm lines.

Click to expand...

Would it be more efficient and convenient to use 50 ohm blocks or to put the components as compact as possible? How would you know if the component are 50 ohm? does the component have to be the same size as the line?