     |
 |
. . . česky . . . deutsch . . . english . . . español . . . français . . . italiano . . . русский . . .
|
|
|
|
| 2011-11-15 ACP proudly presents: the quarterly report ;) |
|
-) Finished developments
-) Parts of the project that are in development
-) Further debates inside the team
-) Public perception of the FireBe -) Further Informations
MiNT Ethernet driverOne of our most important new developments is an FEC Ethernet driver for MiNT. This supports us with fully functional use of networking features via the FireBee's onboard RJ45 connector - even multitasking is working fine. As MiNT is a Posix compatible OS we now can use the wonderful world of networking. While it has been possible with FireTOS to use sFTP and STiK-Clients for some networking services via a proxy we now have the possibility to use all existing MiNTnet clients. The most spectacular feature is to directly surf the net with the browser Netsurf. Further protocols or services are now ready to use. The networking driver has been ported from FireTOS to MiNT by our team member m0n0 during a heavy month long working schedule and thus covers one of the biggest gaps. Networking with MINT is the most wanted feature of the FireBee.MyAESMyAES is finally fixed for FireBee, thanks to the team work of Vincent and Olivier. MyAES is an alternative and quite modern AES, which shows some fresh optical stuff for the world of our Atari machines. MyAES is now optimised for the FireBee, which meant several weeks of heavy programming. Some bugs caused quite intensive hunting and fixing. The cache usage that seemed impossible to do under MyAES couldn't be fixed until Olivier Landemarre teamed up with Vincent Rivière to get it done. Some smaller fixes within MiNT became necessary as well. In our opinion the results of this development are impressive ...
EnclosureThe development of our small custom enclosure is finished. All of your wished changes have been incorporated. The enclosure comes, for instance, with an additional parallel/printer connector, and with a SATA port which can be used either internally or external as eSATA. Furthermore there have been incorporated some optimisations. For example, the enclosure now comes with an internal 2 1/2" slot for hard discs. It will be produced, as requested from you, in four colours. Dark blue, grey, pink and black. The definite price will be confirmed to all advanced bookings. But to hold the price down we decided to make no further prototypes with the changes, but to order directly the first production run. The enclosures will be made in Vienna and delivered to Switzerland in the end of November.DVD drivesThere have been PACKET commands (Atapi for SCSI commands) added to the SCSIDRV of the TOS. Therefore it's now possible, with the use of the software Extendos, to use DVD's with DVD burners.USB Keyboard on MiNTBy patching the MiNT ikbd_scan routines, which on MiNT do not have the same format as on TOS, USB keyboards can now be used with MiNT as comfortable and as fast as with TOS. In the course of these patches there was also an update for the initialisation of AC' 97 via basic system (BaS).NVRAM ClockThe clock inside the NV-RAM is, since the beginning of October, now working very well on the FireBee.New FireTOSAside from the already mentioned ID interrupt routines as well as the AC'97 initialization, the PCI-BIOS got "Bus Support" added, with which the stage for using PCI Express adapters is set. The new limits are: four buses, thirty-two devices at each bus and four features per device. Moreover, all possible IDE-calls inside the boot routines after detection of SCSI devices, which could be not valid, got fixed (attention: SCSI is still not implemented). ATAPI devices got added to SCSIDRV everywhere, IDE devices in contrast were removed. These are now enabled by Inquiry commands at scanning during the boot of SCSIDRV. On the FireBee now, when the boot menu entry "TOS 4.04 for MiNT" is selected, the FEC driver for Ethernet is disabled, to allow use with the new MiNT-net FEC driver.A new FIRECONF.CPXThe control panel FIRECONF.CPX is now available in German language.zViewThe program zView, thanks to the recent work of Miro Kropáček, is now available in a version for the FireBee. That means PDF on the FireBee! zView is software for searching and viewing files from mass storage devices via an overview browser. There can be displayed and converted Bitmap and Vector files like JPGs, TIFFs, Degas, BMPs, Eureka, GIFs, IMG, NEOChrome, PNGs, TGAs, and as mentioned, PDFs. Using zView means that LDG has had to be adapted and updated, and all Codecs had to be recompiled. The upshot is a 100% ColdFire version of this (by Atari measures, comparatively young) software.
New EmuTOS VersionThere was a new release of the GPL Operating System EmuTOS. Version 0.8.6 of EmuTOS comes to us thanks to the sedulous work of Vincent Rivière, the first version with official ColdFire support. Many improvements, like usage of file systems up to 2 GB, have been added to this first release since 2010. EmuTOS is likewise (as a free alternative) preinstalled on every FireBee. The release from November the 8th 2011 is ready for operation on the FireBee.-) Parts of the project that are in development
Linux developmentGreg Ungerer has done several Linux patches that provide support of ColdFire cores and especially for the FireBee. After uClinux is up and running on the FireBee, the aim is then the adaptation of Linux-m68k. These are the first steps for complete Linux distributions (server, desktop, etc.). For more information the linux-m68k mailing list is recommended.NetSurf for FireTOSThere has been some work for Netsurf-browser on pure FireTOS (Single Tasking). This has no high priority inside the team, and shall just be mentioned for the sake of completeness.ATX power supplyAfter some discussion and consideration of the use of ATX power supplies in bigger enclosures together with the FireBee, we decided to develope our own small add-on PCB to use ATX power supplies together with an integral FireBee computer in standard enclosures. This PCB will contain a micro controller, some plugs and few components. The advantages are: the use of the FireBee in different enclosure situations is easier with no need to relocate the solder bridges, also usable with passive PCI backplanes (see below), full functionality of power and reset switches in ATX and mini-ITX enclosures, is immediately usable with no need to burn the new PIC micro controller on the FireBee (also on already delivered computers), and finally, is useful for further ATX power supply tests without ATX motherboards. This realisation has already started. The small PCB will probably be completely made within the Atari community and cost only a few Euros.Onboard PIC microcontrollerThe code for the PIC micro controller residing in the FireBee was recently under development by two people. As the electrical power supply is also operated by the PIC, we are waiting patiently for the official start of delivery, until this development is finished. Presumably there is, besides the expansive ICD-PIC-Burners, a cheap Open Source alternative that is working likewise together with the FireBee, for burning the PIC. We'll keep you informed.-) Further debates inside the team
USB-Floppy DrivesWithin the U-boot source code, which is providing the USB functionality of the FireTOS, there is already support for USB-floppy-drives, which are sadly not normal "mass-storage" devices. Thus we are of the opinion that it shouldn't be too hard to provide a driver for USB-Floppy drives.Infogrames USAWe've been witnesses of a very disturbing initiative from the current Atari IP holders, Infogrames, against several people of the non-commercial Atari community, mainly from the 8-Bit and Console fields. But fortunately after numerous complains from the community they apologized for the dissuasions and wrongful claims. Even though the - especially in that appearance - pleaded incriminations are untenable, and notwithstanding that our project was not hit at all, we hope that Infogrames will in future focus on positive coexistence with communities of all fields and their more as 40,000 members. Be it consoles, 8-Bit or 16/32-Bit. Particularly it is a matter for non-commercial initiatives, that have been partly active since 1994 - meaning 9 years earlier than Infogrames - in the Atari area. We would like to wish the affected persons further enjoyment of their Atari hobby, and want to assure them of our solidarity.Missing VHDL DevelopersInside the ACP team there are recently not enough VHDL developers. We have discussed some possibilities how to get new people on board, who can pursue in their spare time opened tasks like for example the unfinished DSP. Sadly we do not have a satisfactorily solution for this issue, and as well, no resources in this direction. We would feel happy about some support concerning this matter.CompatibilityThe compatibility between ColdFire CPU and 68k CPU until recently was mainly achieved by the freely usable CF69kLib and a few patches inside FireTOS. This is working in many cases, but not in all, and it is partially comparatively slow. From the beginning, it was clear inside the ACP team that the compatibility has to be strongly increased, to be able to let most Atari applications run. Therefore already in 2010 the development of our own compatibility layer for the FireBee was started by Medusa, and it was planned via BaS. But after the successful implementation of CF68kLib directly inside the FireTOS this development stood still for some time. Now we have decided to again go about the further development of the emulation of the missing instructions. This will be mainly realised by branch tables, but for certain instructions - that are behaving differently - like LineA, it will use a JIT-compiler. With our own completed development we expect in the medium term a highly improved compatibility and as well, increasing of the speed of programs that make plenty of use of 68k instructions, no longer present on the Coldfire. It will be realized as a library that can then be used by all Operating Systems like TOS or MiNT if required, and that it is providing the ideal speed of 68k applications on the ColdFire. For this further development the well-known developer Miro Kropáček from Bratislava could be won. Subsequent to its completition, the CF68kLib can be completely replaced by our own development. Theis work will begin in the middle of November.Flash LayoutThere have been some discussions about the layout of the internal Flash Memory, which has an amount of 8MB, inside the team. Recently there is the BaS, FPGA configuration and EmuTOS as well as FireTOS residing inside. We have now talked about how dBug - a free boot loader from Freescale - can be integrated into the system, and how all parts can work together in an optimal way. The integration of dBug benefits are, that there would be some valuably tools onboard at the disposal of developers, and that Operating Systems like Linux could be booted directly with it - allowing, for example, boot over the network. Presumably we will allow the BaS now to take over a few basic initialisations like RAM ad FPGA. Subsequently the BaS will start (or not start) one of the various compatibility layers that have been selected with the DIP-switches, and than pass to one of the Operating Systems like dBug, EmuTOS or FireTOS. This requires some modifications to dBug, but after finishing, should mean an ideal collaboration between all components and a maximum flexibility regarding Operating Systems. Even after the integration of dBug there will be some free space inside the ROM (Flash Memory) for more unforeseen developments.-) Public perception of the FireBeThe FireBee was mentioned a couple of times outside of the Atari community in the last few months.
-) Further Informations
GEM DemoThe GEM Demo is now working complete and without any problems at the FireBee. The values for the graphics inside the FPGA are:
87 (1x), 75 (x2), 18 (x4) at a resolution of 1600 x 1200 @ 16 Bit Passive PCI BackplaneFor people who want to use their FireBee with a PCI-bus, we now have a favourable PCI backplane at our disposal. The passive backplane has four slots, is produced in Germany and will be slightly modified by us. It is tested and working very well together with the FireBee, and the price is below 20 Euros. The availability is assured.Active Radeon Graphic CardsFor people who want to use Radeon graphic cards inside their FireBees, we took a look around for recent manufacturers. At the time being it looks like we´ve found a long-term partner, who can deliver 9200 PC graphic cards for below 40 Euros. The model for 3,3 Volt PCI buses would be immediately available. However, we let produce a card for 3,3 and 5 Volt PCI Buses so that this card can be used at all Radeon-enabled Atari-systems. The specifications are: 9200 PCI card, 3,3 Volt and 5 Volt PCI bus, 256MB RAM with 128 Bit interface, GPU/RAM with 200MHZ/325MHZ, ATI Chip 215 Radeon 9200, low profile, passive cooling! The cards are available for production the upcoming years, and will be built after our orders, within 14 days from a German company. The Radeon 9200 is thereby faster than the 9250 model, which was released later. A first sample is on its way to us.3d LabelsFor the upcoming enclosures, we also have 3D lables provided free by an Atari-User. Tom Kristensen from Denmark has an advertising agency, and is working daily with his Hades Atari Clones. These labels were done with Calamus on a Hades computer!
|
| 2011-08-23 ACP-News-August-2011 |
|
NVDI patched for the FireBeeOn the firebee now a patched NVDI can be used. This makes it possible at all programs that require NVDI (eg Texel, Artworx, Papillon), to test for the FireBee. The GDOS functions of NVDI now enable the management and use of different fonts. Theoretically, the printing functions should also be usable functions. This has not been tested yet. The hack has been tested successfully with the version 5.00, 5.01 and 5.03. Since NVDI no free software, and it is not also not to acquire new, it is not possible to offer a fully patched NVDI (legally) for download. Each user must patch its own version of NVDI. For those who shy away from using a hex editor, Ole Loots has written a patch program. Further informations and dowloads are on the german newtosworld wiki. There is also a german tutorial for it in the how NVDI edited by hand.
cURL on the FireBEEOle Loots has been ported cURL to the Firebee. CURL is a small program that allows it to transfer data from the Internet to your local computer. In combination with FreeMiNT, XaAES and bash it runs very well. Unfortunately you have to make compromises because the direct download is not (yet) possible. As a workaround Ole recommends a local proxy to set up and let it run cURL. The file is availaible for download at curlfire.zip.
GPG for AtariBernd Mueller ported GnuPG for MiNT. GnuPG is a free implementation of PGP (Pretty Good Privacy) and was designed to encrypt and decrypt files and emails as well as create and verify electronic key. Thus, the possibility of creating a modern, secure encryptionsystem to have on the Atari. Evenh ere at the moment is a little manual work is needed so that it runs reasonably well, there is currently no graphical interface to do so. You need an texteditor like QED, joe, nano, ... This sample describes ho to get it with aranym and easymint:
gpg-atari.zip
Bugfix Release of AHCCHenk Robbers has released a preliminary bugfix release which fixes some FPU related errors in the Coldfire target. Who likes to get this special AHCC, should contact him by mail http://ahcc.atari.org.
New TeammembersHeinz Schmidt and Marcel Schön has joined out team this month. Heinz and Marcel are working together to create the howto section and the documentation. Beside that, Marcel is goint to port some Linux Software.
New Betareleas of FireTOS and docs to fireconf.cpxA new public beta release of the documentation & FireTOS FireConf.CPX: On his Web site, Didier Méquignon has published a page dedicated to FireTOS FireConf.CPX. There he describes the settings that can be changed using FireConf.CPX. At it it´s possible to discover the progress he has made with FireTOS in recent months. On his site there is also the current public beta version of FireTOS of 27 July 2011 available. All the enhancements and additions as bugfixes that were described in the last ACP News, are included.
USB improvementsDidier Méquignon has recently fixed a long-standing problem with USB that prevented mass storage devices from working properly. The problem was caused by a hardware bug in the Coldfire chip itself, but could fortunately be worked around in software. As a result, it's now possible to access USB mass storage device in FireTOS and FreeMiNT. With FreeMiNT it's possible to access drives and sticks formatted with FAT16/FAT32 on a PC, which makes exchanging data with a PC or Mac easier than ever.
FreeMiNTWe have now reached an important milestone in the development of FreeMiNT for ColdFire. Vincent has patched all the assembler sources for ColdFire in the kernel, XaAES and the most important modules. Changes were made at 337 places. Everything has been committed into the official FreeMiNT CVS repository. As a result, the current kernel mintv4e.prg which can be found in the daily trunk Archiv, is fully optimized for FireTOS and its 68060 emulation. EmuTOS support will come later.
|
| May 2011 A status update |
|
The past month was characterized by a variety of work on all fronts of the project. Much of this work has been for end-users but rather uninteresting and worked in secret. Nevertheless, it is time to deliver a bigger status update and publish the milestones reached.
The New FireBee LogoThe most important recent event of recent weeks was the final decision on the FireBee logo. After a very lengthy process, the team was finally able to agree on the final appearance, so we can officially present it.
The new FireBee logo is not only our future case decoration, also serves as a brand recognition for the FireBee. We hope it will serve as sign the Atari has not just merely survived, but has continued to evolve with modern technology. We want to warmly thank all who have supported us with proposals, designs and inspirations. The new FireBee logo is a available for download as a vector graphic. color.eps color.svg black.eps black.svg
Welcome to the TeamAnother imporant item is that we would like to welcome and introduce new members to the team. Geir Øyvind Vælidalo - he was a supporter from the beginning of the project. Since he received his FireBee, he decided to join as a developer and porting to TTF GDOS for FireTOS. Jean-Pierre Feicht - a relatively unknown face and late arrival - but just at the right time - responded to a public appeal in 2010. Jean-Pierre is an independent media designer and took over as the artist of the promotional artwork and design of our new FireBee Logos. Currently, he is responsible for the implementation of firebee.org product page and will now also take on a further part of the resulting design tasks. Marc-Anton Kehr - thanks to his work with Easy-MiNT, he will be know by most Atari users. First, a new Easy-MiNT version will be prepared for use with the FireBee. He has also afreed to particiate on the long overdue implmentation of the firebee.org product page and some of the technical details. Ole Loots - actually an "external" developer who cares about the port of Netsurf browser. After delivery and the adaptation of his FireBee Netsurf, it made sense to expand to technical cooperation of OS developers and application programmers. Therefore, the two projects and ACP Netsurf port together are now a little closer. m0n0 now is advisor to the ACP, however, he has also been a very helpful in some important ways beyond the browser.
Technical progressThere has been a lot of work on FireTOS. Since the last news we have received four new updates. FireTOS now offers an improved basic USB support directly in the TOS, which makes use of mass storage devices by hot plug possible. Booting with USB devices is no longer necessary. Bugs caused by GCC problems in implementation of lwip be eliminated. (Now these nasty little bugs have now taken root in the AC'97 initialization). A very large piece of the progress was with the XHDI support. Random FireTOS crashes and incomprehensible loss of data under MiNT NEWFATFS have kept our developers busy for 15 days. After eliminating this problem, the XHDMI implementation can now be optimized. FreeMiNT has been improved to work with the FireBee. This includes, among other things, the optimization for XaAES on V4e processors. This development took more than a year to complete. The 100% optimized version of the ColdFire XaAES kernel modules also brings a great advantage for other processors. Where some parts of XaAES written for pure 68000 processors, all parts of XaAES can be compiled and optimized for each processor type (eg 68030). Internally, of course, there have been many other issues and developments in the pipline. For example, the flash layout, more GCC 4 Bugs, DVI + VGA resolutions and the sync polarity, BDOS and IDE issues, the possible offering CodeHeads GDOS+ as open source, the use of fVDI or oVDI (largely unknown), the possible implementation of a parallel port, debugging HDDriver, preparing a public program compatability list, the long overdue update of the websites, enhanced capabilities of the PIC icrocontroller, etc. In closing, we still have a very positive message. In the very far-reaching environment of the ACP, all programs of Sven Kopac, whose sources are still available under a GPL / Artistic License provided and made accessible. The binaries are now freeware. This includes for example Emailer, PPP-Connect (with CAB.OVL), Income, Alice, 2nd Chance.
|
| 2011-04-13 USB news |
|
ACP team-member David Galvez has for some time now been working on USB-support in FreeMiNT. The result of his work so far can be seen in this video:
As Download as AVI (XviD) What you see here is David's USB stack running as a kernel module in FreeMiNT on a Falcon with CT60 and EtherNat. A Mass Storage Class driver with XHDI-support has been implemented, which in plain words means that FreeMiNT can access USB drives with Atari or MS-DOS FAT16 or FAT32 filesystems, with or without VFAT (long filenames). It can also access Linux ext2 partitions, or any other filesystem that might be supported by FreeMiNT in the future. Exchanging data with mainstream computers has never been easier As the video shows the USB drives can be hotswapped, and multiple drives can be used at the same time. The USB stack has a modular design where both the class drivers (that is, drivers for the USB devices) and USB interface drivers are implemented as external modules. The correct class driver is automatically attached when a USB device is inserted. Thanks to this modular design David's work will benefit all FreeMiNT users with USB hardware.
|
| 2011-04-06 Thing Desktop goes GPL |
|
The ACP Team is proud to announce, that the very popular desktop for Atari environments - Thing which was originally done by Arno Welzel - will be freely distributed. With the cooperation of the Thing authors we have improved the number of available Open Source desktops, which has blossomed into a long-term cooperation. Thing is, and along with jinnee, are the best existing desktops with the largest user bases, and it was not until recently a commercial product. As the sources of jinnee, according to different information sources, are lost, and hence any future development will be impossible. It initially appeared that the FireBee would be limited to the Open Source desktop Teradesk, as far as it concerns native ColdFire versions. The authors of Thing however did support the development of the FireBee very early, and did not just agree to freeing the binary, but also publishing Thing as Open Source. In the meantime there is a version of Thing 1.29 without the registration dialogue in existence, which was built by Dirk Klemmt. The Thing developers work right now on the following approach: First all the sources (1.27, 1.27b, 1.29, 1.30) merged as far as possible, and the code will be revised. Subsequently there will be the publishing of a free version for 68k without registration or any other limitation. Some time later the sources of Thing will be published at our CVS repository at Atari-Forge. Further development by other ACP members is being considered. Even though Thing is already running fine as the pure 68k version, a native ColdFire compilation would again bring a noticeable speed increase. We are pleased about the Thing programmers and their extended cooperation, and that we can together hand the Atari community this gift today. Thing is even working at the very small Ataris and therefor turns out to be a benefit for all Atari users, even without the FireBee. Here now, for the immediate release, the public key: user "PublicACP": ochl Piie juQd Kg6h kfUW ZiTh zbBr UHbx
|
| 2011-04-04 FreeMiNT is coming to the Firebee |
|
Thanks to the work of Olivier Landemarre, Vincent Rivière and Didier Méquignon, FreeMiNT and XaAES are now running over FireTOS on the Firebee. It's far from finished, the kernel is still partially patched for ColdFire, but as the screenshot shows it is already capable of running GEM applications surprisingly well.
This shows a number of applications running under the latest XaAES. All of these applications are working as expected, and despite the relatively slow FireTOS VDI everything feels snappy. BoinkOut2 even produce sounds. ;)
|
| 2011-03-10 Cooperation for using GNU/Linux |
|
The ACP is proud to announce a cooperation with Greg Ungerer to get FireBee running under GNU/Linux operating systems. Greg is a core maintainer of uClinux project and has worked for many years on Coldfire hardware. uClinux, the "Embedded Linux/Microcontroller Project", is a well known port of Linux to processors without a Memory Management Unit (MMU). Greg has taken first steps with a FireBee we sent to Australia, and was able to boot uClinux via Freescale's dBUG loader, loading system images over the ethernet port from another computer. Getting a simple uClinux configuration running on the FireBee was a straight forward job with a basic peripheral set supported. Greg will work torwards getting full Linux with MMU support running. With that in place complete desktop GNU/Linux distributions (like Debian for example) can potentially work on our hardware. The ACP highly welcomes this opportunity to work together with developers of "external" operating systems, especially Open Source ones. We will keep you informed about the progress made in using Linux at the FireBees.
|
| 2011-03-04 Hold yourselves tight! |
|
Within the last ten weeks so many things worked out that we stepped behind with releasing our news. Most of what you'll be reading now would have been worthy of getting an entry for itself just to be appreciated properly. Nonetheless we hope not to beat you down with the sheer amount of text. Enjoy reading!
Delivery of the FireBee in the end of 2010 would have been possible, but unfortunately we're still working on completing the software for the normal user. After a small inofficial enquiry on Atari-Home and via our mailing list we decided to rather wait for some months instead of releasing a product which isn't working all over, and thus might frustrate users. So we delayed the delivery for end users. People who can cope with JTAG and BDS, or who want to join the development certainly may order their machine at any time. Just drop us a mail. And now for some recent developments: Already in November we found a problem with the current supply of the USB chip on the PCI bus. This bug has taken quite a lot of manpower and debugging time but in the end could be solved in a very easy way. Every FireBee will be equipped with a 1000uF condensator connected near the accumulator. PCI is stable now although USB is not perfect due to other issues. Just before the turn of the year solutions for the USB chip (Chip errata 35, I/O access write cycles) and for the Radeon 9250, which also is sitting on the PCI bus, came up.
Another field which we cultivated was the use of different external hardware which is needed for further development, e.g. PIC programmers or BDM adapters. The latter had to work flawlessly via parallel or USB port, which needed a lot of testing in December. Two very important issues which should be compleated before delivery are the flashing software and the code for the PIC microprocessor (for the PS/2 interface and other things). The PIC coding has been taken over by Alan Hourihane who is well-known for "Eiffel". Alan ported a comprehensive amount of code (which had been built by MCS within summer and autumn) from assembly to C. Work on the PIC goes on permanently. The flashing software has been completed in the end of 2010. Didier adapted his flashing software for the CT6x in such a way that a TOS update for the FireBee ROM can be done via TOS. Additionally he adapted Xcontrol and CT60conf which both now work on the FireBee. FireTOS is working flawlessly now with different languages, boot orders and other things. The software can be downloaded from Didiers pages as GPL. Point your browsers to http://ct60conf.atari.org and http://ctpci.atari.org . FireTOS itself has undergone no lesser than five updates which were done after the last news release. Fredis SD-card driver has been implemented into FireTOS, and the BDOS limits got pushed. The SD-card driver is working without the FPGA! TOS got general networking capabilities via LWIP. Protocols such as TFTP, telnet and others are possible. Three flash parameters were added to TOS to be able to configure ethernet if the FireBee is started from BaS (base system) and not from dBUG. The ethernet MAC address can be changed from a CPX now. Further on FireTOS got basic USB capabilities, so that mass storage devices and mice can be used already! By the end of November IDE has been approved too: Now three IDE devices can be used with 32 bits data. The CF-card is implemented at 0xFFF0'0000-0xFFF0'003F just as IDE on all IDE capable 16/32 machines. An FPGA firmware update opened up the possibility to swap the IDE buses. Since mid of December FireTOS is able to use the monitors EDID list of supported resolutions and offer those to the user. TOS can on the fly change between several high resolutions. We published our first 'snippet' showing that. It can be downloaded as MPG4 or viewed as stream. This became, amongst other things, possible because of extensive work on the FPGA, which took place the whole December. The ST-resolutions were improved. "Video clock PLL" was implemented and "Pixel-Clock" is now adjustable in 1MHz Steps. Together with the BaS update this FPGA update brought the possibility to set the Video-base-addresses anywhere inside the ST-Ram. This solves all ST-compatibility issues (as far they are concerning Video). A write operation into the Video-RAM is now divided into a "Atari-" and a "ACP Video-Ram area". This FPGA update brought not only they improved video possibilities, but also the already above brought up complete Ethernet, as the new - from some future users demanded - reset/boot behaviour of the power button (0,5 sec. boot, 0,5 sec. reset, 3 sec. power off). The FPGA is now also delivering all 20 clock frequencies, which are needed at the FireBee. These frequencies became adjusted. The FireBee is now running with 264 MHz. Additionally the FPGA update also brought improved usage of the Floppy. In the meantime reading from HD-Floppies is working under FireTOS. At writing as with DDs there are still problems. Already mid of December could additionally several Floppy bugs in EmuTOS be solved. Of EmuTOS there is, since shortly after our last update, a version available that works completely without MMU. This can be useful for tests as reference or a full optimised ColdFire-System. For Atari compatibility the MMU is naturally needed. A further very interesting development we like to present today is the prototype of our small custom enclosure. The enclosure is just marginal bigger than two side by side lying 3,5" HDs. The aim was to produce the smallest possible enclosure for our PCB. For now it including some LEDs, one Speaker, space for the CF-Card, space for a 2nd battery and there for one hour full usage without external power, notches for all outer ports, as an easy to open upper component. The case can be opened by squaring a single knurled head screw, while the computer stays bolt down with the bottom part. Recently we are considering, if we embed guide rails for 2 1/2" HDs and a SATA port. The enclosure itself is fanless, and with CF-Card also free of mechanical components. It is absolutely silent.
There will be in extenso information about the enclosure in near future. Serial debugging: Now it is possible to adjust the Rsconf Boot-Choice with a CPX ( a list of 16 baud rates like as at the Flacon 030 HSMODEM), and debug with programs like KiTTY, minicom, ZOC oder Hyperterminal, via a simple null modem cable. Beginning of February a NV-RAM Test Bug was detected. Zorro has stopped his work on zDesk. After a public debate with several users, who wished the further development of the never published zDesk, the recent status of the software was published as GPL, like it was done previously with all "z-applications". The Source Code is now available at Atari-Forge. Two new versions of AHCC were published. The GPL Compiler/Assembler is now availableas Version 4.4 from Henk Robbers . There is a new Version of MyAES (0.9.3) which can be obtained at Oliviers page . Mid of January Vincent has fixes GCC, to have a usable math library with m68020-060. For Gem-View there is a patch available, which is based on the work from Guillaume Tello and which was done by Lonny Pursell. With this patch the application should work immediate at the FireBee as soon as MiNT is ported completely. Didier is working a lot recently at the AC'97. This leaded already to several possibilities, that we will present you separately. We'll say this much, MP3s can already be decoded with only the CPU, and the GSBX mixer is working as ACC including all controls at the FireBee.
A further very important point, which hopefully will delight some of you, are the user applications. Recently we are at testing and adapting HDDriver, Calamus and Netsurf. Already adapted were Doom for GEM which is running well, or as well POV Ray. A even more suspenseful point could be the following one: With the recent implementation of the cf68klib already several pure (!) 68k applications are working. Let´s go by the name of QED (3 and 5), Wdialog, ST-Guide, STING, Packman (GEM version), the Atari Mouse Accelerator, or Papyrus. As highlights here a few pictures of strictly 68k programs that work at the FireBee under FireTOS without any patches:
Last but not least two further news regarding our homepage: With Pavel Salač we have a new member in our team, who is now responsible for the translation to Czech language. Pavel has further knowledge, which will the upcoming months expand into the ACP. And the link section of our page got a huge update end of December. Look inside. Even for some Atari netheads there could be something new to recover.
|
| 2010-11-18 As a lot of work has been done or even finished recently, here are some news about the latest achievements: |
|
Vincent has fixed EmuTOS to align programs and Malloc() buffers to multiple of 4-bytes. This ensures that programs always run at optimal speed. Indeed, accessing 32-bit values is a lot faster when they are aligned on 4-byte boundaries, especially on ColdFire. This was already the case with TOS and FreeMiNT. Fredi works on a problem about crashes on long-access on the IDE interface. Henk decided to work on EmuTOS' VDI improvements, e.g. allowing the VDI to use 8bpp and 16/ 24/ 32 bit modes. Didier implemented the I2C protocol into FireTOS allowing automatic monitor detection. FireTOS can be run in 1920x1080! We setup version source control at http://www.atariforge.org/gf/project/firebee where all developers are slowly uploading their work. David is doing nice progress on USB, trying to assure compatibility of his USB stack to the NETUSbee as well! Didier and Fredi have debugged the USB hardware (OHCI1 & OHCI2 interrupts) and got USB mice and keyboards working with FireTOS! Work on USB mass storage has begun now. Vincent and Fredi have implemented a bus error emulation layer, so now it's possible to run programs relying on this (e.g. for getting a list of available hardware). EmuTOS is now able to print nice error messages with a register dump when a faulty program accesses a wrong memory area. Vincent has found the cause of the infamous BaS slowdown: The cache mode was not set optimally (write-through instead of copy-back), finally EmuTOS+BaS are running at full speed! Didier is able to natively run Kronos with all tests apart from AES stuff running natively. Didier and Fredi are debugging power consumption errors on the PCI bus. Didier proved himself being a natural Atari maniac by soldering 240 wires from an unused CTPCI backplane just to give Fredi some test results! Fredi then bought some PCI slots so no further soldering should be needed for debugging the problem. There has also been improvement on development environments: EmuTOS runs AHCC without any emulation layer, and Henk expects AHCC to be able to recompile itself in CF mode! Henk started works on CF-only instruction support for better performance. Thus we have three compilers able to produce fully ColdFire compatible TOS programs. VBCC and AHCC can already be run on the FireBee, while GCC still needs to be run as cross-compiler on another machine.
|
| 2010-10-16 Some News |
|
The remainder of the first series has been finished at the factory, and has arrived at Medusa Computer Systems. There have been some reactions and suggestions from the community in response to the video published by Fredi. Following up on these, the behavior of the power button is now as follows: Power up by keeping the button pressed for 0.5 seconds. Reboot by keeping the button pressed for 0.5 seconds. Power off by keeping the button pressed for 2.5 seconds. The USB driver is advancing: David Galvez has continued the work done by Didier Mequignon with the XHDI mass storage driver to the point that all partition types supported by MiNT are working well (FAT12, FAT16, FAT32, ext2). Alan Hourihane took on the responsibility of implementing PS/2 support, and is adapting Eiffel to our PIC. This means the ground work for PS/2 is now done. Henk Robbers published a new version of the AHCC Compiler, which is running natively on ColdFire. The vbcc compiler has also been ported to ColdFire by external developers. There are now targets for both TOS and MiNT on ColdFire. In the near future, there will also be a native FireBee port of vbcc. At the end of next week we will request a prototype of the Mini-Enclosures from Vienna.
|
| 2010-09-26 TOS in Russian |
|
We are proud to announce that Дима Соболев (Dima Sobolev) finished his work to translate EmuTOS to Russian. Therefore we will be able to ship the FireBee with complete Russian Operating System as also Teradesk is available in Russian language. It is the first time an Atari Clone will be available in series production with Cyrillic alphabet and translated desktop as well as OS. We hope to make usage of Atari environments in Russia, Ukraine, etc. more interesting and even for some people possible at all. This development started inside the ACP, but took mostly place inside the EmuTOS team, and therefore the new language can of course also be used with Aranym or other environments. This is one of the examples like the project likes to improve situation for the whole Atari 16/32 scene. Please enjoy the pictures below done at two different series one board of the FireBee.
|
| 2010-09-16 The summer finally draws to a close, so here is the news: |
|
First the bad news; In addition to the production delays, the first 24 boards have a small fault on the PCB. KiCAD (the printed circuit board application, which we used for the PCB layout) filled an area over the Vias in an undesirable way. Therefor the PCI-Bus does not work correctly. These boards will be fixed by hand. The rest of the series will not have this fault, it has been corrected at the production level. The good news: The first boards have been sent to software developers. Due to the production delays we now have an unanticipated situation, something like a "developer series" ;) On the software side we recently work on the base system, the 68k-handler, the PIC (Real-Time Clock, PS/2, ...), the DSP in VHDL, the USB driver, the adaption of MiNT and the native ColdFire compilers. Our team has grow again. More than 50 people now contribute their knowledge and spare time, to advance the start of the FireBee and to continuous further development. The new members will be presented here soon. A new aspect of the ACP which has moved forward that we can expect real achievements, is our custom case production. This is an often asked question from future customers. And now we are ready to say, that at the end of the year the first mini custom enclosure made exclusively for the FireBee will be offered. The last positive information for today; The question of licenses for the hardware will now be concluded shortly. The schema, which has been available for more than one year freely down-loadable from our site, will be released under GPL/LGPL or a similar hardware license. The layout of the board will be released under Creative Commons BY-NC-SA. Therefor our computer can now indeed be refereed to as Open Hardware. Last but not least, we would like to present a short video in which the initial operation of the FireBee is explained:
The original can be downloaded (size: 220MB):
|
| 2010-09-07 Photos free of charge |
|
Photo 1 high resolution (CMYK, 3,3 MB)
Photo 2 high resolution (CMYK, 6,3 MB)
Photo 3 high resolution (CMYK, 5,8 MB)
Photo 4 high resolution (CMYK, 5 MB)
Photo 5 high resolution (CMYK, 3,9 MB)
|
| 2010-07-09 Production Delays |
|
At the production of the first series in the first half of June, our swiss assembling company had a huge rate of rejects. They got a problem with the very close conducting paths, and had a real production failure at the PCB production. Because Fredi Aschwanden from MCS controlled the boards, we could identify the fault before the parts got assembled. So we have 24 boards working correctly, which got assembled in the last two weeks and will arrive at MCS within the next week. This one will go to developers quickly, after implementing the basic software for initializing the hardware. For the rest of the series we get full compensation. As they made a real fault at production of the first series we get the complete series without any further costs, but with a delay of 4 to 6 weeks. So the first boards for customers will not be delivered in July. None of the defect boards will leave the Factory, they will not be repaired. New PCBs will be produced by our contractor. We are glad that we chose a swiss company which is granting the normal warranty of two years to all hardware production mistakes, and will deliver without problems the complete series as agreed. The warranty of two years also counts for boards delivered to customers. As this question often occures at forums or mailing lists, we like to assure you that you get full hardware warranty for your computer, as usual in Europe. Sorry for the delay!
|
| 2010-05-30 AHCC 4 and TeraDesk |
|
On May 27th Henk Robbers published version 4 of his Pure-C compatible AHCC compiler. This is - to our knowledge - the worlds first development environment which is able to create programs that can be executed on ColdFire CPUs as well as the 68020, 030, 040, and 060 processors. As "proof of concept" Henk Robbers recompiled TeraDesk using AHCC. Therefore, TeraDesk is running ColdFire optimized at the full speed of the FireBee prototype. The same Binaries are executable on the 68020 and up. Both TeraDesk and AHCC are published as Open Source projects.
Here you see a photo from the 21st of May 2010 with TeraDesk optimized for the ColdFire V4e running on one of our prototypes.
|
| 2010-05-17 The first FireBee series is now in production! |
|
A few weeks ago we sent the production order to our swiss manufacturing company to produce the first FireBee series. The assembling company has reported back with a positive report about the part sourcing. All parts will arrive in Switzerland by the 4th of June. Also, the multilayer PCBs will be produced by then. On the 7th of June, Medusa Computer Systems will oversee the boards and the layout on-site, directly at the assembling company one more time. Beginning on the 8th of June, the first FireBee series will be produced. Subsequently, we will need further weeks to implement the software. We will deliver to developers and pre-order persons with a high level of technical knowledge as fast as possible. All other end users will be sent their FireBee in the 2nd half of 2010. We once again, want to explicitly thank all of you, who already by August 2009, through your advance payments made the new Atari-compatible hardware possible, and thereby made the production possible as well. You showed us, that you are willing to push the development forward, together with us. The first series of FireBee computers are still available for Euro 599 directly at our site.
|
| 2010-04-22 Prototypes Update |
|
|
| 2010-03-19 Frank Naumann has passed away |
|
Today we received the sad news that our co-developer, Frank Naumann, has passed away on the 12th of March, 2010. Frank was well known for his brilliant work with FreeMiNT by almost every person inside the Atari community. He supported the Atari Coldfire Project from the beginning, and wanted to help with software porting and optimization for our new computer. The Atari community adopted the idea of Open Source mainly through the selfless accomplishments of Frank Naumann. Now others have to continue his work, what was made possible by Frank's vision. However there will be a big gap inside the small Atari community because of his incredible knowledge and his understanding of the Atari Operating System. Frank left us much too early at the very young age of 34 years. In the Atari community he has made a name for himself and he will not be forgotten, as every machine booted will still display his name well into the future. Frank will be missed by many. Also the Atari Coldfire Project has lost a team member with outstanding knowledge. Our thoughts and compassion go out to his family and his close friends.
|
| 2010-03-01 The Official Name |
|
We are proud to announce, that the community decided for a name of our computer. Firebee is from now on the official name of our hardware.
|
| 2010-02-23 Switch to a New Vote-System |
|
The poll system we´ve been using till now is not the best one, as several people recognised. That´s why the team decided to switch to another vote-system with an extra mailconfirmation, to make it harder for those of you who like some names so mucht to vote hundred times for it. ;) The vote for the name of our new computer is now extended to February the 28th. Please everybody vote again for your favorite name. This time you have to confirm your vote via E-mail before it will be counted. The email adresses won´t be used for anything else than the vote verification. After the vote all of them will be deleted, we do not collect them. The official name will now be published after another week of poll at 1st of March. Sorry for the inconvenience!
|
| 2010-02-11 The voting - till 20th of February |
|
It's getting serious: now you got the chance to vote one out of the 10 following proposals which arrived us through the community to find the name you would like to give the future ACP Atari compatible computer. You have exactly one vote, which you can place at your favorite name. First we will introduce all the nominations to you. Below you can vote by clicking your proper link. All votes will be counted until 23:59 o'clock on 20th of February. On February the 21st the official name under which the computer will be produced es presented. Here are your possible choices:
|
| 2010-02-10 Propositions of names |
|
ACP Team would like to say thanks to all people who joined the process of name finding. Altogether you sent 183 proposals 173 of which we couldn't take into the final list. From tomorrow on you'll find the best ten ideas at this place and have the opportunity to vote for the final name. For the most ideas which couldn't take part in the final voting there already existed brand or product names, or there had already been a copyright on the name. This really was a pity with some of the names, such as 'Leningrad' which is a ZX-Spectrum Clone, 'Colibri' being a Colibri-board or 'AINA - AINA is no Atari' being the IEEE International Conference on Advanced Invormation Networking and Applications. For documentation here are all proposals which did not reach the final voting round: @-R-E, Aarcti, Aarctic, AartiC, Aarctik, Aarctiq, Abroad, ACF 5400, ACM - Atari Coldfire Machine, ACP, ACP 1, ACP 010, ACP 474 ST, ACP 512, ACP-700, ACP 2010, ACP BEE-EATER 010, ACP BEE-EATER 2010, ACP BEE-FORTRESS 010, ACP BEE-FORTRESS 2010, ACP Cold ST, ACP Cougar, ACP Falcon Pro III, ACP Fire ST, ACP Hellfire, ACP Kea, ACP Koldbrann, ACP ONE, ACP ULTOS M 010, ACP ULTOS M 2010, ACP ULTOS 010, ACP ULTOS 2010, AINA- AINA is no Atari, Alinka, Ataconda, Atacondor, Antarctiqq, Antarcticc, Artic Lava, Artic Magma, Artic Vulcano, Arcticc, Arctiqq, ATARINA, Ataspic, Basilisk, bennu, Blizzard, Bluedream, Catari, CGM Coldfire GEM Machine, CGP Coldfire GEM Project, CCS - ColdfireComputerSystem, Cirata, CLDFRE, Colda, Coldac, Coldak, Coldarko, coldbird, coldboard, coldbox, Coldias, Coldiak, Coldibrik, Coldibriq, Coldic, Coldina, Coldinka, Colfir, Coldfirek, Cold Turkey, Colibri, Coldibri, COLTARI, Coltari Falcon, Comet v4e, Condor, Coolfusion, Cooltari, C/Fire, C-LDA, ACP Critter, Difference, Dream, Dreamfire, Eisvogel, Facepower, Falcon V4e, (acp) firebird, Firebird, (acp) firebox, FireGEM, FreeNet, Frenchfree, Fujiboard, Fujijama, Fujiyama, Gather, GemFire, GigaST, GoC - GEM on Coldfire, Goticc, Harrier, Hawk, (ATARI) HCS - HomeComputerSystem, i@Re, iceDream, Icefuji, Falcon - ColdFire Falcon compatible, iceGEM, Icehawk, Ice Kea 7454, IceTos, iDream, Incendia, I N C I T A R I, IrataC, iSpirit, Kestrel, Kolibric, Kolibrik, Kolibriq, Leningrad, LitleMouse, Malinka, MCS Newera V4e, Medusa Coldfire, Medusa Mintari, Medusa Phoenix CF, Merlin, MILAN CF, Mintari, NATARI, Nectaric, Nectaricc, Nektarik, Nectarin, Nektariq, NetCrasy, Newera, Newera ACP, Newera PCI, Newera V4e, Newera CF, NEXTOS, Nextos, Nixiek, Opening, phoenix, Pixelik, Poeticc, Poetiq, Ppoemm, Ppolarr, P-xyp, Pyro60+, Qotiq, Rainbow, Raptor, Sansan, Sierra, Spirit, Spirit v4e, ST(ultra), te ari, The fabulous Ice machine, The Love Machine, Ukeru, Vulcan, Volcano, Vulkan, Widel, Yleqotiq
|
| 2010-01-31 New developers |
|
At latest by the time we announced USB driver development more developers joined forces with ACP. Now we want to tell you who we are talking about. We are very glad about people who couldn't work with us a year ago, or who have been inactive regarding the Atari sector. People knowing the scene will already know most of the names - they speak for themselves! As members we are greeting:
|
| 2010-01-21 Our Board |
|
The following poster provides a complete overview of the board. Anyone who prints it out at A3, will get a 1:1 version of the hardware. The poster itself was first presented, in an older version, at the Eurocon in Vienna, and later at the Chaos Communication Congress (26C3) in Berlin.
Poster's full version in pdf-format (2.8 MB) Because the cost is sometimes asked about, the price is a definitive Euro 599. The hardware will be build by Medusa Computer Systems / Switzerland, who is also providing the standard warranty. It can, since mid 2009, be ordered in advance at the project website.
|
| 2010-01-07 Computer Naming |
|
We decided to let the Atari community find a name for our computer. The current codename of our project is 'Coldari' which some people don't like too much. Thus we want you to find a name for us - you, who supported us, contacted us with hints and sending good wishes within the last year. We want comprehensive debates and brainstormings on Atari-Home.de, Atari-Forum.com and on our mailing list. We ask you to send earnest ideas to 'acpinfo at atari dot org' until January 31st 2010. The team (which has grown to 37 persons up to now) will then find the best 10 propositions and check them regarding usuability. An acclamation until February 20th 2010 will then assign the definite name. Please take care about trademark rights. 'Atari 2010', 'New Falcon' or 'Microsoftkiller' would be discarded at once. On February 21st we will officially nominate the computer to its name, which you chose, and under which it will then be produced. We will nonetheless refrain from throwing a Magnum bottle of Champaigne at one of the prototypes ;) Please send numerous ideas ...
|
| 2009-12-24 Zoom inside the ColdFire |
|
As the name implies, the heart of the ACP computer will be a ColdFire microprocessor. We have chosen the model MCF5474, which is the fastest one. The ColdFire processor family is designed and produced by Freescale Semiconductor, the subsidiary of Motorola on semiconductors. The ColdFire chip is not only a CPU (like the 680x0 family was), it is actually a microcontroller. A lot of useful additional controllers are built directly into the ColdFire chip. They are directly usable on our computer, providing a very high level of integration. This is why the ColdFire is the best choice for designing a modern Atari-compatible computer.
- CPU 266 MHz The main component of the MCF5474 is a ColdFire V4e Core. It is very similar to the 680x0 CPU family, but not fully compatible. Basically, some complex or rarely used instructions and addressing modes have been removed. Thus the ColdFire CPU is more simple and cheaper than any 680x0 CPU, while being a lot faster. With an additional lightweight software layer, the ColdFire CPU can be made compatible with existing 680x0 programs. - FPU The Floating Point Unit is the mathematical coprocessor. When used, it provides a big speedup for programs using intensive math functions. The ColdFire FPU is similar to the 68881, but it has limitations in instruction set and addressing modes, like the CPU. - MMU The Memory Management Unit can be used to provide Virtual Memory or memory protection between the different processes running on an operating system. - Cache 32 KB x 2 The ColdFire cache is divided into a 32 KB instruction cache and a 32 KB data cache. The usage of the caches improves greatly the speed of the CPU, by optimizing the accesses to the external memory. - SRAM 32 KB The Static RAM is a high speed memory embedded into the ColdFire. It is required for DMA transfers, but can also be used for any other purpose when high performance is required. - EMAC unit The Enhanced Multiply-Accumulate Unit is a DSP-like unit supporting integer and fixed-point operations. It can be used to speedup some specific process such as image compression. - PCI controller The PCI bus has become the standard for extension boards in desktop computers. The ColdFire has built-in support for controlling PCI devices. This allows the usage of the hundreds of PCI boards available on the market. Of course specific drivers will have to be developed for every PCI card. - DDR-SDRAM controller The ColdFire can work with external DDR-SDRAM, for excellent performance. The ACP board has 512 MB of DDR-SDRAM built-in. - Interrupt controller The built-in interrupt controller can be compared with the MFP 68901, but a lot more flexible and powerful. It can use up to 63 interrupt sources with a specific exception vector for each source. - Timers 4 General Purpose Timers (GPT) and 2 Slice Timers (SLT) are available. They are comparable to the Timers A/B/C/D of the MFP 68901, but more configurable. - High Speed RS-232 controller The built-in Programmable Serial Controller (PSC) can be configured as a RS-232 interface for high-speed transfers, including at 115200 bps. - Ethernet 10/100 controller A standard Ethernet controller is built directly into the processor. This allows an instant connexion to the local network and Internet. - BDM / JTAG interfaces These debug interfaces can be used with hardware debuggers or external software for monitoring the CPU activity or flashing the EEPROMS. As you have seen, the ColdFire MCF5474 chip contains an incredible list of useful components, and there are even more. The full documentation is available on the Freescale website:
|
| 2009-12-11 First working GEM Application |
|
We have just enabled the floppy drive on our prototype. This means we are now able to run external ColdFire programs. Here is the first one. It's a small GEM test application displaying an alert box. The ACP computer becomes usable, this opens the door to a bunch of new test programs.
|
| 2009-11-26 SDL has been adapted to our machine |
|
By mid of november Patrice Mandin has published a new version of SDL. This 1.2.14 release is the first one containing ColdFire support in order to be able to run on our Coldari to be. We are highly delighted about this as Patrice is not associated with the Atari Coldfire Project in any way, and Simple DirectMedia Layer will be the first external application to support our computer. This support shows - although being theoretical due to lack of testing machines - that we can count on wide acceptance in programmers' circles and that our initiative is a further good way to improve the situation within the Atari 16/32 sector. We hope to see numerous cases of program optimisation within the next months, including some outside the open source area. Further information about SDL on Patrice's web pages.
|
| 2009-11-18 State of the Prototype & First simplified EmuTOS booted! |
|
Since the end of October we are working at the prototype. Following is tested an working:
|
| 2009-10-27 Sound Progress |
|
Since July Didier got one AC´97 codec we are going to use at our computer, soldered to his ColdFire Evaluation-Board. Writing the complete driver consists of 3 parts. The XBIOS, MCF547X/548X driver, and AC97 codec driver. The XBIOS was comparatively easy, as there was the MagicSound (Aranym/MagiCMac), and thus is finished. It uses Linux ControlNames for the ac97_codec.c compatibility. For the progress of AC97 codec part cleaned Linux sources are used, which needs several command replacements. At the moment AC link problems with the Coldfire (writing problems, reset values in registers) are under development and in progress. The XBIOS is compatible with the MilanBlaster/GSXB software and allows to select each source: AUX (PSG on Coldari), CD, LINE, MIC, but also FM -> PC Beep and VIDEO The AC97 native list of frequency is 8000 Hz, 11025 Hz, 16000 Hz, 22050 Hz, 32000 Hz, 44100 Hz and 48000 Hz. As well the driver for the LynxEM (PCI-onboard graphic chip of the evaluationboards) is ready-made. The testing and debuging of the driver will take place in the following weeks. Now you can use the EVBs with onboard graphics and without the need for a Radeon. You can always follow Didiers progress at his page.
At this picture you see the achievement of huge work soldering a AC´97 by hand to a ColdFire Evaluationboard. Didier added a mini-jack sound output on the left between the Ethernet and RS-232 ports.
|
| 2009-10-08 The Prototypes arrived |
|
Picture 1 (high quality)
|
| 2009-10-05 Discount Source for ColdFire Evaluationboards |
|
Our teammember Mark established contacts to a real cheap source for ColdFire Evaluationboards. We got the possibility for at all 20 pieces of M5485 EVBs, for 125 US$ each unit to buy. Original retail price is $850. Through this it will become possible to immediatly (before our project reaches series production) start with the adaptation, and above all, the testing of software which is running at ColdFire. Developers, especially those who care about softwaretitles or who are not part of the ACP, can and shall participate in our contacts. Who wants to order a Evaluationboard for 125 Dollar, should contact us as soon as possible. At this EVBs Didiers patched FireTOS and MiNT is already running. Who want´s one? For those programmers who are not so good with hardware, Mark is offering the possibility of creation of coldfire development machines. They will consist of an M5485EVB which is one of the fastest coldfires available, Eiffel (CAN-Bus) for using PS/2 keybord and mouse, a Radeon graphiccard, CompactFlash. Also HD-Driver, FireTOS and MiNT are ready setup and configured. This ready EVB, which will be 500 US$, comes with the following features:
16MB Boot flash For detailed specifications we recommend the Freescale site for the product and the picture Didier is currently working at a driver for the LynxEM (Graphic PCI chip - which is soldered at the EVBs) and expects to finished to write in several weeks. Please contact Piku directly at http://www.atari-source.org/ or our project mailaderess acpinfo (at) atari.org
|
| 2009-09-18 USB driver development started |
|
Like several times reported, we expect the development of the USB driver to be one of the biggest tasks which is imminent to our project. Today we´d like to acquaint you with our undertaking and everything what happened till now, regarding USB. For about three month we got a new team member - Sascha Heinisch. Sascha decided after a short time within the ACP to become maintainer of the widespread USB-project for Atari. Together with David Galvez he forms a small team, which will realize the long term driver development, among supervision of the rest of the team. The team is, as a matter of course, open for further persons. Especially we would appreciate attendance of those people, who so far already did USB-programming at Atari, and who are not reachable for us within the last weeks! The long way: At present the team familiarize itself with low-level driver development as with OHCI and EHCI specifications. The chip which is used from us, is a further part which needs widespreaded vocational adjustment. Beside the preliminary stages from within the Atari community there are two leads available for us: 1) A GPL Linux-driver for the NEC µPD720101 Because the NEC is used at many PC-USB cards, and also our chip will be attached via PCI-bus, nothing get in the way of using any USB-card with NEC µPD720101 under Aranym to be able to start imediatly with driver development. 2) The since 2009-27-08 freed Amiga stack Poseidon. (The community did buyoff the stack for Aros, and this now was released under the free Aros-licence). With the programmer and former owner Chris Hodges, friendly contact is established for a while. Not least we could possibly learn something from the 8-Bit scene and use reasonable solutions from http://microusb.org/ . The ACP is prepared for a really long-term development. To make clear what that means: we expect USB-Input-Device and USB-Mass-Storage not before summer 2010. Therefore the development can take place at a level which can be kept up over a long time. Our Chip: Contrary to our first announcements - that we like to use the ISP1536, we had rapid to rearrange, because the chip is not longer delivered. Anyhow we think that 95% of our future developments will be usable for all Atari-projects with USB and will leverage the entire platform to utilization of USB. We think about EtherNat, CTLink, NetUSBee und Suska, for which alltogether no expedient USB-stack is in existance. All this projects should, with small adaptions, be able to profit from our Open Source development. Together: Who likes to take a long-term role at ACP is, as every time, called to report, and support the USB-team - even with small contributions! Together many things work much more easier, it simply is imperative that we organize us meaningful.
|
| 2009-09-03 New Pictures |
|
To satisfy your curiosity, waiting for the prototypes arrival, we got again some pictures from a paperboard with laid samples. It is about the last version, which was sent to the factory for production.
|
| 2009-08-27 Official Atari Coldfire Project theme song |
|
We are proud to announce the official Atari Coldfire Project theme song. Produced by Neo Bechstein in July 2009. The Voice "Atari Coldfire" is generated with STSpeech at an 1040 ST and the song was recorded with a Falcon. We like to thank him for his efforts to support our project with a song, with this status. Also we want to show you out there, that there are many possibilities to bring our project forward.
Neo Bechstein - Atari Coldfire Project.mp3 (21MB)
|
| 2009-08-06 Schema & Co. |
|
Today we got the assembly diagram and the schema for you. Below you can download our 5th revision as sent to the factory for the prototype production.
|
| 2009-08-05 Prototypes ordered |
|
At the 4th of august we ordered three prototypes of our new ColdFire Computer. It´s the revision 0.98 of our board with the redesign to a DVI-I connector and some other minor changes. The boards should arrive in around 6 weeks at Medusa Computer Systems in Switzerland. Please everyone who preordered, prepare for the advance payment. ;) Because ordering the prototypes means: the computer will be built!
|
| 2009-07-28 Suska III-C available |
|
Not directly an ACP issue, but very important news is that the Suska III-C board is now available. The Suska is an Atari-ST(e) Clone, with all the modern connectors. It is produced by Wolfgang Förster who is also part of the ACP. The Suska Board was designed during more than 5 years (!) of work done mainly by Wolfgang. Every single Atari-chip was redesigned and reverse-engineered to be configurable in VHDL and to be used inside a modern FPGA. So we now have every custom chip available as open source configuration, to produce further Atari-compatible computers. Also in January Wolfgang agreed to become a member of the ACP and to support Fredi Aschwanden with developing the hardware. The ACP board as far as it depends on the FPGA, is mainly based on the free VHDL core which was developed for the Suska. So we now want to point out the fact that the Suska board is readily designed and that for the first time within the last decade we got a new and available Atari-compatible computer. The Suska is an extremely compatible replacement for our Atari ST(e)s, which might be interesting for people who like to play the old original games or demos. The Suska III-C board is now directly availaible for Euro 619,- (VAT included) from Inventronic. We like to assure you that we do not see our projects as a rivalry. In fact, the huge amount of work done for the Suska lead to a fast developement for the ACP. The Suska and the ACP board are for different needs, and we are proud that our beloved TOS compatible platform soon will get a second deliverable computer.
http://experiment-s.de/en
|
| 2009-07-20 EmuTOS "proof of concept" has been done |
|
Vincent Rivière managed to run EmuTOS on the M5484LITE evaluation board in text mode. He disabled the AES, so it boots into EmuCON, the early console. The console I/O is redirected to the ColdFire serial port. So with any terminal emulator, we can type commands and see the results. Be sure this EmuTOS version is fully ColdFire software and does not contain any 680x0 emulation. Currently EmuTOS boots instantly, the display speed is only limited by the serial port. Vincent also implemented a ramdisk to run ColdFire TOS programs. He is going to cleanly integrate the ColdFire support into the official EmuTOS CVS repository, with the help of the EmuTOS developers. So after Didier Méquignon's FireTOS (patched TOS 4.04), this is the second Atari OS running on ColdFire processors. In this picture (click on it to see the full version) you see a terminal emulator running on another computer, connected to the ColdFire board through an RS-232 cable. The file emutos2.srec, containing the EmuTOS ROM and the ramdisk data, is downloaded into the board through the Ethernet connection via the TFTP protocol. Then EmuTOS can boot and access the ramdisk through the A: drive. This OS can currently run ColdFire programs only (for example, those compiled with GCC and the ColdFire MiNTLib), an additional emulation layer for running standard 680x0 programs will be added later.
|
| 2009-07-12 Hardware transfers |
|
The developments are progressing continuously. Recently quite a lot of hardware transfers occurred in our team. Fredi Aschwanden got a sample of our AC'97 Codec and sent it to Didier Méquignon. Didier has connected it to his ColdFire evaluation board and is currently working on the software part. Markus Fichtenbauer is sending a NetUSBee to Sascha Heinisch. Olivier Landemarre sent his ColdFire evaluation board - which he got from Mark Duckworth - to Vincent Rivière, and we thank him a lot. Vincent also got an Eiffel adapter from Alan Hourihane, so he can use a PS/2 keyboard and a mouse with the board. From now on he will be able to do experiments on the real hardware, for a better understanding of the ColdFire internals. Moreover, he is also porting EmuTOS to ColdFire, and most of the work has already been done.
|
| 2009-06-12 Technical Background |
|
Technical Background about the planned ColdFire-Computer
|
| 2009-06-07 MiNTLib ready-made for ColdFire and a new member |
|
Vincent Rivière has finished to port the MiNTLib to ColdFire, his work has been included into the official sources. First tests on M5484LITE Board show that it works perfectly on TOS. Just after that, he has published a new version of his cross compiler GCC 4.4.0 with full ColdFire support, including the MiNTLib, math and GEMLib libraries. This compiler can be used right now to produce native ColdFire TOS programs that will run at the highest speed on our new computer. The cross-compiler is available on Vincent's page. Dima Sobolev from St. Petersburg/Russia joined the ACP. Dima is well known to STeem and TeraDesk users, and he also russified TOS 2.06. So we from now on got somebody responsible for "translation english/russian & TOS localization". Dimas "work" will reinforce the Atari-Scene from "Kaliningrad" to "Vladivostok" and help russian speaking Atari users to use our latest developements. We now also got a great teammember who got high experiance with lokalization of TOS in general.
|
| 2009-05-23 TOS, MiNT and DVI |
|
Didier Méquignon has adapted Firetos (his already for CT60 use patched version of TOS 4.04) to our new hardware. He has done quite a lot of work, so we can expect to boot our new board as soon as the prototypes are availaible. This work - also an adaptation of the work already done for ColdFire - became possible because of Didiers huge experience at the M5484LITE board since 2006. Vincent Rivière is currently patching the MiNTLib to enable its recompilation for the ColdFire processor, in order to produce fully native ColdFire executables which will run at full speed on our computer. He expects his work to be finished soon. Finally Fredi Aschwanden decided after a discussion on the Atari-Home forum to upgrade the hardware to a DVI-I connector instead of the first planned VGA. Most of the preordering persons would like to see our computer equipped with VGA and DVI possibilities and are also willing to pay more to have the digital video signals. But as a majority wants VGA - which could be seen after a quick poll - we decided to use DVI-I with both signals, and have the possibility to use an adapter from DVI to VGA. The works concerning this matter already began.
|
| 2009-05-19 New development environment and a new language |
|
Besides of the hardware we cared about a new development environment during last week. The developers of vbcc agreed to adapt the MiNT and TOS target of their development environment to the V4e. Thus very soon Crosscompiling will be at our disposal on many different platforms, for example AmigaOS/MorphOS, MacOS X etc. vbcc can be compiled without any changes for each platform on which an Ansi-C compiler is in existance. The vbcc team will change the last two known errors for ColdFire and we will extensively test and send bug-reports, which then again will be worked on. Likewise vbcc will be available on the ColdFire and is the fourth development environment which will run on our computer. Today it was a great pleasure to introduce our spanish news due to further support. We now can report about our progress in five languages. A big thanks goes to Victor and David.
|
| 2009-05-12 New team member for editorial office and translation |
|
After yesterdays request for participation in the ACP, Jens Riemann contacted us and offered help. Jens will immediately assume the editorial office and translations between German and English. Further we need people speaking French and Spanish who are willing to take on a long-term role inside ACP. This would help to release developers and to push on the project.
|
| 2009-05-11 Two of the best Atari developers joined ACP |
|
Didier Méquignon is with us and and will work on the ColdFire board - as soon as his "job" at CTPCI (from powerphenix) is finished . At first he likes to work at the AC'97 codec of the board. He will also bring in his great knowledge about ColdFire in general. To be sure to understand the relevance of this news; Didier patched TOS 4 to run at a ColdFire-board, got MiNT up and running at his developementboard, wrote a driver for ATI-Radeon-cards, and got one of the two only working coldfire-systems which are in existance at the moment. And that was not all! His already done great work can be contemplated at: ctpci-e.htm Lonny Pursell also agreed to join the Atari Coldfire Project as official member. As we already reported, Lonnys aim is to get GFA-Basic ready for ColdFire. Lonnys knowledge will be a big help for the whole team and beside working closely together with the compiler team (AHCC and GCC) he will also help at some minor tasks and give advice to the developers where his competency can be helpfully. Further helping hands, especially translators and developers, are welcome every time!
|
| 2009-05-07 Simplified schematic of our new computer |
|
|
| 2009-05-05 Support of all relevant developement environments assured! |
|
GFA-BASIC Lonny Pursell announced his efforts for supporting the new Coldfire Computer in a first reaction after our public announcement. The maintainer of GFA-Basic prefigured possibility of reworking the GFA-Basic Compiler/Library so it builds CF native apps. Should he succeed then this would allow just a recompile of existing applications, to get even more stuff up and running at improved speeds. Lonney Pursell declared that he´s willing to do a lot of work, how he suspects it, if the machine is released. He explicitely welcomes our decision about which processor we choosed for the future developement of the Atari-ST plattform. "I'd still like to see such a beast of a machine. You would not see me jumping on the support band wagon if it was some other processor though. [...] I see the CF as an old friend who maybe went on a long trip, came back a slightly changed man. A warm familiar face I can easily speak to." he said. Read his entire message at the public MinT Mailinglist. GCC 4.4.0 GCC 4.4.0 has just been released, Vincent Rivière has updated the MiNT patches for it and built the Cygwin binaries. He enabled compiler support for the ColdFire V4e family. So right now, we can run that cross-compiler on Windows systems to write initilialization code for our ColdFire board, and to check the compatibility of other software, even if there is no hardware yet. Since the GCC patches are host-independant, the cross-compiler binaires can be rebuilt for any OS supported by GCC (MacOS X support has already been announced by Markus Fröschle). Your help is welcome ! You can find Vincent's patches and binaries here: m68k-atari-mint tools Vincent has also started to enable ColdFire support in the MiNTLib. He evaluated that is was possible with a reasonable amount of work. When this is done, it will be possible to produce native ColdFire MiNT/TOS software what will run at full speed on our new computer. GCC is the universal C/C++ compiler. Combined with current support inside the FreeMiNT kernel and the MiNTLib, it allows to port GNU/Linux software directly to Atari/MiNT. Moreover, it can be used to write standard TOS software (including GEM) at a very high level of optimization. Pure-C compatible AHCC
Henk Robbers, also part of the Atari Coldfire Project, again declared, that AHCC will be adapted to the ColdFire. So a lightweight C-compiler will most likely be available too. AHCC is verry important for recompiling existing Atari applications - and it´s GEM and single-tasking compatible. Most native Atari software is written using Pure-C and can be recompiled with AHCC. Henk Robbers´ roadmap:
1 Compile and run EmuTOS. As a proof of concept. With this three important developement environments we got the neccessary fundament to compile software native for the ColdFire computer. That would be executable with maximum speed at the ColdFire. For the future we hope thereby for many new developements for the whole range of TOS/GEM
|
| 2009-05-01 News on Italian |
|
Thanks to great support from the community, after our first public announcement, we today could open our news on italian language. From now on we will be reporting about further developments of the project beside english, french and german also in italian. Further languages are welcome!
|
| 2009-04-30 Official ACP email address |
|
We now have an official email address for the project. Please use only this one for any communication with us, including for preorders. acpinfo (at) atari (dot) org
|
| 2009-04-28 Atari Coldfire Project Reloaded |
|
Since the beginning of December 2008 we've been evaluating the possibilities of continuing the Atari Coldfire Project or starting over with it. Our evaluation is now complete and we'd like to let you know that we consider this project basically possible and that from now on we will be working under the name "Atari Coldfire Project". Our primary goal is to build a new Atari TOS clone which is based on a Coldfire processor and to make it available to the public for the lowest possible price. This is a free project that is free of commercial interests of any kind. One of the main goals with our new hardware is to allow further development to be possible in the future in the widest possible scope. "Hardware updates" will become possible by VHDL code for free. The long term goal is absolute compatibility with every Atari TOS machine ever built. In the beginning however, we can assure Hades compatibility. At the moment the following people are involved:
Matthias Alles Further support and advice at special tasks comes from:
Pascal Barlier We are currently in contact with other famous people, this list will soon get bigger! We welcome and request help from every person who can contribute something, no matter how small! Remember too, that since this is a free project, it absolutely depends on community involvement and cannot be accomplished without your help. Because Fredi Aschwanden with the support from Wolfgang Förster directly began to develop the hardware, there is an interim layout of the planned computer available. It got two advantages which could also be interresting for non Atari enthusiasts, we would like to point at. The verry small size of the board and the extremely low power consumption. The actual specifications are:
1. As a independent motherboard. The final price for the board isn't certain yet. For the first 25 preordered boards the fixed cost has been set to Euro557,- (+VAT and shipping). To be able to start with the production of the hardware we need 25 preorders which are fixed through a advanced payment of Euro200,-. Subsequent to that we can begin the prototype production and start with software development. Therefore the first 25 people preordering would then need only to pay the Euro357,- remaining. As much of our work as possible will be published under free licenses. Going forward, we will be continuously reporting further developments at http://acp.atari.org/news.html ACP in April 2009
|
| 2008-12-15 Atari-Coldfire initiative from Switzerland and Austria |
|
According to new impulses from Switzerland and Austria, since the beginning of december 2008 several developers are contacted personally. The aim is to find out which skills and how many support-possibilities in the Atari-Community are in existance. Afterwards can be decided if and how the project can continue, or if it can get started again together.
|
| Atari Coldfire Project Archive | ||
|
You will find here the news related to our first project which took place between 2002 and 2008.
|