At the end of August of 2019 we published the first version of the schematics in pdf format. Then, in October we uploaded the second version and after that the 13th of November we released the Orcad source, accomplishing what we promised.
Schematics Source in EDIF published and ready to be converted to KiCad
Now we have exported it even to EDIF format, to make easier for new volunteers to convert it to Kicad Format. To convert from EDIF to Kicad we have found edif2kicad tools https://github.com/svn2github/edif2kicad but we are sure you will find other tools or even you will be able to create a new one
Few days ago we announced that Slimbook will provide the enclosure we need for our Open Hardware PowerPC notebook. In addition to that, we have published in our repository the pdf containing the new version of the schematics. Further updates will arrive during the following days ( the Orcad source files will be available during October)
In our last post we stated “we hope to publish the initial schematic design before the end of February 2019”. Unfortunately a series of problems arise that we will try to explain in this status update.
As you may probably know, at the end of the first donation campaign we received four incremental releases of electrical schematics from the designer, the last one on the 11th of January 2019. We started a validation task at each release prior to their publication and each time we ended up sending back the schematics because we thought some design decisions had to be improved. These hardware reviews are quite complicated, and we would need additional volunteers, possibly hardware engineers able to properly address the task. Please, contact us or fill the collaboration survey if you can contribute.
At the same time, we were continuously evaluating various notebook chassis options, striving to find a suitable chassis that could fit our motherboard that has a MXM video card. The goal here was to identify an empty chassis that could be bought without a motherboard and with a life cycle of at least two or three years, which is the envisioned time-frame covering our laptop project.
The hardware designer that was paid thanks to donors is stuck waiting for the final notebook chassis as he requires the pinout specifications to finalize the design.
These are the reasons why the development of the electrical schematics is frozen since the 11th of January 2019. As a result, also our collaborators that have volunteered to validate the hardware schematics are stuck, as well as our selected product maker ACube Systems.
Identifying a suitable chassis is taking an unexpected amount of time and it is terribly delaying the second campaign aimed at the PCB design. Last year we had a meeting among both Power Progress Community associated members and the core group that is coordinating the project, and we all agreed to start the second donation campaign only after publishing the schematics resulting from first campaign. However, the amount of accumulated delay arise doubts among the participants on how to solve the current impasse, and we decided to ask the donors to give their opinion about whether to start now the second campaign (link) or keep waiting the publication of the result of the first campaign. If you are willing to actively support the project, feel free to express your opinion about the matter.
Our hardware volunteer team and ACube Systems are still reviewing the next round of documentation for selecting some key components that heavily depend on the selected chassis.
So far we have identified some potentially suitable chassis. However, some of them are designed to host a separated board to manage the batteries, a board that does not exists in our design, and some chassis have a dedicated space for a separated ethernet board, and again, others do not have a thermal design compatible with the heat generated by an MXM video card. The problem here is that the hired hardware designer stated very clearly that he is not willing to make an extra work to alter the initially agreed design features of the main board to accommodate it in a chassis requiring a very different internal layout. In other words, splitting functions in separated boards is out of question.
At this time the key question is: how much time is required to identify the correct notebook chassis?
Unfortunately, given our very limited spending resources and the experience we had so far, we are not able to make any speculation at this point. We strongly believe that it would be better to make a wise choice requiring more time, rather than an hasty solution now, because any wrong decision would risk to compromise the feasibility of the entire project.
We are not a private company aimed at making a profit out of this project in order to survive, in fact we don’t sell anything. We are a group of hardware enthusiasts more or less structured in a non-for-profit association. We are doing our best to face strong limitations to achieve what we knew since the beginning was a very difficult and complex goal.
Taking all of this into account, we think that publishing to a wider audience now a very incomplete electrical schematics could impact way too negatively our project and the donors expectations. On the other hand, we are conscious that we are delaying for way too long the promised outcome of the first campaign and this fact too have a great impact to our credibility.
We are moving forward extremely slowly, that’s for sure. You know the requirements and the extremely limited actions that we can make, mostly due to extremely tight financial constraints.
After all what was said in this status update, we delivered to the donors the current version of the schematics, stressing on the fact that they should be considered only a draft not yet ready to be shared with a wider audience
In the end, we still strongly believe in the project, we are facing a contingent problem (the chassis) that it will be soon or later be solved. If you can help you are more than welcome.
In this article we will report electrical schematics and donations progress and our current activities. Those of you who are following our twitter account and facebook page have already received updates about our work.We are mainly, but not only, focused on software development. In a nutshell, our developers are testing current Linux distributions, setting up the best emulation environment for a Qoriq T2080 machine using Qemu and even experimenting with Yocto Project in order to evaluate a custom made distribution.
Working on PowerPC GNU/Linux distro’s
Let’s start with the distributions first. As you may know, we stated from the beginning that our intention was to help to maintain Debian for 64bit PowerPC (ppc64) and include it as the default Linux OS for our notebook. This idea is still our goal but we have found more good options such as Lubuntu, openSuse and even Gentoo. In relation to this, our testers are creating a list of actively supported Linux distributions working on PowerPC and are evaluating the installation procedure, their performance and keeping trace of the issues found.
Another good finding was done by one of our members when compiling a custom and experimental kernel using 4.14 sources and GCC 7.3 on Gentoo in a PowerMac G5. As he reported, “gcc 7.3.0 is definitely an improvement! Both 2D and 3D acceleration do no longer freeze the card (AGP Radeon 9650 + 9800) on a G5. Also Firefox runs pretty stable now and less sluggish too (could be ’cause it’s running on accelerated Xorg now)”. In addition to this, we are testing newer PCIe video cards on a PowerMac G5 , using the Open Firmware command “boot-device hd0:,\\”, so without using yaboot and without tbxi.
Testing an emulated e6500 core in Qemu
Another effort is being done to setup a virtual environment to emulate our future notebook using Qemu. This is important as it will enable the development of software and to test, for instance, the u-boot boot process. Most importantly, it will allow any interested person to play around with our target platform and configuration, even if owners of x86 hardware only. Sadly there isn’t a ready-made QEMU machine configuration that fits perfectly our specifications, but that’s the reason we are working on this task in the first place. So far, it seems feasible to fire up a machine based on a e6500 CPU using QEMU version 2.11 using the built-in u-boot, but we still have issues making it boot a linux partition. It seems that lately the QEMU development mailing list (https://lists.nongnu.org/archive/html/qemu-devel/) is putting a great effort in putting the PowerPC emulation in a much better shape, making us confident that the upcoming 2.12 will be a life changer for reaching our goal (https://wiki.qemu.org/Planning/2.12). Stay tuned.
Yocto Project for PowerPC Linux Image
Recently, we are experimenting with Yocto Project creating a Linux image. We are evaluating this possibility as it could be an additional option for setting up a fine-tuned OS when the hardware arrives. Currently it runs an LXQT Desktop Environment and includes applications such as Otter browser, The Gimp, kvirc, qBittorrent, Audacious, Mpv, Qps and Speedcrunch. Now we are focusing on having working kernels for a Mac Mini G4 and PowerMac G5.
Custom PPC Yocto PowerPC
Video Card works on NXP T4240QDS devkit
The Borea team tried to setup a desktop based on the NXP T4240QDS development system. After not too much of a struggle he managed to get Gentoo Linux up and running with LXDE, and using an ordinary off-the-shelf ATI Radeon PCIe card and the T4240’s native networking using DPAA, so with no PCIe ethernet adapter. The Borea team used a vanilla Linux 4.14 kernel, a standard PowerPC Gentoo distribution and additionally, Cairo Dock to test the nice GPU effects to check if 2D acceleration was working. The result performed surprisingly well, and without any instability issues! For example, VLC was able to play FullHD Xvid content using 2 cores out of 24 available running at 1,6 GHz with Altivec enabled. LibreOffice 184.108.40.206 works nicely too. Unfortunately getting any web browser running seems to be a bit of a challenge. Firefox 54.x compiles, but gives a segmentation fault at startup, Firefox 58 does not even compile at all, but the team had just a quick try on that. Some packages runs into illegal instructions, most probably due to incompatibilities with the Power8/9 ISA, or probably not all apps correctly support Altivec ABI. Some work has to be done on these issues, but the overall impression using a T4240 in a desktop environment felt quite fast indeed.
Gentoo ppc64 on NXP T4240QDS devikit
We will be attending “Les journées du Logiciel Libre” on March 24th
Finally, we will attend the meeting “Les journées du Logiciel Libre” (“Days of Open Software”) the 24th and 25th of March 2018 in Lyon (http://www.jdll.org/). We will make our presentation named “Portable Open Hardware” on the 24th of March 2018 at 12h30 at the”Maison Pour Tous”, Salle des Rancy 249 rue Vendôme 69003, Lyon. See the map at https://www.openstreetmap.org/way/49680968
Les journées du Logiciel Libre
Don’t miss the latest issue of BSD Mag including an article by Saulo Paiva explaining our Open SourcePower Notebook project and an interview with Roberto Innocenti.
Our PowerPC Notebook project on BSD magazine
On the 29th December 2017 we submitted our project “Open Hardware GNU/Linux PowerPC Notebook” to the Italian “FunkyPrize”. Funkyprize is an award established in 2014 in memory of Marco Zamperini, an Italian professor that knew how to push the younger generations to an informed and cutting edge use of new technologies. The FunkyPrize aims to help filling the gap created by its premature disappearance, encouraging those who intend to pursue its mission, for a more mature, aware and widespread use of the Internet in Italy, and for developing the potential of the Internet as a tool for improving the quality of life and creating new forms of participation in the social and economic life of the country.
Sadly, we have not been selected as a finalist for the prize 🙁
Electrical Schematics Ready! waiting for the last 25% of donations
We just want to remember that our campaign is not targeted to make a product. We aim to create a freely available, well-documented and production-ready electrical schematics, allowing anyone to produce and, why not, customize a PowerPC based hardware.
If you are thinking on helping us, instead of one-time donation, you may want to consider a recursive, monthly donation. This could have a smaller impact to your account. Moreover, monthly donations allow us to better plan our activity, and most importantly, they let the engineer know the remaining period for reaching the final goal.
Please, spread our project and help us to make possible this amazing Linux PowerPC Notebook.
After reaching the 4000€ milestone we are ready to sign our first contract with ACube Systems. This will allow to start the engineering work and we want to thank all the donors for making this possible! All supporters are contributing to keep us moving enthusiastically forward. Of course, we still need to achieve the 12600€ goalto end up the analysis phase and to produce the complete electrical schematics and deliver an extensive documentation, but we feel confident in our long term success.
In the meantime we are making progress on our T2080 CPU based development boards. We are working on booting them with a Radeon graphic card, so please, if some uboot expert like to contribute, please contact us .
Regarding the Radeon card, a few days ago ACube confirmed the inclusion of an MXM connector in the notebook. This is a very good point, as it gives us the possibility to deliver the basis for an upgradeable laptop.
Finally, we would like to end this article reminding that any help is more than welcome. New members joined us to collaborate with the hardware team (i.e. the above mentioned boot process) and many people are contributing in spreading our campaign in the social networks. We need to reach more donors as they will help us reaching our goal and to release a truly Open Source PowerPC Notebook. So please, tell everybody about this project on forums, social networks, and why not, have a chat about it with your friends during these summer evenings.
This first campaign reached its goal! Thank you all!!
And now do not stop donating, even if the next campaign “Printed Circuit Board (PCB) Gerber” is not officially launched yet, any new donations made here will be automatically transferred to it.
The next campaign will be launched when we will publish the outcome of the first campaign that will be soon provided by the hardware designer ACube Systems. See here a dedicated post.
About the Linux PowerPC notebook project
We really want to make it happen: a PowerPC notebook released as Open Source.
To achieve this goal a series of tasks must be performed, such as designing a custom motherboard that would fit into an existing mainstream notebook chassis.
Since October 2014, when the project was launched, the number of people showing interest had been steadily increasing, and we are now quite a few experience volunteers, some of whom are able to contribute in reviewing and checking aspects of the required electronic design. Unfortunately, after all this years we are stuck, because of the lack of either spare time, or engineer capabilities and professional skills.
In order to solve the situation, in 2016 we approached the Italian firm ACube Systems a company that have some experience in designing PowerPC motherboards. We were lucky, as we found a group of passionate people that shared the long-term advantages of the Open Hardware philosophy, and their prior experience in designing a variety of PowerPC motherboards makes them an ideal choice as a partner. Together with ACube we will be able to get the electronic design done, but a professional electronic designer has to be paid.
At this point the group involved in the project decided to get serious and launched a fundraising campaign to pay for the required job.