I am designing a PCB for an open source project using a CC1310. My design uses an SMA connected whip antenna and I plan to use the 915 MHz ISM band. I don't have much RF experience, so my plan was to layout the balun and filter just like on the CC1310 eval kit. TI recommends following the eval kit layout as closely as possible, however that's going to be more complicated, expensive, and time-consuming than I had anticipated, so I'm wondering how much wiggle room I have with the layout. For reference, I have attached a picture below that shows my layout on the left and TI's on the right. They aren't quite at the same scale, but it close enough to see what I'm dealing with. I'm hoping some of you can answer the following questions:

1) The stack-up used on the eval kit has 0.175 mm between the top layer and ground plane. Using that stack-up is going to increase costs, but I believe I can increase the thickness between the top and ground layer to around 0.21 mm (which JLCPCB offers as a standard option). Will that increase in thickness have much impact as long as the traces are still 50 ohms?

2) I selected the same component part numbers as used on the eval kit, but the footprints are a little different (e.g. on the eval kit the inductor footprints have a 0.1 mm larger gap, and the CC1310 pads are 0.1 mm wider, the eval kit uses rounded corners rather than square, etc.). Are footprint differences big enough to make a noticeable impact?

3) As long as I keep the balun and filter symmetrical and aligned, I assume varying the distance from the balun to the CC1310 pins, or the distance between components in the filter and balun within 0.25 mm won't affect things too much, correct?

4) I assume less than 0.25 mm differences in the top layer ground pour edges and fencing vias won't matter too much as long as the vias are much less than 1/10 wavelength apart from each other, correct?

5) I can shorten the distance from the end of the filter to the antenna connector as long as the trace is 50 Ohms, correct?

6) I can reduce costs by using some components from different manufacturers. Is there much risk in swapping out parts as long as the replacements have the same specs, size, performance profile, etc.?

 

If you change your layer stackup you always have to change your trace width to get a defined impedance, e. g. 50 Ohms. Changing trace length may be non critical but the trace can also be used as an impedance tanstormation if the input impedance to the line is not equal to the characteristic impedance of the line. So you have to analyze deeply the evaluation board's layout to check whether it is possible or not. Evetything else is tinketing or try and error.

 

Thanks for the response. I should have stated that I realize I need to change trace widths to keep 50 ohm impedance if I change the distance to the ground layer.

By "tinkering and trial and error" I assume you mean that if I'm going to change anything, I should probably be prepared to test and retune the whole circuit, correct?

 

With "tinkering" I meant if you don't understand in detail how the reference circuit plus layout works with respect to impedance matching, any change may have unexpected results, but you don't know why. So you change something, hoping to get it better but finally you won't get de desired performance.