The goal - some improvements over a common range hood.
aka Smart Range Hood. See blog post #1 for the complete build process.
Cooked Pi anyone?
This will be a quick post of joy, excitement, tragedy, and triumph.
Last night I thought I was ready to install the components into the hood. We only have one week left and I want to get things moving. I finalized most the remaining loose ends and went about the final installation. Once mounted, I plugged in the Pi - and bad things started happening! I hadn’t hooked up any mains voltage yet, but my pcb started to flicker on and off. The LEDs were flashing; the buzzer was going crazy.
I unplugged the Pi Power and started looking around. I checked for sorts between power and ground but didn’t have any. I looked around some more and finally found that the 40-pin header for the screen was plugged in 1 pin over - aaaaaarrrrggggghhhh!!!! How much magic smoke got out???
An artist’s recreation of the mistake.
This happened late at night so I decided to stop and sleep on things. I unplugged everything and went to bed.
Since there is only a week left to finish, I broke down on Friday and took a half day from work. I came home and started diagnosing the issues. I didn’t know how much damage there was. If I was to need any new components, I had to get them on order very quickly. I am actually traveling for work next week so won’t really be able to work on the project after Tuesday; so any delays or setbacks would be very bad...
The first thing I found was that after seating the header properly, the Pi came alive; but only partially worked. The screen was backlit but there was no image. I had temperature data from the DHT but no analog sensor data. The DHT is hooked right to a GPIO but the analog sensor is on the I2C bus. I tried scanning the bus for devices over SSH but got an error instead of a list of devices. See the video below for an example.
Since none of the devices showed, I figured that the Pi itself was cooked and perhaps the I2C pins were ruined on the Broadcom CPU. I pulled the SD card and put it in another Pi. I used a spare blank circuit board to test with. I still got errors trying to scan for devices; with zero and with one I2C devices installed. I switched the screen output over to HDMI and started looking around. I hooked up the scope and found that there was no Data or Clock signals present when I tried scanning; which should not be the case. And this was tested with nothing plugged in; just the pcb with no devices attached.
I finally found that in the Config menu, I2C and SPI were somehow disabled... strange - they worked 10 minutes before I had the problem. I re-enabled them and could see the devices properly again. I switched back over to the original Pi and the full PCB. The Pi was still alive!
The screen was the only casualty. I found that the buttons still work, and I still get a touchscreen interrupt when you tap it, but no video.
I ordered a new screen and will have it next week... could have been a lot worse!
Here is a video of the debug process.
And here is a video of the installed Smarts!!!
(new screen to come)
Yahoo!
So with that, I think I have a complete project! Still some minor loose ends to tie up, but I have done what I set out to do. I wil have a better summary in another blog post to talk about all the individual items.
Oh - one more breakdown- my 3D printed had the new improved cover for the air baffle printing; but the part warped off the bed and broke my Z-Probe mount 
so that is down until I can fix it... but I think I am mostly done with it 
Top Comments