Notecard (WBNAW) opens a new session every ~6 minutes in continuous mode over WiFi — isolated across 3 firmware versions, 3 networks, and 2 antennas

1

Notecard (WBNAW) opens a new session every ~6:03 in continuous mode over WiFi — isolated across firmware, network, and antenna changes

Summary

My NOTE-WBNAW cycles _session.qo (close → reopen) at an almost perfectly consistent 6 minutes 3-4 seconds interval while in continuous mode on WiFi. I’ve spent a full night systematically changing every variable I have access to — host firmware, WiFi network, antenna hardware, and even the Notecard’s own firmware version — and the interval hasn’t moved by more than a second in any test. I’m at the point where I believe this is either a known behavior of continuous mode on this SKU, or something specific to this unit that a card swap would reveal. Posting here before I go source a second card, in case this is already a known quantity.

Hardware / setup

  • Notecard: NOTE-WBNAW (Cell + WiFi), board rev 5.13
  • Carrier: Notecarrier-A
  • Host MCU: ESP32-C3, connected to the Notecard over I2C (SDA/SCL via the Notecarrier’s edge connector)
  • Power: Currently on USB for testing; production deployment is solar + single-cell 3.7V LiPo
  • hub.set configuration:
{"req":"hub.set","mode":"continuous","sync":true}

No outbound/inbound overrides beyond this — using continuous mode’s defaults.

The symptom, exactly as it appears in Notehub events

A typical cycle looks like this:

{"opened":true,"why":"continuous connection mode"}

…followed 10-30 seconds later by:

{"closed":true,"why":"session closed because the notecard has opened a newer session"}

(Occasionally the closure reason is "notecard ended the session" instead — both variants show up.)

The device is otherwise healthy throughout — voltage steady in the 4.0-4.4V range, no correlating _health.qo boot events during steady-state stretches (boots only occurred during physical handling/reconfiguration, not during clean test windows).

Measured interval — consecutive session opens, one representative stretch

5:03:12 PM → 5:08:54 PM   5:42
5:08:54 PM → 5:14:57 PM   6:03
5:14:57 PM → 5:21:01 PM   6:04
5:21:01 PM → 5:27:04 PM   6:03
5:27:04 PM → 5:33:08 PM   6:04
5:33:08 PM → 5:39:11 PM   6:03
5:39:11 PM → 5:45:15 PM   6:04
5:45:15 PM → 5:51:19 PM   6:04
5:51:19 PM → 5:57:22 PM   6:03
6:07:06 PM → 6:13:09 PM   6:03
6:30:22 PM → 6:36:26 PM   6:04
6:36:26 PM → 6:42:29 PM   6:03
6:42:29 PM → 6:48:33 PM   6:04
6:48:33 PM → 6:54:36 PM   6:03

This pattern has now been observed continuously for several hours across the variable changes below.

Variables isolated tonight — none affected the interval

Variable changed Values tested Result
Host firmware 3 different builds (varying only how often the host polls command.qi locally over I2C — every 5s vs. only on telemetry send) No change
WiFi network 3 different SSIDs/APs, signal strength ranging from -37dBm (5 bars, at the router) to -74dBm (2 bars, ~150ft away with obstruction) No change
WiFi antenna Stock onboard antenna vs. a replacement 2.4GHz u.FL antenna No change
Notecard firmware 11.2.2.17633 → 10.2.1.17719 (LTS, applied via Notehub DFU tonight) No change

The host-firmware and I2C-polling-frequency changes were tested specifically because I suspected our own local polling might be nudging the Notecard’s session state — ruled out, since the interval is identical whether we’re polling every 5 seconds or roughly twice an hour.

Other things ruled out / worth mentioning

  • Not a power/brownout issue. Voltage is stable through clean stretches of this pattern; the few brownout resets observed were tied to physical handling during setup, not steady-state operation, and didn’t correlate with the cycling.
  • Cellular fallback works correctly and is unaffected — separately confirmed the device fails over to LTE (Band 2, AT&T, good RSRQ/SINR) when WiFi is disconnected/antenna removed, and cellular sessions don’t show this pattern. This appears specific to the WiFi bearer in continuous mode.
  • Host MCU also runs BLE (to a separate peripheral, unrelated to Notehub connectivity) concurrently on 2.4GHz. Tested across periods where BLE was in heavy continuous-scan state vs. normal connected/idle polling — the ~6:03 interval didn’t change either way, which argues against BLE/WiFi RF coexistence being the driver, though I can’t fully rule out some interaction.
  • Briefly tried periodic mode ("mode":"periodic","outbound":5,"inbound":5) — got one data point before switching tasks that also landed at 6:03, though this wasn’t ran long enough to be conclusive on its own.

What I’m asking

  1. Is a ~6-minute session-cycle in continuous mode over WiFi expected/known behavior for this SKU, and if so, is there a config option to extend or disable it?
  2. Is there a lower-level diagnostic (beyond _session.qo reasons) that would show why the Notecard is deciding to open a new session at this cadence?
  3. Given firmware version changes didn’t affect it, is this more likely a per-unit hardware behavior worth testing with a second card, or something inherent to the WBNAW/WFM200 design that a card swap wouldn’t change?

Happy to provide raw card.wireless/hub.status output or anything else useful — trying to get this solid before it ships in a consumer product monitoring remote generators, so reliability here matters a lot.