Wednesday, April 5 2006
Awhile ago, back in February or perhaps even January, I was working on a project to create an antifreeze pump system for the basement hydronic systems. This would allow the systems to be resupplied with antifreeze instead of water whenver their pressures dropped below 12 psi. In an ideal situation such resupplies would be rare, but not in my setup. Something about the plumbing causes the loop that passes through the boiler to slurp water out of the loop that passes through the solar panel, even when water in that solar loop is at a much lower pressure than water in the boiler loop. The two loops are separated by electric valves that have been closed all winter, but evidently a little can trickle through those closed valves in one direction. The upshot of this is that periodically the boiler loop reaches a pressure of 40 pounds per square inch, triggering the relief valve to discharge water into a waiting collection bucket. Meanwhile, pressure in the solar loop is gradually lost and must be made up by something. With my antifreeze pump system I can dump that released fluid (which contains a lot of antifreeze) back into the antifreeze supply bucket and have it pumped back into the system. The alternative is to have the household plumbing inject pure water into the system, which (over time) dilutes the antifreeze and renders it useless. Antifreeze is not cheap; I spent hundreds of dollars on it this winter.
For weeks now, the problem with my antifreeze pump system has been the pump itself. At first I used cheap drill-powered pumps, but they tended to leak and were unsuited to permanent installation. So then I bought an old piston-style medical fluid pump at P&T Surplus, and it worked fine for awhile. But its motor was weak and its gearbox old, and it proved slow and unreliable. I tinkered with it and attempted to replace its motor, but this only hastened its decline. By the time I returned from Guatemala I'd decided I needed a real pump. I shopped around for awhile but had great difficulty finding anything affordable that could attain the necessary pressures (at least 20 psi). Then I found a $34 pump at TNG Direct with the ability to lift a column of water 40 meters. That translates to about 40 psi, more than enough for my needs. Let me tell you, I looked long and hard for a pump with those specs and would have been perfectly happy with a used pump from a seller on Ebay, but there's just nothing out there at a better price. The pump arrived the other day and today I assembled some of the little bits I'd need to get it working. I needed a check valve at the bottom of the intake pipe in the antifreeze reservoir, and I needed special soldered copper adapters allowing me to connect conventional hoses. These were all things that I worked on this evening.
By the way: I'm continually amazed by the ability of solder to make things that "sort of fit" actually fit together in a completely water-tight manner. This evening I managed to get a brass female-to-female hose adapter over the end of an inch-to-half-inch copper fitting by beating it with a hammer, and once it was in place I completed the connection with solder. Then of course I had to sand the dents out of the face of the adapter so a rubber gasket could mate with it without leaking.
Lately my computer has been crapping out on me whenever I have too many windows open and I try to launch Photoshop. I suspect it has something to do with a full scratch disk, but do I really want to muddle through the diagnostics of such a problem? I have to say, having used Windows 2000 regularly for five years and Windows XP regularly for only one that Windows 2000 was a lot more reliable. I get funky crashes and strange behavior all the time from Windows XP that never happened in 2000. The most maddening of these involve crashes in Explorer, a process that often cannot be killed in the Task Manager. This violates what I'd thought was a cardinal rule of the Task Manager (from at least as far back as the Windows NT 4.0): that there is no process the Task Manager can't kill.
For linking purposes this article's URL is:feedback
previous | next