SparkFun Forums 

Where electronics enthusiasts find answers.

All things pertaining to wireless and RF links
By reklipz
#123361
[I apologize for the wordy post. I just want to ensure all of the details are here and clearly explained so as to garner the most input.]

Hopefully there's an RF guru lurking out there somewhere who is able to answer my questions.

I'm trying to design a PCB dipole strip antenna for use in the 915 MHz ISM band. As the half-wave length at 915 MHz is reasonably short, I'm going to attempt printing/etching the dipole on the PCB. Also, because my front-end is a differential port (100 ohms), using a dipole (rather than a monopole) allows me to eliminate the balun from the design.

What I'm having trouble with is designing the feed lines for the dipole. In a monopole design, a grounding plane is necessary, and the monopole strip is located as close as possible to the ground plane. Ideally, the monopole strip is perpendicular to the ground plane, however, this makes the design bulky and more costly to manufacture, so most designs use a coplanar layout. Having the monopole strip right next to the ground plane means that the feed line can be a simple microstrip (feedline on one side of pcb, and ground pour behind it). The calculations for the impedance of the microstrip are simple.

However, with a dipole, there is ideally no ground plane present, and the dipole should be placed as far as reasonably possible from the components and ground pours. This means a microstrip feed can't be used (at least as conventionally known). One paper I saw used a microstrip feed for part as far as possible (or needed, to bring the feed to the edge of the ground pour), and then simply kept the same width for the rest of the feed line until reaching the dipole strips. I initially layed out my pcb in this manner, with the two feed lines side by side. However, due to how I laid out my board, a microstrip feed from the IC to the pour edge is not necessary.

Then, it was suggested that perhaps by placing the feedlines on opposite sides of the board, because they are anti-phase, they will still each act as a microstrip. This is how I have currently laid out my design, which can be seen here:

gEDA pcb
Gerber
top + bottom PNG
top PNG
bottom PNG

I have placed some passive pads for impedance matching, as I don't think it's likely the dipole will be very near 100 ohms, as required by the differential port.

I'm not certain that having the feedlines opposing eachother like this means they will act as microstrips; mostly because a microstrip has a ground plane that is wider than the strip.

How do I determine the width / spacing of the feedlines when laid out in this manner? Perhaps this layout isn't a good idea, and some other feedline layout should be considered. Please, share your experiences!!

And finally: There is a lot of "wasted" PCB with a printed dipole design. I know there is a folded dipole layout, but the size isn't what worries me, it's cost. I would be just as happy using a piece of solid copper wire in place of the PCB dipole strip. I think this might even be more ideal, as the wire is thicker and uniform; am I correct in this understanding? The only problem is, I don't know how to determine the impedance of the wire dipole (And then again, subsequently determine the proper feedline for it, although the feeding of this and the pcb strip should be the same).

I appreciate your thoughts!

--
Nate