Thanks everyone. I take the cautions to heart, esp. Roger's as I've followed all his stove articles on BPL with interest. Rest assured, I don't intend to use this windscreen again, although I have used it before without issue. And of course, I take full responsibility for anything that happens when I use a stove.
However, it's worth pointing out what I used appears to be very similar to a windscreen design published on BPL by Ryan Jordan:
I thought I followed that article pretty closely in making my own. Looking at the article more closely now, it appears RJ used a larger pot and canister, perhaps creating more air space around the stove burner. Both of us have about a 1/2" gap between pot and screen, about half the space Roger recommends. It may be unclear from my photos, but my windscreen only extends halfway up the side of the pot, and the valve body of the stove is entirely outside the enclosed space, and a good 3/4" (2 cm) from the bottom of the reflector.
So I don't know. As I said in the original post, I monitored the canister closely throughout, as I do any time I use a windscreen. Flame level was moderate, for good fuel efficiency. In fact, my hand was touching the top of the canister and the valve body of the stove when the valve failed. The canister and valve body were warm to the touch--warmer than usual, but I had no problem keeping my hand in place. If the valve did melt, obviously something at some point was "well over 100C" but nothing I could get my fingers on. There's no evidence of any heat damage to the thread o-ring or stove base washer. The valve still works smoothly. No evidence of heat oxidation on any of the brass parts. Obviously there could be damage I can't see, but the stove seems to work fine on a new canister--much better in fact than I remember the canister that failed working (most of that canister was used in a [well-ventilated] cabin, without windscreen). However, when screwed and unscrewed from the new canister there seems to be a point where the Lindal valve is depressed before the thread connection is fully sealed, and a bit of gas is vented. O-ring replacement seems prudent. Roger, do you happen to know the size off-hand?
Is it possible for the flame to have burned back into the stove valve body and melted the Lindal valve (if it indeed melted) without raising the temperature of the canister and stove beyond "touch" levels? Due perhaps to a too-tight windscreen, low flame, and I-don't-know?
Edit: Was still writing this as you posted, Ken. As noted, that article was my reference and inspiration. Be interesting to see if RJ still uses that style windscreen. But again, I take full responsibility for the risk with anything I make, regardless of the source.
Edit 2: Will Reitveld, in another BPL article
has a photo of a windscreen he made for the Coleman Exponent F1 also very similar in appearance to mine. He also references the Jordan article--in fact, I believe that's the link I originally followed. Again, pot size may make a significant difference in the actual air space around the stove (hard to tell from Will's photo what size pot/canister he's using. And it's ultimately my risk and responsibility. But I don't think I was building anything *apparently* beyond the scope of what's been presented in past BPL articles.