I did some early prototype parts using JB weld. Most of the parts ended up working out ok, BUT: One prototype appeared to have good, fully cured joints. I pressurized the part with an air compressor, submerged it, and it passed the test with no bubbles. I ran it with coolant for a couple days without any power applied to the electronics, no problem. 2 days after I started using the JB welded part with power applied to the electronics, the bloody thing developed a small leak and literally fried a dual CPU server motherboard, complete with slag surrounding the VRM that blew up as result of the puddle. Fortunately both CPU's and the RAM survived, but it was a hard lesson to learn, so, no more JB Weld type compounds for doing these joints. Thus my transition to brazing and now adding TIG variations to my fabrication skill set. Theoretically the same problem could occur with TIG-tacked and brazed joints, but my confidence level is much higher with an all metal joint that passes the same pressurized air submersion test for water-tightness.
Plus, I must admit that part of my desire to go the TIG route is simply that I've wanted to buy welding equipment to add to my shop for a while, but lacked sufficient justification for the purchase. Still, for this project a nice and simple solution would be preferable, but it seems the TIG route would get the most reliable result with equipment purchases that are more general purpose, general use welding equipement compared to something more specialized with little use for me other than this specific project.