OK I have an update.
First, I knew I don't drive by temperature. That I knew. But being obsessed with collecting neutral, I had this idea that, if I let the temperature rise too much in the elbow, then I would be collecting tails. I was misinterpreting (not his fault at all) the advice about keeping an eye on the thermometer to watch for the tails to come, and not let it raise, but I did that "too early". I do presume that if one controls the dephlegmator in keeping a constant temperature one can "squeeze" or "sift" the heart while having the tail components reflow, but that should be just the very end. I am supposed to have a long intermediate run with the vapour temperature basically constant. For me, that vapour temperature was 73°C or so, which makes me suspect the thermometer well is too thick.
Another source of mistake was tasting the product often and always finding it too strong. That was probably a mistake. Had I just collected the product in many numbered jars, while keeping a constant product flow (reducing the cold water in the dephlegmator when the collection slowed or stopped, and going on) and made all analysis later, I would have thought less and made less mistakes.
Yesterday evening I resumed the stilling, but the leakage in the glass window was showing off brutally with a stink of tails which was all around my kitchen. I definitely know, now, why it's described as "wet dog" or "wet cardboard". It's the only thing which I learned for sure, how to detect the smell of tails
. Leakage is not good, obviously.
So yesterday at 4 AM I stopped the work and this morning I woke up with the idea of solving the leakage first. I dismounted the contraption, I tried to open the glass element and I saw that it was somehow stuck, too much tension applied, the "prisoners"* were skewed. So when I reassembled it, I just tightened the screws by hand, telling to myself that less is more. Another mistake. When the still was hot again, all the screws were lose. I tightened them again but there was a tiny little leakage nonetheless. That's certainly due to the inner gaskets being covered with Teflon tape, making the surface not perfectly even. Visual inspection of the gaskets doesn't show any irregularity, though.
Next time I will simply tight the screws by tool, and before mounting the glass on the still I will test the leakage by filling it with water.
Today after two hours of "equilibrium building" I reduced the cold water in the dephlegmator very, very slowly. It took another hour before I begun collecting in order not to disturb the equilibrium of the column. At that point, the temperature at the elbow was 92°C. The probe doesn't go inside the vapour flow (it goes inside a closed well), maybe the vapor is a bit hotter. Maybe it is a lot hotter.
I tasted the product and it was very, very, very smooth, no pricking, no nothing... too smooth... really too smooth, so I measured the alcohol content, and it's a whopping 0 %, as in "distilled water".
I collected overall around 4 litres vs theoretical 8 litres in the pot. My supposition now is that the remaining alcohol escaped through the leak during these three days of run. The stink of heads and tails is detectable, and probably the pure ethanol was not, but it was going away with the rest. The "mirror test", which I applied several times, failed miserably: the mirror would not get steamy even when the glass element was leaking in front of me.
Well, learning is this, correcting one's mistakes.
Now I will have some fun in taking all 30-something jars and noting down first the ABV of each jar. Then I'll do some tasting with some more exact dilution. Then I'll decide what to do with this stuff, there might be some good product in it.
Next move is to make one single stripping run and immediately a spirit run, because the stripping is stupid-easy, while the spirit run is where I have to iron out all the wrinkles.
* I don't know how do you call a
prigioniero in English. It's a bar which is threaded at both ends and which serves as element of union between two or three objects, as in a sandwich. This is e.g. used in engines to couple the cylinder head and the oil pan with the main engine block.