General

Improving on the mask of the calibre biometrics

3 weeks ago (edited)

Hello Calibre Bio users. Was thinking about this for a long time. Finally let a friend , 3d scan the calibre mask and 3d print an adapter to use it on a elevation training mask I had around. No more problems with sweat and less breath condensation in the calibre mask itself. I noticed more condensation in the elevation mask. Im an amateur mtbiker. So I train a lot with the calibre indoors, but every time I take the mask of to wipe my face off sweat, it skrewup the measurement, a little bit. But now I noticed a more presicely measurements. Today I did a 45min session and have Chatgpt calculate my oxigen uptake at 180w (last 10 min of the ride) and gave everything to Chatgpt (weight , room temp , humidity) and loaded my ride fit file and the openspiro CSV file , chatgpt told me it is reasonable in the range. So I think this is an improvement. The only con I can think of is the CO2 rebreathing , but on the indoor trainer I didn't felt anything , it felt great. What do you think?

Thread image 1
Thread image 2
Thread image 3
Thread image 4
Thread image 5
Thread image 6
/
Swipe left/right to navigate
Exercise
Diagnostics
Natural
2

Please sign in to post a reply.

· 3 weeks ago (edited)

@brgmn This topic is your zone of genius, correct? What do you think?

Thanks for sharing, @rubencito-specht

2
· 3 weeks ago (edited)

Hey @rubencito-specht, thanks for sharing - always great to see people experimenting with their setups!

A couple of technical points worth flagging on the Calibre Bio mask, though, because they make most physical modifications harder than they look:

1. It's an open-mask design by intent. The mask is built around a fixed inlet geometry that relies on the surrounding air being at typical indoor conditions. The whole device is calibrated against that assumption. Any change to the inlet geometry or how ambient air enters shifts the operating point it was characterised against.
2. There is no flow sensor. This is the part that surprises most people: the mask doesn't actually measure your breath flow. The device only measures O2 at a single point and infers VO2 via a fixed model that combines that reading with assumed ventilation from population-level reference data. So modifying the mask doesn't unlock better flow data — there was never flow data to begin with.
3. The firmware is closed. Even if your modification were mechanically sound, you'd have no way to re-characterise the mask against the device's internal model. The conversion from O2 reading to VO2 lives in firmware, and the constants in there assume the original mask geometry. A modified mask + unchanged firmware = systematically wrong VO2.


So unfortunately the mask, the firmware, and the reference data are a tightly-coupled system — you can't really change one without owning the others. Honest about it being a limitation, but it's why I'd be cautious about reading your post-mod numbers as physiologically meaningful.


On a related note — this is actually one of the reasons I'm building OpenSpiro OSM-1, an own open VO2 analyzer prototype. The whole design philosophy is the opposite of the closed-mask approach: real per-breath flow measurement, real per-breath O2 measurement, standard interfaces so off-the-shelf parts fit, and open firmware so calibration constants are something you can read and edit rather than a sealed binary blob. We're still in prototype phase, but if tinkering with this kind of hardware is your thing, you might find it interesting once we're further along. Happy to share more once we have something to show.

Keep experimenting!

3