AI is going to help me be a true success! While a large part of the population is afraid that AI is going to take their job, I remember a statement from Tony Robbins, the Self-Help, Success-Guy, “you are not going to lose your job to AI, you are going to lose your job to someone that knows how to use AI.”
Here’s my prompt (what I ask AI) today. “I’ve chosen components and calculated resonances, etc. for a 20 meter band pass filter using SMD components, air-core inductors (Coilcraft), and 0805 high-Q capacitors with female SMA connectors on the ends. Can you talk to me about things to consider as I design my PCB board to be really compact and use RFI shielding clips for an aluminum enclosure? Anything that I should focus on specifically?”
And, here’s the response. Remember, it’s mine to use, author, etc.
You’re in a really fun part of the project now — the physics starts to show up in the copper. Let’s talk specifically about what matters for a compact 20 m BPF PCB inside a shielded aluminum can.
1. Board stackup and grounding
This is the most important foundation.
- Solid ground plane:
- Use a continuous ground plane on the layer directly under the filter components.
- Don’t route signals in that ground layer under the filter if you can avoid it.
- Short return paths:
- Every cap/inductor junction should have a very short path to ground (via right next to the pad).
- For shunt capacitors, put the via as close as manufacturable (0.2–0.5 mm away).
- Many ground vias:
- Stitch ground around the filter area (via fence along the edges and around the SMA footprints).
- Aim for via spacing on the order of a few mm in the RF path region.
The idea: the filter lives in a “RF cavity” defined by the ground plane and the metal lid; your PCB layout is the boundary condition.
2. Signal routing and physical layout
At 14 MHz, you’re not in microwave territory, but layout still matters a lot for coupling and stability.
- Keep the filter linear and symmetrical:
- Lay out the network in a straight line: SMA → series L → shunt C → series L → shunt C … → SMA.
- Avoid dog‑legs and wildly asymmetrical element placement.
- Keep traces short but sane:
- Use short, reasonably wide traces between components (you’re roughly in 50 Ω land, but not truly transmission‑line dominated at 14 MHz).
- Don’t obsess over exact 50 Ω microstrip, but avoid long skinny runs.
- Minimize parallelism:
- Don’t run input and output traces closely parallel — route them in a way that minimizes direct capacitive coupling (no “looping back” alongside each other).
- Inductor orientation:
- Rotate adjacent air‑core inductors 90° to each other wherever possible.
- Maximize physical distance between inductors that aren’t meant to couple; even a few extra mm helps.
You’re trying to preserve the intended L/C coupling and avoid creating a mystery extra coupling path that changes your filter shape.
3. SMA connector integration
Female SMAs on a compact board inside a shield can get tricky mechanically.
- Grounding the connectors:
- Ensure the SMA’s ground tabs or edge‑mount ground are well tied into the ground plane with multiple vias right at the connector.
- Don’t rely on just the mechanical tab; stitch it with vias all around.
- Mechanical registration with enclosure:
- Think in 3D: where the SMAs sit relative to the enclosure wall or hole.
- Make sure there’s enough clearance so the cable can mate without pushing the PCB or stressing solder joints.
- Shield continuity at the connector:
- Ideally, the SMA body and the enclosure are at the same RF ground potential.
- If the SMA isn’t panel‑mounted but just pokes through a hole, you may want extra ground fingers or copper pads near that opening.
4. RFI shielding clips and aluminum enclosure
With PCB‑mount shield clips + aluminum lid/box, some specifics matter.
- Placing the clips:
- Put clips on grounded copper regions, not over signal traces.
- Use ground via stitching under/near the clip pads so each clip connection has a low‑impedance path to the main ground plane.
- Shield “wall” on the PCB:
- Run a ground copper ring around the filter area inside where the shield will sit.
- Stitch that ring with vias to the ground plane (this becomes your RF “gasket” under the lid).
- Gap control:
- Try to avoid large gaps between the lid and the PCB copper ring. Even at HF it’s not critical, but better contact = less chance of weird coupling.
- Aluminum is not solderable:
- Your electrical connection is via the clips to a copper pad, not the lid itself.
- Ensure that the lid or box is mechanically pressed against those clips or edges so there’s real metal‑to‑metal contact.
If your board sits inside a larger aluminum box, you can use both the PCB shield can and the external box as nested shielding — very effective.
5. Compactness vs coupling
When you shrink everything, coupling wants to creep in; this is where most “mystery behavior” comes from.
- Watch in/out isolation:
- Keep input and output physically separated by the filter network and some ground copper.
- Don’t let the SMA inputs “see” each other over the top of the filter via stray capacitance.
- Use ground “fences” between sections:
- You can place thin grounded copper strips between sections of the filter (with vias), acting as mini bulkheads.
- Guard traces:
- Put grounded copper between sensitive nodes if you’re worried about coupling — especially between input and output ends.
Compact is good, but don’t make it so compact that inductors and traces are stacked over each other.
6. Practical tuning/debugging considerations
Even with careful calculation, the real PCB will need a little love.
- Leave some tuning flexibility:
- Use parallel pads so you can add or change capacitors (e.g., room for small parallel trimmers or extra 0805s).
- Make it easy to snip out an inductor turn or stretch/compress them if you ever go custom‑wound later.
- Test points:
- At minimum, expose input and output SMA and make sure you can easily connect a VNA or sig gen + power meter without contortions.
- Ground test pad:
- Provide a good ground pad near the filter for probing, not just the SMA body.
This lets you actually see how your layout choices affected the filter response — and refine future versions.
7. Specific “watch this” items for your setup
Given what you described (SMD, Coilcraft air‑cores, 0805 high‑Q, shield clips, aluminum box):
- Air‑core inductors:
- Keep them slightly away from the enclosure walls and lid; very close metal will change inductance and Q.
- High‑Q capacitors:
- Great choice. Just keep their leads short and put ground vias right beside shunt cap grounds.
- Aluminum box:
- Plan screw locations or standoffs so the PCB ground has several solid mechanical + electrical bonds to the box.
- Avoid routing high‑impedance nodes right under screw heads or standoffs.
If you’d like, describe (or sketch verbally) your basic layout: SMA on each end, number of poles, and board size target. I can then walk through a more concrete “place this here, avoid this there” pass tailored to your exact filter.
It’s me again! That’s quite some education there. Go learn AI.
My About page provides the background of my project, the Freedom7 HF Transceiver.
If this story resonates, comments are welcome. You can also reach me at david [at] kr4bad-dot-communications. no com
And if you believe understanding matters more than black boxes, you can subscribe to my WordPress https://kr4bad.com/?subscribe=1.
73 KR4BAD David
Leave a Reply