Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Regarding RF PCB layout design in my company product, some questions

Status
Not open for further replies.

simbaliya

Member level 4
Joined
Feb 1, 2011
Messages
74
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,943
I am a PCB design engineer, the product from my company uses BLE solution, so there is a RF chip, followed by BAW filter and antenna.

Regarding the PCB design for this part of circuit, I have some questions that my wireless colleague can not explain to me clearly, I believe the good practice of RF layout design he taught me is from previous experience, but he can not trace the real reason.

The PCB layout is designed like this, the PCB ground plane is split into two(not difference layers but different territory), one is RF ground, one is antenna ground, means the second ground plane is used as RF antenna. In between these two planes there is a 22nH inductor electrically connect the two. Our main circuits including RF circuit are located on the RF ground plane side, while a few other circuits are located on the other ground plane as RF antenna due to some physical limitations. So we are using part of the PCB ground planes to behave as RF antenna, my first question is:

1. what is the purpose of the 22nH inductor between RF antenna plane and RF ground plane?

Besides of that, I was told to add a lot of via along the surrounding RF ground planes of the RF transmission line, this is my second question:

2. Why do we need via along the RF ground plane that surrounding RF transmission line?

There is more, at the RF ground plane side, since our PCB is 4 layers, I was told of stitch via wherever possible on the ground planes, my third questions:

3. Why do we need so many via on the RF ground plane? Is there via here serving different purpose compared with via in question 2?

Additionally, one the antenna plane side, there are also many stitching via, my forth question:

4. Why do we need so many via on the RF antenna plane? Is it correct? Is the antenna still emitting RF EM wave as expected?

At last, remember I mentioned that there are a few circuits in the antenna plane side? Some nets from the circuits will route from antenna plane to RF ground plane, when every of such trace reaching the gap of the two planes, a 22nH inductor is placed in series, my last question?

5. What is the purpose of the 22nH inductor here? Is it serving the same purpose as in question 1?
 

To comment the "RF antenna plane" point and related questions, we should have more information about the implemented antenna type, the role of the "antenna plane" etc. At first sight, this sounds strange.

Regarding stitching vias or via fences, I presume that the basic transmission line topology is microstrip. If you embed the microstrip with ground traces or copper pours in the outer layer, the topology is changed to coplanar with ground. It gives the option to make smaller transmission lines compared to pure microstrip. But it's necessary to tie the coplanar ground closely to the ground plane, other wise they form line stubs coupled to the micro strip. Possible effects are cross talk and unwanted resonances.

Similarly any ground stubs on the signal layer, e.g. all SMD ground pins, need a short, low inductance connection to the ground plane, which is best implemented by multiple vias.
 

To comment the "RF antenna plane" point and related questions, we should have more information about the implemented antenna type, the role of the "antenna plane" etc. At first sight, this sounds strange.

Regarding stitching vias or via fences, I presume that the basic transmission line topology is microstrip. If you embed the microstrip with ground traces or copper pours in the outer layer, the topology is changed to coplanar with ground. It gives the option to make smaller transmission lines compared to pure microstrip. But it's necessary to tie the coplanar ground closely to the ground plane, other wise they form line stubs coupled to the micro strip. Possible effects are cross talk and unwanted resonances.

Similarly any ground stubs on the signal layer, e.g. all SMD ground pins, need a short, low inductance connection to the ground plane, which is best implemented by multiple vias.

Hi FvM, thanks for the reply.

Indeed, the transmission line topology is co-planar with ground, the RF trace in at outer layer. So you mean the ground via along the pour copper surrounding the RF trace is to avoid those pour copper become line stubs. I am not an RF expert so I don’t quite understand how line stub will cause cross talk. My understanding is that, even I only add one via here, it will act as ground shield to protect RF receiver signal or to avoid cross talk of RF transmitting signal to near by circuits, isn’t it?

And when you say the stub will also cause resonance, may I know how does this resonance happen? And Why the resonance is bad?

Regarding more information of the antenna, in old times we were using some metal part attached to the product housing as a dipole antenna, recently, we gave up the idea since it is expensive. Instead, we use part of the PCB ground plane(separate from RF ground plane by a 22nH inductor) as monopole antenna as I mentioned earlier, thus we save the dipole antenna cost. Doing so we compromise our antenna performance, but it is still with spec.
 

Regarding more information of the antenna, in old times we were using some metal part attached to the product housing as a dipole antenna, recently, we gave up the idea since it is expensive. Instead, we use part of the PCB ground plane(separate from RF ground plane by a 22nH inductor) as monopole antenna as I mentioned earlier, thus we save the dipole antenna cost. Doing so we compromise our antenna performance, but it is still with spec.

My guess would be [in the absence of specifics] that the antenna ground plane section plus inductor is acting as a half-wavelength (free-sapce) resonator, where most of the phase is being provided by the inductor. This arrangement isn't very efficient (as you indicated), but it will radiate slightly. As to why it's connected the RF ground and not a signal line is ambiguous at this point, but is most likely the most practical solution given the layout.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top