Battery_Status messages come in pairs - one with correct values and one with V=0 and I=0 (but same ID field)

Battery_Status messages seem to be sent in pairs from my Holybro PixHawk 6x running PX4 V1.14. I get one with ID=0, V=0, and I=0 and then immediately I get a second message with ID=0, but then with correct V and I values. I can see in the QGC MAVLINK Inspector that I am actually getting a pair of messages since the message counter increments by 2 about once every second. If I graph, for instance, the current value it looks like a sawtooth pattern because it jumps between 0 and the actual value on each update. I see essentially the same behavior with my own MAVSDK-based software… two battery updates always come in back-to-back with the same ID field, but one has V=0 and I=NaN. I can try to catch and ignore the bad packet, but I would rather get to the bottom of why this is happening in the first place.

Looking at my power configuration, I have Bat1 configured (12S setup) and I use a Holybro PM02D high-voltage power module for monitoring. I have Bat2 set to 0 cells and source is set to Disabled, which as far as I can tell is correct.

Is this some kind of configuration issue? If the ID fields were different I would have figured that this was the cause, but since the ID fields are the same it kind of looks like a bug. Anyone have any thoughts on what might be going on?