exxosuk

There is no more room to make the port "longer" without case re-design. Considering the quotes I have had are around the £12,000 mark for tooling and £300 per case.. I don't think that is gonna happen.

I used Ghostlink long time ago to transfer my falcon hard drive over to my pc as a backup These days we have cosmosex and wireless networking

Always the same though. I can sell something for years and nobody buys it, then the week it vanishes from sale, everyone wants it. This board isn't likely to be any use to people who want a stock machine. My thoughts are that, if you buy a motherboard for £50, then my 4MB kit and CPU booster, then thats £180 easily just for basics. Add re-cap kits etc etc, and a new board could be done for around the same price. Plus rather than me stocking various kits, I just wouldn't need to anymore. If people want a upgraded new machine then they buy this new board, or buy re-cap kits to keep older machines running. It just takes to much time to re-invent kits all the time for various machines, I just don't want to do that anymore. Time better spent just producing a new motherboard then anyone can easily update it. For example, much space can be added for a 030 card. Space is a huge problem in the STF machines, PCB layouts make it 90% impossible to even fit in some such machines. So whats the point it designing a 030 addon where almost nobody would be able to physically fit it. There is also the same problem with the cartridge port as the CPU. Yes the 030 is nice, it has caches, can run fast, I want to add it into the board eventually, but as a option. I'm not going to force people to have a 030 only system as 50% of people want a 68000 system. The 030 can be a option to addon if people want it. So everyone is happy. The cartridge connectors are $10 each from BEST. Pretty high price. I don't want to spend $100+ on 10 connectors. But again mixed views. 50% want new port, 50% don't. But if its something new, which can be backwards compatible with original carts. Then really everyone should be happy. But still that connector will be needed on a adapter board. Or as I i hope, just to add both options on the motherboard and let the user decide which port to add. At least it can be changed for the "Super IO port" if someone designed a cartridge for it. Though as said before, there is basically nobody designing new hardware these days.. other than myself and a few others, even so, I think its unlikely new cartridges would get created. But I don't wan't to rule out the possibility, and don't want to force some new connector onto people where nobody will ever use it. So bottom line, I will likely use original port, it will have to be for beta testing anyway. Then I will try add in option for the "Super IO port" in a way people can just plug in a card to adapt the port.

Yeah, I would like to create everything everyone wants. Though even simple things, like re-creating the simms, it took 2 months on and off until that was done. That is a super simple project in itself. So even simple things people want, take a long time, anything more advanced, well, basically forget it, I just don't have the time. People just don't realise the work involved. Most people think something will only take a few hours to do, well maybe so, but more hours to design PCB, build test, debug, takes weeks, months or even years. Multiply that by 100 people wanting 100 Simple things" and thats more like 10 years of my life gone. Work on my boosters has hit a standstill because I am struggling to keep up with stock in my store. So designing anything else, well, not enough hours in the day. There is also the fact that no matter what I develop, likely only 5-10 things of them will ever sell. So I am reluctant to keep spending months of time to develop something which hardly anyone will use. More to the point (again) with all my kits, I have pretty much re-designed the motherboard anyway.. Future kits will become a lot easier to make and for people to fit. Anything is possible if enough people work on things. Just people have to realise, I am not "super man" and can't create the 200+ hour day. So 99 projects won't get done, thats just how things are. I do know what everyone wants, and yes it will be done as addons eventually, and yes they will be included as "Stock" on the new board, but that is likely some years away as there is a lot of work to be done.

The new port with 32MHz CPU can be run at 32mhz as well. So there are future options as well. As mentioned before, Likely kickstarter will have to fund this. I was using profits from my store to fund things, but last year I have lost over £2,000 so I just can't keep doing that Then it will be up to the community if the new board gets done or not. I am working on the design with Rodolphe's help. I have done page 1 , about half way through the second page now.

In light of cartridge connectors being hard to get and the fact people at least want the RW line on the port.. I put this to a vote. New cartridge port, likely 60 way header (will fit basically the same as the existing cartridge port) which will have more address lines, and /DTACK, /AS and whatever to create a really good expansion port. Or keep the port as it is, no changes. Though possible a GND can be sacrificed to add the RW line in there. Old carts would still work, new carts can use the RW line. The exception is that if a new port is created, a adapter block can be made to convert the new IO connector to the old format, so people could still use the old cartridges. Though this will be hard since the cartridge connectors are near impossible to find now. From the motherboard point, I am hoping to be able to add both connectors, but obviously only one can be used. Then the user could choose to solder the new port, or the old one. In anycase, I thought I would put this to a vote and see what people think about such a change. While I am all for better upgrades, it does beg the question if anyone would even bother to develop or use the new "super IO" port.

Yes believe me, I have every suggestion at least 100 times But I don't have a flash solution, or IDE, or fast-ram, or RTC, etc etc *yet*. I can only work on 20 things at a time, so a lot of stuff has to wait until I get around to it. Meanwhile, once the board is done, nothing to stop people from developing their own solutions.

Some lines sound like what I read in your blog I really want to re-create the MMU in a new PLD. Though ditch the DRAM and use SRAM. While I understand basically (like everyone else probably) the "jist" of how it works, there is no real information out there on it. The internal design of the MMU is to unknown to replicate. The ASIC designs seem lost unless by some huge stroke of luck you have them still ?

Caches give about 25% speed, at least it does on the falcon with caches. You would have to benchmark the MSTE vs a 32mhz STE to see what the difference is speed wise. All instructions on the 68k will run double fast, but the problem is, ST-RAM holds up most instructions, so they don't run any faster. ST-RAM runs at stock speed, so does instructions running in ST-RAM. Instructions like int-div run faster because they can take 100+ internal cycles inside the CPU. So there more MHz the faster things goes. Just look at the STE booster figures... http://www.exxoshost.co.uk/atari/last/V1STE/ste32mhz.jpg

Just so people are aware, "16mhz" in steem is 200% "across the board" increase in speed. A "16mhz booster" cannot increase ST-RAM speed and is a bottleneck on real hardware. So while you may get 200% at 16mhz in steem, realistically you will only likely get 25% overall speed boost on real hardware. The exception is said above, 3D games as long divisions can run at near 200% speeds as they do not access RAM. Most instructions which do not access RAM will run faster. TOS will run at higher speeds due to 16MHz access to ROM. Though games do not generally only use OS calls. Most game CPU time is accessing RAM.

I have had a lot of suggestions for this new motherboard. Some will be added (basically easy ones for now). Such as someone requested a jumper link to swap drive AB select wires around. So that type of thing can be easily added. I will post a better "overview" of the board on my blog soon. Though work is ongoing, a few hours a day or week. There is actually a lot of "background" changes to the design. Though to many to list at the moment. I have probably wired up about 80% of the design so far. I am really busy next week with work, but I hope to get the schematic done in some form over the next couple of weeks. Then after a lot of double-checking, we can look into getting a prototype board designed and built http://exxosnews.blogspot.co.uk/2017/07/stfm-rebuild-overview-of-basic-changes.html

Did you work on the ST MMU ? Finding information about its internal works is near impossible.

I had similar issues. My hard drive used a normal SCSI drive, so with the right cable can be connected to the Falcon's SCSI connector. Internal IDE to SD works on the falcon. I use one in mine. Might be worth checking you have the latest HDD driver software as older ones probably won't work with SD cards.

For the moment I am not really considering RF. I asked on another group and people generally said they didn't use it. So it may just be left off. Not sure yet. As for video fixes. That solution doesn't likely exist. Something I don't have time to do. My hands are full with running my store and I hardly have any time to work on my boosters anymore, nevermind anything else. My work will probably take 2-5 years to complete. So if people want such things then someone will have to design it. I will do motherboard, add in some needed fixes, and put expansion ports. If people want to make it better, then people will have to do the work themselves. I really don't want to spend the next 20+ years designing Atari mods. This is also why I want to do a motherboard which people can upgrade easily in the future. As said above. A lot of solutions are already done, my store is full of them for example. So better to build motherboard with current "addons" in mind. All the popular ones, like IDE etc. My IDE solutions isn't built or tested, and could be a year or more befores I get time to do it. Though someone else could design a addon anyway. All this stuff doesn't need my own personal attention.

Yes. Probably a expansion port over DMA IC near the back of the case would be dedicated for hard drives.

Likely the port will be similar design as falcon expansion ports. I aim to re-use those old 4MB falcon RAM cards for this new motherboards. As for expansion design, I don't have one yet. mostly all CPU pins routed to the port, as most upgrades need address and databus, and only CPU really offers that. Expansion port design itself is for another time. Like I said in my blog, there could well be 3 expansion ports depending on space, maybe even some smaller ones. All IC's can have breakout headers to replace IC's with FPGA if someone wants to develop that. All my upgrades will be taken into account and to add booster will be as simple as to plug in a PLD. Similar with floppy upgrades, 4MB likely be standard anyway. Anything can be done. As said before, its just to much work to continue developing upgrades for various motherboard revisions. Spending time tracing bugs in motherboard, spending time trying to route PCBs to actually fit some style motherboards. It would be less work long term to create new motherboard and just have proper expansion connectors like on falcon. While board may be expensive, upgrades would cost many times less. Bottom line for a upgraded machine cost wise would not likely be much different than new motherboard. I'm not building in 030 32mhz system as standard or any upgrades which will break any software. There are enough machines , clones, and emulators doing that already. This is a stock machine with some much needed changes without breaking anything. Then if people want to upgrade, they can. If they do not, then they dont. Idea is to allow easy upgrades for future and fix all bugs and old designs on motherboard and have a new standard design , not 20 versions of motherboard which Atari seem to have. First step is to re-create motherboard and see what interest is like to continue or not after that. If the community don't like this idea, or don't want to support this project, fair enough. Saves me a lot of un-paid work. Everyone has this one time chance to help create something really good. I think it would be a shame not to do this new board, but its not really up to me.

If you replace all IC's with FPGA, and add all the changes, then you basically end up with a MiST. I am not looking to re-create it. If that is what people want, then they go buy it. As for using Atari IC, no choice. If someone wants to design cycle accurate IC and make them better, sure I will add them. But for now, there is nothing. As for addons, sure I already mention this. But if people want everything standard, like IDE, fast-ram, new DMA, new video, new everything, not realistic project for one person to do all. Sure I am working on my own IDE design, needs time to debug and test. Same with other 20 projects I am working on. My aim is to create a new motherboard which is more stable with expansion ports. If people want more features, then the people will have to design it as I dont want to spend the next 10 years only working on Atari stuff. In time the board will have all my upgrades as optional extras, but I am still working on a lot of designs which may take some years to complete. As said before, main issue with Atari design is slow MMU, this needs to be updated with SRAM and working at 32mhz with the CPU. If someone designs that, then it can be plugged into new motherboard. If nobody designs it, then it will never get done. Same with video etc. But also it moves away from legacy software and compatibility. I want 100% clone. So likely only using original Atari chips is ever going to happen. Also I can't give any more detail than I already have. I can't comment on a project which has only just been started. I only plan to create a stable STFM clone, one easier to add upgrades for future. Beyond that, others will have to do some work if they want more features.

Its a STFM clone. Though someone asked if ATX holes could be placed into the board also. I guess holes are not a problem to add if there are cases which will physically fit the board. Incidentally, I have almost done page 1 of the schematics for interested parties. Just started on page 2. There is some mods going in, for example, I am adding resistors into clock lines, the STE has some resistors on some buffers like on the 1772 circuit, so I am also adding those to the STFM. The audio circuit differs from ST to STF, so ran some simulations on both and tweaked it a little. Similar with the video circuit, its being more like the MEGA ST version, a lot similar, Though again I have simulated it and tweaked it. Actually overall taking a lot longer than I thought, so as I am really busy next week, likely be a couple of weeks before the schematic is finished.

As there is talk about motherboards lately...This isn't to discuss "addons" and what the new motherboard should or should not have, or what motherboard is to be done. This project is to create a clone of the STFM ONLY , which basically has an array of expansion ports to upgrade easily. There will also be some general improvements and modifications, such as moving away from 6 chip ROM to a single 4096 PLCC ROM. It's clear with so many motherboards going faulty, and so many revisions, that its just not viable to create a "one size fits all" upgrade for all revisions. Not only that, I really am not motivated to create 20 versions of kits to fit all motherboard revisions. Its just not a good use of time or funds. Something just has to change. The idea of doing a new motherboard has been talked about for some years already between myself and Rodolphe Pineau. Please note, this is a STFM clone project. I'm not going to consider a 060 falcon motherboard, MSTE, STE, TT etc etc or anything like that. I want to create an easy open source STFM platform which is easy to upgrade and people can develop their own addons easily for it. If there is enough interest and enough people willing to put cash into a kickstarter campaign, then the motherboards can be prototyped and eventually become a reality. More info see my blog post. http://exxosnews.blogspot.co.uk/2017/07/super-stfm-motherboard-project.html

Try this demo. http://www.pouet.net/prod.php?which=66787 That is up to date to take into account various timing issues on various ST's.

Only things which break that I am aware of are games which use rasters. There was 2 games on the STE which broke, but I can't remember which ones now. 3D games generally run faster because the CPU can do math divisions a lot faster. Other games in general maybe around 25% booster on 32MHz booster. It depends what instructions are used in the game. Games which use ROM or do some long divisions will run a lot faster. GEM apps will see a huge boost in speed as access to ROM is at 32mhz not 8mhz. Overall, only demos are going to break because they all need timing accurate system to work. The exception is there are the odd game. I doubt anyone has compiled any lists though.

VHDL is code running in FPGA stuff to emulate the chips. Such as code emulating the 68000 instructions. It breaks timing dependent things like rasters, fullscreen demos etc. PLD that I am talking about, is "gate level" stuff, I mean it would be a direct clone of the blitter, just running in a PLD which can still be obtained and programmed. So it would be identical to the original Atari blitter.

I'm going to start work on the blitter soon as the design has been recovered. http://exxosnews.blogspot.co.uk/2017/07/blitter-thoughts.html As for MMU, GLUE, SHIFTER, DMA etc, there are only VHDL cores, and that is already built into the MiST, Suska etc. I would love for a MMU re-deisgn, just using SRAM, though nobody has ever been willing to spend the time into developing one If the IC design files could be found, at least I could clone them into a modern PLD and make changes.

+1.

Not that I know of. Mostly I don't use less than TOS104 though anyway.