Not so much a question as a comment on building your own receiver break out board.

+4 votes

Earlier this year I decided to design and build a breakout board for the Neo format u-blox modules. I had looked at many of the offerings from e-bay and other sources. I have many of these boards, and some are better than others, some use "fake/clone" modules, some are downright junk.

I had no help from u-blox, and I have no association with them except as a customer/user. I am a retired EE a ham and have a passing knowledge of RF pcb layout. It took me about a week to do the actual layout (many many revs). I had the boards built in China, (cost about $US2.00/board) and bought all the parts from DigiKey, including the modules. I assembled the boards myself. It took me about an hour to build 3-4 boards. I test their power supply before soldering on the u-blox module.

I have built and tested boards with the Neo-7N, Neo-M8N, Neo-M8T, Neo-M8U. The boards support the Neo-M8P as well.

The attached pictures show some of the boards and the antenna, the layout, a screenshot of u-Center. You might notice that the signal levels are a bit low. That's because the antenna is a GPS only antenna located inside about 5 feet below a metal roof that shields 1/2 the sky. Most of the signals are from multipath. With a good antenna (Tallysman 3710), outside, the signal levels are very, very strong.

To say that I am pleased with the performance of the Neo series parts is an understatement. They are easy to work with, the boards worked the first time, and yes, there is one red wire that was a feature that I decided to add after I sent the boards off for fab.

by KenMcGuire asked Jun 5, 2017
0 votes
I'm sure that the folks at u-blox would appreciate your comments.
As a former FAE in the GNSS industry and now retired, my most painful memories are of small-shop customers who designed their RF like a bulb and battery circuit, who almost come to tears when I told them they would have to redesign; their response was that their startup would then go under.
Thanks for being so positive.  
Maybe we can collaborate on an "RF for dummies"  practical guide book ... fuggedabout Friis and Maxwell.
I once was an RF dummy, lived thru it, and now am almost competent (---another 50 years as EE should do it.)
Thanks for your story.
by grampy answered Jun 6, 2017
Let's see if I remember this correctly...
Ground everything.
If you can't ground it, shield it.
If you can't shield it, bypass it.
If you can't bypass it, filter it.
If you can't filter it, forget it.
If you can't forget it, you shouldn't be doing RF design.

50 more years, and I might be able to write a C/C++ program from scratch that actually does something.
Your RF "waveguide" goes under the module PCB with the steel cover on top. Is it really a good idea ?
The steel cover of the Neo module isn't as much concern as the ground plane of the bottom of the module which is very much closer to the antenna trace. And yep that isn't as good a layout as I might have liked but I looked at a lot of implementations and the one that was generally regarded as a good performer was the CSG-M8T. Its' antenna trace followed the same path as mine.
The trace is also on the same side as the SMA connector and passes close to its' grounded body, not optimum either.
With all that said, what I didn't want is a via carrying the antenna signal. If I were doing this for a production design, I would have tried several different routings of the antenna signal and evaluated each with a network analyzer. As it is, I know that the impedance is not 50 ohms, I know that the circuit board material does not have great RF characteristics (cheap Chinese FR4 is not Duroid), I know that there is stray capacitance between the module ground plane and the signal. I know that there is stray capacitance between the body of the SMA and the antenna signal.
This is a case where I chose to forget it and see how it performed. It was a conscious decision.
For helge ---
With a good active antenna and reasonably low loss in its cable, Friis comes to the rescue and diminshes the effect of slight impedance bumps.
website banner