Edwin Wise

Coilin' Fool

Woot! I just coiled about 1,000 turns (12 inches of 30 gauge... or was it 32? I should check) in maybe 15 minutes coiling time.

That's some kind of new record. And, the coil is pretty good. Only had to stop twice for gaps (no overlaps); left one gap not fully closed, so it's not perfect yet, but pretty good. And, 15 minutes! That's like eight times faster than the first one.

Now to automate the tensioning/guide process that I currently do with my hand (oooo hand cramps); and THEN to make a ponoko pattern and rebuild it into a nice version. With a turn counter; I'm thinking embedded magnets and hall effect sensors in a quadrature format. Or maybe optical. The pedometers tend to bounce really badly (one step for the lift peg, one or two more steps when they fall off the lift peg, very sad).

Progress

The tin shed is up in the backyard; the doors worked a LOT better once I stuffed a rock under one corner of the foundation to torque it all into proper alignment. Bah! The big rain (previous to THIS big rain) made all my cinderblocks settle unevenly. Humbug!

I spread the lumps of dirt (most of 'em; Marla swears she'll move the rest this week or something; she wanted me to stop, 'cause I guess I looked liked was going to die out there in the humidity), and I laid most of the paving stones around it, and it's all kind of nice. It will be even prettier once the grapes grow back, or something, to fill in the stark nakedness of that corner. I'll post pics later.

Oh, the rain, I need to see how much water leaks in... may need to assault the whole thing with a caulk gun. Wheee! Or maybe glow-in-the-dark gluestick. That would be amusing

I just ordered from ABC fill-in-the-blank services a "full exclusion" rat service, for a tad over $500. Ouch. But then again, I just got a reimbursement on my property tax due to my finally getting off my ass and putting in for a homestead exemption, totalling a tad under $400. So that helps a lot.

July is ... tomorrow! Egads! And I have a lovely four-day weekend as part of it, during which I want to settle down a connect-the-dots design (and start the prototype implementation), and I want to get starting schematics and SPICE simulations on the new Tesla coil. And maybe some parts ordered.

I'm going to go wayyyy off into high-risk territory and put some International Rectifier IGBTs into parallel conduction -- a task fraught with negative feedback and runaway current loading, which can lead to some lovely pyrotechnics. However, with the soft-switching I have planned, and running them actually SLOWER than rated (for a complete turnabout from normal Tesla practice), I think it will be just fine.

Plus, these parts are (a) cheap and (b) accessible. The part I _wanted_ was (a) expensive, and (b) really hard to actually GET. In fact, I got a call from the corporate rep and I had to tell him my plan... to us IR's part. Because it's cheap. And I can buy it. He was all, "but it shows here that it's in stock in distribution!". I countered with the hard-to-refute knowledge that of the 2 (two!) distributors that claim stock, one doesn't have stock, and the other has a horrible and useless website without any ordering forms.

Sucks to be him.

Dear God, these are educated people, right?

What a week... trying to get a fix from my H/W guy so I can freakin' TEST the changes that have come down the pike, so I can get a decent update to the rest of the team so THEY can test, so we can be done with this damn project.

Of course, we have three revisions of the device and the newer images don't work right on the older hardware (e.g. Rev C firmware on Rev B device; not my fault, it's hard-wired that way; if my H/W guy had used reverse logic on the sleep line it would have worked, but noooo, probably not enough room for an inverter. Or the '!' sign in the VHDL. Who knows.)

So yesterday I send out an e-mail that includes the note that the modified image I was sending FIXES the Rev-B sleep issue, but MAY BREAK four specific modules due to creeping incompatibilities. This gives us a lot more devices we can use for testing, with the effect of just losing 4 of 36 modules... no big deal, big improvement overall.

I get a ping from one of the guys downstairs saying that the RevB sleep hack doesn't work, they can't use modules! Okay, so my H/W guy messed up (don't we all?) and the hack doesn't work. Bring one up and I'll verify the problem.

Umm. Guys. This works. I see the module, I see data...

Oh! It turns out it only fails on THIS module. And this second one as well.

Both of which were listed as possibly broken in my e-mail.

These are smart guys. Did they not read my e-mail at all? It wasn't even a long e-mail! I kept it short!

::headdesk::

Thank god it's Friday; I _may_ get through this day without killing someone.