The plan... 2018

New sub forum for version 2 of the STF remake board.
User avatar
exxos
Site Admin
Site Admin
Posts: 5109
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

The plan... 2018

Post by exxos » Fri Dec 07, 2018 11:18 am

The mongrel edition board is the planned version 2 of the remake board. Now before anyone gets excited, this is probably going to be more like towards the end of next year before this board sees the light of day.. There's a lot of work to be done yet..

The majority of the current fixes are listed in the revision G thread below..

viewtopic.php?f=19&t=1230


Because the GLUE & MMU are basically impossible to find, I have purchased a fairly large stock from Atari freaks of the SMT version. These will have to be soldered directly to the motherboard but as there is a shortage of these chips, is the only real option.

The mongrel edition got its name because we are actually going from a socket to a SMT chip which cannot be updated or changed in the future, at least with the socket it could open the door to the FPGA world. Though as the FPGA stuff is looking like it has a bleak future, the project is basically going on hold while the mongrel board is done. At least it will have all the updates some other things added on as well.

Things to do circuit wise is ...

To fix the reset switch footprint, I still haven't done this!
Add the video DAC circuit and remove the old RGB drivers.
Change the GLUE & MMU to the SMT versions.
1772 & DMA circuit needs a total rebuild to include current modifications.
I plan to make the next board full-sized, but I need to work at all the hole locations etc which will take a great deal of time itself.
I also plan to add internal DMA header to fit the internal ultra Satan.
There will be a DIP 68000 connector somewhere so my SEC BOOSTER can plug into this board eventually.
A lot of the decoupling capacitors will likely be changed to SMT versions.
The reset circuit needs updating to the new chip.
IDE - Assuming the project is up and working at that time.
"Super GLUE" - Possible to add in a PLD for TOS206 decoding. Be the same PLD which does the floppy GAL logic as well. MAybe even the "DoubleST" mods. Needs a lot of thinking about yet though...
LS chips replaced - A lot of the chips will be replaced with single gate "tiny logic". A lot of gates are not used in some of the LS chips. Also there is a lot of "double inversions" in the circuit which can be changed to a single buffer chip. The inverters were most likely used in that way because of the need for a open collector high-output drive current gate.. But with modern parts we can do this with a single noninverting buffer. So a lot of the chips can be removed and compacted into smaller gates. Also possible if we have a PLD on there, a few functions could be rooted through that as well.


Other possible thoughts...

Socket to add my HD V7 floppy GAL mod. This will allow 1.44 floppies, does the media change fix, and hopefully with a little recode in the same chip, drive A/B swap can be added. But that is a subproject needs to be tested out first (lack of time again for that).

Also similar I want to try and add the "Double ST" speed up. Where the MMU & RAM can be double clocked for a across the board 200% speeds. The problem with this project is that the shifter gets double clocked also, and is not very stable in mono mode. I hope the FPGA HDMI board can be adapted to fix that problem in the future.

I was of course looking towards the FPGA blitter. But of course this project is turning into a epic amount of messing about. So that project will just have to wait until I have more time for it all.





RAM..

This of course is a problem.. At the time I created the original schematics Falcon RAM boards were available very cheaply and easily obtainable. Though of course as a year passed before the board was produced, these pretty much vanish totally leaving everyone in a bit of a fix.

I was planning on adding DRAM directly onto the motherboard, but as these chips are hard to solder and hard to diagnose, they would be better being on their own board so they can be tested individually. But this basically means my MMU STFM board, why don't want people to be forced to buy this rather expensive kit. Creating a new board take a great deal of work, but it basically would end up being the Falcon RAM board again. It still is rather expensive.

I see the overall cheapest and easiest solution is to use right angle simm sockets. This way the simm and replaced towards the front of the motherboard, under the keyboard, out of the way, where there is not much height available for much else anyway.

The sockets are fortunately expensive, they seem to be £15 in RS.. But it is still half the price of my Falcon memory board. It rules out some connectors and soldering and another PCB, so likely this will be solution done.

Other thoughts..
It is likely the new motherboard will have to be done on 4 layers. This will double the price of the board unfortunately. Also as the board will be larger, assuming I do the full-size board, that is doubling the price up as well. Though I have been using JLC for some boards lately and they do not seem to bad overall. They are the cheapest for projects like this.. I do not think the price will end up being more than the original board anyway. As always import taxes and charges tend to add a large lump onto all this stuff :(

If I have forgotten anything, then shout out will update this post..
4MB STFM 1.44 FD- VELOCE+ 020 STE - 4MB STE 32MHz - STFM 16MHz - STM - MEGA ST - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - HxC - CosmosEx - Ultrasatan - various clutter

https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
PaulJ
Trusted Guru
Trusted Guru
Posts: 311
Joined: Sun Apr 08, 2018 1:14 am
Location: USA

Re: The plan... 2018

Post by PaulJ » Fri Dec 07, 2018 2:54 pm

Exxos, this sounds like a great plan but I believe it would be worth while to create a gold version of the current board to eliminate the unfu*kit boards, errors and alignment issues. Not making any major changes I believe minimize the time required to fix the eagle Gerber files. If we were to:

1: Add the missing pull-ups.
2. Fix the reset circuit pc board error. (On my board the reset circuit works flawlessly)
3. Add the mono driver.
4. Fix the video output connector errors.
5. Fix the floppy connector wiring errors.
6. sync fix
7. 16MHZ fix
8. Glue data line fix.
9. MFP Fix
A. Ram RW Fix
B. 1772 IRQ TYPO fix
C. Keyboard resister fix and keyboard missing connector pin fix.
D. ACIA databus fix
E. Floppy missing resister pull-ups.
F. 1772 DMA Bus Short Fix.
G. Parallel port fix.
H. Move connectors to align with case holes.
I. Move power supply holes to fit actual supply.

Making these changes to the existing board should be fairly quick. Even though I'm pretty proud of my atari box case its really not a long term solution. If you don't want make the effort I believe I could get the changes made but I suspect you don't want to let the eagle files out. I'd be willing to help in any way. What do ya think Exxos?

User avatar
exxos
Site Admin
Site Admin
Posts: 5109
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: The plan... 2018

Post by exxos » Fri Dec 07, 2018 3:17 pm

Those things are already done in the REVG board (aside from the extra logic for the DMA/1772 stuff). But it is not routed.. If there is enough people to warrant the board being produced then I can do that. But I wouldn't be doing any kits or parts for it like last time.

IMHO I am not sure its worth the trouble as it doesn't bring anything "new" really to the table over what people already have.. But if people want the REVG produced I can produce it.

EDIT:

The REV G does need actually checking that I haven't missed any mods or messed something up.
4MB STFM 1.44 FD- VELOCE+ 020 STE - 4MB STE 32MHz - STFM 16MHz - STM - MEGA ST - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - HxC - CosmosEx - Ultrasatan - various clutter

https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
PaulJ
Trusted Guru
Trusted Guru
Posts: 311
Joined: Sun Apr 08, 2018 1:14 am
Location: USA

Re: The plan... 2018

Post by PaulJ » Fri Dec 07, 2018 3:35 pm

I didn't relize you had put the fixes in. I'd be willing to buy a few of the boards. I would also be willing to help in any way possible. Maybe a parts list from mouser, I have most of the info. I would think Forgotmyname would rather build a board that doesn't need the unfit*kit boards and fits in a normal case but that my opinion. I'd vote for this. What is the price structure getting these boards made?

Forgottenmyname
Trusted Guru
Trusted Guru
Posts: 708
Joined: Fri Mar 23, 2018 8:22 pm

Re: The plan... 2018

Post by Forgottenmyname » Fri Dec 07, 2018 6:20 pm

PaulJ wrote:
Fri Dec 07, 2018 3:35 pm
I didn't relize you had put the fixes in. I'd be willing to buy a few of the boards. I would also be willing to help in any way possible. Maybe a parts list from mouser, I have most of the info. I would think Forgotmyname would rather build a board that doesn't need the unfit*kit boards and fits in a normal case but that my opinion. I'd vote for this. What is the price structure getting these boards made?
I'm actually going to be looking at the whole unf*ck bit with an open mind and will see what I can do with this next board as I slowly build it up. For instance, fitting headers instead of sockets in some cases and not fitting half of the original RGB circuit and instead fitting the unf*ck board I have and connecting that direct etc etc.

Whilst I appreciate PaulJ's opinion and he makes some good points, personally my vote would be for the Mongel board but from the point of this is the next step in the evolution of the ultimate STF remake and with it would come new challenges etc.

What we have now with the STF remake is a rather ugly but functional PCB. I'm all for us developing the new VME Bus expansion and then figuring out what we can do with the "new" STF, whilst keeping it compatible, etc.

But as Paul says, if I had a board that was easier to put together then yes, I would be overjoyed to assemble it but still would vote for the next Step. Even got some plans as to how I'd use the expansions etc.

User avatar
exxos
Site Admin
Site Admin
Posts: 5109
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: The plan... 2018

Post by exxos » Fri Dec 07, 2018 6:55 pm

The quickest step overall without adding to much development time, is to take the REV G board, add the new video DAC circuit, add the new DMA stuff, new reset circuit. As Icky has already confirmed those working now.

I would also add the DIP 68000 so I can develop the SEC booster with it in mind.. Pretty much leave everything else "as is". Though I really need to sort out the hole locations. They likely won't be perfect, but if we are talking about a spin of boards, then it makes sense to at least get something done with the hole to see how far out they are (i'm talking about screw holes more than anything here)

I would probably do the 72pin simm in place of the falcon RAM board as well actually...

Everything else needs development time and will take a long time to complete.

I do tend to agree that it will be better to produce a board with all current fixes.. But I just think it is a whole lot of work and expense for you guys to basically just have a neater version of the board which you already have.

In terms of expense, I think something like 30 boards were made at about £35 per board. I did work this out before, and I think 20 boards was more like £45, and 10 boards £60 or something like that. I'm going to use a cheaper fab house this time, will likely the board will be four layers, so the price doubles, but I'm hoping the price will be half the different fab house.. ( End result is the PCB be more expensive before worst-case, But if you we went for a full-size board, then this is going to increase costs anyway. But at this moment in time I don't think there is any point) The PCB price itself is actually rather low, but the setup charges, postage, and import taxes what ends up ramping the costs up considerably. :roll:

But bear in mind, it will be a couple months before I really get chance to do anything with the design. I've got these STE boosters to finish, plus the rest of my PSU's stock should be here next week, plus I have the V2.2 booster to build and test, plus I need to actually finish the SEC design.. So I have a lot of work to catch up on yet. I don't want to start mixing in the new motherboard into all this as this is how mistakes end up happening because I am doing too much all at once again.
4MB STFM 1.44 FD- VELOCE+ 020 STE - 4MB STE 32MHz - STFM 16MHz - STM - MEGA ST - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - HxC - CosmosEx - Ultrasatan - various clutter

https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.

Forgottenmyname
Trusted Guru
Trusted Guru
Posts: 708
Joined: Fri Mar 23, 2018 8:22 pm

Re: The plan... 2018

Post by Forgottenmyname » Fri Dec 07, 2018 7:05 pm

Ultimately it's your baby, so we will all help however we can.

If the boards were cheap enough I'd probably buy two. Only problem is if we use the same chips etc, as you've already said, some are now expensive or hard to get, so why not go to the SMT Glue etc but use DIP chips for the easier to get hold of stuff.

You can work out all the logic change bits when you get more time.

User avatar
exxos
Site Admin
Site Admin
Posts: 5109
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: The plan... 2018

Post by exxos » Fri Dec 07, 2018 7:30 pm

Forgottenmyname wrote:
Fri Dec 07, 2018 7:05 pm
Only problem is if we use the same chips etc, as you've already said, some are now expensive or hard to get, so why not go to the SMT Glue etc but use DIP chips for the easier to get hold of stuff.
Is not so simple just to swap to the SMT stuff as yet... I need to build and test the ,MMU I think, adapter first, and I haven't even started on the GLUE libs yet. The adapter board needs to be tested to actually make sure the chip is wired up correctly.. Obviously just plonking this chip on a motherboard with a untested library isn't a good idea. But again, those projects I haven't got time to work on at the moment. It would actually save likely several weeks of time easily in just keeping to the PLCC versions.

Forgottenmyname wrote:
Fri Dec 07, 2018 7:05 pm
You can work out all the logic change bits when you get more time.
Yes, the logic alterations really need my full attention, and I want don't want to do this (or any huge changes) while I am simultaneously working on several other projects at the same time as it will end in disaster. It's all stuff which generally should not take a lot of time, but I have been off work today all day, I've been going flat-out doing stuff, and all I have actually managed to do is change 2 GAL's on the STE booster boards.. Its constantly getting distracted onto million other things which need my attention also, Its not just Atari stuff either. :roll:

It is like any work I do , should take half an hour for example, actually ends up taking the entire day.. Its actually really annoying as I really do not know where all the time goes, I got up this morning and thought to myself I'm going to do these STE booster boards and get them finished, and it is probably the one thing I haven't actually had time to do all day :(

It's like yesterday, I thought I would test the GAL's on one machine, while testing the boosters on another machine to speed things up.. And what happened, I spend the entire day diagnosing those floppy faults instead.. So basically nothing got done booster wise yesterday.. This is just how my days always go, Rinse and repeat daily, then six months have gone and I don't seem to have actually done anything :roll:
4MB STFM 1.44 FD- VELOCE+ 020 STE - 4MB STE 32MHz - STFM 16MHz - STM - MEGA ST - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - HxC - CosmosEx - Ultrasatan - various clutter

https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
PaulJ
Trusted Guru
Trusted Guru
Posts: 311
Joined: Sun Apr 08, 2018 1:14 am
Location: USA

Re: The plan... 2018

Post by PaulJ » Fri Dec 07, 2018 8:02 pm

Forgottenmyname wrote:
Fri Dec 07, 2018 7:05 pm
Ultimately it's your baby, so we will all help however we can.

If the boards were cheap enough I'd probably buy two. Only problem is if we use the same chips etc, as you've already said, some are now expensive or hard to get, so why not go to the SMT Glue etc but use DIP chips for the easier to get hold of stuff.

You can work out all the logic change bits when you get more time.
Although its always nice to have it all I was pretty happy with the current Exxos Atari. I wouldn't want Exxos to spend a lot of time rolling some improvements in which would cost him a lot more time since we knows he satuated. Since he has the rev G layout which has the fixes its a minimal effort to finalize and get it spun instead of generating a new layout with major layout changes.

Forgottenmyname
Trusted Guru
Trusted Guru
Posts: 708
Joined: Fri Mar 23, 2018 8:22 pm

Re: The plan... 2018

Post by Forgottenmyname » Fri Dec 07, 2018 8:22 pm

Totally agree Paul, I'd just like expos to do things that he wants as well as the things to please others if you know what I mean.

ie... not doing the Falcon repairs, STE boosters, etc and working on what he wants.

I'm up for an Exxos STunF*ck remake regardless :D

Post Reply