• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Reverse Engineering the Macintosh LC III Logicboard

max1zzz

Well-known member
Indeed 31.3344 seems to be the correct value, it just seems to be such a odd value to be generating the system clocks from.....
I'm wondering if that clock chip is actually internally generating the system clock (Which is stated to be 50mhz in the devnotes, the VLSI divides in by 2 to get the 25mhz CPU clock) and the video / audio clocks and that he crystal is solely used for the serial clock (Which is 31.3344mhz according to the devnotes)

Might be intresting to try removeing the crystal form a III and seeing if it still boots....
 

mmu_man

Well-known member
The fact that the overclocking mod changes R14 which is right next to the chip and the quartz and that it makes it go to "33" (but anyone actually measured it?)… I'd bet there's a scaler (or PLL then, if it needs to go to 50) in this chip that is configured by this resistor.
 

mmu_man

Well-known member
Google drove me back to 68kMLA 🙃 :
Here is the text from the LC III's Developer Note:

"Omega is a custom programmable phase locked loop clock generator that provides the primary clocks required by the Macintosh LC III. Omega produces a hardware-selectable system clock for the microprocessor bus, a programmable dot clock for the video circuits, and a fixed 31.3344 MHz clock used for the serial ports, sound circuits, floppy disk, and various timers".
🤷‍♂️
 

mg.man

Well-known member
it's already documented in this article
Yes, well, my reason for posting the video with the "hack" was to highlight that the crystal frequency <> the CPU clock frequency... It looks like everyone is on that page now - but the question remains... what *is* that crystal frequency?... Perhaps @Bolle is correct? Anyone know to help out @max1zzz?
 

mg.man

Well-known member
Ah... looks like it's in your last post - didn't see it at first, had to "click to expand" 😀
 

mg.man

Well-known member
So... kinda out of my depth here... but if the CPU clock is "generated"... might it be possible to find a combo of resistor config that pumps 40Mhz or even 50Mhz into it and thus build a super-charged LC III++(+)? 😀 I suppose you'd need super-fast RAM / VRAM?
 

max1zzz

Well-known member
@mg.man Its might be possible there is, but looking at the schematic it is unlikely, the speed is controlled by a single pair of resistors that pull a signal on the cock chip and vlsi either to ground or 5v there doesn't seem to be any other configuration options
 

max1zzz

Well-known member
Doesn't that give 4 'possibilities'?
Given the config resistors are just 0 ohm links I would advise against installing both at the same time :)

There is a mystery chip that can sit between some signals of the clock chip and he vlsi, but it's not clear what that chip could be or what it's purpose was
 

mg.man

Well-known member
I would advise against installing both at the same time
But are they tied to different input pin? Thus giving a "0" or "1" option for each input? I'm away, so no access to schematics... just thinking "logically"... 😉
 

max1zzz

Well-known member
So how does that work? Does one go to 5V and the other to GND - so that there are three states? "Open", "GND" or "5V"?
Screenshot from 2021-08-29 15-43-16.png
R74 and R14 are the ones your looking at

I think it's pretty unlikely the chip will be able to detect a floating pin, and given the chip is likely custom to just this mac (or maybe this one and the 475) it seems unlikely apple would have built in support for running it at a significantly faster clock
 

mg.man

Well-known member
Interesting... so you recon it's a case of "force it high" ["1"] or "force it low" ["0"]?
 

mmu_man

Well-known member
As explained in the dev notes, the "Omega" chip as they call it is a PLL, which generates different frequencies from a base clock by monitoring the drift of the phase of the clock wrt the base one.
The config pin simply chooses one of two built-in multipliers, and I guess it's also connected to the VLSI so as to be probed to detect the model number.
It should be possible to make a replacement chip from some PLD/FPGA it seems as it seems it can be done with pure logic.
 

mmu_man

Well-known member
If anyone is looking for something to do, I think a website to select PLDs based on pinout compatibility with an existing chip would likely be very useful to many :)
 

max1zzz

Well-known member
What the hell Apple! Why are there traces bridging C39 and C37???
30-08-21_wtf.PNG

And yes, there defiantly where caps installed there on the original board

30-8-21_Caps.jpg

Guess that's two less caps I have to install on the reproduction board :)

It would be need if you reformulated the board into an ATX for factor offshoot.
It certainly possible and is actually a very interesting idea, a mini-itx board would be really cool, defiantly something to consider for the future!
 
Top