2

I have a little project that's been experiencing component failures, and I'm having trouble figuring out what to do about it. I think the issue is ESD but that's just a guess. Also I'm a total noob so, whatever's happening, the root cause is definitely that I did something dumb, but hey, "experience is a series of non-fatal mistakes."

I've got an SBC and a few breakout boards in an aluminum enclosure with rubber feet. The breakout boards are a segmented LED backpack, an LCD character display, a differential I2C driver, and a LoRa radio. Power comes from a two-prong DC wall wart into a plastic barrel jack; from there a buck converter/switching regulator produces 5V.

There's an environmental detail that may or may not be important: the box is sitting on top of an air band (VHF, AM) radio transmitter—a fixed unit, not a mobile/handheld—connected to an antenna outside.

The first failures I had were the differential I2C driver getting fried. This happened twice. On a lark I added a connection from the case to ground at the barrel jack, along with a diode to make sure transients go to the wall wart and not to the ground pin on the buck converter, which is also connected to the jack. I haven't had any issues with I2C chips dying since doing that. (I'm not sure if this setup makes sense. Power kind of scares me, and like I said, this was a lark. But it seems to have worked?)

Now I've got a problem with the LoRa radio. It went very suddenly from working fine to completely unresponsive, and I suspect it's the same thing that was killing the I2C chips. It's connected to an antenna on the outside of the case, and the connector is a metal through-hole type. In other words, there's an electrical connection between the connector and the case. I don't know if that was the right thing to do either; you can add RF to the list of things that scare me.

In any event, I'm wondering if maybe whatever was zapping the I2C chips has now gotten the LoRa radio through the antenna connector. Before I start futzing with things and buying more parts though I'd really like to nail down exactly what's going on here, and clean up any blunders I've made along the way.

Does anything stick out to you experienced folks that might be the cause of these failures?

[Edited to add:] Rough schematic

danavee
  • 135
  • 6
  • Can you post the schematics? – Justme Mar 24 '21 at 19:57
  • I'll do my best to whip some up. I don't have physical access to the project at the moment, but I have some helpers there I can enlist. – danavee Mar 24 '21 at 20:06
  • Well grounded shielded long cables are a must when near antenna cable and antenna but supply issues are an unknown in your schematic for overshoot on startup. – Tony Stewart EE75 Mar 24 '21 at 21:01
  • Do you know the RF power output and frequency of that “air band” transmitter? Or just manufacturer+model number? – Mark Leavitt Mar 24 '21 at 21:10
  • 1
    It's a FlightLine FL-M1000AB. Aviation voice comms frequency band (118-136.975 MHz), though almost always tuned to 122.8. Output power is 9.44 watts. https://www.aircraftspruce.com/catalog/avpages/vhfradiostation.php – danavee Mar 24 '21 at 21:15
  • @danavee I think it may well be ESD, brought in to your circuit from the 60m sensor cable and/or the LoRa antenna and cable (if it's outside). Could also be RF from the aviation radio, but start the same way to fix it. The metal box should be grounded. The internal ground connections from your various modules should be connected together and tied to the metal box at one point. The LoRa antenna cable shield should be grounded to the box too. The 60m sensor line should be shielded. You may need transient suppressors on the i2c lines, and maybe on the LoRa antenna line. – Mark Leavitt Mar 25 '21 at 18:03
  • Thanks Mark. For transient suppression on the LoRa antenna line, how do folks typically do that? The LoRa antenna is affixed directly to the box with an SMA connector (lock washer/nut type). From there it's just a few inches of line to the module. Also, I assumed the SMA connector itself would be sufficient to ground the line to the box. Is that true? Lastly, about grounding the box: do you mean to earth (mains), or is wiring the box to the - pin on the barrel jack sufficient? (Maybe I should post a more general question about safety and protection for DC electronics in metal enclosures) – danavee Mar 25 '21 at 21:34

0 Answers0