Well, let me try…
With the original evaporator plate the refrigerant went through the entire channel system of the plate.
The increases the ability to cool a bit faster in return for a less than ideal performance to reach very low temperatures.
A bit of icing up is fine as this SHOULD drop off as water during the off cycle.
A cold finger acts as the evaporator while also doubling up as a reservoir for the coolant.
So rather than spreading the effect over the largest possible area it is concentrated on a very small footprint - the pool of refrigerant at the bottom the keep boiling and the cooling from the refrigerant coming from a tiny outlet (the metering capillary pipe) to go into a rather oversized low pressure environment.
Like that and if kept mostly ice free the cold finger can reach temps as low as minus 42 degrees Celsius using just propane.
With the finger being made from a copper pipe it conducts heat very well, resulting in this cooling effect quickly getting the entire finger down to below minus 30.
The idea behind the use is that a liquid will condensate on a surface if the temperature of this surface is below the dew point.
We dry filament at 55 or more degrees, so based on the hygrometer reading we could calculate the required dew point.
Under normal conditions and just considering the water in the air the ideal dew point for highest condensations levels is usually between 4 and 6 degrees.
We however dry filament, which means the moisture is released rather slowly and the rate goes down the dryer the filament gets.
Dry, hot air does not really like to take on much moisture.
Rather than fine droplets forming it more a matter of individual water molecules.
In those conditions and cold finger or cold trap would not be very efficient in terms of attracting this moisture.
You would require a rather huge surface area with a very slow air flow rate.
Convection would also be a problem as you want to limit and control the cold air exiting the cold trap.
Ideally you have the heating on side and at the bottom while cold trap is on other side and high up.
This way natural convection does all the hard work and takes the guessing out of the flow rate requirements.
The intentionally way too low temperature for the cold finger has a dual purpose:
First to ensure best performance and then to allow for a much longer off cycle to save energy.
You want to slightly overcharge the system:
During the start up phase the compressor has little bit more work to than before the mod - it is just harder due to the excessive pressure on the high side.
But once the finger reaches operational temperatures more and more of the refrigerant will fail to boil off or evaporate inside the finger.
The load on the compressor goes down with the rising refrigerant levels in the finger.
The low side then goes down from the initial 35-40 PSI to operate close to zero.
Like this the boiling temp of the refrigerant is much lower and the temperature of the finger reaches the maximum low.
In a working model it would go like this:
Drying is activated and the heaters turn on to get the chamber temp up to specs.
A hygrometer switch/controller is used to determine when the humidity levels inside the box get above, let’s say 20%.
If so the compressor is turned on to cycle in the required intervals to allow for a good water drainage.
Once levels are below of what these sensors can handle (around 10%) the cycle is kept for specified time to be determined based on testing and filament requirements.
A bit like freeze drying just without the vacuum.