Reverse Engineering the Macintosh SE PCB & Custom Chips for 1:1 reproduction

[Kai Robinson]

I checked the presence of the RESET and HALT signals - both are pulled HIGH, although i was measuring from the SWIM Chip socket, and although it was high, it looked like it was dipping for exactly 1 clock cycle to ground, then back to being pulled high. 

[techknight]

Are there Address/Data pull-ups somewhere on that board? you verify they have their proper VCC/GND connection? 

[Phipli]

Are the bus arbitration pins like /BR functioning? Does the CPU think a peripheral is using the bus?

[Kai Robinson]

/BR is pulled up by R43 to VCC

BGACK is pulled up by R45 to VCC

DTACK is pulled up by R46 to VCC

 

RESET and HALT are connected together - but also to the 6522, SWIM, 5380, SND & BBU...might be one of those causing the (potential) wierdness on those pins...?

[Kai Robinson]

Out of curiosity, to see if there were any buried traces, I sanded around U4E with 80 and 100 grit, to get down to the ground plane... If the area is anything to go by, the inner two layers are power and ground only, no traces missing. 

 

 

[quorten]

11 hours ago, Kai Robinson said:

I checked the presence of the RESET and HALT signals - both are pulled HIGH, although i was measuring from the SWIM Chip socket, and although it was high, it looked like it was dipping for exactly 1 clock cycle to ground, then back to being pulled high.

The problem could be that RESET is not held for long enough for the processor's internal state to stabilize.  According to the MC68000 User's Manual, page 10-12, note 4: "For power-up, the MC68000 must be held in the reset state for 100 ms to allow stabilization of on-chip circuitry. After the system is powered up, #56 refers to the minimum pulse width required to reset the processor."

 

Specification #56 indicates that during a normal operational state after power-on, a RESET only needs to be held for 10 clock cycles (1.25 microseconds).

 

Did you try pressing the programmer's RESET button after power-on?  If this isn't actually directly connected to the processor RESET line, you can trigger it yourself using jumper wires on the PDS slot connection to RESET.  BUT you would need to ensure this is low-pass filtered, so best just to use the programmer's RESET button as it already is filtered.  (N.B. HALT and RESET are wired together at the PDS connector to achieve a proper external reset condition for the MC68000.)

 

After ensuring that RESET is being properly asserted, check that *AS is asserted.  This at least ensures that the CPU is attempting to fetch the RESET interrupt vector, which is the very first memory operation performed after RESET.  Then if the BBU is responding with *DTACK and *ROMEN, and the ROM is responding...

Edited December 28, 2020 by quorten
Also note next debug steps after proper RESET

[cheesestraws]

14 hours ago, quorten said:

The problem could be that RESET is not held for long enough for the processor's internal state to stabilize.

 

Yeah, being pulled low for one cycle doesn't sound like anywhere near long enough to me...

 

(I'm hoping to finish my build today!)

[Phipli]

Folks, look at what someone posted on the Facebook "Vintage Macintosh Restoration and Preservation" group. It is a 3rd party BBU that allows for a higher bus speed on the SE!

[maceffects]

@Phipli I forgot about the Brainstorm!  That would be a brilliant thing to reverse engineer!  A bit out of my domain, and obtaining one would be very difficult.  Thanks for sharing.

 

Update: I replied on the post to see if I can obtain that SE board so it can be reverse engineered.  I am already making the PLCC 84 breakout adapter anyway. 

Edited December 29, 2020 by maceffects

[Phipli]

@maceffects

 

These folks seem to have one. Anyone live close? Do you think they'd let someone borrow it?

https://www.freegeektwincities.org/

[SlickClick]

On 12/28/2020 at 6:10 AM, Kai Robinson said:

Well, i was a very lucky boy for Christmas and santa brought me a new DS1102 Oscilloscope! Along with a Yihua 959D Hot Air station - I've just bought a Meanwell RT65B which has all the right voltages with enough current to set this up as a bench PSU for the board. I can monitor the video out signal on the scope. I tried in a separate SE/30 chassis but still no dice. I'm not getting anything. I think @techknight is right - there's a clock signal missing or not quite right somewhere. 

Pretty sure I mentioned that it was a Clock Signal Issue in My Prior Post... JS...

[maceffects]

2 hours ago, Phipli said:

@maceffects

 

These folks seem to have one. Anyone live close? Do you think they'd let someone borrow it?

https://www.freegeektwincities.org/

Our very own @ScutBoy happens to volunteer there and if the other individual doesn't snatch up this machine I'll see if he can help us out.  I think with the PLCC 84 breakout board designed by @Kai Robinson I'm building, it should be able to be cloned.  I suspect the PDS card is also part of the system but that (should) be easy to do unless it has PAL in which case I'd have to send those out for to be reverse engineered. 

[Kai Robinson]

@maceffects reversing a PAL is 'easy' if the PAL isn't registered!  Can use the PAL dumper PCB here and read it as a 27C020 and reverse it from the .bin file it produces. 

 

If anyone wants money for that SE - i'll donate - and have it shipped to whoever is going to reverse engineer that custom BBU!

 

PALDUMPERV2.zip

[maceffects]

Thanks @Kai Robinson I'm hopeful that with your breakout someone can reverse engineer that and I'd clone the PDS 16mhz upgrade board.  I'd assume it needs a logic probe but that is, I'm afraid, outside the scope of my knowledge. 

[Kai Robinson]

I'm willing to ship my entire Philips PM3585 64 Channel Logic Analyser to someone, if it accomplished the task. 

[Kai Robinson]

...i just found something i can try...

http://www.vcfed.org/forum/showthread.php?64106-Mac-SE-Checkerboard-pattern-after-cleaning-battery-leak&p=520560#post520560

[Kai Robinson]

And funnily enough - i just found a place that's selling NOS of the 343-0045B Sony Sound chips...

[ScutBoy]

6 hours ago, maceffects said:

Our very own @ScutBoy happens to volunteer there and if the other individual doesn't snatch up this machine I'll see if he can help us out.  I think with the PLCC 84 breakout board designed by @Kai Robinson I'm building, it should be able to be cloned.  I suspect the PDS card is also part of the system but that (should) be easy to do unless it has PAL in which case I'd have to send those out for to be reverse engineered. 

 

OK - I purchased the Brainstorm SE from FreeGeek. Had no idea it was something pertinent to this project otherwise I could have snared it before it went to the store, but as a volunteer and board member I don't mind giving them the money.

 

I should have the machine back in hand by the weekend; if it needs to then go somewhere else I would probably just ship the logic board, unless the whole machine is really needed.

[ravuya]

The other nice thing about having an aftermarket BBU that we know about is that it opens up another set of engineers that we could potentially track down and ask questions of.

 

AF had some drivers here, but the link seems to be broken now.

Edited December 29, 2020 by ravuya

[maceffects]

@ScutBoy thank you so much for your help!  I’m happy to pay whatever the charge was. I’ll start work on the PDS first as I think that will be easiest. Meanwhile, I’m probing around China to see if someone is skilled enough to reverse engineer the BBU reasonably cheap.  I also found a nice stock of 16mhz 68000 processors which should serve nicely.  Taking this side project wasn’t really a goal of mine, however, now I see it as a challenge that must be overcome. How cool would it be to have brand new SE boards with all new parts?  

Edited December 29, 2020 by maceffects

[quorten]

It really isn't hard to adapt the original BBU design to support a 16 MHz CPU.  I'd guess the purpose of the drivers is primarily to support a soft switch for "turbo" in case you want to run at the normal speed.  I must admit that the back side of that PDS circuit board is quite funny looking, though.

 

@Kai Robinson @maceffects If you looked at the Unitron's PAL equations closely, reverse engineering non-registered PALs is not as easy as you'd think.  There's nothing that prevents a PAL equation designer from referencing the same output in its input equations, and in fact Burrell's design explicitly uses self-referrential equations to achieve some desired state machine effects with the non-registered PALs, thereby turning them into "registered" PALs.

 

Also, keep in mind that the BBU naturally doesn't provide as many "test points" as the Macintosh Plus cluster of PALs does.  If I should mention it again, it would have been a good idea to obtain a more comprehensive understanding of the Macintosh Plus before going for the Macintosh SE, but well, we have what we have.

 

EDIT: Brainstorm accelerated Macintosh Plus too:

 

https://lowendmac.com/2000/mac-plus-with-16-mhz-brainstorm-upgrade-benchmarked/

Edited December 29, 2020 by quorten
Drivers, Macintosh Plus

[ScutBoy]

3 hours ago, ravuya said:

 

AF had some drivers here, but the link seems to be broken now.

 

As best I could find when I was researching to refurb this machine, there are no "drivers" necessary for the Brainstorm. There was a firmware upgrade that I applied (1.0 -> 2.5) but there's nothing in the System Folder I could find, nor did an Internet search turn up any mention of drivers or Control Panels/Extensions. It seems to "just work"

 

Speedometer tests showed it as performing at approx 2X a Mac Classic.

 

There was also a Mac Plus version of the Brainstorm. I think that was one of the Killy Clip style of accelerators. Not sure if it replaced ROMs or anything like that.

[ScutBoy]

8 hours ago, maceffects said:

  I suspect the PDS card is also part of the system but that (should) be easy to do unless it has PAL in which case I'd have to send those out for to be reverse engineered. 

 

Yes - the PDS card is part of the Brainstorm system. You need both the PDS card and the BBU chip as far as I know.

[ScutBoy]

Info from Macworld back in the day:

 

Sometimes, the less said about a product, the better. In the case of the Brainstorm Accelerator for the Macintosh SE, that's praise. This upgrade does precisely what it should: make your SE faster and otherwise remain invisible. In this day of the Power Mac, accelerating an SE might seem absurd. Yet many SEs still lead serviceable lives in schools, homes, nonprofit organizations, and small offices, for which acceleration may be more practical than replacement.

The Brainstorm consists of an accelerator card that fits into the SE's expansion slot, and a Brainstorm Bus Accelerator (BBA) chip. The card contains a 16MHz 68000 processor—twice the speed of the SE's stock 8MHz 68000.

The unique part of the Brainstorm package is the BBA. It replaces the SE's 8MHz bus controller with a proprietary 16MHz chip, increasing the speed with which data shuttles between memory and SCSI peripherals (such as hard drives) on one side, and the processor on the other. The result is an SE that approaches a Classic II in overall performance.

  — Macworld (March 1995)

[Phipli]

https://macintoshgarden.org/apps/brainstorm-accelerator

 

I read older firmware versions needed drivers, but later ones didn't I think.