Peter wrote:Nasta wrote:As the author of said vaporware, I can only agree. When I did Aurora I hoped it would get the QL out of it's old case but no such luck...
If the announcement was about a variant of the GoldFire, then why does it say "Miracle Systems" and not "Nasta"?
Are you sure this was not sort of a competitive project from Stuart Honeyball?
Quite sure.
In fact the reason the Masterpiece graphics card never materialized from Stuart was... well, not exactly simple and I can't be completely sure, but I think Stuart was at the time slowly getting out of the QL world. One thing I know for sure his partner in Miracle Systems (can't recall the name at the moment, the script 'To Dawn' written on SGC PCB artwork was a dedication to his wife) left Miracle and actually got a job designing motherboards for the then new Pentium Pro.
At the time Stuart was looking at several things like the Masterpiece, and an adapter that would make it possible to connect the SGC to the PC in a similar fashion to QXL, but it would be cheaper (and slower).
In any case, at the time I was experimenting with a LCD adapter card. I was at one of the Italian 'Garage meetings' at Davide Santachiara's garage when Stuart talked about the Masterpiece so I asked him about some details so I could do a compatible LCD card that ran in parallel (so you could get a monochrome mode 4 display on the LCD).
Then there was the meeting near Munich (don't remember the name of the place), I met Marcel there for the first time
and I actually had a prototype of the LCD card running on a monochrome LCD recycled from an old laptop. It was based on the MAsterpiece specs I got from Stuart but a completely stand-alone design. Masterpiece was originally ment to go inside the black case, plugged into the 8301 socket, with a ZIP cased 128k VRAM facing down between the bottom row of the old RAM chips and the front of the case.
At the next meeting I saw Stuart, I think in Eindhoven, he and Tony Firshman pulled me to the side and I was given a large folder, a bunch of AMD MACH CPLDS, a set of MACHXL floppy disks and a hardware key. It turned out to be the work Stuart had done on Masterpiece. Basically, I was told 'You seem to be competent enough, so finish it or redesign to your liking'. Needless to say I was very surprised and actually a bit sad, as I could see Stuart was really losing interest, which was later confirmed by Tony F, he said that he kept dragging Stuart to meetings since quite a while before that. So, Stuart helped me immensely by providing a huge leap with Aurora. Most of the spec and logic was already done on paper but having a tool chain (which he paid for a LOT) and actual chips to work with was... well, a quantum leap, no pun intended. I ended up not using anything of the Masterpiece logic. The first Aurora prototype used 3x 64 macrocell MACH130 CPLDs, and 2x 256kx 4 VRAM chips, and if I remember correctly I was showing it at the next Italian meeting in Regio Emillia, using Minerva and regular PTR_GEN having poked proper values for bitmap start, stride and limits in to the con driver linkage block
and in mode 4. 16 and 256 colors were just demos in superbasic, no drivers back then.
That was before the German meeting where we first met, and to be honest I don't remember if I was showing off the perfboard prototype or the very first real Aurora board there.
Even back when Stuart was thinking about a SGC to PC bridge board, there was some discussion on the next 'GOLD CARD' accelerator. Both Tony and Roy Wood were subtly pressing him to do it as he was then the go-to guy for that sort of thing. I know he was reluctant and at the time he said they did look at getting the 040 onto such a board but the power budget was simply not there.
Later on, when the first MCF5102 ColdFire happened, it was Stuart that introduced me to it, and in fact gave me a preliminary data book he got from Motorola. I think it was just before I moved to the US, and I think I also got a 20 or 25MHz sample chip from Stuart. I did get the 33 and 40MHz version while in the US, and again I think at the Vermont meeting Stuart gave me a CD and book he got from Cypress for the new CPLD line he was considering for such a project, but it was quite clear he was not very interested in doing it himself. I remember already back at the German meeting Tony F. told me Stuart was involved in some sort of EMS (Electric muscle stimulation) job, if I recall correctly, Laurence Reeves was also involved, programming the uC that ran that hardware, and it either was already or was becoming far more lucrative than anything QL. So, I was the one that got subtly pressured by him, Tony F and Roy Wood to do a successor to the SGC. What I also recall is, the last project Stuart did, that Tony F, sold, was the ROMdisq, and I also remember the CPLD logic for that was done on an older software that could use a schematic capture tool Stuart already had, so it was originally designed using 74 series logic macros
It should also be noted that Ron Dunnett (Qubbesoft) distributed QUBIDE and Aurora, and later got involved with manufacturing of either the last batch or the one before that of SGCs, as Stuart did not want to get involved past getting the parts, I don't exactly remember why but I recall it was something to do with a person or persons who did hardware assembly (soldering etc) moving to another job and the replacement did not do a good job of it, because on my visit with Ron I got do do a lot of re-soldering of DOA SGCs. Also taught Ron how to solder SMDs properly
In any case, it was a much more expensive project and Ron was getting out of the QL scene as well, so one more reason for Stuart, Tony and Roy to subtly and not so subtly pressure me into a SGC replacement, and were probably willing to put up the money.
What I can tell you for sure is that I was asked to spec it and the article was mostly actually written by me (the 'I' in the article towards the bottom half of the page was me, actually).
Of course, the Q40 appeared - and I am sure you remember us all talking over dinner and Roy deciding to start initial production (which I remember was rather ill-fated until Derek took over). I made it clear I could not finish the (then still un-named) SGC successor soon since I was leaving for the US soon. I actually met with Tony and Stuart the first time I want and Tony and Roy when I actually moved as I was flying out through London and had long layovers both times.
After that, what happened is what I wrote above... the sad part was, I actually had a perf-board version of a MCF5102 GF made, with the logic spread into 2 MACH 231 chips as used on Aurora, with a 72-pin SIMM for memory, even before I moved to the US, hence the specs in that article. Once there, it soon became clear 68060 was the way, and in fact Tony F. helped me finance a batch of 68EC060/66MHz... I still have them
though all the logic done for the MCF had to be substantially redesigned and the whole thing re-done for a non-multiplexed bus. There was even a PCB design done half way and later re-done again half way for SDRAM (SO-DIMM as used in laptops back then).
Of course most of this was not common knowledge in the QL community back then and was discussed in private, but I suppose no reason to not set the record straight now. Can't even begin to say how much I was sorry to hear Stuart passed away. His nudging me to do Aurora and more actually helped me a lot professionally later on, for which I will be eternally thankful, ditto Tony F. without whom it would never have happened in the first place.
