Hi guys,
after an afternoon of experiments and formulas, I have some good and bad news...
The good first: using the calculated "corrected" TCReff with my dani, greatly improved the temperature displayed during the water test (constantly hovering around 94-96 °C)
The bad now: In my first study of the formula to find a TCReff, I'm ashamed to admit I screwed up and it is actually impossible to calculate a TCReff that doesn't depend on temperature.
I'll try to explain: you can only calculate a TCReff that will make temperature measurement accurate at a certain temperature, and you have to adapt your TCReff depending on the temperature you're interested in.
Ok, doesn't sound much clearer, let's try with some math, it's simple I promise, and someone can also double check I'm not screwing it up again...
My reasoning is, for a certain dT (i.e. the delta-T, i.e. the temperature difference between the calibration temperature T0 and the final temperature T), in an ideal world where all resistance is temperature dependent (i.e. non static resistance Kr), given a base resistance R0 at T0, we have
R(T) = R0 (1 + TCR * dT)
When Kr enters the scene, what we actually have is
R(T) = Kr + R0(1 + TCR * dT)
NOTE: for Ni200, TCR here is 0.62/100, i.e. a relative change of 62% every 100°C
Now, what we can do is find an TCReff such that
R0(1 + TCR*dT) = Kr + R0(1 + TCReff*dT)
But unfortunately, if you don't screw up this simple equation as I did in my first analysis, what you get is
TCReff = TCR - Kr/(R0 * dT)
That means that the best you can do is, say you're going to vape at 220 °C (i.e. dT ~= 200 °C), calculate a TCReff that will let the temperature accurate around 220 °C.
That is what I did with my water test.
I built a coil which steam-engine estimated being around 0.14 ohms, when on the characterized K4, my dani is measuring 0.16 ohm (which is quite OK with my findings of Kr ~= 0.026).
Now, if I do the water test with the nominal TCR of 0.62/100, my temperature hovers at around 78 °C.
So, since I want the measurement to be accurate at about 100 °C (i.e. dT = 80°C) I, calculated
TCReff = 0,62 - (0.02 * 100)/(0.14 * 80) = 0,441
So I dialed in 441 on the dani, waited for everything to cool down and repeated the water test, and now temperature stayed pretty much constant (while there was water on the wick) at about 95%.
So, the concept of a TCReff is still something that can give some more accuracy, but unfortunately I'm not sure it's as useful as I initially (and mistakenly) advertised as now you have kind of "plan for" which temperature you want your measurement work better...
Sorry again for misleading you with my mistake in my first post, which I apparently cannot edit anymore to put a warning at the beginning...
