Next, I brazed on the evaps to the condensing unit. The dual-feeding filter/drier
was brazed on as well.

As seen in the picture, the filter distributes the refrigerant to the 2 evaps
and the refrigerant flows back and meet together into one suction pipe which
goes back to the compressor.

Another view:

Next, I charged her up and checked to see the speed of frosting on the
2 evaps to determine whether the cap tube are short enough and proportioned
out well. Its hard to pin point, I go by experience judging how fast the evaps
frost up and back on the suction hose and how fast they frost in relation to
each other. In general, the frost on the CPU suction hose should move faster
than the frost on the GPU hose as the flow here should be faster.

I charge my rig like this: Static charge in about 40 PSI of R290, start the
rig up, put in some R22 till the evap gets a little cold, liquid-charge in R410A
bit by bit until high side pressure is near the highest pressure I am willing
to run the system with, and then charge up the rest with R22 if the refrigerant
is not enough. This gives me a safe pressure I’m willing to go when working
with R410A and a low boiling point at the same time. I always plant a temperature probe at the accumulator to help with charging and I try to keep this temperature low enough so that there is enough cold refrigerant going back to help cool the compressor.

Next, I start to insulate the evaps with neoprene, foam pieces and bass wood
pieces cut to size:

From this picture, you can see that the evap top is spread with silicone first
and a foam piece is stuck on and then silicone spread again and finally bass
wood. This is one of my favourite way of insulating. The temp probe runs out
and the cap tube is wrapped around the suction hose for subcooling.

I use only foam pieces for the side of the GPU evap due to space constraints.
The top of the GPU evap is insulated similarly: