1
\$\begingroup\$

I am a software engineer, but new to hardware hacking, so I apologize if this is missing any detail, I will be thorough.

I am working on a small tracking system as a personal project. I am using a SparkFun Pro Micro board based on the RP2040 that I had laying around as a microcontroller, and I want to use the SparkFun Ublox NEO-M9N as the GPS unit. I had a spare ceramic antenna that I attached to the GPS unit's antenna socket. For the software side I installed MicroPython on the MC board.

When I tested the components separately, they all work. The GPS unit works well with my macbook, and with gpsmon I can verify that it has a lock and the location is correct. It seems like it defaults to GLONASS, but once I figured that out it wasn't hard to translate the sentences.

My problem is that when I soldered the boards together and try to run code on the board to print out the serial communications across UART if gives me junk data. I expected to see the GPS sentences you might expect i.e. $GNPLL ..... output

The PPS LED still blinks indicating that there is a satellite lock, and the power flows freely - no suspect sparks or smells indicating a bad burn or anything.

I have been looking around to find any support, but it doesn't look like it's out there or I know the correct search terms.

I was sure to do the following:

  1. Attach Rx to TX and TX to RX on the boards

  2. Check that the baud rate of the GPS and the board match (baud rate is set to 115200, compatible according to both board's data sheets. In the below example I use 4800, but 115200 gives the same junk output)

I have considered that it could be an encoding issue, I have found some references to a UBX format online, but can't find the documentation.

I also considered that I could have soldered poorly, could poor solder connections explain the output above? I also accidentally soldered the CS port closed, but didn't cause any shorts, I had assumed it would be okay.

Soldering work here: Soldering

I redid the soldering on the Yellow RX cable to improve the connection. Checking the output with a baud rate of 115200 as specified in doc sheets yields: enter image description here

\$\endgroup\$
12
  • 1
    \$\begingroup\$ Can you post a diagram of what you have connected together? \$\endgroup\$
    – Aaron
    May 19 '21 at 18:10
  • 1
    \$\begingroup\$ Sure thing! Sorry, that's definitely helpful. \$\endgroup\$
    – TomM
    May 19 '21 at 18:16
  • 1
    \$\begingroup\$ The Yellow cable is a little loose on the GPS port, does that imply a poor connection? \$\endgroup\$
    – TomM
    May 19 '21 at 18:25
  • 1
    \$\begingroup\$ does that imply a poor connection? Yes! I would fix/solder that RX wire before much more testing is done. \$\endgroup\$
    – Aaron
    May 19 '21 at 20:18
  • 1
    \$\begingroup\$ Roger! I will fix the soldering there and check the data again \$\endgroup\$
    – TomM
    May 19 '21 at 21:12
2
\$\begingroup\$

Just in case you wonder how to do this "by the book". Assuming these vias are plated on both sides and connected, then the proper way to solder them (as per IPC standards) would be:

  • Peel & cut the wires so that no more than 0.5mm protrudes on the solder side. About as far as an insulated AWG24 wire is thick, so you could use such a wire as template.
  • Measure and cut before soldering.
  • Ideally, fixate the wire to the board using tape or similar before soldering, rather than bending it on the solder side just to hold it in place.
  • The solder should wet on both sides, forming a cone shape around the wire on both sides. You shouldn't be able to see the via on either side, it should be completely covered. You might be using too little solder or you remove the solder iron too early, leaving a "cold joint". The blob on "A3" on one of the boards would be an example of such a cold joint.

For general experiment board stuff, I'd also recommend to use multiple strand wires since these are less prone to silently break.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ This is really great advice, and the amount of wire and the gauge. Tape is a really good idea too. I am worried to touch too much stuff to the microcontrollers because I don't want to damage them. My only previous experience of hardware is 2000's pentium processors with those oh-so-delicate pins. \$\endgroup\$
    – TomM
    May 20 '21 at 17:59
2
\$\begingroup\$

I had the incorrect baud rate, and a loose connection on the RX pin hole of the GPS Unit.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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