A puff of smoke, bad smell, and now I can't connect to the printer

You've shorted your board somewhere or somehow - that is the AVCC and Ground return area of the IC Die, and is about the location of its protection diodes. You've dumped many many amps onto the AVCC pin one way or another (For the purists - yes conventional current flow is actually GND to VCC, but lets net get bogged down in semantics ), and its blown clear through the encapsulation (they survive for a good while, but the power dissipation of a short condition will rapidly overwhelm it, hence the 'it worked but then failed' scenario). In IC terms - this is kinda rare and a bit of an effort to achieve in normal operating conditions - so I can't hazard a guess as to how its occurred - it could be as simple as shorting the pins underneath against the case through to other possibilities.. but unfortunately - its not a 'defect' thats caused this..

As for solving your situation - You are better off ordering an ATmega1284.. since its cheaper anyway : http://www.botronicz.com/atmel-atmega-p … ololu.html

Buying from mouser, you will need a ICSP as it comes sans-bootloader. The URL above is for a 1284 (more RAM, Program Mem) and includes a bootloader.

Removal and replacement is simple - its in a DIL socket.. it just plugs and unplugs....

But before just replacing the IC.. I'd be using a DMM to 'buzz test' all the VCC lines on the board, as something has or had a nasty short somewhere ...

And that 'bad smell'.. well.. Its a unique smell that relates to inherent electrical componentary death.. and is a sure sign of things are dead and almost universally includes a component or 3 having 'released their magic smoke' (as everyone in the industry knows, all electronic components don't run on electrons, but rather magic smoke... and once you let that smoke escape from a component, it ceases to work.. Q.E.D.. electronics are made from magic smoke... )

Thanks for the responses.  Chip came out easily.  Best I can tell the buzz test passed.

The board was in it's original mount position (SD3) with the pins held away from the case.

I did have the thermistor positioned a little different in that last run and a bit contorted.  I took another look at it and it seems  one of the lead wires is loose, moving inside of the bead.  It's possible the leads were touching outside of the bead as well the way I had it taped.

This all happened shortly after I ordered an E3D.  So I'll be out of commission until I receive a new ATMEGA1284P as well as hotend.

Anyone know if the E3D lead time is still 4 weeks as it reads on their site?

I have 2 ATMega1284P with marlin preloaded. I will sell you one for the cost of shipping and a beer :-) ~$8

3d-oodler
We have had most parts arrive early for the hotends, we are packing them now.

We are waiting on the final part - heaters. They are in the hands of FedEx right now - but there are a LOT of defective cartridges on the market right now, so I won't drop the lead time until I have parts in my hands, inspected and tested.

Best case scenario is that we will be shipping on next Monday (12th Aug) - all depends on the heaters coming in and being within spec. I refuse to ship substandard product, it comes back to bite me long term.