Gregory's Serverheh, they even put the bus sizing features in this chip. motorola removed that in the '040 (the '030 had it) and it caused innumerable problems for 68k computer designers.
|
26 comments
  @tubetime This post suggests power management. https://68kmla.org/bb/index.php?threads/where-to-find-341s0851-c-1990-apple-lciii-boot-issues.37152/ based on the pinout i think it's a 24C01 serial EEPROM, probably for a unique serial number or something. @tubetime that's 343S0120, often found alongside the customized 68HC05 Apple calls Egret (Cuda in later models) used for ADB, housekeeping tasks, PRAM, reset (and NMI) management, etc. IIRC '0120 serves as a power monitor, switching the HC05 to battery power when system power is cut. ...as seen here on a 630 DOS/LC580 board. yeah i'm wrong, this is a battery switchover circuit much like the MAX703. it provides voltage to the Cuda chip from the battery, and automatically switches to the 5V rail when power comes up. I've 3D printed a drive bracket for the Performa. it's designed to hold the SCSI2SD. it's a little loose but not a bad fit. I designed this based on photos of a real bracket, so the dimensions are probably slightly wrong today's status: i'm having issues with the joystick port. the machine thinks it is pegged to the lower right even when it is centered. the design is not an ADC. it uses a quad 555 timer with RC circuits composed of the external joystick resistance and 22nF capacitors. the Apple IIe just measures the time it takes to exceed the threshold. it's just barely pegged at 250 or so. 2.8ms. the resistance is 87.8K, so this RC curve implies a capacitance of 68nF. that can't be correct. inspecting the board reveals corrosion on the pins of an 'LS125, some of whose pins are wired into the joystick inputs (for some reason). i think the capacitor (already recapped C2) spilled electrolyte all over this area. i removed the 'ls125. there was definitely a puddle of electrolyte underneath (it crackled and fizzed during desoldering). note that the potentiometer connection (red arrow) is right next to a pin tied to ground (blue arrow). electrolyte is conductive, so maybe it was creating a leakage current. removing and cleaning the area helped quite a bit! it's not perfect so i think i have more cleaning to do. i'm also thinking that the etched/oxidized solder joints in combination with the spilled electrolyte formed an electrolytic capacitor with about 45nF of capacitance. yeah that would do it. so what does the 74ls125 do in this circuit? it's just discharging the timing capacitors. the NE558 also does this, so the 74ls125 is connected in parallel. I guess it... helps? regardless i went and cleaned the board again more thoroughly, and it's working a lot better now. i'll replace the damaged 74ls125 later.
[DATA EXPUNGED]
@tubetime Did you figure out what the ‘LS125 is doing there? I looked it up and guess it’s just a buffer. https://www.ti.com/lit/ds/symlink/sn54ls125a.pdf @kgutwin it has the ability to pull the joystick lines low. no idea what this was for, none of the other apple II compatibles have it  @tubetime Ahhh, so IBM _wasn't_ the only company that used an RC circuit as an integrating ADC :'D!  @tubetime I feel like that's the spiritual successor of the SE's BBU which turned a half dozen or so PALs into a single chip @tubetime I wonder why a similar situation didn't cause problems in PC land when going from 8088 to 286 bus (and again from 286 to 386)? 8-bit ISA works (mostly) fine on newer machines. I wonder if this is a port-mapped I/O accommodation? I remember hearing- emphasis on _only hearing_- the 286 has some _slight_ incompatibilities when talking to 8-bit devices with an I/O port at an odd address. But most cards in practice didn't run into this. @cr1901 hmm, i don't know. intel was all about backwards compatibility, so probably the CPU handled it automatically. |
the Commodore folks solved this on their '040 Amiga CPU card by using a raft of ten (10) GALs, which burned a ton of power.