5
\$\begingroup\$

I have a fair bit of industry experience with Electronics Design, as I'm a design engineer for a company in Australia. However, over the last 5 years I've only really specialised in power supplies, control systems, analogue/digital signal processing, and wired communication (RS-232, RS-485, and CAN).

Before I do any professional work with radio, I wanted to do a little personal project at home. Something like two PCBs, one controller and one relay board which turns on a 240VAC relay or something, so as to remote-control a light or other home appliance - with the communication between them / to them being via bluetooth or via some generic bluetooth application on a phone. That scope is for me to worry about.

As someone who has a lot of experience in board design except radio (Wi-Fi, Bluetooth, and RF in general), I had some brief questions for anyone who is experienced in board design for those applications:

  • I have decided to go with the Nordic Semiconductor NRF52810. This was recommended by a colleague, as the Nordic support community is great (a lot of noob questions, with a lot of FAE answers containing example layouts/parts/code for specific cases). Any thoughts on using this as a first timer? As much as I shouldn't be picky, it would be nice to have an integrated transceiver solution such as this so as to reduce the possibility of error.
  • Assuming my said Micro layout is perfect, I had some sub questions:

    • I planned to use a pi-layout (C-L-C) for the tuning circuit, with some rough starting values. I know these need to be empirically tuned, but do they matter a heap if I only plan to want 2-5 meters of range?
    • For the antenna, I'm planning on using a SMT one (not a board-copper-track one - keep it simple). I've read that the choice doesn't really matter (so long as it can handle the maximum power you plan to transmit). The main thing is that it is designed for 2.4GHz, and has 50 \$ \Omega \$ impedance. I have chosen the Johanson Tech 2450AT07A0100T. Does this seems okay?
    • Finally, the trace from the Micro -> Tuning Circuit -> Antenna. As long as this is 50\$\Omega\$ controlled impedance, has via stitching to ground along the edges, and is a reasonably short length, is there anything else I need to worry about?

    • In terms of firmware, that's for me to worry about. The main thing is getting the hardware right. So my last question would be, is there anything I seemed to have skimmed over/missed? Total noob in this field so just be honest.

Thanks so much. If you have any good resources on PCB Bluetooth design from scratch please feel free to comment links. I have read a lot of resources on different parts (chip choice, antenna tuning, antenna placement and routing), but haven't really found a good one on "the whole picture".

Cheers.

\$\endgroup\$
  • \$\begingroup\$ Try a module like BL-652 first. It has the exact chip you want and is very easy to design with. \$\endgroup\$ – filo Jul 31 '18 at 12:45
  • 1
    \$\begingroup\$ The main purpose of experimenting is to learn new skills. How to measure Return Loss on PCB's, design CLC filter, microstrip, controlled impedance tolerances, antenna gain/loss factors from RL, aperture size and orientation, ground plane proximity. How to use a VNA, SA, Return loss bridge or a splitter or directional coupler, design RF antenna pads, how to choose best DC filter components, effects on SRF, PCB dielectric differences, RF design tools. Rather than just a cookbook recipe for a design. For example, I might choose both a half-wave rigid wire and the Johanson chip to compare. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Aug 7 '18 at 17:45
  • \$\begingroup\$ Yeah and I'm keen to learn all this as I go - hands on is the best way. Just want to make sure the PCB actually works (even if it's poor power output), otherwise there's no point in even sending it off \$\endgroup\$ – DSWG Aug 7 '18 at 23:06
1
\$\begingroup\$

Make your PI matching circuit heavy on the Cs and light on the L, so any PCB parasitic does not cause major mistuning.

And go for low-Q, so you don't need precision or adjustable components.

What is low-Q? Given the circuit functions in a 50 ohm network, having Z(L) = Z(C) = 50 ohms would definitely be low-Q. Having Z(L) = Z(C)= 5,000 ohms would be high-Q. I'd pick a Q of 10, thus your 3dB points will be +_5% of carrier. This may require careful selection/measurement of values. Thus Q=3 is easier for you.

Regarding stitching, for serious isolation use 2 rows of stitching to restrain the circulating currents right in those vias.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for the reply. This is my stitching as is, with tuning circuit halfway between the Micro and the Antenna: i.imgur.com/UUmJJqs.png. When you say Low-Q, what do you mean sorry? \$\endgroup\$ – DSWG Aug 1 '18 at 5:13
  • \$\begingroup\$ To localize stitch-via-long-range-circulating currents, add another row of vias on each side. Check with the experts about this. I saw these dual-rows used on 900MHz video-TX & RX circuits, made by Recoton, decades ago. \$\endgroup\$ – analogsystemsrf Aug 1 '18 at 5:25
  • \$\begingroup\$ A thinner board makes it easier to increase C in pF/mm and lower Q but use a low loss tangent substrate \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Aug 11 '18 at 20:34

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.