The copper thickness-->affect the PCB antenna and RF trac

[wccheng]

Dear all,

For the original design, the copper thickness at toppest and lowest PCB is 1.5oz. If we change it to the 0.5oz or 1.0oz thickness, what were the effect for the PCB antenna and the RF trace path? (The PCB antenna and the RF trace path are on the toppest and lowest PCB) Do I need to re-design the PCB antenna dimension and the width of the RF trace path after I change the thickness of the copper?

thanks

wccheng

 

[vfone]

Re: The copper thickness-->affect the PCB antenna and RF

Theoretically if the loss resistance of the copper trace on the PCB is not changed too much, you don’t need to redesign the antenna.
From my experience in small PCB antennas, changing from 1.5oz to 1oz is not big difference. This changing is frequency dependant.

 

[sergio mariotti]

Re: The copper thickness-->affect the PCB antenna and RF

From a physical point of view there is no reason why the losses changes.
Losses depends on Cu roughness, penetration thickness (very smaller that Cu thickness) and Dielectric losses.
Also characteristic impedance is very insensitive to Cu thickness.

So, from a Physical - RF point of view, changing the copper thickness don't change significally the antenna performances.

 

[ADS_RFMW]

Re: The copper thickness-->affect the PCB antenna and RF

Use EM simulation tool to get confirmation

 

[smithchart]

Re: The copper thickness-->affect the PCB antenna and RF

Hi,

I do agree with Mr. sergio mariotti that "losses depends on Cu roughness, penetration thickness (very smaller that Cu thickness) and Dielectric losses etc." However, I don't quite agree that "characteristic impedance is very insensitive to Cu thickness". Below are my comments:

1. Losses due to the penetration thickness: skin effect.

According to the formula: δ=sqrt(1/(σ*pi*u*f)). The skin effect of copper @ 10MHZ is about 20um, while the thickness of 1OZ copper is about 35um. Thus, the resistance is skin depth limited for traces working at 10MHz and above. Here, as a rule of thumb, you can calculate the skin depth of copper based on: δ=1/sqrt(f)*2um, where f has the unit GHz. (For example, δ=2um @ 1GHz).

From this point of view, change the original 1.5OZ design to 1OZ or even 0.5OZ is not a big deal. (1OZ=1.4mil=35.6um)

2. Impedance change due to the change of Cu thickness.

Since Z=sqrt(L/C), when the Cu thickness becomes thin, we would expect the following two effects:
a-- L could be slightly decrease or increase depends on the operating frequency range(usually increase a little bit at most of the cases);
b-- C will decrease slightly due to the fact that contributions of the fringing capacitance decrease slightly, especially for the uStrip cases.
As a consequenc of a&b, the impedance will increase slightly. This "intuition" is also proved by the below simulated results from a 2D-solver shown as below:

uStrip w/o soldermask
H W t DK Z
8mil 14mil 1.5OZ 3.9 52.3
8mil 14mil 1OZ 3.9 53.2
8mil 14mil 0.5OZ 3.9 54.3

uStrip w/ soldermask (0.5mil/DK=3.9)
H W t DK Z
8mil 14mil 1.5OZ 3.9 49.3
8mil 14mil 1OZ 3.9 50.7
8mil 14mil 0.5OZ 3.9 52.4

In addition, if the traces are coupled uStrip or differential pair, the impact of Cu thickness change will be much more higher due to the fact that the thickness change directly changes the area of the side wall that the coupled traces facing to each other.


Suggestions:
If you are looking for a higher accuracy or seeking a more firm confidence with your final design, you'd better give your design a quick simulation using some EM solvers.

Best regards,

 

[elek_fyc]

normally you should not notice a big difference.
However this depend of the operating frequency. So may be for high frequencies there is some skin effect.
On the other hand, the thikness can modify the capacitive effect and hence modify a little bit the impedance.