2
\$\begingroup\$

I'm trying to interface a GPS receiver with a Minnowboard Turbot Single Board Computer via a ttl serial UART on the boards expansion header but there is a substantial amount of noise on the UART and power pins coming from the Minnowboard and it is causing very low SNR values in the GPS device and the GPS never gets a fix. Moving the GPS device or antenna further from the board doesn't work, the noise is being transmitted through the header pins themselves.

What kind of circuitry do I need[0] to build to isolate the GPS from the interference on those pins?

enter image description here

[0] I've already tried putting ferrite beads on all the connections tx, rx, pwr, gnd to the GPS but it didn't improve the situation enough to get a GPS lock.

\$\endgroup\$
21
  • 1
    \$\begingroup\$ Its hard to imagine the GPIO pins are injecting so much noise as to prevent locking. Have you double checked your grounding connections? \$\endgroup\$
    – MAM
    Mar 18, 2017 at 18:51
  • 3
    \$\begingroup\$ I'm positive it is coming through the pins. The processor runs at 1.46 Ghz and the USB3.0 also appears to add noise if I have a device plugged in. I used a process of elimination leaving the GPS in the same spot relative to the board and connecting it to the pins on the header vs connecting it through a FTDI USB -> serial converter. using the FTDI all the interference disappears and the GPS locks in less than 60 seconds. If I use any of the pins on that header though SNR values are about 50% worse and it can sit all day w/o getting a fix. \$\endgroup\$
    – Pete
    Mar 18, 2017 at 19:31
  • 2
    \$\begingroup\$ @Adil Malik It is interfering with the GPS which operates around 1.5 Ghz. I also attached the pins to a spectrum analyzer using a DC block and there were spikes in that range. I also attached the GPS via a FTDI chip instead of to that header and the SNR values reported by the GPS were almost twice as good and it locked immediately. \$\endgroup\$
    – Pete
    Mar 18, 2017 at 20:01
  • 2
    \$\begingroup\$ @Adil Malik Yeah, I tried a separate power supply but that alone made no difference because interference was still coming in on the serial lines. I need something to isolate the GPS from the EMI on that header. \$\endgroup\$
    – Pete
    Mar 18, 2017 at 20:11
  • 2
    \$\begingroup\$ @MarcusMüller - the issue would not typically be the UART signals, or even their edges, but rather that those conductors (including the signal grounds) come from a board chock full of radio frequency digital noise sources. Hence it's typical for them to couple out interference even when the serial port is idle with no signaling in either direction. \$\endgroup\$ Jan 4, 2018 at 16:46

1 Answer 1

1
\$\begingroup\$

How high a resistance can you place in those UART lines? 1Kohms? 10Kohms?

Notice I've created a low-resistance (but not ZERO-resistance) ground path, at bottom of the schematic.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
0

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.