Maker Pro
Maker Pro

Grounded co-planar waveguide design

Hi all,

I recently took a sneak peek in the RF world of black magic and pixie dust..
What I made is a device (two devices I suppose..) for my old man. The sensor part can monitor temperature, humidity and water level, in his greenhouse, which is then transmitted to another device that prints the information out on an LCD display.

The thing works and im pretty happy with it. But im not so sure about this waveguide I made.

Not knowing much RF stuff, I just went with off the shelf components. What im using are two SX1272 LoRa transciever modules from Semtech, link down below. Working at 868 MHz, which is an open frequency here in Europe. The antenna output is going through my waveguide into the SMA connector, into the coax cable and out into my external antenna.

http://www.farnell.com/datasheets/2162976.pdf
http://dk.farnell.com/te-connectivi...ial-sma-straight-jack-50ohm/dp/1248990?st=sma
http://dk.farnell.com/lprs/rg174-150mm/cable-rg174-150mm/dp/2096227?st=sma
http://dk.farnell.com/lprs/ant-ss900/antenna-stub-straight-868-915mhz/dp/2096218?ost=2096218&ddkey=http:da-DK/Element14_Denmark/search

So, 50 ohm output, into 50 ohm transmission line, into 50 ohm connector, into 50 ohm cable, into 50 ohm antenna. That's how it should be, as far as I read..

Not much to this schematic, but adding it so it's out of the way. A1 is the SMA connector ofcourse..


PCB layout. Ground plane on both top and bottom.


Topside of etched PCB


Think I traced the waveguide on the wrong side. I soldered that SMA connector on the top, seems it's supposed to connect on the bottom instead, woooops...


Now, anyway.

I got the trace width and ground plane distance by playing around with the values on this site: http://chemandy.com/calculators/coplanar-waveguide-with-ground-calculator.htm

Here was my results. Not 50 Ohm exactly, but pretty close I think, I had to be able to actually etch it too..
The PCB was some stuff we had at work, 1.5mm thickness, it said 4.1 ER on the package. So, assumed that was it.


The width of the gap it seems was this. I made that 0.25mm


So far so good I think? Feel free to murder me if it's wrong at this point.

But to my actual question. What really bugs me is, nowhere I look, does it seem to specify a length of the trace, or care about the frequency been used. Am I missing something? Or does this simply not matter.. I would love to know if this is correct, I dont have the equipment to actually verify the signal strength and losses.

Hope you guys can sort me out as always. :cool:

Best Regards
 
Nope you are not missing something. You will get some losses due to frequency but for your application it should be fine. For long traces you will get resistive losses and dielectric losses but they will be small for you.
Thanks
Adam
 
Nope you are not missing something.
Well, great, glad to hear it :)
You will get some losses due to frequency but for your application it should be fine. For long traces you will get resistive losses and dielectric losses but they will be small for you.
Thanks
Adam

So. Keep traces as short as possible, got it.

When you say I will get losses due to frequency, is there a way I could calculate those? Now, I dont know if something like that gets too hairy, and requires a grey beard and white coat. But I'd love some pointers.
 
It's not black magic what ever anyone tells you. That's just them telling you that they think they are better than you. That's bollocks. It may involve quite a bit of Physics and math which I don't have to hand. What I will do is try to get some stuff together to show you.
Thanks
Adam
 
It's not black magic what ever anyone tells you. That's just them telling you that they think they are better than you. That's bollocks.
You are probably right. Im only a student. What usually happens when we venture into RF territory. Is that teachers and other people in the know always go, "Dont do it, just dont". It is kind of discouraging, but I dont give up easily.
It may involve quite a bit of Physics and math which I don't have to hand. What I will do is try to get some stuff together to show you.
Thanks
Adam

Wow. I would really appreciate that, thanks alot Adam!
 
The reason why the teachers say that is because they don't know. They don't want you asking questions they can't answer :)

I can totally see the look on their faces. if I was to even insinuate, that they didn't have the answers. :D

Actually. Coming to think of it. It would probably be quite amusing. Probably best not to try something like that before graduation however.. :rolleyes:
 
I have found some interesting information about Transmission lines.
http://www.feynmanlectures.caltech.edu/II_24.html

http://www.standard-wire.com/coax_cable_theory_and_application.html

And some more information about coax cables which you might find interesting. I know you are using a PCB WG and the signals propagate differently versus a transmission line but the information about characteristic impedance and the lack of a mention of frequency might give you the answers you have been looking for.

See if you can work out why frequency is not a huge factor in determining CI.

Hint: Although CI has the units Ohms it is not considered a resistor but more of a ratio.

Thanks
Adam
 
Top