Hello,
One thing that should be addressed is that if you only read one measurement
and go to sleep then your VOC is incorrect as it needs 5 mins or something
to warm up. Of course I can read very rapidly to make that work and I assume
that thread in mailing list with continous stream is all about that.
5 mins ? Are you sure ?
Datasheet:
---------
"The heating duration is specified by writing to the corresponding
gas_wait_x<7:0> control register. Heating durations between 1 ms
and 4032 ms can be configured. In practice, approximately 20–30 ms
are necessary for the heater to reach the intended target temperature."
For ensuring that the heater sink is heated to the target temperature, I
added the following in the driver, and if it fails to do that we just
abort:
<snip>
729 /*
730 * occurs if either the gas heating duration was insuffient
731 * to reach the target heater temperature or the target
732 * heater temperature was too high for the heater sink to
733 * reach.
734 */
735 if ((check & BME680_GAS_STAB_BIT) == 0) {
736 dev_err(dev, "heater failed to reach the target temperature\n");
737 return -EINVAL;
738 }
</snip>
And now I wonder that -EINVAL isn't the best way to represent such a
failure. -ERETRY or something more appropriate should have been used.
That thread was about adding power management support + triggered buffer
support.
Device is already power managed as it automatically goes to sleep mode
soon after a single TPHG cycle is performed.
OTOH, I'm not brave and smart enough to add triggered buffer support
even after the long string of email threads.
I don't know about data-sheet but when you rapidly read BME680 then
resistance in VOC reader tends to rise something like 3-5 mins and then
you can get 'correct' reading (rising stops in somepoint). To have
something in kernel under public eye is brave enough :). But I have to
investigate data-sheet better and see if I am doing something very very
stupid and heater just heats up.
Bosch claimed that new datasheet shall be soon be available with all the
missing info, and then anyone wouldn't need to reverse engineer their API
or BSEC to get the desired info :P
Hopefully this comes sooner than later as it would make things more easier
Thank you for using the IIO driver and I'm glad that there are users out
there testing it. Please let me know if there are any more issues :)
No problem.. I'll try to make it work as that Python library but I have to
over come these few problems.
Sorry! I will try to recticy those and others that I have in mind, but
might not quick about it.
No worries, take you time as I have 'working' solution and I can test
new features as they appear as I have personal interest about this.
Sincerely,
Tuukka
