The Mask is Off By One

A common software mistake, off by one errors happen occasionally. Indexing starting at 0 instead of one, you got ten things, so the last on is 9, not 10. What about hardware, you go pins, and they are numbered, or sometimes labeled. Dang it, wouldn't you know it, the labels on the board were off by one.

If you look at those labels, you can see the ground and 3.3V labels and that is where I am connected. Trouble is, if you look on the board this shield plugs into, the labels are off by one.

Starting with the pin labeled 5.0V on the shield, that is connected to RESET on the MEGA128 board that I am using. The next pin, is labeled  3.3V, but the traces in front of it are connected to the 5V. The next pin, labeled GND but the trace in front of it is really 3.3V. See these two pictures.

I believe I may have cooked my chip. All I could get out of the code was a 0. Over and over again. I should check the other writing to make sure I am connected to the proper pins. 

I have a spare MAX31855, so I am not totally dead in the water yet. I want to connect it to the right voltage, incase it does a shutdown on over voltage, and not leak smoke only. (I never saw any smoke).

Unsoldering the chip may be possible as well.

I really am trying, I just kept fiddling with the software, until tonight when I checked the voltage. (trust me the software is monkey see monkey do. I am using the SPI library included with the Arduino code.

I am an idiot. The 50, 51, 52 ... pins at the end start with 2 ground pins. I didn't count. So it is completely wired wrong. I need to take another whack at that part before I get around to calling my chip bad.

I am blaming it on being out of hardware practice. Can you take 5 years off from something and pick it up in a weekend.

Ever be so dumb yourself?

I re-did the connections. I am, using the ICSP header for both the SCLK and MISO. I've got it connected to 3.3V and it seems to be outputting data. 0 with nothing connected and some random data when the thermocouple is connected.

I've found the Adafruit libraries for the MAX31855, and want to try them, but am out of time for tonight. I'll get to looking at them another evening when I have more time. 

Chips Are Getting Smaller

I started building with integrated circuits in about 1976. Radio Shack started carrying the 7400 series chips sometime around then. In school I had learned about half and full adders, so I knew what AND, OR and NOT gates were. I imagine they were a couple bucks each back then, but I wanted to build something. I quickly soldered up a half adder using these chips on a piece of Formica, dead bug style.

Over the years I built many projects hand soldering huge boards of chips. In high school, me and three other guys (Scott and Steve) built Don Lancasters TV Typewriter II. We etched the board our selves, but without plated through holes, ended up soldering sockets on both sides of the board (Steve pulled his hair out troubleshooting the thing, eventually removing about half the sockets, soldering the chips in, on both sides). I built many expander boards for the various computers I've had in the last 40 years.

Tonight I soldered in the MAX31855KASA chip I got as a sample a couple months ago. Fun, cool, challenging. I cleaned the heck out of the soldering iron, had to refile the tip to get it sharp enough, and it happened. The first time, I got 4 of 8 pins, the second time I got 6 of 8, and finally with some heat, I got the other two. It doesn't look pretty, but it works, and no shorts. Mass production will require solderflow, if I am to use these chips.

Yesterday and today, I did some research. Yes, the Arduino supports SPI. There is a nice SPI reference library as well. The MAX31855 only does serial out, so I connected the pins like:

   44  - IO44  - Chip Select of the 31855
   50  - MISO  - Serial out of the 31855
   52  - SCK - Serial clock input of the 31855

I used my standard color coding, that I have been doing for over 30 years:
   Blue - Ground
   Red - Vcc
   White - Signal

Then the rainbow ribbon cable is:
   Red - T+ Thermocouple +
   Orange - T- Thermocouple minus

The results look a little messy:
   

I am a little nervous that the chip is at one end of the board and the SPI is at the other end, but that is the layout of this proto board. There is only one thermocouple input in this chip, so that also disapoints me, but who knows, working in eagle, and maybe it won't matter. 

I tested the K-thermocouple that I have using a soldering iron and my vold meter. It is hard to read the voltage in the meter, but it looks like 15mv:

I haven't written the code yet, but I'd like to as soon as I can. Maybe tomorrow.