Again, I find myself excited at the arrival of parts for the immersion cooling system. Again, I'm blissfully unaware of pending trouble.
I decided to start stocking replacement parts... I decided this too soon.
I cleaned and taped each piece that needed to go together--the bulkhead core was the only repair, but I had a couple minor upgrades to make. One of these upgrades was to add valves immediately outside the tank on the intake and output ports. I installed two ball-valves before the wye fittings between the tank and the pump intake, and the tank and the radiator.
Once I had everything together and felt confident that I did not over-tighten the bulkhead core, I grabbed the little 3gpm pump I used for testing and set it up to pump from the reservoir below the rig to the tank on top. I forgot the little pump had a very limited head-height (the maximum height a mump can carry fluid).
Not a problem, I thought, I'll simply shut the valve from the tank to the pump intake, remove the pump intake hose from the valve, dip the hose into the reservoir, and run the 22gpm for about half a minute. Then, I'd just lift the reservoir and pour in the remaining oil.
This is exactly how I installed the valves...
I carefully loosened the worm-gear clamp and slid the tube to the intake off of the valve. I had to fight the tube's coil to get it to sit at a good place in the reservoir--not too high that it will quit too soon, and not too close to the bottom of the tank that it will suck up any contaminants that may have settled there.
I made sure the pump power was off, then flipped the breaker on. I opened the valve to the pump intake. I knelt down beside the rig and flipped the pump on. I heard the grinding settle a little--a good sign since the thing that prompted the over-tightening of the bulkhead was air getting into the intake line. I heard fluid, but it sounded wrong.
Remember, the output valve is between the radiator and the tank, there is no valve between the pump and the radiator yet.
I popped my head up and my heart dropped as I immediately understood the situation. I flipped the pump off, but the pressure in the line was still high enough to continue spraying the far wall--5 meters away--with mineral oil from the overflow tube on the radiator. As the pressure behind the closed output valve had nowhere to go--and the 22gpm pump is more than capable of generating some serious pressure--the pressure cap on the radiator was doing its job. I opened the output valve and the system made a satisfying hiss and gurgle.
Well, that was fun...
I took a deep breath, checked all the seals and fittings, triple-checked the valve positions, and engaged the pump a second time. The grind started briefly, then mellowed, but did not go away. The oil churned from the output and frothed at the surface. There was still air in the system. I shut it down and stared at it for a solid five minutes--mentally exhausted.
This was after only a few minutes with the pump off, the air did start to dissapate from the tubes some.
I decided I needed to take a break from the rig to gather my wits and not give in to feeling defeated. Everything was perfectly still, and there was no indication that there was any other problem than the air in the lines--and pressumably the pump.
Twenty minutes later, my son runs into my office and says, "I think your thing is leaking." My first thought is that he's seen the spray from the closed valve incident...
Four... Four gallons of mineral oil...
FOUR! Four GALLONS of mineral oil!
Apparently, the core was able to very slowly turn in the tube, despite the clamp, under the weight of the oil and unscrew itself to the last couple threads before just snapping off at the end. To beat all, it was the same fitting I had just replaced and tried not to over-tighten.
My initial reaction was building dams all over the basement (in case you hadn't noticed that in the pictures) out of towels and sheets--when I ran out of towels. After I was satisfied that I had isolated the spill to only most of that side of the basement, I spread cat litter to absorb the oil so I could easily pick it up.
As I moved around the various obstacles, I came across the computer that used to read the intake and exhaust temperatures. It was sitting on a small piece of cardboard that had become saturated with oil, so I sat it in the utility sink to drain a bit before I threw it away. I was beginning to get a tension headache.
I threw all the towels in the washer to be rinsed at least three times, and went upstairs to find something to eat. A half an hour later, my son comes to me and says, "I think your thing is leaking, again."
I rush the top of the stair to see a strange sight at the bottom of the stairs. The liquid looked pearlescent. I rushed down and found the oil level in the tank was the same, but the entire floor was wet and bubbly. I quickly realized that the cardboard I had sat in the utility sink had slid into the water during the draining cycle and the sink overflowed onto the floor.
The litter where the water didn't infiltrate is drak grey in the upper-right. It served as a perfect damn at some point and prevented the water form going that way--instead, it went everywhere else the oil hadn't gone yet.
I flipped on the dehumidifier, turned around and walked back upstairs to finish my food before dealing with it. I decided to read. I opened my Kindle, and the screen would not respond. I am very glad it is not yesterday anymore...
Thanks to my mining partner--and the fact that he owns an industrial supply company--I will be getting some metal fittings (that will require some hack-sawing) to replace the polypropylene bulkheads. The fittings will be here tomorrow, but the nuts that go on them will not be here until Monday. I've decided to focus on some work and some unfinished writing projects in lieu of crying into some whiskey while I blubber to a stranger about miners that aren't working and my personal experiences with oil slicks.