Machines Now Hate Me

Naturally, just when I need it most, a machine starts hating me and gives me crappy data. It's never liked me much, I think. So now, despite going to lab early and staying late, my productivity hasn't increased any. If anything, it decreased, since now I have days with zero usable data points collected.

The problem is that the problem itself isn't reproducible, either. I'm essentially running the same sample over and over (with possible minor tweaks) until a lucky strike hits and the data turns out alright*. Then I agonize about whether to trust the data because none of the runs are consistent with each other, and the sample changes a bit after every analysis.

Usually, when things are unreproducibly wrong, I blame malfunctioning electronics. Maybe something's loose, connections aren't quite meeting, something shorted out, or something like that. However, by inspection (I daren't take it apart**, and most of it is visible as-is anyway) things look ok. Also, other people don't seem to have problems. It's just me.

It's not a sample-inherent problem, either, since I ran similar samples before and, while the data weren't great, they were more-or-less decent. (I had exactly one run where the data was actually pretty. Sadly, it wasn't reproducible.)

I'm slightly tempted to make my supervisor do a run from scratch (sample prep and all) and see if the machine hates him too. But not that much, mostly because I don't trust him. Among other reasons, sample prep is slightly finicky and I have quite a bit more practice than he does.

So yea, the whole thing is just one big mess. Sigh.

Not quite as important but annoying all the same: my cell phone's dying, too. All the buttons stop working unless I plug it in to an outlet, in which case half the buttons work. Argh. Nor haved I gotten around to replacing my laptop yet***. Dammit, I'm too poor for this - need to pay tuition in the next few days, too!

[Edited]
Cherry on the cake: the other machine I regularly use is now toast. By "toast" I mean "needs to be sent out for repair". Arghh.

*All I ask for is a differentiable-looking curve, and an acceptable "QC" (quality control, though the QC itself isn't that great). The machine fails even that. Sigh.
**Not a good habit; it comes from industry where, if something doesn't work and I can't fix it readily, I get my boss to call the manufacturer and the professionals can deal with it. Though "readily" may well involve taking everything apart, poking around, then putting everything back together. But I only do that when I'm responsible for the machine, and not someone else, as in this case.
***I'm seriously tempted to get a Mac - even my computer-science acquaintances succumbed (though most of them seem to run some variety of Linux or BSD on the Mac). Now I'm waiting for Snow Leopard, since I don't like Vista and can't wait that long for Windows 7.