MultiNav Pro+ BLE - interference from the sensors module during active GNSS fix?

@Christina Scott yeah that timing correlation is a real clue — 18Hz is fast enough that the GNSS serial bursts could absolutely be causing noise on shared power rails or even coupling into the BLE antenna path if your layout has them physically close together.

 

First thing I'd try is dropping the GNSS fix rate down to something like 5Hz or 10Hz temporarily just to see if the dropout frequency changes proportionally — if it does, you've basically confirmed the link. From there the usual suspects are power supply ripple (the GNSS module pulling current spikes at fix time) and shared ground plane issues, so adding some bulk capacitance close to the GNSS module's VCC pin and making sure your ground return paths are as clean as possible would be my next moves.

 

Also worth checking if you're running the BLE and GNSS comms on the same bus or sharing any interrupt lines — sometimes it's not RF interference at all but a firmware scheduling conflict where the GNSS data handler is blocking the BLE stack long enough to cause a dropout. If you're on the same STM32WB07 handling both, that's worth looking at pretty closely.

Yeah I ran into something pretty similar on a tracker build a few months back - what fixed it for me was staggering the GNSS poll interval so it wasn't firing right on top of the BLE connection event window, basically just adding a small offset in the scheduling loop. Also worth double-checking your power rail if the GNSS and BLE are sharing one, because the current spike during active fix can cause enough noise to mess with the radio side of things.

Jumping on what Shannon said about the power rail - this was actually a known quirk we saw during internal testing on a similar config, and adding a small LC filter on the GNSS supply line made a pretty noticeable difference in how clean the BLE side behaved. Also if you haven't already, try dropping the GNSS to 10Hz instead of 18 during active BLE connection events and see if the dropout window shifts or disappears, that'll at least help you confirm if it's a timing/scheduling issue vs actual RF or power noise.

Both of those suggestions are solid - I'd actually do the 10Hz test first before going down the LC filter rabbit hole, just because it'll tell you pretty quickly whether you're dealing with a scheduling collision vs actual electrical noise, and those are two pretty different fixes. If dropping the rate shifts the dropout window rather than eliminating it, that's your confirmation it's timing and you can just dial in the offset Shannon mentioned without touching hardware at all.

Good call on doing the rate drop test first - one thing worth adding is to also log timestamps on both the BLE dropout events and the GNSS fix cycles if you haven't already, because that'll make the correlation way more obvious and give you a cleaner picture of whether the window is shifting or just narrowing when you change the rate. Saved me a ton of guesswork on a similar build.