So far, it’s here. This URL might change if the name changes:
http://bazaar.launchpad.net/~toykeeper/flashlight-firmware/convoy/files/head:/ToyKeeper/bistro/mini/
Name changed. The code is here:
http://bazaar.launchpad.net/~toykeeper/flashlight-firmware/convoy/files/head:/ToyKeeper/bistro/biscotti/
It’s also in the main trunk repository now too.
Oh, if I understand correctly, the driver isn’t part of the price change. The difference is the other upgraded components, like a copper DTP MCPCB and AR lens and probably a couple other things. But if you just want the body, he’ll probably have empty hosts available eventually.
I only mention the lack of promo deals because Simon doesn’t act like a typical businessman. Usually sales people have a mindset of price being elastic, trying to get as much money from the customer as possible, playing with margins and loss leaders and other games in an attempt to inflate sales… but Simon doesn’t really do that. He typically just sells things for slightly above his cost, without the usual pricing games.
It makes no difference to me what you do with it. I reflash most of my drivers too, usually even if they shipped with my code on them.
There’s also the BLF discount code, but the process of redeeming it is a little awkward and I always forget about it. I also kinda feel bad using it because it probably reduces his margin to zero.
When it’s in the “mode group select mode”, it writes a value to eeprom every time it blinks. Basically, it flashes once, sets the mode group to 1, then waits. Flashes a second time, sets the mode group to 2, then waits. Repeat this 12 times then loop. If you leave it doing this forever, it could potentially wear out an eeprom cell in under a month. However, if you merely changed the mode group once per day, it should last somewhere between 50 and 500 years. And for most people, changing the config only rarely, it should last well over a lifetime.
So, it’s basically saying… don’t use the mode group select mode as a signal blinker mode. Use it only when you want to actually change mode groups. The battcheck mode works far better as a signal blinker.
Oh, what happened in that revision is that I actually measured it.
In r217, the numbers were generated by a calculator (bin/level_calc.py).
In r218, the numbers were adjusted manually until my power meter told me the current draw was correct. It was an excuse to use my fancy new ACME Power Cape, and kind of fun because I could script a lot of it, like automating the button presses to change modes and enter config mode and such. I’ll probably measure again after I get one of the new C8s, to make sure the calibration is still good on production samples.
It’s weird that the numbers shifted the way they did, but I guess the 7135’s response curve isn’t as linear as I expected.
So, the values should be pretty close to spec on the drivers Convoy is using, but they might be a little skewed on other drivers using components from other sources.
Is that on a new Convoy driver, or on something else? It’s calibrated based on the hardware Convoy sent me, which had different response curves than I normally see.
It sounds like people are leaning slightly more toward biscotti, so I’ll plan on that for now.