Mac SE/SE/30 Ethernet Card Recreation

[cb88]

Obviously a SCSI based Ethernet card caters to a different problem than built in Ethernet or add-on boards.... and doesn't have the advantage of being able to use a DeclROM driver..

That said it is a solution that doesn't require opening your Mac or interfering with other modifications... so there is a case for both (the built in or add-on on board is going to be better/faster most likely... but less flexible in what other mods you add).

 

[Johnnya101]

Count me in for buying one if you can make it!

 

[Unknown_K]

If you can Ethernet a C64 you can make one for the SE, you just need to make a driver for a ready made Ethernet board.

In this day and age SCSI to WIFI would be a better alternative to Ethernet.

 

[techknight]

Cool thing about a SCSI to WiFi adapter would be for the early powerbooks. 

Imagine an all-in-one storage/WiFi solution for the powerbook considering the early ones never had an expansion bus other than the Duo. 

 

[IlikeTech]

Techfury, any news? How goes it?

 

[trag]

I like Techfury90's approach but it's been over a year.     I was just wondering, for those of us who are less software oriented and more hardware oriented...

How hard would it be to put an SE/30 (or SE) ethernet card next to an LC-slot card and compare contrast until one could convert the LC card into an SE/30 card?    Ultimately, this might devolve into using the LC-slot card as a source for parts, but they're still pretty cheap and common and I don't think the SE/30 card has anything on there that isn't also on the LC card -- but I'm not certain.

Anyone know of any giant hordes of LC-slot ethernet cards?    Piles at recyclers (Ole Pigeon?) .   Might want to preserve them for the old parts.

 

[techknight]

I think the slot address on the LC card is fixed, and it falls in range of the SE/30 video. but its been so long since I have studied this I forget. 

 

[Bolle]

We had that discussed in another thread not too long ago I think. You have to swap at least a few address lines to shift the address into another slot range and connect the LC cards SLOT_IRQ to the corresponding IRQ on the SE/30 side. Nobody ended up with actually trying though I think.

EDIT - found it: https://68kmla.org/forums/index.php?/topic/30295-030-pds-lciii-pds-adapter/&do=findComment&comment=574104

 

[Trash80toHP_Mini]

IIRC I've got simplified DELTA versions of that rat's nest diagram or I could cook one up if/when I find the original file.  Should I update that thread or post it here. Does someone wanna do a reboot thread? Mine tend to get ignored for the most part.

I've been looking ants' WiFi projects and using the Pi Zero Two(?) with a ROM card based adapter or a CPU socket setup for multifunction I/O, but I'm in way over my head yet again. :/

 

[Trash80toHP_Mini]

IIRC there's another thread of mine with IIci Cache pinout flanked by LCIII/IIsi connectors.

 

[Trash80toHP_Mini]

We had that discussed in another thread not too long ago I think. You have to swap at least a few address lines to shift the address into another slot range and connect the LC cards SLOT_IRQ to the corresponding IRQ on the SE/30 side. Nobody ended up with actually trying though I think.

This pic of awesome LC PDS multiple slot adapter came up in the eBay finds thread.

Looks like it would have to go in one of the expansion boxes to get the lid on it. At any rate it got me thinking about memory mapping.

1 - Each Slot ID has its own layer in the I/O memory mapping chart, so that shouldn't be problematic.

2 - Control line remap appears to be simple enough:



2 - In that case, only the Slot ID remains to be spoofed on an adapter.

3 - The LC NIC assumes it's addressed at /SlotIRQ.E but has no way of knowing which line is pulling its chain.

4 - Remapping /SlotIRQ.E to any of the three available 030 IRQ lines (bold black) requires only the two jumper setup of 030 cards.

5 - As I understand it, the system detects a card on any available interrupt at startup and Slot Manager takes care of the rest.

6 - Slot Manager shouldn't care which line it needs to yank, just that it's servicing whatever was detected at startup.

7 - NIC drivers are commonly set up to handle PDS and NuBus setups alike, so maybe?.

Dunno, looks simple enough in my caffeine deprived musings this morning. Anybody got a comment on any of these points?

 

[Trash80toHP_Mini]

I haven't looked at the SE PDS pinout, it's a subset of the 030 slot as is the LC/LCII PDS. It also uses the same connector as NuBus and LC. That connector fits within the 10x10cm square for someone to run a batch of inexpensive test PCBs. Such a board would also have space for a couple of other test setups. [}]

edit: that would be someone else doing the PCB/testing, not set up for it ATM. But with help, I could map out the connections for a test PCB in Illustrator for someone else's playtime.

 

[techknight]

Slot E falls in a specific address block so I can only assume that hte address decoder on the LC Nic is going to want to decode for Slot E. So I think bolle might be right you have to rejigger the address lines. 

 

[Trash80toHP_Mini]

Don't think so, the NIC shouldn't care where it is or what interrupt is handling it/it's using. The address lines remain the same across all slots, most card connectors AFAIK. NICs seem to be 16bit to me. That appears to be the setup for all NuBus architecture PowerBooks, even the Blackbirds LEM mistakenly reported as having faster networking on a wider bus.

IIRC, Slot E is on the top of the slot space memory stack. Is that still within the first 16bits or does it trickle into 24bit territory?

Slot Manager assigns a specific block of memory in that stack depending on which slot the card shows up in at startup. The Slot ID/Interrupt is jiggered on the adapter board. If the driver for the card is the pretty much the same as the NuBus card by the same mfr. there may be no problem at all. I've got an Asante LC NIC somewhere around here.  []

 

[Trash80toHP_Mini]

eBay finds thread discussion about the 100/Duo FDD shook what might prove interesting about the Duos and this NIC. ISTR the DuoDock is a multifunction Slot $E PDS card implementation. Floppy MicroDock PCB connector's thruhole so that board, an LC NIC and ants' WiFi project seem like an interesting avenue of exploration.

 

[Trash80toHP_Mini]

Found a parallel example illustrating such a FloppyDock PDS adaptation, just a few more wires required.  :lol:



http://www.linkclub.or.jp/~maakun/PC_Dock/PCUltraDock03.html

On the SE/30/LC NIC front:

Similarly with no PowerCache connectors required: patch wiring directly into the thruholes of the passthru on one of Bolle's adapter PCBs seems like the obvious way to approach this hack to me. :approve:

On the SE/LC NIC front:

Does anyone know offhand if the SE PDS is addressed at $E, a quick look at the docs turned up empty. Wondering if drivers for any of these cards might be backward compatible to the 68000 instruction subset of the LC's 68020 PDS setup? I remember the Epson Stylus Color drivers having an unimplemented trap schiznit fit running on my PowerBook 100.

Somebody jump in here before EudiG accuses me of thread-jacking again. :blink:

 

[Trash80toHP_Mini]

I think the slot address on the LC card is fixed, and it falls in range of the SE/30 video. but its been so long since I have studied this I forget. 

Forgot to confirm this, it's the LC PDS that's located at Slot $E which is the home of SE/30 and IIsi video.

An expansion card's nondenominational, it's fixed in the 68000 PDS Slot, LC PDS Slot and NuBus Slot connectors themselves or the interrupt hardwiring is card configurable in the SE/30/IIsi Slot ID daisy-chain capable PDS configuration.

 

[techknight]

Dont think you follow me. Or maybe I just dont understand. 

I know video cards and some Ethernet cards designed for PDS slots have physical jumpers on them to configure the card for its slot address. So once the card knows what slot address its at, then the card itself does the address decoding for its designated slot address.

The LC card may be hard-configured for decoding slot $E. I dont know. 

I always thought with Nubus systems the slot address is configured based on which slot your using. 

 

[Trash80toHP_Mini]

Yep, each NuBus slot on has a four-bit slot identifier at the hardware level and each Slot's hooked up to the proper Interrupt line for that ID, IIRC.

PDS is a bit different, but it's still the Slot itself in the LC & LCII that's hardcoded, being connected to /STOTIRQ.E only. The LCIII slot extension added (unsupported of course) /STOTIRQ.C and /STOTIRQ.D as in my diagram above. SE/30 and IIsi PDS would be the only configurable multiple card setups I know of.

The driver works through Slot Manager which handles decoding, I don't think the PDS card's registers care how they're filled or where they're dumped back into the system, but I could very well be mistaken. Driver/ToolBox/Slot Manager do the address line hokey-pokey and the NIC is just a marionette AFAIK.

Which isn't all that much in this instance BTW. Somebody familiar with Drivers/ToolBox/Slot Manager from the Inside Macintosh angle needs to join in here. I just putz around with diagrammed blocks of hardware like they're miscellaneous bits and pieces from a box of borken bar puzzles for fun.

 

[Trash80toHP_Mini]

Does anyone know of a manufacturer who did NICs for SE, SE/30 PDS AND NuBus? Asante comes to mind, but I don't know if they did an SE NIC?

A direct connect toledogeekPDSextender(c) using M->F breadboard wires between card connectors just popped up in the morning haze. :blink:

View attachment 13271

IIsi + RCPII/IIsi <-> M-F breadbiard wires <-> LC NIC = cheap as' semi-permanent direct connect Test Setup for plug and play testing. /SLOTIRQ.E pin on LC NIC is easily plugged into any of the three slot interrupts on the 030 PDS. []