4
\$\begingroup\$

I have been tasked with designing a low power sensor system and the customer has decided location would be a great option. I have done GPS before but only with complete modules (Ant. + IC). To keep costs down I was looking at basic GPS Receiver modules such as the SE4150L-R by Skyworks Solutions and adding a chip or ceramic antenna.

My concern is the external SAW filter and antenna matching network. I'm assuming for the antenna line and network there the board should be 50 Ohm controlled. The biggest problem is I lack a network analyzer to check the board. I'm concerned that with a cheep fab, over time if the FR4 changes or with component tolerances that even if it works now, will the 1000th one work?

Does anyone have any experience using GPS RX modules and can input how it worked out for them? Or any suggestions such as stay far away from this RF black magic?

Please don't think I'm ignorant, I have done basic RF work in the past, but this GHz stuff at-least from what I hear can be a different animal.

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ You're brave :-). How much cost difference is there likely to be between module and DIY? The net potential savings would need to be significant to merit the risk. eg you ask about te FR4 changing. Where would these made and what control do you have over the QC of sourcing. FR4 is notionally a fairly generic material - can & will manage aspects like that? | Then there's the actual design :-). If you email me I can provide the name of a person who may have some insights - I don't know if he's done GPS but he on occasion spends a day + of Cray 1 equivalent time modelling arcane RF structures. \$\endgroup\$
    – Russell McMahon ♦
    Commented Oct 22, 2015 at 1:22
  • \$\begingroup\$ my email address is on my profile page. \$\endgroup\$
    – Russell McMahon ♦
    Commented Oct 22, 2015 at 1:23
  • 1
    \$\begingroup\$ I guess I am brave haha, I could decide to go from the ground up with a LNA, RF Frontend, etc... At-least this is almost a finished solution. @RussellMcMahon thanks for the offer but the more I read the more I am leaning towards the 4x more expensive module. Then I can sleep knowing I delivered a product that will work. I'll wait a day an see if anyone else has anything to say. \$\endgroup\$
    – MadHatter
    Commented Oct 22, 2015 at 2:19

1 Answer 1

Reset to default
2
\$\begingroup\$

You'll just need to tell your board house that the antenna trace needs to be 50 ohms @ 1575.42GHz (I assume you're using L1). +/- 10% on the trace impedance should be fine.

You may run into problems with antenna detuning and radiation pattern. These problems can occur due to the case (plastics, screws, etc), board layout (ground plane, stitching vias, antenna placement), and board components (battery clips for example), and anything near the case wherever it's installed/used.

You can assess radiation pattern by putting the unit in an open area for 24 hours and collecting data on altitude, azimuth, and C/No. However, to measure detuning (and check performance of a correcting pi network) you'll need a network analyzer.

I'd recommend working closely with an engineer from both your antenna manufacturer and your GPS receiver manufacturer. For example, they may be able to correct detuning for you on a service basis for a lower cost than buying a network analyzer.

I'm not sure what level of accuracy / acquisition time you'll need, or what environment you'll be in, but I can tell you I know for a fact some $7k+ GNSS-aided INS units use a uBlox MAX7 or MAX8 (which run roughly $10 at quantity).

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.