@Cody Murphy yeah, 4-5 meters under dense Pacific Northwest canopy is actually pretty consistent with what others have reported — that's not really a module problem, that's just physics. Heavy tree cover causes multipath interference where signals bounce off branches and trunks before hitting the receiver, and it also straight-up blocks weaker satellite signals, so your constellation count drops and the geometry gets worse.
The 18Hz fix rate helps here more than people realize — because the module is updating so frequently, the drift tends to average out better over a moving track compared to slower modules. But it can't fully compensate when you're losing line-of-sight to half the sky.
A few things worth trying: if you're not already logging raw position data and doing any post-processing, that can recover meaningful accuracy even from noisy tracks. Also check how many satellites you're locking onto under canopy versus open sky — if you're dropping below 6-7 birds, that's where the geometry degrades fast. Some people also get marginal improvement by mounting the module as high on their body as possible (chest mount vs. hip) to maximize sky view angle, which sounds obvious but actually makes a real difference under heavy cover.
What does your satellite count look like when the drift gets bad?
