Rapid-acting insulin amounts incorrect with fslibre2 driver

[edquist]

Hi there!

With much gratitude I've been using glucometerutils since 2019. Really great having a way to extract the data to do my own analysis & plots.

With the original freestyle libre, using the dump command, the BG, Rapid-acting and Long-acting values came back accurately. But with the libre2, the BG and Long-acting values appear to be correct, while the Rapid-acting values are incorrect and all over the place.

In case it helps, here are some sample Scan results from running glucometer --driver=fslibre2 --device /dev/hidraw6 dump

"2026-06-10 06:25:15","145.00","","CGM","(Scan); Rapid-acting insulin (12.5)"
"2026-06-10 08:38:38","128.00","","CGM","(Scan); Long-acting insulin (14.0)"
"2026-06-10 10:22:44","185.00","","CGM","(Scan); Rapid-acting insulin (11.0)"
"2026-06-10 14:56:49","135.00","","CGM","(Scan); Rapid-acting insulin (28.0)"
"2026-06-10 16:09:31","182.00","","CGM","(Scan); Rapid-acting insulin (4.5)"
"2026-06-10 17:16:47","207.00","","CGM","(Scan); Rapid-acting insulin (8.0)"
"2026-06-10 18:37:22","139.00","","CGM","(Scan); Rapid-acting insulin (18.5)"

And here are those rapid-acting values vs what's show on the device logbook (as I entered them)

csv   device
12.5  2u
11.0  2u
28.0  6u
 4.5  4u
 8.0  3u
18.5  2u

... No obvious pattern. Seems like the fslibre2 driver is somehow reading garbage values for Rapid-acting.

Any thoughts?

Many thanks for providing this driver!

[edquist]

I should mention that this has been the case for me with multiple FreeStyleLibre2 readers over the last year+, both US and Greek, which have had this same problem with the fast-acting insulin values specifically.