2
\$\begingroup\$

I am designing a device with two stacked parallel PCBs of which the top one serves as a front panel (with touch pads), while the bottom PCB holds all the components. I need to add a Bluetooth LE module to the bottom PCB and since the top PCB is mostly covered in copper, and the enclosure will likely be metal, I need to find a way for Bluetooth to radiate.

I see two options here:
1. Choose a Bluetooth module with U.FL connector. Mount it together with all the components on the bottom side of bottom PCB. Mount another U.FL connector on the bottom side of front panel. From there I could draw a PCB antenna on the front panel either on the top side (and make a via to U.FL) or on the bottom side.
2. Choose a Bluetooth module with chip/PCB antenna and mount it on the top side of bottom PCB. Make a opening in the copper layers of the front panel large enough for bluetooth to work.

1st approach - U.FL cable 2nd approach - opening in copper

The second approach will be much cheaper than the first (U.FL connector and cable add up to $7.50 on Mouser) but I'm worried it will make my bluetooth radiation pattern too directional, and I can't imagine what is the reasonable size of the opening in copper layer? I can't afford to spend much front panel area on that feature. OTOH, if the opening area required would be about the same as area required for a PCB antenna + keepout area (approach 1) then I can do it.

\$\endgroup\$
1
\$\begingroup\$

I've dealt with RF for about 10-12 years now, and my best recommendation is to spring for the extra $$ and use the antenna mounted to the top PCB. Putting the radiating element under another layer of copper, silicone, and other materials will attenuate the signal by (very exact answer here) a lot. On the other hand, if you plan on using the Bluetooth interface in very close proximity (less than a meter), it may not matter much. You could always build a prototype of each one and use a spectrum analyzer to measure the performance before putting your design into production. Also, if the whole thing is going to be encased in metal, the BT antenna mounted to the top PCB will probably be severely attenuated as well. You may want to consider a material with better RF transparency for the case. Plastic, maybe?

\$\endgroup\$
  • \$\begingroup\$ Thanks for the answer. The device itself is a module, which is going to be put into the the case together with other modules by customer, so I can't control the enclosure material (it would be either wood, metal, plastic, even cardboard!). I just know that there will be other modules co-planar to this one. \$\endgroup\$ – Volodymyr Smotesko Jun 13 '17 at 14:17
  • \$\begingroup\$ Also, consider the wavelength of a signal at 2.4GHz is about 12.5mm, so if you want to go the route of having an opening and using the chip antenna, 12.5mm is probably the smallest I would recommend. \$\endgroup\$ – GroundRat Jun 13 '17 at 14:20
  • \$\begingroup\$ that actually sounds promising! It's certainly acceptable for me to even make it 12.5x25mm, and yes, this is going to be used at distance about 1 to 1.5m, so I may try it. \$\endgroup\$ – Volodymyr Smotesko Jun 13 '17 at 14:24
  • 1
    \$\begingroup\$ To radiate in numerous directions, you need numerous slits. With enough slits (5 or 10), you'll get enough overlapping patterns the nulls tend to fill in, but multi-pathing and precise positioning will be a bother. You'll need slits in both X and Y directions. Speaking as non-antenna guy. \$\endgroup\$ – analogsystemsrf Jun 13 '17 at 14:26
  • 1
    \$\begingroup\$ I said mm? nope. cm! I meant centimeters! lolz \$\endgroup\$ – GroundRat Jun 13 '17 at 14:31
2
\$\begingroup\$

I would consider mounting the A20737A BT module on the bottom side of your top PCB. Follow the datasheet instructions for the clear of copper area on all layers of the top PCB. Then test the antenna pattern for your application.

You will need to interconnect the data lines of the A20737A to the logic on the lower PCB but this is a much lower cost connector than a $7.50 ufl connector plus antenna and you already have switch contacts that require board to board interconnectivity anyway.

The ideal design would be a single board construction with all of the components in the back side of the switch board.

\$\endgroup\$
  • \$\begingroup\$ I was considering this option, and what bothered me is this will introduce the first "active" component on the top PCB. At the moment the top PCB only has touch sensor copper pads and a SMT connector, so it's totally passive. This allows it to be assembled with more loose requirements to the soldering process / temperature profile. However, this is not really a benefit, so I may as well put some active components there. \$\endgroup\$ – Volodymyr Smotesko Jun 14 '17 at 7:44
  • \$\begingroup\$ Unfortunately, I can't make a "ideal design" with just one PCB because I've got some through-hole components, which are not acceptable on the front panel PCB. \$\endgroup\$ – Volodymyr Smotesko Jun 14 '17 at 7:47
  • \$\begingroup\$ Depending on the component, some through hole devices can be reformed to be gull wing mounted. But I understand your challenges. All the best with the project. \$\endgroup\$ – Glenn W9IQ Jun 14 '17 at 19:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.