Thanks for your interest in this.
I’ll post some pictures tonight along with the code. My code is the same as the instructable with some changes made for my thermistor settings. I am not that good at this type of code to get in and change much else.
I’ll build another couple over the w/e to see if it still replicated in the IMPs that I have just obtained.
Following another report of this problem from another customer, we continued our investigation and did finally replicate an issue on their hardware. This issue is likely to share a root cause with the one you have been (patiently) reporting from your project. To quote our engineering team:
"[…] Suspicion is falling on a change in the way we use the ADC peripheral [in impOS 36] which is not documented as affecting its input impedance, but appears in practice to do so, to the detriment of users trying to feed it from a high-impedance input.
We’re looking into a fix, but have already identified a potential workaround: if you use hardware.sampler on that pin, even if you don’t take many samples, then this uses the ADC peripheral in the original manner and so should produce the “release-34” values even on release-36."
Thanks for the headsup…
My Imps are no longer reporting as I have not found another host after saprkfun closed down their servers.
I will probably build my own phant server to capture the data… a project for the Christmas break…
I do intend to get them back online… just a matter of priorities…
Please note this behaviour was fixed in impOS 36.12, making measurements consistent with impOS 34.11.
ImpOS 36.14 was released to developer imps a few weeks ago and is now rolling out to production devices.
Looking forward to try out the new OS… once I get a server up and running to capture the data…