...Um. Not so much.
Along about dinner time last night, I had a call from work.
"Along about dinner" being that point in the process where I had two New York strip steaks and a batch of mushrooms in a cast-iron grill pan on the stove, having decided it was too hot to grill outdoors. My right knee was hurting and while I hadn't quite been in pajamas all day, what I was wearing wasn't suited to anything more formal than weeding a garden, and maybe not even that.
Dinner was at that point where you either finish cooking it and eat it, or you throw it away. Dinner for two.
So the phone rings and it's one of the Master Control techs, reporting a pressurization issue at the unstaffed North Campus. We monitor a couple of things up there that are kept dry by keeping them full of dry air (from a sophisticated compressor-dehydrator*) or, as a backup, dry nitrogen from tall, 3500-psi tanks of the gas. The system only runs 5 psi above ambient, but it's a critical 5 psi; what we're keeping dry runs a thousand feet, carrying high voltage, high-current VHF and UHF energy and if it gets wet, the connection points (one every twenty feet) can begin to arc and heat up, melting their support insulators. This doesn't end well.
Dry air is fed into the system from a single point but there are two different monitors. I am rewiring alarm connections, and one of them failed shortly after my last work. It had been intermittent; I thought I had found the wiring problem but I had not. Hey, no problem, we have two alarm sensors; I asked Master Control be especially vigilant until I could return to the North Campus.
And that was why the tech had called: the second alarm had tripped.
Maybe it was a false indication, maybe not. There's a while left once the alarm goes, at 2.75 psi. Nothing for it but to finish cooking, eat dinner in some haste, change into slightly more civilized attire (it's always 65°F at the North Campus, unless something has gone very wrong), grab a few bottles of water, and head out.
* * *
Arrive to discover the compressor was dead. D-E-A-D, no breakers tripped, motor windings reading less than half the normal DC resistance.
System pressure was just under 2 psi. I started up the dry-nitrogen system and the regulator was wonky: the tank was low. I set flow rate with the valve so it was sitting at 3.5 psi to repressurize, loaded all one-hundred and fifty pounds of the compressor-dehydrator onto a cart, and hauled it to the the workshop in the far end of the building.
A few more checks later, it was obvious there was no fixing the compressor this side of an electric-motor shop. I have a spare, of course. It's tricky to change; you can't get at the mounting bolts, so you remove the support plate from the shockmounts. The plate is held by bolts through the rubber shockmounts from underneath and, you guessed it, they won't support the 60-some pounds of the thing upside down or sideways and the feet of the unit aren't tall enough to let you reach under. Nope, one hangs the enclosure over the edge of the cart and removes one bolt at a time. A bit of un-plumbing, some electrical, and it was out. The same thing in reverse to install the replacement (after a spine-freezing moment when I read "230V/50Hz" at the top of the label on it, until I read on down to the description of how to strap it for 120V and the maker's admission that it would run on 60-cycle power, too). Checked it was set up for wall-socket juice and proceeded to test the finished installation. This is quick to write but it's nearly two hour's work.
It ran fine!
Hauled it back to the far end of the building, plugged it in, and started it up without connecting the output. The compressor feeds a finned cooling section and then a pair of "molecular sieves" that remove the last bit of moisture from the air. There's a clever arrangement of valves, one of them run by a solenoid controlled by a thirty-second timer: every half a minute, one of the sieves gets backflushed to clear any captured moisture, while the other is online. Thirty seconds later, they trade places. If the thing has been shut down, you need to run it to clear them both out, a process that can take up to an hour.
The dry air is held in a small tank at about 70 psi, and fed out through a regulator at 2 to 10 psi, so the compressor can run at a reasonable pressure and comfortable duty cycle, and so there's sufficient pressure for the backflush. Drying it out, all you need to do is let it spill air into the room and make sure compressor run time isn't excessive. --For an hour or more. There's a nice front-panel light to show when it's making dry air again.
I passed the time checking nitrogen pressure and flow, and pondering if I was too sleepy to change out a big tank if necessary. Cleaned up the shop, puttered around at this and that, and an hour later--
An hour and a half later, there was still a big "Humidity Alarm" light. It hadn't sneezed any water out the drain line, either. And the more I listened, the more that solenoid valve cycling sounded just a little wrong.
Replacement valves are amazingly expensive. Rebuild kits for them aren't and I keep a couple on hand whenever the budget permits. (Okay, I admit it: I sneaked the second one in years ago. There are a lot of fiddly little bits to replace and "Two is one and one is none," if one of the pieces goes missing!)
Unplugged the compressor-dehydrator and horsed it back onto the cart, all 175 pounds of it. Back to the shop, cover off, the valve is buried deep in the guts of the thing but I have been at it before; there's a trick to it.
Two of the air connections are "prestolock" types and there's no removing them in situ, period. Out of situ, maybe, but it's not worth it. The drain line is nothing, just a short section of flexible tubing that pokes through the back panel. The fourth connection is a section of flexible line I replaced a decade ago, softer line connected with a hose clamp and barb. That, you remove with a razor blade, and lose a short section every time.†
That gives you one half still plumbed into the machine, with one dual valve seat and four O-rings on it; and another half with the four-part solenoid plunger assembly -- no, make that five parts -- a complex double-ended commutating inner valve, a single valve seat and two more O-rings. Except for the threaded outer solenoid tube (got your 1" box wrench handy?), it's all force-fit and stuck together with a couple of years worth of gloop.
Take everything apart, clean it with alcohol, dry, lube the O-rings with silicone grease and the valve seats with a tiny dab of molybdenum disulfide in silicone, keeping any excess away from the rest of the valve. The solenoid plunger has a key that engages the commuting valve, and an internal spring and brass tube that complicates the task. The other end gets rebuilt in place; the seat is very tight and has to be prised out with tiny screwdrivers, and the far side takes a pair of tiny O-rings that you'd never find if they hit the floor and rolled away.
(It was at this point that I noticed my right knee wasn't taking weight very well. Oh, well, the job still has to be done.)
The two halves get mated up and awkwardly bolted tight, it's moved close to position and the coil assembly is slid back on, and a new retaining clip is installed.
New retaining clip-- Rats, I'd left it on the workbench.
Left it, in fact, right next to a short coil spring that goes inside the body of the valve, around the solenoid plunger right where it engages the commutating inner valve. You know, the thing that's not all that easy to put in even when you follow the correct procedure.
|The solenoid plunger tube sticking out of the valve body and the spring that's supposed to be inside the brass part.|
Nothing for it but the valve's got to come back apart. That takes time but eventually the big half is back on the bench. Rather than unscrew the plunger tube and risk tearing up the O-ring, I used thin needlenose pliers to hold the plunger as I slid the commutating part out, put the spring on, and slid the inner valve back in place. It only took three tries to get everything all lined up.
Next, reassembly, with all the fun that entails, and by then I was getting a little punchy, having to slow down and check myself at every step.
Got it all back together, tightening down the last hose clamp, plugged it in and hit the switch. The compressor started up, gauges fluttered up the scale and thirty seconds later, the drain line sneezed out a couple of tablespoonfuls of dirty water as one of the molecular sieves ran its first purge cycle.
I put the outer cover back on, hauled it back to the far end of the building, wrestled all 200 pounds of the thing off the cart, plugged it in and started it. An hour and a half later, the "Humidity Alarm" light went out for good (after a half hour of off again/on again: one sieve was a lot wetter than the other) and I reconnected it, shut the dry-nitrogen valves at tank and manifold, then watched, adjusted, waited and adjusted again to get the system back to normal pressure.
I limped around putting tools away while getting it all stabilized and finally left after another hour. Total time from arrival to departure, eight hours. On no sleep. My knee was hurting bad enough to make up for it, at least.
Driving back home just ahead of sunrise was, well, "interesting." I texted my boss, ate a snack and took ibuprofen, and made it to the bed right before I fell soundly asleep.
That's why there wasn't a post this morning.
* You can think of it as being like the thing in your garage you use to air up tires and maybe operate small tools, but it's not. The compressor itself and its motor is a single unit the size of a sewing machine and twice the weight; the output is cooled, regulated and run through a "molecular sieve" to remove any trace of moisture. The whole assembly is about the size of three sewing machines. They cost about what I usually pay when I buy a used car.
† There are various barb designs and some of them purport to be more removable than others. In my experience, after a year or more of being compressed, the hose cannot be removed by anyone of normal strength. YMMV.