This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CC1350: Balun and chip antenna recommendations for CC1350

Part Number: CC1350
Other Parts Discussed in Thread: CC2650, CC1310

I am working on a project for the Hackaday Prize 2017, you can see it here

Right now i am using CC2650, but for the final prototype i will use CC1350.

I have gone through almost all the questions and answers regarding antenna and balun for CC1350 and CC1310. However they are scattered and sometimes inconclusive.

Since right now CC1350 modules are not available, I am forced to make my own design from scratch. But because of size restrictions, I would want to use external chip antennas. Also I would like to use baluns instead of discrete passives.

Also I am under a deadline for a proof of concept so I want to make “single ended design” like in CC1350 Sensor Tag Kit for simplicity.

Right now what I have to do is (according to my understanding) the following –

1. Select any 2.4GHz antenna that has 50 ohm impedance.

2. Select any SubGHz antenna that has 50 ohm impedance.

3. Get a balun for CC1350…..unfortunately its not available. So I do the following –

According to some suggestions for questions, one was to use the CC1310 SubGHz balun for SubGHz antenna for CC1350 and the CC2650 balun for 2.4GHz antenna for CC1350.

Q1) Is it a good idea to do this or are there dedicated baluns for CC1350 (preferably, currently available for purchase)?

According to TI App Note for IPC for CC1310 http://www.ti.com/lit/an/swra524a/swra524a.pdf

The SubGHz balun for CC1310 from Johanson Tech is https://www.johansontechnology.com/datasheets/chipset-specific/0850BM14E0016.pdf

The 2.4GHz CC2650 balun from Johanson Tech is https://www.johansontechnology.com/datasheets/baluns-matched/2450BM14G0011.pdf

Are there any other alternatives for the baluns or dedicated baluns for CC1350 that are currently available for purchase?

Q2) Are there tested(recommended) antennas for CC1350 too because since the modules will come in “end of 2017” , they will be using chip antennas as far as I can tell, and I am guessing that TI engineers must have already tested some antennas by now.

  • Antenna: For a module we are looking into a different solution than a chip antenna but this work hasn't concluded yet and the solution we are looking at is not a off the self solution. Note that if you are using a chip antenna this will give poor efficiency due to the size of the board. (or to be more precise, the size of the ground plane)

    Baluns: We are working with a balun for CC1350 with both 868 MHz and 2.4 GHz in but that will still take some time until finalized.

    A balun has a differential input, in your post you write that you want to use single ended design. If you use a balun you also have to use a RF switch. Could you clarify?
  • Oh, i realize my mistake now..
    So now i have two options ....
    1) Get a RF switch, get the two baluns i mentioned above and use the PCB antenna.
    2) Don't use a RF switch or baluns, rather get two IPCs and use PCB antenna.

    Q1) Is this correct or am i again making mistakes with IPC and balun?
    Q2) BTW if i decide to go with option 1 , then is it a good idea to use those two baluns ?
    Q3) I would probably want to use option 2. Can something be done about it?
    Q4) I really need a solution by July - August as the contest ends in October, can you please give me some advice ?
  • A balun is a passive circuit that convert two differential signals with 180 degrees phase difference into one signal. The typical way to do is by using discrete passive components (capacitors and inductors). After the balun a filter has to be used to attenuate the harmonics. For most of our radios it's also possible to get a IPC (integrated passive component) for one or more frequency bands. The IPC integrate the balun and filter into one device and makes it possible to save space.

    so in your case:
    - Use CC1350 single ended and implement the required matching/ filter with discrete components
    or
    - place a RF switch ( one differential input, two differential outputs) close to the RF_P/N pins as on the CC1350 LP and place a 868 MHz and a 2.4 GHz IPC on the switch output. This is a solution we haven't tested and I see a couple of issues. The 868 MHz IPC is designed for external bias but since the bias is a DC voltage it's not possible to do after the switch. You can try to leave the RXTX pin floating, not sure how that will influence the RF performance. The switch also introduce a delay/ phaseshift that could give potential non optimal reflection of the 5. harmonics giving some VCO pulling. On the latest version of the CC1350 LP a couple of caps has been added to counter the phaseshift.

    From this using single ended could be the solution that has less risk. If you are space constrained look into the use of 0201 components. Do a mock layout with the different options and see what would be possible on the available area.
  • Thanks for clearing that out.

    Since i don't own any measurement equipment, the best solution right now is to blindly follow CC1350 STK schematics for the discrete components ?

  • Yes. And if you don't have any means to tune antennas I would look into chip antennas but you have to check how dependent they are for the size of the ground plane.
  • I am actually thinking of not using chip antennas after the discussion. I will make PCB antennas. One more final question--
    Do i need to bother about the copper thickness and board thickness for PCB antenna, and trace width for the traces that connect the RF pins to discrete components to the antenna finally ? (Because for 50 ohm trace impedance these things matter but nowhere in TI's notes i see the mention of 50 ohm impedance tracks. Just to let you know that the RF pins, discrete components, and the PCB antenna would be as close together as possible, do i still need to worry about those parameters?)
  • For antennas: To get optimal performance you should get access to a VNA or get www.wimo.com/minivna-network-analyser_e.html (or similar)

    If a trace is longer than about 1/10 wavelength of the frequency you are operating on the trace will behave as a transmission line and you have to calculate to get a 50 ohm trace.

    For a helix antenna the board thickness will come into play since the vias is part of the antenna length. If you get access to a VNA you can make the antenna a bit longer than you think is should be and use a prototype and cut the antenna until you get a good match. Note that you can do this on a PCB without components.