Darklord wrote: Thu Jul 19, 2018 6:42 pm
I've moaned and groaned about the lack of a "one size fits all" upgrade for ST's and
specifically the STacy for years and years. There are certainly some great products
out there and people like Chris (Exxos) are doing wonderful things.
Yeah , there are so many revisions of the STFM its unreal. Because space is limited the board has to be specifically designed for that one revision to fit, which means probably 90% of people wouldn't be able to fit my V2.2 booster for example.
The V1.5 booster "fits all" but sales wise, it was a huge disaster
The only extra option was DUAL-TOS on the V2.2.. but expensive way if people wasn't using the booster. So "one size fits all" actually doesn't sell
Ultimately I won't be doing any other PCB layouts other than STFM, STE. STE is easier as mostly all the same layout.. STFM, I'm talking the one particular revision I always use. but if I ever get the design finished, I may be open to deals for others to use my schmatic to create new PCBs for other machines like STacy etc. But of course we get onto licencing issues and risk of being "ripped off" like what happened with TerribleFire, PCB layout is not easy with this stuff either.
Darklord wrote: Thu Jul 19, 2018 6:42 pm
Some people would say I'm being unrealistic to want one board that gives:
dual-TOS (at least 2 versions)
Alt-RAM
accelerator
IDE mass storage
FPU
Not unreasonable, but takes maybe 6-12months or more to develop and test each item easily. FPU I am 99% sure it won't be added on my stuff. Basically no software will use it, and it takes up huge PCB space which is at a premium. Plus huge time and cost invested in pcbs, time to create and test all.. just don't see it worthwhile spending the time on it.
FPU stuff works different on the 68000 as it has no proper FPU support. So its called "IO MAPPED". on later CPU's like 030 it has proper LINEF support but works totally different. So a program written to use a FPU on a 68000 isn't going to work on a FPU in the falcon for example. (at least AFAIK). When d.m.l and I talked about FPU support for GB6, we decided not to even bother coding test routines for it. Because if you tested a FPU in a 68000 system, it would benchmark way slower than the same FPU in the falcon. So the test would be meaningless (which what happens in GB3,4 with crazy results)
IDE I am still in 2 minds over, but firstly need to see if there is any space on the PCB left... but really I don't want to re-create another IDE solution. I'd rather develop some internal interface like a internal ultrasatan which solders over the DMA chip.. There are other reasons for that, but won't go into it all now.
