Skip to content

Build 843 G7 Replacement Field Observation - 2026-09-13

Field Value
Classification Supporting predeployment field evidence; not formal verification
App identity BionicLoop 1.0 (843)
Test subject identifier JRC-FORMAL-VV
Evidence source Retained BionicScout DEV device telemetry plus operator observation
Reviewed interval 2026-09-13 4:09 PM-5:37 PM EDT (20:09Z-21:37Z)
Review date 2026-09-14
Reported field issue After a sensor change, the app could remain bound to the previous G7 instead of acquiring the replacement G7
Issue result Closed for the observed old-to-new replacement scenario in Build 843

Question Examined

Can Build 843 continue using the still-working prior G7 when a replacement G7 is not yet advertising, then adopt the intended replacement after it begins advertising even though the prior G7 remains nearby and active?

Operator Context

The operator confirmed that the first acquisition attempt occurred while the new G7 was not advertising. Returning to the still-working prior sensor DXCMbT was intentional. The operator then initiated the actual transition after the new sensor became available. The prior sensor remained nearby, advertising, and functional during the successful transition to DXCMsW.

Correlated Timeline

All local timestamps are EDT (UTC-4).

Local time UTC Observation
4:09 PM 20:09Z Prior sensor DXCMbT remained active and supplied a reliable reading.
4:14:32 PM 20:14:32Z A device-lifecycle suspectedEndOfSession replacement-acquisition episode began while the intended replacement was not advertising.
4:19:25 PM 20:19:25Z The app re-adopted DXCMbT, matching the operator's intentional return to the still-working prior sensor.
4:29:32 PM 20:29:32Z A second suspectedEndOfSession replacement-acquisition episode began after the intended replacement was available.
4:30:07 PM 20:30:07Z The app recorded authenticated adoption of DXCMsW, with DXCMbT identified as the previous sensor. Adoption occurred 35 seconds after the second episode began.
4:30-5:30 PM 20:30-21:30Z Replacement-sensor warmup readings were classified as unreliable and were not supplied to the dosing algorithm; the app showed CGM-unavailable handling.
5:30:04 PM 21:30:04Z First reliable DXCMsW reading was processed.
5:31 PM 21:31Z CGM-unavailable state cleared after reliable replacement-sensor data arrived.
5:35:33 PM 21:35:33Z Algorithm step 5793 used a reliable DXCMsW value.
5:36:50 PM 21:36:50Z Meal step 5794 used the replacement-sensor value.

No telemetry from DXCMbT appeared after DXCMsW was adopted. Fifty-six reliable replacement-sensor readings were present in the reviewed post-warmup interval, and no return to the prior sensor was observed.

Result

The reported stuck-on-old-sensor behavior was not reproduced in Build 843 when the intended replacement became available. The app:

  • retained usable prior-sensor service while the replacement was unavailable;
  • adopted the intended replacement promptly once it advertised;
  • did not return to the prior sensor after replacement adoption;
  • excluded warmup/unreliable replacement values from algorithm input; and
  • resumed automated and meal-step use only after reliable replacement data.

This evidence supports notifying the study team that the submitted old-sensor replacement issue is handled in Build 843 for the observed workflow.

Evidence Boundary

This is a field observation supported by DEV telemetry and operator context, not a formal verification result. It does not establish every condition in the accepted D06 predeployment exercise. In particular, it does not establish:

  • rejection of another person's active G7;
  • behavior after reopening the CGM setup screen during acquisition;
  • every concurrent-candidate callback ordering; or
  • the complete acquisition-timeout/retry matrix.

Those limits remain under RA-023 and do not negate closure of the specific reported old-to-new replacement issue documented here.

Package Effect

This post-approval operational evidence does not change Build 843, the frozen software baseline, approved IFU, formal test results, or the authorized release Core and Technical Appendices PDFs.