Okay, so honestly I'm attempting to do a large amount of work with D-DAQ being so small. A little ambitious, yes. By nature of that, I've not gone into looking at every fine detail of every component and every function simply because my head would explode.

Today's reason for this post is the MPL3115A: my barometer and ambient temperature sensor. Someone brought to my attention cold junction compensation for my thermocouple connection. I've intentionally not included any circuitry for this and will not use a specialized IC for thermocouples. The latter option is 2x to 4x the cost of the MAX4239 op amp which is more than precise and accurate enough for the job and is a little costly in and of itself. A quick read up on cold junction compensation shows that a thermocouple output is basically a relative measurement between the temperature being measured and the temperature that the amplifier. Well, I'm already measuring the ambient temperature of the mainboard so what's the big deal?

Heh, well, guess what is next to the MPL3115A? My 14V LDO. My hot 14V LDO. I'm considering this the final straw. I'm not going to delay prototyping as a result, but the "final" design of D-DAQs mainboard is going to be moving the power supply subsystem to a separate daughter board that will plug into the backside of the mainboard. I'll be moving to a different SMPS chip, most likely the ZSPM4022, and use and LDO for the 5V rail.

That's back to the original configuration. I previously failed to note that one of my specialized ICs is 3.3V driven, not 5V driven. This leaves just external sensors and various amplifiers needing a 5V supply to avoid rail-to-rail operation and allows software + hardware auto-detection on the sensors to work as well.

Supplementing the redesign drive is the PIC32MX chip. I'm using the 12 mm x 12 mm TQFP packages which requires 7 mil trace with and 7 mil spacing. Advance Circuits charges a premium for anything below 8x8. I originally chose to use this size package to conserve board space due to how small I wanted D-DAQ. Well, there is plenty of room for the 14x14 package. Changing over to the different package requires me to reroute nearly *everything* on the board. I don't have time for that right now, and because it's not a show stopper and the current boards will work, I'm filing this for a later date albeit at a high priority.

Having a 2 board design is conducive to fitting a fan on the final enclosure to help with cooling in really hot environments. As such, there are simply more benefits to this redesign and move to two boards despite the added cost.