Does anyone makes new replacement pcb for 2600?

[TwoShedsWilson]

I would love this. I have a stack of otherwise dead 2600s that would be excellent to refurb with something like this especially if you're embedding an A/V mod and a modern power supply. I'd be thrilled to get them out of my basement and back into use.

[calcmandan]

It's already been done multiple times with all revisions of the commodore 64. Don't see why the 2600 can't be done by some brilliant engineers with a fondness of retro gaming on the atari.

Sent from my brand unspecific tablet

[SamhainTM]

I've been slooowly working on a schematic and board design, based on chregu's KiCad drawing for the 2600 Jr. I'm also working on a 4-switch version. Work in real life has been overwhelming, so time is a luxury for me right now.

 

Some of what I'm trying to incorporate:

* Direct replacement for existing boards in a 4-switch and Jr. 6-switch will come later.

* Modernization of components. Not sure if it's going to be mostly through hole, SMD, or a combination of both. I will try to source modern components for everything except the 3 main ICs.

* No RF. All output will be via composite AV. S-Video is also an option if I can figure out the mixing/de-mixing. I'd like to use a chip similar to the UAV, but those components are difficult to source now. Due to complexity, I will not be taking a stab at integrated HDMI output. It's way above my paygrade. There are plenty of external devices to do that now and I may look into wiring one up inside the case.

* Built in pause circuit

* USB-C power. I'm looking at a ZY12PDN power adaptor to supply initial voltage and pass it on to a 7805 for regulated power. 

 

I'm primarily working on a main board to fit inside a rackmount 2600 case, but it shouldn't be too difficult to remix it to a standard board layout for any of the current 2600 form factors. I really want to do a few what-ifs with a rackmount design, mostly for personal interest. My plan on those is to use soft-switch style controls similar to high-end audio gear. Behind the scenes, it will use flip-flop type latching circuits to drive the switch states. I may even wire up bluetooth based controls if I can figure it out. It will definitely be a Frankenstiened system though. Plans are to do other systems also up to possibly a Sega Dreamcast. I'd like to get everything into a 1U (1.75") rackspace for each device, but I may make the 5200 a 3U or 4U height unit to highlight the size of that beast.

 

 

[+wongojack]

Following - I think it would be fun to build a 2600 from available chips and parts.  As others have noted, this has been done a few times with the C64.

 

The discussion about which chips to use is interesting.  Ideally, the board would have room for all the original chips AND the large single chip that replaces them.  That way it would leave the option for some versatility and incorporate the possibility for using future replacement chips.

[SamhainTM]

The all in one "unicorn" chip is really rare, so I won't be incorporating that into the design. The pinouts between the 3 chip set and the unicorn are vastly different and would be difficult to incorporate into a single design on a two layer board. 

 

As for something like an FPGA replacement chipset, it's already been done and is out of scope for this particular project. If someone wants to design a replacement FPGA based RIOT/MPU/TIA set, it can be done as a direct chip replacement and designed in a way that it can fit into the three chips via dip sockets. If they want to go the route of the unicorn chip, then it should be easy enough to redesign the board to accommodate it. Either way, it should be a drop-in chip or set that emulates what the original chips do in order to maintain the "original-ness" of a 2600... otherwise, it would just be another Retron77.

[SamhainTM]

This is the current A/V circuit I'm working with. It outputs Y/C (S video) natively with a mixdown to composite. I'm trying to make the setup modular in nature so that other video circuits such as the UAV can be used if desired. Output lines can be configured to support connecting to normal RCA jacks, S-video connectors, or TRRS connectors.

[SamhainTM]

This is the control panel/switch schematic with integrated pause feature. This is mostly for my rack mount idea and I may end up moving the pause circuits onto the main board to help with portability across board revisions. The switches are all momentary switches connected to flip-flops that latch open or closed when the switch is activated. It will be easy to rehome to the main board in order to use original Atari switches (without the flipflops) though. The only things that would need a case mod will be the pause switch, unless we replace the color/bw switch function, but that would break Space Shuttle, which is one of my favorite Atari games.

[TwoShedsWilson]

I had never really considered rack mount options before. Out of curiosity, what made you want to go that route?

[SamhainTM]

51 minutes ago, TwoShedsWilson said:

I had never really considered rack mount options before. Out of curiosity, what made you want to go that route?

The challenge mostly. I think it would be cool to have bunch of systems rack-mounted along with the rest of my AV gear. I could also hook up way more systems in a limited space and be able to play them with a couple of button clicks to line up the TV input.

 

It also doesn't help that I have about 30 years in the server admin field. To me, everything should be rack-mounted! 

[SamhainTM]

Just had another bright idea... instead of making multiple revisions for each form factor, I may explore a modular approach with a central "CPU" card and form factor specific sleds that it can plug into. That way it could have a standard CPU module and basically tailored I/O sleds for all of the interface components (power, controllers, AV). This would allow for reusing existing switch components and AV/Power connectors as desired. The sleds would have minimal active components, primarily the AV and power circuits only. The CPU module would host the three ICs and the osc circuit. I may move the pause circuit on there also.

[+wongojack]

I am really interested in progress on this.  I think the world needs more of these boards and chips.  I know there is a lot of work to do, but how would this move to production?  Have you considered a kickstarter or anything like that?  

[SamhainTM]

1 hour ago, wongojack said:

I am really interested in progress on this.  I think the world needs more of these boards and chips.  I know there is a lot of work to do, but how would this move to production?  Have you considered a kickstarter or anything like that?  

Honestly, I haven't considered anything beyond today at the moment. I wouldn't really want to get into production right now, but it may be something I would consider down the road. I have a pretty tight schedule with real life, like everyone else that does these kinds of projects, but I'm also all for open-sourcing when possible. Maybe something like the Opentendo project for the internals. 

 

My biggest concerns though are cost and my skill level. I'm not a complete beginner, but my professional career in discrete electronics ended back in 1996 when I left the Navy. I just recently picked it all back up and continue to improve. As for cost, circuit boards are easy to get printed now. The new components can easily be sourced from Mouser or Digikey. The only things that need to come from donor machines would be the ICs and switches (BEFORE people start crying about leaving old machines alone, keep in mind these systems are typically already dead. As for my rackmount version, that's on me to do what I want. I have a couple of systems with cracked/smashed cases and I'm sure there are a ton more in the same condition and beyond rejuvenation).

 

Any way that I proceed, I will be posting my files up on GITHub in the near future. Circuits & boards, case designs, 3D models, artwork, etc.. I'll make it available when I have more progress than what I currently do and I would most assuredly accept any and all help with the project.

 

[SamhainTM]

Forgot to mention the rack mount cases... those can be expensive for a custom designed case and faceplate. Cheap generic cases are available, but still require a ton of work for adding holes for ports and connectors. Some of it can be 3D printed, but will still require time and effort to clean up and label properly. If I end up building and selling, this will be the deal-breaker for most people as the $300 custom case may not appeal to them. This won't apply to the internals though as they will be designed around existing Atari cases.

[+SvOlli]

Fine, now I can dump my effort on doing a 2600 board. The CPU board idea is quite interesting, since I wanted to do a version that's as compact as possible.

What also could be nice would be an option to use a 6502 instead of the 6507, since those are easier to source.

[SamhainTM]

1 hour ago, SvOlli said:

Fine, now I can dump my effort on doing a 2600 board. The CPU board idea is quite interesting, since I wanted to do a version that's as compact as possible.

What also could be nice would be an option to use a 6502 instead of the 6507, since those are easier to source.

Don't give up! Collaborate! Let me get what I have into GITHub and we can work together if you want. 

 

As for the 6502 vs 6507, it's technically feasible, but I may need to learn a few more things before I drop it into the mix. 

[+wongojack]

Oh - I read @SvOlli's comment to mean he will put more effort into working on a 2600 board.  As in I dump/put more effort into it.  Yes - Collaborate, GitHub, go go go

[SamhainTM]

As for github... I will, I will. I promise! Need to set up a repo for this project first and get my junk sorted out. I've got files in 16 different places. 

[StephenJ]

1 hour ago, wongojack said:

Oh - I read @SvOlli's comment to mean he will put more effort into working on a 2600 board.  As in I dump/put more effort into it.  Yes - Collaborate, GitHub, go go go

I agree!

[jhd]

On 4/28/2021 at 9:06 AM, SamhainTM said:

Just had another bright idea... instead of making multiple revisions for each form factor, I may explore a modular approach with a central "CPU" card and form factor specific sleds that it can plug into. That way it could have a standard CPU module and basically tailored I/O sleds for all of the interface components (power, controllers, AV). 

 

 

This looks like the Pioneer LaserActive. It supported various modules for different media formats (and video games).

https://en.wikipedia.org/wiki/LaserActive#PAC_modules

 

This may provide a model or some implementation ideas.

[NuClear235]

Excellent idea!

[+SvOlli]

I'd rather support one step at a time: first build a good baseboard, then see what it can be plugged into.

 

And with the baseboard utilizing a 6502, just using it as a 6507 replacement would be plain simple: connect IRQ and NMI to +5V, and keep A15, A14 and A13 unconnected. A bit more interesting would be adding a GAL chip creating the chip select signals in a way that RAM could be placed at $1000-$EFFF. Or even $1000-$FFFF that can be loaded (with an UART for example) by a small bootrom that then turns itself off. At least that was the kind of expansion that I looked into with my try to build a PCB. But that's something open to discussion once we've gathered enough substance.

 

The other idea I had was to utilize a 6502 B which is capable of running at 3MHz. This could be overclocked to 3.57 MHz then running at the same frequency as the TIA. But this will almost certainly not work, because there's a feedback from the CPU frequency back into the TIA. Also the 6502 B is way much harder to source than the 6507.

 

Rereading what I've just written, I noted that if done right, the "RAM expansion" can be done _outside_ of the baseboard, if the baseboard has slightly more output that just the cartridge port pins. If we use just standard 2.54mm connectors, the other side could be either a cartridge slot, or that vaporware RAM board, if not only D0-D7 and A0-A12 get put on this connector, but (at least) A13-A15 and R/W as well. That way we'd still have the "pure-with-6502" option, as well as the RAM board one, based on the same baseboard.

 

I am absolutely positive that a PCB with two CPU sockets (one for 6507, the other for 6502) would be possible, if only one of those two has a chip put in it. A bit more tricky will be trying to allow both PAL and NTSC TIA fitting. The solution I came up with so far would need at least some jumpers.

 

Looking forward for this project to throw my 2c and work in it.

[SamhainTM]

keep in mind that the original scope is to create modern replacement boards for original hardware. The "CPU sled" being a common component at this time. I'm not opposed to modifications to support upgrades and alternate components though as that's the magic of branching in GIT. My rackmount VCS prototype probably being the first alternate configuration.

 

I do apologize for any delays though as the real world is quite demanding of me at this time. I'm reworking some of the schematics right now and will upload them to Github soon  (in a few weeks hopefully) and share the repo. 

 

Stay tuned though... more to come.

[+SvOlli]

Whow... I might have just found a shortcut:
https://github.com/baldengineer/bit-preserve/tree/master/Atari/Atari2600/C012283.RevB

Spoiler: schmatics recreated in KiCAD, no PCB, but still better than converting on your own. Seems to be from a PAL machine.

[SamhainTM]

4 hours ago, SvOlli said:

Whow... I might have just found a shortcut:
https://github.com/baldengineer/bit-preserve/tree/master/Atari/Atari2600/C012283.RevB

Spoiler: schmatics recreated in KiCAD, no PCB, but still better than converting on your own. Seems to be from a PAL machine.

Yeah, I've run across that one before. I'm doing up the NTSC version, which has some differences, along with modern components where possible. I'm also integrating a built in pause circuit along with USB-C power and integrated composite/S video (no RF). I've got the CPU module schematic worked up in KiCAD, but I'm still working on available components so I can begin footprinting for the PCB design. Maybe next week for the first upload to GITHub.

 

Still have a ways to go with "sled" design though.

[StephenJ]

13 hours ago, SamhainTM said:

Yeah, I've run across that one before. I'm doing up the NTSC version, which has some differences, along with modern components where possible. I'm also integrating a built in pause circuit along with USB-C power and integrated composite/S video (no RF). I've got the CPU module schematic worked up in KiCAD, but I'm still working on available components so I can begin footprinting for the PCB design. Maybe next week for the first upload to GITHub.

 

Still have a ways to go with "sled" design though.

Really looking forward to this!