Actual news and a Balthazar development progress
A short history of Balthazar:
Development of Balthazar as a possible and realistic device has started at the end of April 2019 as a result of the long sessions that lasted a bit more than a few weeks of conferencing and emailing while discussing "whitepapers" content and realistically putting it into the context with the various sides involved.
We have talked to various people, representatives and teachers in MIT, SiFive, TUFTS, Bauhaus University in Weimar and LEGO on a joint support in creating a working committee and a case for Balthazar.
Balthazar as a name for a laptop draws its associations from the Three Wise Man known as a Magus from Babylon and a Professor Balthazar, a 1970's cartoon character that creatively solves problems in cooperative and inventive ways, being helped out by picturesque early depiction of an AI machine, "Veronica", in a process.
The conclusion was that the Balthazar laptop has to be a magical and beautiful personal computing device incorporating ethos, cooperative, inventive, and innovative characteristics as a combination of those two historical characters whether they were real or drawn.
We are developing the final case design, also creating documentation and instruction manual assembly. All modules are now in delivery and a production to be put in a real working prototype. It will take some time to assemble and show in a video a real “alpha” working prototype, but mostly all our PCBs and other adjoining elements are ready. We have also updated our Github pages with new design and images of the case.
And here are PCB modules that we received, Unifying board, keyboard and PSU are produced. Those are ready to be built and connected to a working state, also as it is envisaged in the case layout. Keyboard is being assembled. More pictures of interconnected PCBs to come real soon. In the meantime, in the gallery below there are also tests of Balthazar's modules, screen, and a keyboard working with RP3. Production process of separate modules are to be found on our Documentation page, updating regularly and as things progress.
We will develop, once when ready, the final case design into rapid prototyping print-outs so we can have it 3D printed and ready for connecting and testing. Documentation and some nice new animations and drawings will be done.
Ordering all of the PCB pre-prototypes to construct first working models in various combinations. There are some issues with chips supply chain, but we do hope we can adopt. Also we are working fast towards clear design and positioning inside the case.
The major design of Unifying PCB has started and we design it to be an expansion and a crucial element to our device. It enables Compute module sized cards to be sandwiched on, giving some reasonable expansion abilities to Balthazar. Also we modified, debugged and updated all our previous modules.
We did a third version of our ergonomic laptop keyboard, still updating QMK and adding features. It is already on our GitHub page. There we also have a preliminary Unifying HAT board. It is done as a basic in its design, but we need to complete BoM and adjoining the documentation. This is due during this month or at least early next month. Issues with the case design we are solving on the go and waiting for final precise dimensions.
Starting the third year of Balthazar's development while getting preliminary case designs soon to be published on our GitHub page. Also, we are sorting out and updating all the documentation on a GitHub. We are also readying the third iteration of a keyboard with standard slim Cherry style click switches. Ports and I/O HAT-on module is still being deep in the design process. We are lacking many options there and are forced on changing basic design due to the continuous shortage, lack of a wide availability and a narrowed choice of chips. But, we will definitely get there.
We started to design a HAT module that will accomodate any standardized HAT-on CPU such as FPGA based such as ULX3S, ULX4M or Raspberry Pi Compute module, or any ther SoM, lining a groundwork to equip or co-CPU Balthazar with emulated cores of user's choice and also have better GPIO bridge to connect to cameras, sensors, Arduinos, synthesizers or any other IoT "things".
While design of the housing is being developed, we concentrate and work to get ready for all the internal elements. That includes battery, I/O power module, keyboard gets proper Cherry slim switches. We also see that to get towards designing HAT module we need to redesign back and upper right corner of the housing - also there is an idea to accomodate ThinkPad's keyboards for up to models series X2xx. QMK software is being updated and upgraded to a fully working version to accompany our secure keyboard.
We work hard on developing the proper multimodal housing, choosing natural, recyclable materials. So the primary focus starting this month also for some time to come, is on pure design, aesthetics, haptics, colours, look, and feel.
Very soon a new website connected to this wiki will be done for better communication to a wider public and decision-makers. 10, 20, 30 — no less, but no more.
As we partner along with Radiona from Zagreb, Croatia, and will design and produce an extension board based on FPGA ULX3S as our main coprocessor, we will also produce all our modules with them. Probably we will opt for a more capable chip there, but it remains to be seen. Exciting stuff, nevertheless. On the other front of various security issues and how to really trust the devices, we also talk to some SmartCard solution that would be inexpensive to build in/on and would bring Balthazar closer to a really trusted device. Also, Alibaba rocks as always, so we follow their RISC-V from T-Head Semiconductor Co., Ltd. chips.
We have committed to cooperation with rapid prototyping and PCB design Radiona makers place from Zagreb, Croatia. The ULX35 FPGA is in one way or another going to be added to our I/O board in form of a co-processor expansion to better handle GPIO, add WiFi and BlueTooth, and some other funky stuff like Amiga and C64 retro gaming, RISC-V cores, some rudimentary Linux and lots of other cool stuff.
Also with the help of Radiona, we will develop the "add_the_other_keyboard" module board - the idea there is to have compatibility with the ThinkPad series way up to the model X230 since those keyboards are rather inexpensive, they perform well and are plenty of on the market.
In August we have bundled our modules, keyboard and screen together and got them pretty much a working system running on NanoPi from OpenPi, Raspberry Pi 4, Odroid and Beagleboard. QMK software needs polishing a bit, also a small OLED screen for the keyboard we now need to test and to opt for the best solution. So far we are happy with how it looks (but we will need to work on a better versatile casing/housing design) and how it all performs.
Technical drawings are done for moulding/cutting and we will collect and choose a final candidate for speakers, mic and camera. Also, a keyboard will get its final QMK file with an early password entry system enabling boot into OS. Still waiting for RISC-V boards to be delivered. EU laws are now in full force of delaying and taxing imports prolonging delivery times, meanwhile, we search for other options - dreaming of a European chip supply.
We have done CNC/3D pre-prototype case elements getting to the phase where we can incorporate the screen and all the modules in. Also during the next few weeks, we will look into options having those mass-produced - choice of materials being most focused on. Happy with the overall progress, we now are closing in on finishing the design, colouring scheme, ergonomics of the keyboard functionalities and if not final then close to the final QMK files. The Prospect of having a real RiscV SoM running full-stack Linux is coming closer and we will also work on that making sure everything fits and works together.
While at the final construction and the design approach we try to fix some issues with connecting external modules. Also, we wait for the response on StarFive SoM that we might find as a first choice in the scarce market of RiscV SoMs.
The first casing elements were printed/CNCed. We need to pair and align them, determine proper screw holes, see that cooling is viable without the fan, see that cabling is neat and practical. This is the last of the pre-production design phase and we seek quality and simplicity.
We got both I/O USB and keyboard modules working and paired with a screen panel that we would use. Screen panels are tested and those seem like a good choice. QMK firmware is getting the last check-up and will be soon done and uploaded on our Github, however, there was a temporary lack of chips that we want to use, so we are running a bit later than expected - we are trying to catch up so we can move to another phase - which is to decide on a final housing design and hinge system. We have some mock-up layouts and speakers and touchpad layouts and internals.
We made a preliminary mockup with positioning PCBs into the first proper dimensions within the housing and now we basically can see the whole Balthazar slowly coming to a shape. When QMK firmware is finished and tested we will upload it to Github. Expected in this month is to have first screenshots of an interconnected keyboard, power supply, I/O USB, screen and password input before the boot into the system is done in QMK.
Power supply and I/O USB PCBs came in a bit later than expected, but that was fine as we were busy with QMK firmware and perfecting our keyboard. The screen is being mounted into the prototype frame with the hinges, also tested for resolution. We had borrowed 3rd party HDMI PCB drivers for testing the whole system using Raspberry Pie 4 and some other more open similar modules.
We wait for the first batch of PCBs that is merged into one - the power supply and I/O hub - as a single test module. Keyboard PCB is now in the testing phase and we will publish the results. While there are no real issues encountered in the first testing and firmware setting phase, we also wait for the silicone cover keys, but we were told that it might take time until we get our hands on it. Housing is getting milled, changing and adopting minor dimensional changes and so far we are also rather happy with the preliminary results and overall achieved dimensions. Also, a rather good choice is made on the speakers and a camera/mic module, so those are ordered as we wait for the packages.
Screens arrived and also some of the PCBs are on the way back to us. There is a preliminary casing being drilled and milled and also a hinge system looked into with few simple yet robust solutions underway. Soon we will publish photos and a first look at the casings and the other possible designs we play with and about. Also, the tests and results of our freshly arrived keyboard PCB and adjoining prototype modules will be published.
We have continued the development of a unified PCB that will serve as a hub to support various additional Systems-on-Module delivering power and all the I/O connectors needed. Also, we still wait for the first batch of prototypes of 2 PCBs and a special keyboard PCB. On those, we think also in merging those into the one as it will reduce the costs. Very soon we will have the first bottom chassis CNC-ed housing and cabling designed and incorporated into CNC housing. Screens are on the way as we have ordered various resolutions and types. We need to evaluate those first and see the larger batch availability and that depends also on our production partners. We do expect to have all those PCB received and tested during the end of this or the following month.
We have ordered a couple of screens that we might want to test and choose the one to use. Also, we will be travelling to Italy and Spain to meet with our development partners to discuss some aspects of the design and some clearance in the technical solutions. After mid-September, we expect to have the housing case and its elements and get all the PCBs ready for fitting and testing. We also look now into the options of some SoM to use for testing the modular elements. During the next couple of weeks, we intend to CNC some crucial parts of the housing from plastic and aluminium, adjusting the external and internal dimensions of built-in elements. Naturally, photos and detailed descriptions, with all the results and measurements will be published here.
While waiting for PCB samples to arrive we have been working on a few 3D models and overall designs for the housing. Also, we started to design further PCBs to accomplish and complete the innards of Balthazar. We will try during August and early September to CNC some possible designs and see what would be the best fit. Also in the coming weeks, we will look into the screens that we might want to use.
We had a short summer break. PCBs are back-ordered and now we are waiting for the prototype boards to be made and arrive apparently in late August. Also, we will see into the camera, speakers and other peripherals and order a few samples to test and see what would be the best possible fit as a prototype.
June is a revision month and we are applying to FSF for a check-up on our hardware compliance, also the first breadboard/prototypes will be made. And we are going to start working on the overall design incorporating also CNC prototype of housing with a fitting keyboard and a space for other modules.
May has started with finishing the designs for the I/O hub and a board that will enable a connection to our keyboard and power supply. I/O USB page was set up and files were ready to publish. We now have three original PCBs freely downloadable and ready for prototyping.
It has been one year that our idea has been born. We have opened a Github page with KiCad files ready to download. Also, our licence is now CERN OSHW instead of GNU v3.0. There we will be updating files and all our "deliverables". In March and continuing this whole month we have started to develop an I/O USB board and work on it to deliver files for our internal I/O USB docking board for the keyboard, power supply, camera, microphone, somewhere in mid-May.
Two PCBs keyboard and a power supply are ready real soon, we aim at or around March 15th and we will upload all the relevant data and files. So thereafter bit of C-19 crisis delay as we were caught with a quarantine we have delivered both PCB designs and schematics also worked towards opening a Github page.
We have started to develop 2 PCBs - a keyboard and a power supply. It is expected to be ready in a month also with PCB schematics, a Bill of materials and design guidelines.
With New Year comes also the new prospect of cooperation with Slimbook - a pure Linux laptops manufacturer from Spain. We are researching to see how we can incorporate our design proposals and align to have a joint development. We have also released the power supply concepts, looked into the price and the availability of the parts for it and started our encrypted keyboard PCB development. Also, schematics and PCB for the overall "modular upgradeability and add-on" concept is in development.
Some case design proposals are now in and we evaluate those, as we met with a few LEGO design people and got a good insight into plastic casing designs and heat issues. Also, hinges are a bit of an issue, but till the end of the year, there will be some viable solutions. We prepare the 2-day meet-up in Weimar, somewhat later than anticipated, but will happen in the next 2 weeks provided that all steering members can commit to attending. As it is an end to one very exciting year, we work hard towards having everything ready before the holidays. Some boards are in and received and some are still coming, so with the help of Armbian team, we will see and choose what the best options are there. Also thanks to IBM 's Open Sourcing move in Power9 ISA field, we started to look hard into that direction, too.
Somewhat longer emailing between SiFive and us on adopting RISC-V and ISA making it work with GPUs. We met with Armbian people that might help us with setting up RISC -V build tools and work deeper on HW encryption. Visited Tipro a well known European custom and industrial keyboard company with proposals and basic HW design and software. Keyboard's design and FW driven bus we will base on smaller RISC-V controller for bus-to-motherboard encryption and we will work on it together over the coming months. We have drafted a mutual agreement to publish it all publicly under CERN/GPL3 licences, more to confirm on that later in the year. More serious RISC-V development boars namely Freedom Unleashed are ordered and those are on the way, as well as a few of EOMA68 evaluation boards. We are also off to Denmark, Billund to meet with Moulding people from LEGO and also brief some representatives of LEGO Education on our current development. In the Weimar end of November latest mid-December at the Bauhaus Uni, we will gather in a formal first meeting with steering committee members and set the strategy and a path for the next few months of accelerated development. We are looking into CrowdSupply support and options we have there. One thing we do know for sure is - we do not want to over-promise.
We did the first renders and a few design studies of the casing, but more we will discuss with LEGO Moulding people in December. Also, some 2D and 3D drawings were made, still to be detailed before it goes to the CNC alpha case phase. Presented project by mailing a Turkish high-quality moulding company that is specialised in "electronic enclosures" for internal plastics and guidelines, cooling etc. Collected and decided on a few software models that are Open Source and rather powerful when combined. A German screen and eInk company was questioned for a second screen choice, with our requirements. Looks good, so far. We also got support from NLNET in a shape of technical support and a financial donation to speed up R&D and come up with tangible prototypes faster. What a great push that is.
A second public presentation, this time at an even more appropriate event, was made in Venice, Italy, during the ESC - End of Summer Camp where advanced computer and internet users, some HW developers, as per default, offered their participation in various aspects of development. An enthusiastic and very knowledgeable crowd. We collected ourselves an old and new bunch of Linux capable boards that we used over the years, some new micro-computers, SoC's and SoM's and started looking into software features and encryption modes and options that we can make use of it. Keyboard design and a secure keyboard to motherboard concept was done, even more drawings were made and put up for UI and UX also for the proposed educational interface. We talked with HiTechic again on the development of a secure USB dongle and GPIO boards design and they gave us an estimate and some overview of expectations for prototypes. Power supply design requirements are made and also battery and other issues (magnetic vs. mechanical switches etc.) were worked on until we all agreed on how what and when.
prepared a primary presentation, description and imaginary nice to have specs and a Bill_Of_Materials. A first semi-public presentation was held at the end of August on premises of Bauhaus University in Weimar to a working group consisting of 30 + Master and Bachelor students of Media Systems and with the help of Charles Wuetrich and his Chair for Computer GFX. There were tons of questions in the discussion and some valuable points were incorporated into the further strategy of the development.
All the security wish-features were added to our whitepaper style webpage, professor Chris Rogers from TUFTS helped us to make the case in educational aspects of Balthazar, HiTechnic's owner Steve Barker gave his views in designing and building Balthazar with tons of useful tips and contacts. Michael Anderson, ex-LEGO electronic engineer and HW designer guided some technical decisions sharing his production knowledge experience. We also applied for some minor external funding to NLNET, as it is not easy to make a new approach in designing the laptop investing just our financial resources.
A concept for a form factor came up and the first drawings and sketches were made. Deep dive into RISC-V ISA and foresight of viable future processors came up. Keyboard design was in the making, also thorough research was carried on, under the wise guidance especially in some technological approaches of EOMA68 developer Luke Kenneth Casson Leighton. That collaboration gave us some useful tips and helped us to find good reasons not to reinvent the hot water. We won't.
started preliminary talks to Tripodo started on Intellectual Property licences and issues for a famous XO PixelQi screen with a dimmable backlight and sun-readable features. That is still pending. The domain is registred and a website goes up with a rather white-paper content leaving an input open for contributors that came from all sides.
a day Zero, where we realised that it can be done, it should be done and we even kind of knew how-to.