Falcon Designer

[nativ]

I would like to know more about the Falcon 040 MicroBox ?

Me too... :Df

I barely can wait fir book "Atari Corp. Bussines is war"

find an old playstation 2

two reasons to own a PS2 Rtype and microbox design

[tresas]

first thing that cross my mind is: "Blossom" - ATW800/ABAQ video card. and ATW was presented at Comdex, November 1987.
Kodak buy 100 pieces of ATW probably because of good graphic capabilities.

beside, there are infos that Falcon video is based on Blossom.

Hmmm...you have a good point but I'm not very sure... I think that if it is 1987 its too late to be the "Blossom" since just a few months later it was presented complete in Comdex I am wondering (if it doesn't refer to Falcon), if it refers to Panther, since we know that in 1988 Atari was working on two next gen consoles, Panther and Jaguar (has also both DSP chips).

In any case, Atari Falcon is a riddle to me. We know a lot of things for the story of the STs but nothing about Falcons. We know that in many cases the companies bought ready ideas of third party manufacturers, which was cheaper for them that develop their own ideas. Is it possible for Atari to have designed it from scratch without any cooperation?

[DarkLord]

Hopefully Curt and Marty will be able to shed some light on this in book 2, Business
is War.

[bid]

... Atari Falcon is a riddle to me. We know a lot of things for the story of the STs but nothing about Falcons. We know that in many cases the companies bought ready ideas of third party manufacturers, which was cheaper for them that develop their own ideas. Is it possible for Atari to have designed it from scratch without any cooperation?

We know that Flare1 and Flare2 were ex Sinclair Research engineers. Richard Miller is thought to have designed the Falcon (but this is not proven).

Flare1 was developed by Flare Technology. This formed the basis for the Atari Panther 32bit console http://en.wikipedia.org/wiki/Atari_Panther

We know that Flare2 became the Jaguar. The same team developed the Blossom chip for the Atari ATW Transputer, and according to this unsouced article, the VIDEL (Falcon) is a crippled and hence more compatible Blossom http://retrobrothers.hubpages.com/hub/Atari-Falcon

It seems to me that Shiraz Shiva designed the ST (obviously), and UNIX expert Roy Good took his place and designed the Atari TT. After Roy, then Richard Miller came along (just weeks before the TT launch) and he was responsible for the design of the Falcon, and sub contracted the Jaguar to his friends at Flare (who were very close and ex-colleges).

Roy Good is an unsung here IMHO, and the hardware was excellent. The only thing wrong with the TT I can think of is the Video-Shifter, which is quite basic.

Most Atari 68k machines seem to be a CPU stuck onto a video-shifter, and some clever cost saving measures to reduce chip count. The TT was great, and the Falcon030 in the ST case was definitely (IMHO) a very cut down and cost reduced item. I would guess, that they wanted to use up stock for Falcon, and had cut the budget, and originally had intended to produce (a) a 030 MicroBox, in which tooling was complete, and (b) a 040 mini tower, in which was at prototype stage, and could take ISA type cards. My guess is Atari decided to leave the computer business, and concentrate on the Jaguar as a last ditch effort. Sony stopped them. The stock for the future Falcons ended up in the Falcon 030, and I would guess there was quite a lot, as CLAB could also continue making them.

Loose theories seem to hint that the Falcon 030 started out as an accelerator board inside an ST. I would guess that this is partly true. I think that the MicroBox was fully designed, and I think there are 5 prototypes, and also the injection mould tools were 100% definitely made. The Falcon 040 prototype board was almost complete, and the case designed but not tooled (only pre-production prototypes).

Interestingly, the development platform for the Jaguar was on the Atari TT. So Jaguar was in the works for a loooong time. There is now lots of new info on the Panther uncovered. Including prototypes, and Cybermorph, the awful Crescent Galaxy and even Raiden were apparently ported from the Panther system.

I personally find it strange to have a radical departure from the TT, and clearly the Falcon was quite different it seems. I would guess that Atari had initially been making a big push, or else, why would they not use the TT/MegaSTe case and create tooling for a deskop 030 and 040 based Falcon. The 040 would have been more PC like and take standard cards by the looks of it. The MicroBox was complete I believe, so a shame it was never sold. The difference with the MicroBox I think, was that it had a 32bit x 32bit architecture, and the cut down Falcon030 what we all know has a 16bit x 32 bit architecture and is thus quite limited. Shame, when the TT managed to run most software, so I am guessing that it was a 'kludge' to use up old cased, components and liquidate stock. ... Now if we could find the tooling for the Microbox???!! Who has it? ... The Jag tooling was sold to a medical manufacturer.

Personally, I cant wait for Rudolph Czuba to build the PX60. If he does, then he will inherit the crown from Richard Miller, and create the next and possibly ultimate generation of the ST series. ...

[calimero]

Great article bid.

[bid]

Great article bid.

Cheers Matey,

I realise that some elements might not be accurate. I have in places added (IMHO - in my humble opinion), or I have guess certain elements and mentioned where I have guessed and filled in between the lines.

No one really truely knows about the Falcon, as stated. And also the TT we can guess was Roy Good, who was a UNIX Guru and took over from Shiraz Shiva. ... We can see a flavour of UNIX in the TT, especially the unreleased TTX (which was a TT Tower, and there is a leaked specification document I think on the GFA Basic site from a forum member here). The TTX would have had, I think 4-8 VME slots. The UNIX System V was made for the TT, and to be honest looked excellent. Never took off though, which is a shame. However, maybe if it ran LINUX and they had Ubuntu or Debian at the time, maybe then Micro$oft's evil dominance could have been thwarted, to our mutual benefit.

I read stories about the TT's development. And certainly the best and most interesting inside knowledge comes from Dadhacker's blog. Very interesting, if you have not read it! ... But the TT seems to have nothing Falcon, nor ATW (Transputer) about it. Roy Good left, as Richard Miller came in to replace him. Richard Miller was ex Perihelion Software and / or Flare. They were basically the same group of people. They made Blossom, which was the Atari ATW Transputers graphics system. And they also made Flare1 and Flare2, which was Panther and Jaguar basis. ... But of course, Flare1 was also supposed to be the Konix multisystem, took some Amiga ideas (which were actually ex-atari in the first instance, and not Commodore engineers) and made the custom chipset for Panther. However, the TT video shifter looks to be very basic, and maybe like the ST's but just adding video memory. I guess they thought video cards could be added to the TT, and there were some very good ones.

Anyway, VIDEL is supposed to be based on a crippled version of the Blossom. I have no idea how much Blossom and Flare shared similar technology. They were designed by the same people, or at least colleagues. But I dont think the Falcon has any Jaguar tech (thankfully, as it was rushed and flawed). It may have some Flare1 or Panther tech in it though. Panther was a big old sprite mover, old school and 32bit

It seems to me that the future for Atari graphics lay along the lines of the TT / Falcon 040 idea, of adding industry standard graphics cards. ... Certainly, the way that GEM, TOS and VDI / AES was compartmentalised (right from the first ST) is genius!!! ... This means that VDI / AES can be replaced for example.

Even more interestingly, the TOS and TTP is some kind of cut down CP/M, which is quite DOS like. And TOS can run command line style, just like UNIX and Linux, even in my 520ST. GEM is a great layer over the top and replaceable. You can also program right down at hardware (mainly for games, as the ST structure is so rudimentary). But with the ability to add new VDI and graphics, it seems that anything is possible. And the code is so small on the ST/TT/Falcon that is just so fast and non-bloatworthy. ... Really what computers should be like!!!

I hope that Rudolph Czuba makes the PX60, which will be a 68060 32bit Falcon-a-like on a ITX board, with USB2, maybe SATA, graphics and the full shabang! ... There is even talk of adding an Arm Co-Processor. So maybe in Gem, would you be able to gave ATP (Arm Takes Parameters) or even implement Java using existing code. ... We know that the ST architecture is very modular, and well designed (thanks Gary Kildal RIP), and the power of the DOS/CP/M mix and UNIX'y flavours already expressed, means I think that having a general purpose modern computer, with portable software from other platforms, may not be as impossible as people don't even dare to dream.

Imagine taking GPU software source to the Atari. Once again power without the price (well strike price, and replace that word with bloatware)

Long live Atari!!!!

[nativ]

Great article bid.

I hope that Rudolph Czuba makes the PX60, which will be a 68060 32bit Falcon-a-like on a ITX board, with USB2, maybe SATA, graphics and the full shabang! ... There is even talk of adding an Arm Co-Processor. So maybe in Gem, would you be able to gave ATP (Arm Takes Parameters) or even implement Java using existing code. ... We know that the ST architecture is very modular, and well designed (thanks Gary Kildal RIP), and the power of the DOS/CP/M mix and UNIX'y flavours already expressed, means I think that having a general purpose modern computer, with portable software from other platforms, may not be as impossible as people don't even dare to dream.

Imagine taking GPU software source to the Atari. Once again power without the price (well strike price, and replace that word with bloatware)

Long live Atari!!!!

I think that it would be a missed opportunity not to add an ARM interface with a Rasberry Pi or that other device mk808 android thing...

As you say Bid GEM rocks for modularity

[calimero]

Hmmm...you have a good point but I'm not very sure... I think that if it is 1987 its too late to be the "Blossom" since just a few months later it was presented complete in Comdex

maybe Steve Jobs planed to use _finished_ Blossom as graphic card for NeXT computers...

I am wondering (if it doesn't refer to Falcon), if it refers to Panther, since we know that in 1988 Atari was working on two next gen consoles, Panther and Jaguar (has also both DSP chips).

Commodore also have prototype of Amiga 3000 with onboard DSP chip. Never released. They focuse on A4000 instead.

Is it possible for Atari to have designed it from scratch without any cooperation?

they took ST and add 030 and DSP as CPU add-on board. later they add Videl...



let me try to make timeline, with references:

1985. ST went on sale
1986. there are rumors that Atari working on 68020 machine
1987. Atari hire Roy Good to lead development of 32-bit successor to Atari ST, ultimately named the Atari TT
1987. Atari get involved with inmos (they are quite popular at time; Commodore also presented A2000 with inmos transputers) - present ABAQ at Comdex
- Richard Miller create Blossom video card for transputer
1989. (feb) Richard Miller joined Atari, replacing Roy Good (may)
- all three of them were from Sinclair: Richard Miller, Martin Brennan and John Mathieson.

• Richard Miller called Martin Brennan and John Mathieson since he got task at Atari to make 2D console (based on his experience with Blossom). As Martin Brennan say:
  
  "He (Richard Miller) became a director of Atari in Sunnyvale and he had a project called Panther - It wasn't called Panther when I joined."
  
  so Panther is not external project! Atari want game console and R. Miller bring two friend that was up to taks.
  
  what happend next is that Martin Brennan and John Mathieson bring idea for 3D games at Atari:
  
  "But while I was over in California in '89, I actually convinced the bosses at Atari that 3D was the way to go, with the experience we'd gained on Flare one - if you didn't just do flat rendering, but shaded rendering you got a 3D appearance."
  
  so Jaguar project was born!
  
  "So the Jaguar project was born from the Panther project.
  In essence Atari looked at the Panther and looked at what we were promising for the Atari project and said can the Panther project."
  
  http://www.konixmultisystem.co.uk/index ... rtin#start

back on Falcon: R. Miller was in charge with Falcon. As Jeff Minter sad: [q]"One day I went in to Atari and there he was (Richard Miller), having just designed the Falcon."[/q] ref

what references I can not find right now are about CPU boards for ST! This is how Falcon conceive its life: as 020, 030 + DSP boards inside ST.

just dig: "Bill Rehbock told me at the Sacramento show that they were making 350
Falcons a day. x20 working days a month comes to 7,000/month. They
should be showing up all over soon." http://www.atarimax.com/freenet/freenet ... le.php?360

this is not official but:

JUN
17
The Life And Times Of The Atari ST (part 3) - Crowning glory, fading twilight...

By the end of 1990, reading between the lines now you can clearly see the first signs of Atari being in disarray internally. The STe launch had been horrendously mishandled, the TT030 had been hacked to run faster due to criticism and to top it all a batch of STe computers had made it into the retail channels with a faulty DMA chip that could corrupt data on attached hard drives. This was not the best way to launch a machine designed to rival the Commodore Amiga, but worse than that, there were persistent rumors of a machine in development with over 244,000 colours, which many people had assumed would be the STe.

But Atari were still developing the current ST line and announced some interesting new products for release in 1991. These included the Mega STe to replace the now decidedly underpowered Mega ST, the ST Book to replace the STacy laptop, and a curious piece of hardware known as the STylus, an ST machine based around a pen interface. At the same time it was also announced that Atari were developing new games consoles and that a new version of the Lynx colour handheld games device. The games consoles were called Panther and Jaguar, the former was a 32 bit system while the latter was a 64 bit system. Panther was to be launched first with Jaguar a few years later.

While the outside world saw these devices being announced, inside the labs of Atari, a couple of STe computers had been butchered and their processors removed. The processors were replaced with a board containing not 1, but 2 processors, a Motorola 68030 and a Motorola 56000 Digital signal Processor, the codename for the board was Sparrow. This was to be a test bed for what was going to be a new generation of home computers and was in part done to see how feasible such a dual processor like this was in the ST's architecture. http://subvertedsoul.blogspot.com/2011/ ... art-3.html



I will try to put all this data into database and make some visualizer for it...


btw if someboy missed: http://www.youtube.com/watch?v=9CMNFo8zIEM Richard Miller and other ex-VML employees on youtube

[nativ]

what references I can not find right now are about CPU boards for ST! This is how Falcon conceive its life: as 020, 030 + DSP boards inside ST.

While the outside world saw these devices being announced, inside the labs of Atari, a couple of STe computers had been butchered and their processors removed. The processors were replaced with a board containing not 1, but 2 processors, a Motorola 68030 and a Motorola 56000 Digital signal Processor, the codename for the board was Sparrow. This was to be a test bed for what was going to be a new generation of home computers and was in part done to see how feasible such a dual processor like this was in the ST's architecture.[/i] http://subvertedsoul.blogspot.com/2011/ ... art-3.html

I'll check the board for dates, but there was a DSP board available for the MEGA ST that was used with the PRO Tools setup.

[tresas]

Wow!!! Thank you guys for the interesting historic information and for your time!

[rabindranath72]

Atari manage to bring unix like OS, but still TOS compatible while Apple fail! (after failing, they bought NeXT)

Ehm no, not at all. Have you ever used A/UX? A/UX failed not because of lack of compatibility (which was just a plus for the typical UNIX user) but because of the huge royalties that AT&T wanted. And the fact that on the UNIX market there were much stronger contenders like Sun.

[calimero]

Atari manage to bring unix like OS, but still TOS compatible while Apple fail! (after failing, they bought NeXT)

Ehm no, not at all. Have you ever used A/UX?

I only learn recently learn that A/UX could run Mac application in parallel!

[shoggoth]

Who knows what Atari had in mind for the Falcon and future machines. If you look in the leaked VDI sources for example, you'll see that the TOS4 VDI supports 32-bit and 8-bit chunky modes, even though those modes aren't available in the Falcon. I exploit this in the SuperVidel drivers, because Atari actually even had a "nice" (as in unfinished, but it's there) way to install graphics drivers from the "ground up" just by replacing VsetScreen(). It is then evident from the VDI sources that Atari had to "wrap things up" pretty fast to produce a TOS version for the Falcon where stuff is still (mostly) hard coded.. I guess they were in a hurry since things were already going downhill fast..

[Dal]

Getting back on topic, this guy from Atari Age is selling a Sparrow prototype. He claims to be a member of the Atari development team so should be able to fill in the blanks with regards to who is behind the Falcon:

http://atariage.com/forums/user/35790-elig/

[calimero]

Great finding Dal!

Who will contact guy (seller) from atariage?

What should we ask him?

It would be great if he would wrote inside story from atari when they worked on falcon...

[Cyprian]

Who knows what Atari had in mind for the Falcon and future machines. If you look in the leaked VDI sources for example, you'll see that the TOS4 VDI supports 32-bit and 8-bit chunky modes, even though those modes aren't available in the Falcon. I exploit this in the SuperVidel drivers, because Atari actually even had a "nice" (as in unfinished, but it's there) way to install graphics drivers from the "ground up" just by replacing VsetScreen(). It is then evident from the VDI sources that Atari had to "wrap things up" pretty fast to produce a TOS version for the Falcon where stuff is still (mostly) hard coded.. I guess they were in a hurry since things were already going downhill fast..

that is really interesting, can you share more details?

[Knezzen]

Bumping an old thread a bit. Any more findings on the history of the Falcon? It's all very interesting to read, and I can't wait to get to read some more

[calimero]

nothing new from me...

EXCEPT (!) that I come to clash with Marty Goldberg (co-author of book: "Atari Inc. - Business is Fun") at Wikipedia (see "talk" page on Atari Falcon page, section "16 bit bus" link)

it looks like that we need to wait for "Atari Inc. - Business is War." by Marty Goldberg and Curt Vendel to learn something more about Atari Falcon history...

[calimero]

btw regarding wikipedia and talk page:

"6 I have to wonder..."

copy&paste:

Why did they go and base it on the ST specifically, when the capabilities and hardware were being largely expanded - when the TT and indeed affordable Mega STe already existed? IE, Atari already had a machine, from many years previous, with a 32MHz 68030 on a 32MHz, 32-bit bus, and the bitplane-pushing power - even if not the full palette - to exceed the actual Falcon's highest resolutions... it wouldn't have been too expensive a thing to make, surely, especially as it was ultimately intended to go in a standalone system box, rather than an outdated keyboard-computer box?

Just imagine if they'd released THAT instead - essentially a TT with a 56k DSP, 18-bit (VGA-like) palette, hi-colour mode (at higher resolution thanks to the faster bus and RAM access), HD floppy drive and a modulator, aimed at the home market instead of the professional one. Maybe with a CD drive too. It would have done a much better job of disrupting things versus what they released - the overall power would probably have been comparable to a mid-lower end "multimedia" 486 of the time, for less than half the price. It almost seems like they took a very early stage, proof of concept prototype, and sent it into series production...

(Ah, alternate realities, how I love you. You're where I'm a viking!) 193.63.174.211 (talk) 12:17, 4 June 2014 (UTC)

---
1. TT (32bit ST) was in development for more than 3 years - and it was riddle with many difficulties.
I suspect lack of Jack Tramiel leadership and lack of Shiraz Shivji involvement entire "TT" project took to much time until final product. Take a look at: http://www.atari-computermuseum.de/tt030.htm (Atari EST)




on other hand, Falcon project was based on ST since ST was proven technology and it made solutions by it's own for graphics and sounds independently from TT design.

I suspect that two totally different teams work on TT and F030!
their goals was different, same goes to technical specification (TT trapped in late 80s (lack of 100% VGA compatibility; "double Shifter design" for graphics, moving from 020 to 030, from 16MHz to 32MHz in middle of design process... TT had lot of problems at birth :/), Falcon with new TC chunk mode like "modern" VGA Mode 13h...)

TT already doubled Shifter access to RAM in comparation to ST - twice as ST (64bits). If Falcon follow this logics than you will need 4 banks for ST-RAM to achieve even greater resolutions.

It almost seems like they took a very early stage, proof of concept prototype, and sent it into series production...

Falcon looks like this but since it was obviously in fall of '92 that you _can not_ compete with WinTel power it is good that we have Falcon at all! Real shame is that there was no 32bit ST (TT or EST) in ~88 - 89 (around same year as Macintosh II).

[martyg]

EXCEPT (!) that I come to clash with Marty Goldberg (co-author of book: "Atari Inc. - Business is Fun") at Wikipedia (see "talk" page on Atari Falcon page, section "16 bit bus" link)

What do you mean came to clash with me? You tried posting some stuff there using this forum as a reference, I stated forums aren't considered valid references there and it would need a different source. Then you made a weird claim about me not wanting the info there because we're saving it for our book, and I explained again that had zero to do with it and that forums aren't a valid reference there. The "clash" is purely in your mind and that's the second time I've seen you make more out of it here than it was.

[calimero]

What do you mean came to clash with me? You tried posting some stuff there using this forum as a reference, I stated forums aren't considered valid references there and it would need a different source. Then you made a weird claim about me not wanting the info there because we're saving it for our book, and I explained again that had zero to do with it and that forums aren't a valid reference there. The "clash" is purely in your mind and that's the second time I've seen you make more out of it here than it was.

Dear Marty, let me quote your entire post.

Question is simple:

do you know something about this topic "Falcon designer" ?

[martyg]

What do you mean came to clash with me? You tried posting some stuff there using this forum as a reference, I stated forums aren't considered valid references there and it would need a different source. Then you made a weird claim about me not wanting the info there because we're saving it for our book, and I explained again that had zero to do with it and that forums aren't a valid reference there. The "clash" is purely in your mind and that's the second time I've seen you make more out of it here than it was.

Dear Marty, let me quote your entire post.

Not sure what that's meant to imply, but ok.

Question is simple:

do you know something about this topic "Falcon designer" ?

As I mentioned to you on Wikipedia, we're still focusing on the earlier years right now and haven't organized all the info for the 90s period yet (or done the full interviews related to these projects). Curt does have the original engineering log for the Falcon project and early notes for the Sparrow project as well. I asked him to look up the team names for you (Curt mentioned Richard was involved with the Falcon but not the sole person responsible for the design) and he'll have to dig these out of their storage area and get back with the info. We'll be out in California doing former employee interviews en masse in September as well.

[calimero]

Thank you !!!

[dhedberg]

As I mentioned to you on Wikipedia, we're still focusing on the earlier years right now and haven't organized all the info for the 90s period yet (or done the full interviews related to these projects). Curt does have the original engineering log for the Falcon project and early notes for the Sparrow project as well. I asked him to look up the team names for you (Curt mentioned Richard was involved with the Falcon but not the sole person responsible for the design) and he'll have to dig these out of their storage area and get back with the info. We'll be out in California doing former employee interviews en masse in September as well.

Just want to tell you that I appreciate the time you put into writing the books! I bought the first book and it was a great read. Looking forward to the second volume!

[martyg]

Just an update, Curt had another heart surgery (this is hopefully the last) so he hasn't been able to dig into the archives (literally dig in). But in the mean time in honor of the creation of Atari Corp. 30 years ago today (July 2nd) we've released the cover of our forthcoming second book, Atari Corp. - Business Is War: