RF PCB design
MadhuWesly , 09-22-2017, 10:36 PM
Hi Everyone,
I am working on an RF PCB design. Since I am new to RF(around 900MHz) I have few technical queries, please can any one clarify them:
1. What is the best way to take an RF line from top(top side) to bottom(other side) using via. What should be the via considerations there to minimize losses/reflections? - Consider 4-Layer PCB, with 0.8mm final board thickness.
2. In an application circuitry he has mentioned to maintain spacing between components in between RF in and RF out line. It is told in the data-sheet that the additional components(capacitors and inductors) are for 50 Ohm length matching, should we maintain the spacing? Is that for being safe-side or the spacing matters in achieving 50 Ohm impedance?
Ref image:
Thanks & regards,
Madhuwesly.
I am working on an RF PCB design. Since I am new to RF(around 900MHz) I have few technical queries, please can any one clarify them:
1. What is the best way to take an RF line from top(top side) to bottom(other side) using via. What should be the via considerations there to minimize losses/reflections? - Consider 4-Layer PCB, with 0.8mm final board thickness.
2. In an application circuitry he has mentioned to maintain spacing between components in between RF in and RF out line. It is told in the data-sheet that the additional components(capacitors and inductors) are for 50 Ohm length matching, should we maintain the spacing? Is that for being safe-side or the spacing matters in achieving 50 Ohm impedance?
Ref image:
Thanks & regards,
Madhuwesly.
Luca , 09-24-2017, 01:16 PM
Which IC are you using?
Yes the distance is important for the matching.
Yes the distance is important for the matching.
robertferanec , 09-27-2017, 08:52 AM
I do not have much experience with RF 
If I would do it, I would have a look at the reference board and would do it exactly the same way as they did it. Especially because from the picture the requirements look very strict. But I guess, you can not do it same, otherwise you would not be asking.
Possibly, I would contact the chip manufacturer and maybe asked for their opinion.
If I would do it, I would have a look at the reference board and would do it exactly the same way as they did it. Especially because from the picture the requirements look very strict. But I guess, you can not do it same, otherwise you would not be asking.
Possibly, I would contact the chip manufacturer and maybe asked for their opinion.
MadhuWesly , 09-27-2017, 11:28 AM
Very Thanks for your response @Luca and @Robert.
@Luca,
1. For the impedance matching, there is the circuitry with the lumped elements(Inductors, Capacitors, etc), you can observe it in the provided circuit diagram. Though, we are trying to achieve impedance matching why should we maintain the spacing between the components placement, can you please explain me and please do share good resources related to RF PCB design if you have any.
2. Please let me know how to calculate and maintain the spacing while dealing with other RF circuitry if there is not mentioned anything about spacing.
@Luca,
1. For the impedance matching, there is the circuitry with the lumped elements(Inductors, Capacitors, etc), you can observe it in the provided circuit diagram. Though, we are trying to achieve impedance matching why should we maintain the spacing between the components placement, can you please explain me and please do share good resources related to RF PCB design if you have any.
2. Please let me know how to calculate and maintain the spacing while dealing with other RF circuitry if there is not mentioned anything about spacing.
Luca , 09-29-2017, 12:01 AM
For the lumped components you mean only the capacitors? Like the 9pF, 14pF and so on..? Because the "inductor" with microstrip that you rappresent aren't lumped components but the are "distribuite elements".
The spacing between components, especially between the components pads, above a certain frequency (and also above certain RF signal level) must be considered for matching (and also for other stuffs) because as you know in RF matters the effective capacitive (and not only) coupling between traces.
What im saying is a little bit differences between the crosstalk, but the basic concept is roughly the same.
Actually i can't share any PDF or kind of resources because i'm out ot time and place, but i never do manually the design process of track and pads (for components and other stuffs) in RF design but i use ADS from Agilent Technology and also AppCAD.
Designing in RF are very nice, but also very very complex because everything you do with tracks, components placing, any "phisical" layout action MATTERS ... especially with higher frequency and/or RF signal.
I agree with @robertferanec, you have to know how the IC manufactor treat the layout and components.
Without more information it's difficult to say more.
The spacing between components, especially between the components pads, above a certain frequency (and also above certain RF signal level) must be considered for matching (and also for other stuffs) because as you know in RF matters the effective capacitive (and not only) coupling between traces.
What im saying is a little bit differences between the crosstalk, but the basic concept is roughly the same.
Actually i can't share any PDF or kind of resources because i'm out ot time and place, but i never do manually the design process of track and pads (for components and other stuffs) in RF design but i use ADS from Agilent Technology and also AppCAD.
Designing in RF are very nice, but also very very complex because everything you do with tracks, components placing, any "phisical" layout action MATTERS ... especially with higher frequency and/or RF signal.
I agree with @robertferanec, you have to know how the IC manufactor treat the layout and components.
Without more information it's difficult to say more.
MadhuWesly , 10-06-2017, 04:22 AM
Thank you @Luca.
Luca , 10-06-2017, 04:41 AM
Oh for nothing, don't say thanks 
Without more information is impossible to say more.
But i'm sure that you can solve the problem also with the help of software, for example the two that i've mentioned before.
AppCAD are "easy" but still usefull and i use a lot especially at the early stage of designing..
ADS is more more complex but you can really do everything and i use for final stage of the RF designing.
We are here
Regards, Luca
Without more information is impossible to say more.
But i'm sure that you can solve the problem also with the help of software, for example the two that i've mentioned before.
AppCAD are "easy" but still usefull and i use a lot especially at the early stage of designing..
ADS is more more complex but you can really do everything and i use for final stage of the RF designing.
We are here
Regards, Luca
MadhuWesly , 02-28-2018, 03:49 AM
a. 4-Layer stack-up 1: GND(Top) - Signal - Signal - PWR/GND(Bottom).
4-Layer stack-up 2: Signal(Top) - GND - PWR - Signal(Bottom).
4-Layer stack-up 2: Signal(Top) - GND - PWR - Signal(Bottom).
On my 4 layer PCB, I am placing patch antenna on Bottom layer, all other components on Top. When I wanted to use 4-layer stack-up 2, most of the Antenna manufacturers suggest to use layer 3 as Ground plane(Just below the Patch antenna), but due to having high speed signals on top layer I can't shift GND to third layer. So, if I change stack-up and go with
4-Layer stack-up 1 then will it satisfy their recommendation? by using stack-up 1 I will be keeping the Top layer as solid Ground plane directly under the Patch antenna. Will this work good?
b. Will there be any degradation in antenna performance If I draw Data signals(Not RF and Power) on the top layer (under the patch antenna)?
Luca , 02-28-2018, 07:47 AM
0: difficult to answer well....
a: if the antenna manufacturers suggest a ground plane below the antenna track, do it.
a-tips 1: remember, the phisical dimension of the antenna track and of the dielectric material affect the performance for RF signal, so you have to be sure that your track dimension, and stack up dimension (and dielectric material) are good for the application. What manufacturer says about this aspect? It's very important
b: if you are using GND - SIG - SIG - PWR/GND, the first layer (top layer) should be GND... why you want to draw data signals there?
a: if the antenna manufacturers suggest a ground plane below the antenna track, do it.
a-tips 1: remember, the phisical dimension of the antenna track and of the dielectric material affect the performance for RF signal, so you have to be sure that your track dimension, and stack up dimension (and dielectric material) are good for the application. What manufacturer says about this aspect? It's very important
b: if you are using GND - SIG - SIG - PWR/GND, the first layer (top layer) should be GND... why you want to draw data signals there?
MadhuWesly , 02-28-2018, 08:15 AM
b. In case of b: The stack-up assumed is Signal(Top) - GND - PWR - Signal(Bottom).
Luca , 02-28-2018, 08:33 AM
I'm not sure but if you place the antenna on bottom side, you can place data signals on top side without degradation of RF signal.
If i've understand correctly, this is true only in theory.
The data signals could affect RF signals, and viceversa, depends on the "strenght" of that kind of signals.
In generally you can say, for low poer application, that GND and PWR plane shield enough the data/RF signals.
Just a curiosity, which device needs an RF antenna with GND plane underneath ?
If i've understand correctly, this is true only in theory.
The data signals could affect RF signals, and viceversa, depends on the "strenght" of that kind of signals.
In generally you can say, for low poer application, that GND and PWR plane shield enough the data/RF signals.
Just a curiosity, which device needs an RF antenna with GND plane underneath ?
Use our interactive Discord forum to reply or ask new questions.