Sunday, February 06, 2005

Day 2 of a week of evening shifts

Maybe I should describe the building, so my readers (are there any out there?) know the context of jargon.

The building sits next to the B0 collision region at Fermilab's Tevatron; a 10 minute walk from the highrise. The collision point is, of course, underground and buried under a large berm to boot. The building exists to maintain the detector around the collision point. It is 3 stories tall and looks like a large warehouse from the outside. Inside there's a ground floor with a 3 story-deep pit in the middle of it, and 3 stories of rooms along the side next to the berm.

The pit is where we assemble the detector pieces. When the 100 ton door (actually a wall in the side of the pit) is moved out of the way (can you say "musical chairs?") we can move the central detectors to and from the collision hall. There's a 30-ton door as well, which allows people to move Genies and other small hardware in and out, and a baffle-passage which lets people go in when the other doors are closed. Needless to say all this stuff is interlocked with the beam seven ways from Sunday so you can't get in while the beam is on.

The first floor rooms contain a machine shop, a long room full of relay racks that hold our readout system, a small kitchen, a large room full of very noisy pumps, and a small office for the building manager.

The second floor rooms contain a large video conference room, a small conference room, another smaller video conference room, and then a long room divided into a section with relay racks surrounding desks and then the control room. After that you find the offices for silicon detector experts, and then a set of electronics testing labs (right over those noisy pumps. Well, there used to be a conference room over the pumps!). That section with the relay racks contains a lot of our triggering system, in particular the parts that try to reconstruct tracks online and associate primitives with tracks. By primitives I mean things like energy in the ElectroMagnetic energy calorimeter with no energy in the Hadronic energy calorimeter behind it, or pairs of hits in aligned muon chambers that might indicate a muon stub.

The third floor contains offices, a long hall for computers (including our Level-3 Trigger computers, which try to reconstruct those events that have filtered through the lower-level triggering systems to see if there's anything we want in them), and some more offices and electronics labs.

I forgot--there's another way to get into the collision hall if you're a Rhesus monkey who is smart enough to pull out bags of lead shot blocking what little openings remain in the cable penetrations. Even that would be a tight squeeze.

Right now I'm sitting waiting for the beams division people to figure out where the quench was this morning so they can put another shot of anti-protons in the machine. I'm pondering my jaw and wondering if I dinged it without noticing or if I've a toothache starting (but the dentist didn't see anything three weeks ago). The room is crowded with silicon experts working on something I don't understand, and experts on other things (like the Time Of Flight counters) whose calibrations didn't slide into the database the way they were supposed to this afternoon.

They've injected a few protons into the line, and are going to try the squeeze sequence again to duplicate the morning's failure. Waiting.

Still waiting. I sent off a few math drill problems for the youngest daughter to chew on. And I'm trying a minimization problem for a metric space question I've been working on. The silicon crew is trying all sorts of tests: they want the Consumer Monitors in cosmic mode for them. It looks like they're getting a few events.

Beam division has some new messages: "Linac quad problem holding off beam. Experts enroute."

OK, so I've gotten it reduced to {q^2 + {q^'}^2 = \alpha h^2}, where {q \equiv 1/r} and {\alpha \gt 0}. If {h \equiv 1} that reduces to a nice straight line as desired; so I haven't made too many mistakes. I presume this is nicely indexed someplace, but it takes longer to find the result than rederive it. For the h(r) I'm playing with that isn't strictly fair, since its derivative has a singularity, but . ..

Grabbed some soup and peas. They estimate shot setup in half an hour. We'll have to move quickly when they do: the silicon testers are hanging on to RunControl for as long as they can. Can't say as I blame them. They're trying to get the new Pulsar board working as a replacement for the slower Alpha. Long ago the Alpha was blindingly fast, but DEC got sold, twice, and Alpha chip development went by the wayside. We've faster chips now, and we need them. As the luminosity goes up, so does the number of hits in the silicon, and so does the processing time to piece out the tracks.

OK, they promise us beam in about 10 minutes. Time to get the consumers set up properly. That was about 10 minutes ago, now. . . (22:17).

23:03 and they've gone back from "porch" to "tuneup." I think "porch" means they've got beam in position for injection. I think this shift isn't going to get anything to tape tonight. Even if they inject quickly, it takes a while to stablize the beam, and we don't turn on HV until the beam is stable. There's no point: high losses mean high current draw and the chambers trip their power supplies.

Now that I think of it, it's kind of funny: we use these deep heavy relay racks to hold the PCs and the deep heavy computer monitors (mostly Trinitrons). But we've added more computer monitors (and TV monitors); mostly flat panel, hanging easily in whatever blank space was handy. The big old Trinitrons are still going strong, and they have the significant advantage that the glass screen will stand up to physicists jabbing fingers to illustrate problematic parts of the displays. I remember when these were all VAXes.

Chung! "L2 duhcision Tim-out" I wish they'd use a more intelligent voice program. Given how many non-native English speakers we have (and how many of them have extreme accents), you'd expect that we'd try to make our own English as clear as possible.

Running cosmics again. 23:47 I think this is it for the night.

No comments: