obsolete's SE/30 Reloaded Build

[obsolete]

Is it still cool to document reloaded builds? I was lucky enough to get a board from zigzagjoe's recent batch. This is Bolle's Rev04 with croissantking's fix, green solder mask, and HASL finish, which solders like a dream. Also has all of the "birdseed" components on the bottom pre-populated, which is a huge time-saver.

Most of my parts showed up last weekend and I spent about 5 hours assembling. Here's how far I got:

All the SOICs, all the fuses, all the inductors, most of the caps, most of the connectors, some of the PLCCs, and a couple of DIP sockets are populated. My order from Phoenix Enterprises should arrive tomorrow, which has the CPU socket, RAM and ROM slots, SCSI and floppy connectors, and the rest of the DIP sockets.

I am using my board from this thread as a donor. Some might cry foul because it wasn't hopelessly battery bombed, and it should have been possible to save it with a couple more bodge wires, but I just got tired of beeping out traces. I'm having a lot of fun building the new board up, and isn't having fun what this hobby is about? I'm excited to have a fresh start without blackened, degraded traces from cap juice corrosion creeping under the solder mask. Thanks to everyone who contributed to make the Reloaded boards a reality.

 

[Kouzui]

Is it still cool to document reloaded builds?

I think it is! It's a lot of work at least.

I'm looking forward to the Mac II boards, and seeing what people do with them.

And good luck. I did my own Classic and will probably be doing a iisi and iici in the future.

Edit: do you have any before pics of the logic board? I didn't see them in the other thread.

 

[JC8080]

I'm at almost the same spot on my build. I think you're going faster than I am, I look forward to seeing the progress.

 

[JC8080]

Which RAM SIMM sockets did you get from PE? I assume HWS5040? I need to step up and buy those, I've been putting it off since its $40 for 8, but I know I'll regret reusing the old plastic clip sockets down the road.

 

[zigzagjoe]

Looks good to me! If the old board isn't too far gone, you might find someone willing to buy it once stripped.

 

[croissantking]

Is it still cool to document reloaded builds?

Absolutely! Entertaining to see the different approaches to building these up as time goes by, and also the Rev04 boards entering into propagation. I see you have newly sourced VIAs and SCSI chip on there. Interested to know more about the HASL finish?

Some might cry foul because it wasn't hopelessly battery bombed, and it should have been possible to save it with a couple more bodge wires, but I just got tired of beeping out traces. I'm having a lot of fun building the new board up, and isn't having fun what this hobby is about?

I built three boards up using non-bombed donors, so I am more guilty than you. For two of them, I desoldered everything very carefully so that I could reuse the Apple PCBs if desired - and I left most of the through-hole parts on them other than the PALs, Sony chips and RTC. I did indeed end up restoring one of these donor boards, using parts from a fourth, battery bombed board. So I now have a sort of Frankenstein-ish original SE/30 board as part of my collection.

 

[zigzagjoe]

Absolutely! Entertaining to see the different approaches to building these up as time goes by, and also the Rev04 boards entering into propagation. I see you have newly sourced VIAs and SCSI chip on there. Interested to know more about the HASL finish?

I built three boards up using non-bombed donors, so I am more guilty than you. For two of them, I desoldered everything very carefully so that I could reuse the Apple PCBs if desired - and I left most of the through-hole parts on them other than the PALs, Sony chips and RTC. I did indeed end up restoring one of these donor boards, using parts from a fourth, battery bombed board. So I now have a sort of Frankenstein-ish original SE/30 board as part of my collection.

I agree, the new chips are cool to see.

HASL: I used the leaded finish, as long as it's clean I find it solders very nicely. Here's the JLC description:

 

[obsolete]

Edit: do you have any before pics of the logic board? I didn't see them in the other thread.

Yeah, I wasn't really good about taking pictures of that board. When I get it under some good light/magnification I'll try to grab some. I went over most of the areas of corrosion already but there are still a couple left.

 

[obsolete]

Which RAM SIMM sockets did you get from PE? I assume HWS5040? I need to step up and buy those, I've been putting it off since its $40 for 8, but I know I'll regret reusing the old plastic clip sockets down the road.

Yep, those are the ones. The RAM slots on my donor board were flaky, so that made the decision easy. It's also a time vs. money thing; 8 RAM slots is 240 pins to desolder. That's a lot of work to save 35 year old plastic parts. Plus Phoenix is the only place I know of to buy a ROM slot, so I was going to order from them anyway, and they have the CPU socket too. My only complaint about them is that they do not ship fast. I think it took 4 or 5 days from when I paid until my order actually shipped.

I've never had a board with metal clips on the RAM and ROM sockets before. I'm looking forward to not cringing every time I install or remove a SIMM.

 

[obsolete]

Looks good to me! If the old board isn't too far gone, you might find someone willing to buy it once stripped.

Thanks! When I get to that point, I would happily give it away for free.

 

[obsolete]

On the new chips: yeah, I got pretty into the "treasure hunt" of sourcing all the parts and ended up deciding to use as many new (and non-Apple) components as possible. So I have the W65C22N6TPLG-14 VIAs on there since croissantking confirmed them to work, and the LOGIC SCSI controller since I found it and the AM26LS30JCs in stock at the same eBay seller. I also have a new 68030, AM85C30, and TMS4461 VRAMs coming from Eric Woo. I got a new FPU and sound chips off AliExpress a while back.

I also got pgreenland's ATTinyRTCModule from CayMac Vintage, and I'm planning on using tashtari's adaptation of Apple's ADB code on a PIC16F88 with max1zzz's adapter from the QFN (ugh) to a PLCC footprint. I ordered 3 of the adapters from OSH Park for $0.95 shipped, which feels a little wrong. I have no idea whether the castellated edges will come out properly, but for that price, I'm willing to find out. I think I'll be able to solder it down either way.

So, the only parts I need to use from my donor board are the ASC, GLUE, SWIM, and...the video ROM, I guess? I suppose I could put the video ROM on a new EEPROM just out of spite. I'm also going to use the original PALs just because it's convenient, but may eventually try burning replacements from Bolle's files.

I know we're not there yet, but I like the idea of the SE/30 being an open source computer, no longer dependent on Apple ASICs. That may never actually happen, but I guess this will be as close as I can get for now.

 

[Kouzui]

On the new chips: yeah, I got pretty into the "treasure hunt" of sourcing all the parts and ended up deciding to use as many new (and non-Apple) components as possible.

I might have to do this with the IIsi I have on the way. And yeah, gotta love proprietary chips. Maybe one day!

 

[croissantking]

So, the only parts I need to use from my donor board are the ASC, GLUE, SWIM, and...the video ROM, I guess? I suppose I could put the video ROM on a new EEPROM just out of spite. I'm also going to use the original PALs just because it's convenient, but may eventually try burning replacements from Bolle's files.

Spite
If you’re doing the treasure hunt thing then for sure the VROM needs to go on a brand new EEPROM. I’ve successfully burned a few onto AT28C64B-15PU chips.
I’m curious if you can get new GALs (to replace the PALs), I’ve been sourcing some from AliExpress but they’re obviously harvested from old circuit boards.

I know we're not there yet, but I like the idea of the SE/30 being an open source computer, no longer dependent on Apple ASICs.

Me too.

 

[obsolete]

Spite
If you’re doing the treasure hunt thing then for sure the VROM needs to go on a brand new EEPROM. I’ve successfully burned a few onto AT28C64B-15PU chips.

If a guy is willing to settle for an EPROM, it looks like 27C256s and 27C512s are still out there for around half the price of the 28C64, and programming voltage for them is only 13V.

I’m curious if you can get new GALs (to replace the PALs), I’ve been sourcing some from AliExpress but they’re obviously harvested from old circuit boards.

Seems like ATF16V8Bs should work, no? Bolle has a warning to avoid "low power" ATF16V8Bs, which I think is referring to the ATF16V8BQL. Regular (non-QL) ATF16V8Bs are still available new and cheap, so that's what I'd try first.

 

[croissantking]

I know we're not there yet, but I like the idea of the SE/30 being an open source computer, no longer dependent on Apple ASICs. That may never actually happen, but I guess this will be as close as I can get for now.

What would be the process for producing new ASICs? I know that @Bolle mentioned that he'd had the GLUE imaged.

 

[obsolete]

What would be the process for producing new ASICs? I know that @Bolle mentioned that he'd had the GLUE imaged.

I believe the GLUE and SWIM have both been imaged. It may be possible to reverse engineer them with Degate but that's beyond my skill level. I think Kai Robinson started down this path with the SE's BBU, of which the SE/30's GLUE appears to be a later version, but I'm not sure whether he ever got anywhere with it. It may also be possible to infer some of the functions of the GLUE from the PALs on Plus and older Macs that it replaced.

Once reverse engineered, I think the idea would be to program a modern FPGA to replace the original chip.

 

[obsolete]

As promised, pictures of the donor board. Sorry it's kind of dirty, I already started pulling parts from it. There is corrosion like this under the solder mask in several areas. I already fixed the worst spots, and all the traces below have good continuity, but clearly there is still one or more broken traces somewhere that I couldn't find.




Here is an example of one I fixed by scraping off the solder mask and corrosion and soldering over the remaining copper:


Here's the whole board. Looks pretty good from a distance.

 

[Callan]

Heya Obsolete. I have LOTS of 27C512's (256's too, but did I mention I have a lot of 512's?). Send me a message if you need a couple and I'll send them your way gratis.

 

[Melkhior]

What would be the process for producing new ASICs? I know that @Bolle mentioned that he'd had the GLUE imaged.

(a) figure out what the chip is doing, and the required timings
(b) recreate the functionality in some modern language (some form of verilog)
(c) validate some key characteristics in simulation
(d) validate in-situ with a FPGA
(e) do the design and validatation of the ASIC itself

(a) to (d) are largely achievable by a motivated hobbyist today. What the GLUE is doing functionally is somewhat documented in Apple's documents like DCDMF3 or Macintosh Hardware Overview, timings can be observed with an oscilloscope if needed.

(e) currently requires very expensive tools (I'm talking 6+ digits worth of annual license for some of the tools) and a lot of expertise The costs of making an actual ASIC run, even for an old process in a MPW, are also very high. Even if one can leverage initiative like Europractice (for Europeans), the minimum MPW cost for a chip in GloFo 130nm is 47500€ (25mm^2 minimum at 1900€/mm^2).

As @obsolete said, FPGA is the most likely option for now (FOSS/H tools are progressing, open-source PDKs are helping democratize chip design, maybe someday MPW cost will go down) unless you can convince a professor somewhere it could be an interesting project for chip design students (or convince them to create FOSH CPLD/FPGA that are 5V-compatible as it would already help).

I believe the GLUE and SWIM have both been imaged

My suggestion would be to target the IIcx GLUE first. It is the simplest, as it doesn't need to deal with an internal video, only with the standards Mac I/Os. And today, it's the only mandatory chip in a IIcx with no easy replacement. By mandatory I mean a recreated IIcx could live without a SWIM (storage can be handled by SCSI) or a NuChip (video and ethernet can go on the CPU bus via PDS devices, my own IIsiFPGA and @halkyardo's SEthernet are already there). ADB and RTC have been recreated at least in prototype form, SCSI is still being made, serial has available substitutes.

 

[cheesestraws]

My suggestion would be to target the IIcx GLUE first. It is the simplest, as it doesn't need to deal with an internal video, only with the standards Mac I/Os. And today, it's the only mandatory chip in a IIcx with no easy replacement. By mandatory I mean a recreated IIcx could live without a SWIM (storage can be handled by SCSI) or a NuChip (video and ethernet can go on the CPU bus via PDS devices, my own IIsiFPGA and @halkyardo's SEthernet are already there). ADB and RTC have been recreated at least in prototype form, SCSI is still being made, serial has available substitutes.

Agreed; this sounds a very sensible potential path. (Relatively) low effort, relatively big results.