Releases: dcmini-org/dcmini-hw
Releases · dcmini-org/dcmini-hw
SR7: All-in-one
Rather than spin two boards -- one DCMini and one frontend electrode board, what if some clever design allowed the DCMini PCB itself to function as a full biopotential recording solution including electrode snap connectors?
This revision supports three primary transformations by folding/snipping off parts of the PCB
- Larger footprint Referential recording: Snip off 1P/2P; populate all snap connectors on rear for a 4-channel referential Biopotential setup. Don't like where the reference is? No problem, snip off RB2 and put it where you want, connected via U.FL coax cable
- Snip off 3N, 4N and RB2; populate all snap connectors on front and fold all electrodes back under PCB for a small-footprint 2-channel differential setup
- Snip off all electrodes and PPG tongue; fold board in half and put into a box.
This change drastically reduces price, assembly burden, and thickness of configurations that only need 4-channels of electrophysiology (such as forehead sleep configurations)
Changelog
Breaking Changes
- FH35C 45-pin connector for FPC daughterboards has changed to a FH35C 51-pin connector to support actively driven shield output on this interface. Pinout adjusted slightly with better layout of AGND and DGND pins.
- BIASINV pin adjusted to buffered SHIELD output on AFE header
- Previously reserved unconnected pins are all now tied to ground. The tiny pads kept lifting off the flex PCB. This solution keeps them anchored.
- Moved nPM1300 PMIC I2C to a dedicated pair of pins (SDA @ P1.14 and SCL @ P1.13) so that communication with PMIC isn't affected if some other I2C device isn't properly powered up (solves a chicken/egg problem)
New Features
- Included electrode frontend
- Actively driven shield plane for better dry-electrode recordings
- Supported via a Buffer amp connected to the BIASOUT rail. Tuneable with unpopulated RISO footprint.
- Added a MAX30101 PPG for heart-rate measurements on a fold-back tongue
- Interrupt on NRFGPIO4; snip this off if you intend to use NRFGPIO4 on daughterboards
- Added a 2-channel 16-bit DAC for two analog outputs available on the EXT port.
- This was added on the same SPI bus as the ADS1299s for better or for worse.
- Interrupt on NRFGPIO7; Depopulate if you intend to use NRFGPIO7 on daughterboards
- Supports LIR2450 rechargeable coin-cell battery OR any battery pack you wish to use via a JST SM02B-SURS connector (tiny tiny)
Other changes
- Proper F.Stiffener and B.Stiffener (0.5mm FR4) layers for more reliable assembly and more strength when used without being folded up and put in a box.
- Removed most test points on rear to support this change. Many of these testpoints were on nets that are exposed via AFE and EXT headers; board test can still be performed using a custom breakout/test board that attaches to the DF40 connectors.
- Still have VBUS/VSYS/VLIPO test points but they're moved near the debug connector.
- Board programming can also be done via the EXT header further supporting that interface as a better test-fixture attachment point.
- USB-C port must be populated on front; as nice as it would have been to have a centered USB port when device is folded and put in a box, the port MUST have stiffener applied underneath which complicates layout and assembly.
- Removed an unnecessary 100 uF bulk decoupling capacitor on VBUS_ISO to make room for a battery connector, allowing the whole coin-cell/NFC/debug-port part of the PCB to be snipped off for some configurations
- Attached one of the PMIC's GPIOs to the nRF for battery monitor interrupt
- BOM Optimization: (Changes that may affect electrical specs in bold)
- Feedback divider on TPS61030 adjusted from 2.49M/270k (5.11V output) to 1M/(100k+10k) (5.09V output)
- Replaced high-side switch transistor for EN5V with a dedicated level shifter IC which is already being used to drive Neopixel. Part is overkill for this application, but this change reduces assembly cost and final unit price.
- Consolidated all 100uF tantalum caps on board to a 10V Tantalum Polymer (1.2mm height)
- Removed 10k ohm resistor array and replaced with 4x 0402 10k ohm resistors to remove a line item
- In patient bias drive, removed dedicated 1000pf 0402 capacitor and makes use of an existing capacitor on an existing cap array.
- Shield makes use of an existing ESD protection clamp as well as an existing 2.2k ohm resistor on a resistor array, with an additional footprint in parallel to adjust the resistance down
- Only one OPA196 is needed on the main board, but an additional footprint exists on Daisy for the buffer amp even though it isn't used for anything. It should be marked DNP, but Kicad doesn't support different DNP settings for the same part on a duplicated heirarchical block.