Added an RLD circuit for better ambient noise suppression

Moved to something closer to semantic versioning (ditching hotfix because it doesn't make a ton of sense for hardware in my opinion).

In addition to RLD, we've also changed some of the chips and the power supply is now a tighter design. There are already a few issues that I have found in initial testing, so I wouldn't immediately adopt this design. Once I get the rest of the parts in for it, and more fully test this design there will be a new release with all the fixes. It should be pretty quick after this one.

Full Changelog: va.4-dev...v1.1

Built in electrode dc offset removal, stabilized gain resistor, added current limiters

DEVLPR:dev.a.4

This release is mainly minor fixes and improvements. The most significant is a redesigned reference for the instrumentation amplifier with the goal of dynamically removing the offset introduced by the EMG electrodes. This is essentially a high pass filter targeting ~1Hz. We have also added 100k Ohm resistors between the electrodes and the power source, more strictly following medical instrumentation guidelines. This should allow for users to safely power the board off of a mains supply. Lastly, we have added more test points so that the board can be debugged easier and the raw signals can be used by other, external devices.

Reworked Power Supply, New Instrumentation Amp, Fixed Bad Math

The global IC shortage has necessitated a redesign on a couple fronts. The power supply on the previous release was still giving us trouble, and it recently went out of stock so it seemed like a good opportunity to re-examine what we needed from it and try something new. Now it is a -3V line, using a voltage pump followed by an LDO.

The second major change was for the primary amp, the INA333. It is now out of stock, so we used a pin-for-pin replacement: the INA818. It also has a gain of 10000 V/V, hopefully an upgrade over our previous limit of 1000 V/V.

Lastly, we're still testing designs for our silkscreen to create something that's fun but also reflects our commitment to open source hardware.

Improved power supply, silkscreen, and filtering

This release brought together the results of filter tests done on previous iterations. We were able to merge the 2nd order filter into the actual level shifter, removing the need for an additional op amp and making the design feasible both from a space and cost perspective. We also incorrectly had an LDO attempting to output -5V from a 5V supply (they need a little head room!), and we fixed this by changing R2 to 182k Ohms giving an output of -4.5V. Lastly, the silkscreen has been updated to better highlight API documentation as well as emphasize our commitment to open hardware.

Our previous board, dev.a.1, was the first that actually detected and amplified EMG successfully, so this board is mostly improvements on that design.