[Rack] [Noisebridge-discuss] Access

[Jonathan Lassoff]

I'm not suggesting that this *needs* to come off of minotaur now, I'm
just hypothesizing how we could make it even more awesome and
reliable.

On Thu, Jul 4, 2013 at 3:19 PM, Jake <jake at spaz.org> wrote:
> I think minotaur has been doing a fine job for two years or so.  Any
> computer will eventually fail if its hard drive or RAM start to fuck up. I
> think the solution is to replace the failing component rather than start
> from scratch.

That's exactly my point: we've added a lot of moving parts to
minotaur, when really all we're doing with the keypads is super
simple. Looking for certain combinations of keypresses and sending
several bits of state back to the keypad.
Running this on something more solid state and less-consumptive should
make it easier to add battery backups, and isolate the few components
that are needed to make the door work.

> As for the holding the door buzzer on thing, i think that was a rare fluke
> (it must have crashed up while it was opening the door?) and i doubt it will
> ever happen again ever.  If it does, i will certainly install a diode, a
> resistor, and a capacitor at the place where the USB parallel port plugs
> into the door-open button of the intercom, and it will be a fail-safe in
> case the computer goes nuts.  But i doubt it is needed.

Probably a good idea. It's still just attached to the Parallel port,
right? Something is latching down those pins, or the UART goes
offline, uncontrolled or something.

> so i'll say again:  minotaur crashed a couple times.  Let's fix whatever
> caused it and move on.

I still have no idea what's causing the lockups. We'd need to see
what's getting logged on the kernel's console.
Any idea what's wrong with it?

--j