The general purpose micro-controller board is based upon a Freescale (now NXP Semiconductor) MK20DX128VLF5 part. The part we believe provides the right balance in features, I/O, processing, and memory. A resident boot-loader occupies 16kB of the 128kB of program Flash. When the loader aspect is invoked the board appears to a USB connected host as a generic USB MSD and permits uploading an application to the board via a simple file copy. It is of note that being a 'DX' type part, there is also an additional 32kB of data Flash (aka Flex NVM) which is separate from, and not part of, the program Flash. The boot-loader is capable of supporting uploading to this area as well.
USB access is provided through a standard USB micro type B connector. For all other I/O, access is provided by way of plated through holes (aka PTH). The PTHs permit direct wire attachment if so inclined. Alternatively, they are arranged to support common 0.1" spacing headers. The bottom-side of the board provides easy identification of PTH I/O connectivity.
A crystal for the primary clock is provided. The main reason for this is to ensure a quality and compatible clock can be easily generated for the USB module within the part. It is possible to create a USB module clock from one of the resident internally generated clocks, however it is not straight forward as there are considerations that need to be taken into account. We feel the benefit of this crystal being on-board outweighs any cost/board-area penalty.
A (watch) crystal for the real-time clock module is also provided. We have found that the real-time clock module is being used more often than not so we feel it makes sense to have this capability right from the start.
The external reset is accessible via a PTH and also an on-board momentary push-button switch. The momentary push-button switch greatly eases use of the resident boot-loader (as opposed to the common tweezers shorting method of reset).
Power can be sourced in one of two ways. First from the USB connector. Second from a PTH (labeled VEXT). There is a diode in-line with the USB connector source. In the event power from both the USB connector and from the PTH (labeled VEXT) are present, the diode prevents any potential back powering into the USB connector source. The part has a wide operational voltage range. Depending upon the voltage of the power source, some features of the part may not be available (eg. USB).
2 years ago (Spring of 2018) we embarked on creating our own take on a general purpose micro-controller board. Our initial plan was to begin offering it in the Summer of 2018. Sadly at that time unrelated business came up and required our full undivided attention so the general purpose micro-controller board was put on the back burner. Summer of 2019 a low-volume production order came up for which this board is an excellent fit and so it was dusted off and brought into production. To date production orders for a specific client have consumed any and all units. Our intention is to also begin offering them here in the near future.
Stay tuned, we'll be posting more details about this incredible little work-horse.
We just received 3 lots of box style switches and 1 lot of speed style switches from Kailh!
Presently we are registering 0 for each style. That should change over the next day or so as we package them into bundles of 10.
I hope folks can turn an eye and pardon my product pictures for now. Photography is not my strong suit and the team member with such photography skills is currently tied up.
As the community knows, Kailh offers several varieties of mechanical keyboard switches. Being that we are just starting out, we had to narrow down what to initially offer. Our hope is that the choices we made for our initial offerings are correct. It seemed people would be interested in the box style switch, and to that we selected: linear, soft tactile, and click tactile types. Further we added to the mix a linear type speed style switch as well.
When the MCUs are assembled onto the rev 02 boards they are blank. Furthermore, no code, be it the boot-loader or the application has been developed. So with that will comes debugging and with that debugging comes the need to load the MCU Flash on multiple occasions, not to mention access to the MCU during debugging. Soldering wires directly to the target board test points of the MCU SWD interface is certainly possible. However in our experience it is a royal PIA and should be done as a last resort sans other options. Further one would need to do this to all ten prototypes. To address this we decided upon creating two jigs to support these tasks. The jigs permit interconnection of the J-link pod with the MCU SWD interface on the target board. We chose the SWD interface over the EZ-port interface because of the debugger support. We chose the SWD interface over the JTAG interface because of the lower pin count.We're using actual Rev 02 boards as the base template of the jig. Which is perfect because it ensures proper alignment of the jig pogo pins to the test point pads of the target board. Alignment is conveniently achieved utilizing the key switch center holes which are non-plated through. Thus there is no possibility of shorting layers on the target board. Phenolic washers are used as spacers to ensure spacing which is compatible with the height of the jig pogo pins. Further being non-metallic, there is no risk of shorting anything on the top layer of the target board (target boards are put upside down for mating with the jig). The hot glue gun got another opportunity for use. Hot glue makes for a reasonable stress-relief on the wiring at the J-Link pod ribbon cable header and at where it transitions with the jig board edge. Simple M-F standoffs are used and threaded onto the alignment screws to put slight positive pressure on the target board to ensure good continuity with the pogo pins on the jig.
Sounds like the name of a pub but it isn't. It's one wall of the dining room where we've set up shop with our manual SMT equipment to build the prototypes.
Yes you see it right there in the picture friends, high-end manual SMT equipment in the dining room with a shitty bench and stool. What gives? Well let me tell you what gives, Orlando area shop rent pricing. We even looked at purchasing instead. Nada. Unless of course one wants to overpay. What about proper benches and chairs? Yeah we have them. Disassembled and in storage. I was given an envelope I had to stay within or else. So under the advisement of my wedding tackle, the proper benches and chairs were sidelined and the shitty bench & stool called up for duty.
Where to from here? Well, the plan is to build up two units as well as two programming jigs. The two units will be utilized to validate the electrical part of the design before moving on to the software and mechanical fit checks.