AVR Freaks

Thanks, Smiley. That's certainly been my experience, although I usually don't leave my little concoctions running very long.

I'll put the Butterfly Scorcher Project on my (probably) somewhat shorter list of things to do. I'll let everyone know what I find out.

Rain away. We need (and appreciate) the rain.

It seems odd to me that a "demonstration and evaluation kit" that pulls those multipurpose (but dedicated due to hard wiring) pins out to the headers wouldn't have a comment in the documentation about any restrictions of this sort. That's not to imply that there aren't restrictions; just that if you're trying to hook people (and engineers, too) on your product line, you might want them educated. Or not, says the cynical Uncle Bob in me.

Color me embarrassed if there is such a note in the Butterfly's user manual.

Anyway, thanks for the comments. I was able to solve the immediate problem within specs, but the general problem persists. Keep them cards and letters coming.

But anyway, I was thinking it would be nice to see the progression of LCD deterioration, if any, so I should burn the glass for several hours, put up all segments, turn on a light, and snap a picture. Rinse and repeat for a week or two, or until the camera's SD card fills up.

So that's where the project is now. Engineering still has it. Marketing will probably come in and screw it up, as usual.

I personally wouldn't use 5v on the dataflash (or the Butterfly), but there have been threads here in which people have stated it works for them. You might do a search and see who it was to track down more details.

As for the LCD, here's my current theory, and it's only a theory. The real test has yet to start because I can't get my silly camera to focus tightly enough on the LCD to show minor changes.

There are 4 COM lines to the LCD on port A, and the 25 segment lines are on ports A, C, D, and G. Ports B, D, E, and F are brought out to headers on the Butterfly (directly or indirectly), and COM0 comes out indirectly.

I think (again, opinion) that if you leave the COM lines alone (leave them as inputs), you can diddle with the segments (that is, all the "normal" I/O pins) and not hurt anything. I think the problem is going to arise when the COM lines are one DC polarity and the segments lines are the other.

I have two tests devised, one which leaves the COMs floating and powers the segments, and one which powers COMs and segments in the opposite polarity. They'll get run any day now, as soon as I get a round tuit that will do a macro shot and can be automated to run unattended.

Smiley has a tutorial on the alternate pin usage of the Butterfly if you want to have a look at that. It lays all this out.

OK, first round of tests done.

I didn't see any damage if I only used the I/O pins that are brought out to headers, and did not apply an opposite voltage to the LCD COMs outputs. However, I didn't run this test very long, so that part is inconclusive.

Then I ran the full cook test (segments high, COMs low) for about 33 hours, with the results shown in the photo below. The left photo is of the LCD at rest (disabled), and the right photo, taken at the same time, is of the LCD in use.

But, I waited an hour or so and snapped another picture (the single below), and there's no sign of persistent damage. I can't physically inspect the Butterfly without disassembling everything, and I'm not quite ready to do that yet. But it looks like there has been no permanent damage, although 33 hours is hardly a significant amount of testing.

So yes, the LCD doesn't like it, but I'd say it's too early to tell whether it's fatal.