Close

Troubleshooting

A project log for GPS Clock

A simple desk clock that gets extremely accurate time from GPS

nick-sayerNick Sayer 05/25/2017 at 15:330 Comments

Here's a troubleshooting guide for the clock:

"no GPS" - FIX LED on at half brightness

This means that the GPS module has been fried.

"no GPS" - FIX LED blinking at 1/2 Hz

This means that the GPS module is working, but for whatever reason the microcontroller is either not getting the PPS or not getting the NMEA serial data. This likely means a poor solder joint on the relevant pads on the bottom of the GPS module. You can tell which by using a scope to examine the pins on the controller. If you see lots of square wave activity on the RXD pin most of the time, then the serial pin is working. On the PPS pin you should see a ~800 µs positive going pulse once per second. A lack of either causes this symptom.

"no GPS" - FIX LED on solid

You'll usually see the FIX LED off at power-up and then come on solid after about a second. If it never starts to blink after that, then you are not getting GPS reception. Check the center conductor of the SMA jack for 5 VDC. If you don't get that, check the AP2331 (the SOT23-3 nearby) to see if it's hot. If it is, then there's a DC short on the center pin. Check the soldering of the TVS diode (the 0402 part nearby), and if not that, then check your antenna. If your antenna presents a DC short, then you must disable the active antenna power by opening all 3 pads of the power selector solder jumper. If your antenna requires other than either 3.3 volts or 5 volts active power, then you must open the power jumper and supply whatever other power is required with an external bias-T.
Active antennas are highly recommended. Passive patch antennas will only work outdoors and only barely.

No display at all - not even self-test

If the display never does anything, but the GPS shows signs of life, then check the soldering of the MAX6951 chip. The pins are very fine pitch and it's easy for them to short to their neighbors. If this happens with the chip select pin and Vcc (which are neighbors), then the display will never wake up from sleep mode.

Display self-test passes, but weirdness ensues

If the display self-test immediately after power-up shows all the LEDs on, but in operation LEDs operate incorrectly, that likely means they're mounted backwards. If the AM/PM or colon LEDs are backwards, they'll come on in conjunction with other segments rather than when they're supposed to. If the tenth digit is installed backwards, it will self-test correctly, but act weird. This should be obvious by the fact that the decimal point will be in the top left rather than bottom right corner.

As a reminder, the 7 segment displays all have their decimal points at the bottom, and the colon and AM/PM LEDs have their anodes at the bottom.

Power up self test works, then the menu comes up, or the buttons don't work

Make sure that the buttons are not jammed against the back of the cabinet. When assembling the case, you should make sure that they're freely moving inside of their holes.

Discussions