SparkFun Forums

[shraken]

I'm using a GPS chip antenna (model A10137) for a GPS tracking application.

I'm using a Trimble Copernicus GPS module the states I should expect a cold start lock in around 45 seconds. I've used this module with an external SMA passive antenna before and had lock times as stated.

I'm looking to switch to the a chip antenna for size reasons so I've adopted the Antenova A10137 in my project.
http://www.antenova-m2m.com/documents/d ... 187d51cfa6

I attached a picture of my schematic and layout for the part here:
https://imgur.com/a/kIFEg

With the A10137 the GPS cold lock takes a very long time (around 5 to 10 minutes) and the chip often loses lock intermittently. When the chip loses lock it will drop to 1 or 2 satellites in view indicating to me that its a SNR problem.

I'm not using any matching network although the PCB allows for a PI (L-C-L) type network. I've simply shorted the C component so that the antenna feed line is connected directly to the RFIN of the GPS module.

How can I improve the performance of this project? I'm new to designing matching networks and simply thought that if my transmission line is sized to 50-ohm that all would work. I know the aperture of the antenna is small but 10 minutes with a dropping lock is not good enough for my application.

Thanks for your assistance,
Nicholas

[falingtrea]

Did you follow the design recommendations in this document?
http://www.antenova-m2m.com/documents/d ... 0a69a200c9
Chip antennas are very sensitive to placement and routing.

[shraken]

Hi Falingtrea,

Appreciate the response. I have reviewed this document in the process of building but just read it again.

I think I'm having problems because of a poorly design co-planar waveguide (CPW). I was tight on space in this design as the board is small (50 mm x 50 mm) and kind of thought I could get away with a poorly designed CPW.

I've decided to fab a couple different PCB designs for the Antenova A10137 chip antenna and the Johanssen version that Sparkfun sells. I was thinking of making them with a proper CPW and connecting to a female SMA. I plan to take this test board to a VNA and analyze the frequency response for a 1.45 Ghz stimulus. I'll take the smith chart and design the matching network based on the short circuit response and retest on the VNA.

I think that's the best I can do because I cannot afford to have Antenova do a matching network design. I'll keep this thread updated with my progress for those who are interested.

[shraken]

Hi, I laid out a test GPS board with the patch antenna and SMD connector. I have 3 designs on it: one to test the Antenova Patch antenna, one for the Johanson sparkfun GPS antenna, and another for the Trimble module.

I think I got the CPW forum right. I tried to use a 50 mil trace width with 10 mil isolation which is the result I found from AppCAD. If anyone could take a peak at it and tell me if they see any thing wrong. There is a top and bottom ground plane tied together with through VIA's.

The matching network is a capacitor in series with two inductor shunts. I also included a DC blocking capacitor.

Appreciate the help, I'm going to try and send away to oshpark and make a measurement on a VNA In the next few weeks. I'm going to try and measure at 1.575 Ghz for L1.



Oh it's one main board but I'll break it into 3 boards, 2 to test with the VNA and one to try the improved matching network with the actual Trimble module.

[darkzim]

Shraken,

I'm trying to prototype a small GPS tracker using that antenna as well, except with a Linx FM GPS unit and an EFM32 ARM microcontroller. I found this thread when I was searching for examples of how to layout the PCB using a chip antenna (I have zero RF design experience).I'll be interested in hearing your results, especially with the VNA... I was just about to give up and go with one of the integrated solutions offered by Antenova (the giganova series).

One thing though -- I plugged your 50mil width/10mil spacing (1.27mm/0.254mm) into the following calculator, with a 4.6 dialectric constant and 1.6mm thickness from osh park and got an impedence of 52.1 ohms. http://chemandy.com/calculators/coplana ... ulator.htm

[shraken]

Hi Darkzim,

I think the design should be independent of the GPS module in question. Although, the modules may differ in their LNA topology and gain control but the RFIN should be 50 ohm input.

I don't have much RF experience either which is a real hindrance. I took an Antenna design course in graduate studies where I got to work with CST to make Dipole Antennas and other varieties but it's a bit more difficult when you are working with real copper.

The test board I mentioned in my first post is poorly matched. I played around with it this week on the porch. It takes a solid 2 hours or so to get lock indicating to me that the SNR is very poor require long acquisition time. I assume this is because the amount of power reaching RFIN is low.

I'll be sure to post a link to my blog and my results in this thread when I get the board back and have a chance to use the VNA.