Techknight™ - LC PDS NETWORK CARD IN A SE/30?

[Trash80toHP_Mini]

Dunno, but the keying may still not be right, check it out. I know trag's prices are right for the left parts.

I've been playing in AI again:

[attachment=0]Rosetta Stone.000.2p.jpg[/attachment]

 

[techknight]

Yea that could be true. I dunno. Anyway.... I need to find those connectors regardless. Right angles, streights, 96pin/120pin male and females.

 

[Trash80toHP_Mini]

If you spot any EuroDIN 120 in wire wrap, please let me know, the largest I can find are NuBus size so I'll need to cut two of them down, that means three connectors for two hacks.

Digi-Key has a lot more soldertail EuroDIN-120 in different orientations (direction and sex) in stock than I'd expected.

 

[techknight]

So far, I am designing the 120-pin PDS extender, so i can plug my video card in without lopping things off, bending 90-deg, etc..

Now I didnt make any offset measurements, I just simply through 2 eurodins in place and routed the PCB.

Here is what i have so far.

[attachment=0]expander.png[/attachment]

 

[Trash80toHP_Mini]

Are you thinking of adding a connector for the LC NIC to that card?

 

[techknight]

Maybe.

But, I may not. I might be using that just as a proper-offset riser, and then ill roll another small PCB to connect to the passthrough on the pivot card, and that will be the crossover-bus adapter card.

ill probably open A24 to A31 as jumper holes, and the IRQ1/2/3 as jumper holes as well. this way, I can swap address lines and set the IRQ for Slot 9 from the LC's E. the pivot has its own jumpers and its default at 9, ill put the LC card on A.

Then i could put a IIe card on B.

But that leaves an open question. how to design the LC board so that the ethernet card sits in there correctly. if i just make a streight-through top mount eurocard adapter, the RJ45 port will face the front, and not the back. lol.

but if i design the adapter that the eurodin goes on the bottom, this means the LC card will mount upside down allowing the RJ45 to face the back.

ugh decisions, decisions....

 

[uniserver]

Then i could put a IIe card on B.

awesome!

LC board so that the ethernet card sits in there correctly.

as far as i'm concerned that is the least of your worries.. i couldn't care less how the card fits in there… ill run a 9foot ethernet cable right out of the back!

I mean unless you are trying to be a perfectionist!!!

 

[uniserver]

so i can plug my video card in without lopping things off

awesome!

 

[Trash80toHP_Mini]

I got your good news and your bad news kind of situation here:

Bad news: It's looking to me like we've got to reverse engineer the function of the PowerCache Adapter chip to get the LC NIC to work.

I started out mapping the similarities between the IIsi/SE/30 and the LC Slot NIC and hit a brick wall, so I attacked it from the other side. I've restarted the Rosetta Stone project, couldn't find the spreadsheet, so I'm working it out visually, before I try to locate or rebuild that crazy piece of work.

Good news: It's looking to me like we've got to reverse engineer the function of the PowerCache Adapter chip to get the LC NIC to work. [] ]'>

The IIci Cache Slot is the key to understanding the differences between Apple's several incompatible variations on the 68030 PDS theme. The machine specific PowerCache PDS adaptations for their IIci Cache Slot Accelerators provide clues as to the differences between these PDS implementations.

I've haven't got a PowerCache Adapter for either machine in question, but I do have one for the LCIII. Since it looks like it's pretty much a straight thru (right angled and flip flopped as it may be) PDS shoehorn with no active components on board, the LCIII's PDS looks to be as similar to the IIci Cache Slot as anything I've seen:

There are only seven tiny SMT resistors on board, eight of the same package type, but capacitors and then five large SMT capacitors that I'm guessing condition the power translation from the five power lines (only three of which are +5V pins, BTW) in the LCIII slot to the eleven Vcc connections on the IIci Slot/PowerCache Receptacle.

That's a total of just twenty capacitors and resistors needed to turn an LCIII slot into a PowerCache/IIci Cache Slot!

Once we've figured that control line translation stuff out it'll be lots of fun designing all manner of crazy adapter concoctions for PowerCache accelerated IIsi/SE/30 boxen. [}] ]'>

 

[Trash80toHP_Mini]

Here's to hoping that the NIC's control line requirements won't exceed what's readily available without the conversion!.

 

[techknight]

Why do we have to reverse engineer anything? you lost me.

the connections are the same, only difference is SE/30 has slotspace for 9/A/B, while LC is E. and the LC has slot space addressing in 24bit like the SE, but super slot space/32bit is signified by the single A31 line on the LC. Pretty much the same on the SE/30 except you have all address lines available.

So i dunno what the big deal is?

if anything putting LC PDS cards in an SE/30 should be easier for the fact that the card was designed with more restrictions in mind. Its going the other way around thats difficult. (SE/30 cards in LC).

 

[Trash80toHP_Mini]

if anything putting LC PDS cards in an SE/30 should be easier for the fact that the card was designed with more restrictions in mind. Its going the other way around thats difficult. (SE/30 cards in LC).

Not necessarily, direction of the conversion is irrelevant, whereas functional requirements are a make or break situation..

It boils down to which of the control lines you need to get the NIC to function. They're not the same between the SE/30 and the LC, whereas they are identical between the IIci and the LC.

If you're lucky, you won't need any of the control signals that differ. If not, you'll need to tweak/synthesize the required control signals from the batch that differ and the ones that don't. That's why the PowerCache adapters are so different. Some need those conversions done in the PALs you see on the SE/30 PowerCache connector.

The requirements for a NIC as opposed to an accelerator are trivial, hopefully the basic set of control signals that are identical in all the of Apple's 68030 PDS variations will suffice.

LCIII <-> IIci = Trivial

LCIII <-> SE/30 = not necessarily trivial, depending upon which control signals are required.

See post with the two diagrams above.

 

[techknight]

I did look at your diagrams. What differs? its all there... Your not thinking 4th dimensionally.

 

[uniserver]

Your not thinking 4th dimensionally.

 

[Trash80toHP_Mini]

AHA! 8-o Thanks for that tidbit, that's what I get for doing work in AI until the wee hours . . . I'm not really supposed to do that.

I went back to look at the data again. Thank goodness I wasn't following it or the discussion linearly. I think I've finally got a handle on what those little black boxes on the PowerCache adapters do to the IIsi, SE/30 and the other PDS Slots which aren't signal for signal compatible with the IIci Cache Slot.

Gotta get probes on one of those suckers ASAP! :approve:

 

[techknight]

Off topic:

The Powerbook Duo is setup for Slot.E as well. hmmmmm. That means I can tie an LC ethernet card directly to the 152 pin docking connector of the PB Duo.

Humm.... I need a dock connector now. Maybe a scrap floppy dock.

 

[Trash80toHP_Mini]

That's exactly why I've been studying every permutation of the 68030 PDS since about the time the MLA was founded. My choice was adapting the LC version of my NuBus PAS16 for my DuoBoomBoxDock™ entry for a free computer contest over on 'fritter.

I was wondering when someone was going to take this into the Duo realm! Good thinking there, tk!

 

[techknight]

Well in the case of LC to Duo, its not hard. its basically direct-connect. no fiddling with addresses or anything.

 

[Trash80toHP_Mini]

I dragged out the Mac II PowerCache adapter mcd donated to my project to check out before I buy, borrow or steal any of the other PowerCache adapters. I figure having the discrete control signals coming from 68020 and MMU sockets available for analyzing how they play out in the two PALs on board that particular adapter may tell me more that just jumping straight into the 68030 converters with integrated CPU/MMU signals. Dunno, it's worth a shot and an adapter on hand . . .

Long story short, it's likely the missing cache a/o /cache.hit control lines that are synthesized from other signals in the little black PAL boxen on the IIsi/SE/30 PowerCache Adapters. Not jumping to conclusions, but that's the bet bet from where I'm sitting.

Found a couple of pics over on 'fritter just now: NuBus Mafia

[attachment=1]iisi_pds_adapter-10199.jpg[/attachment]

[attachment=0]daystar_what-20222.jpg[/attachment]

....

 

[techknight]

That chip U1 is the decode logic specifically for allocating the cache correctly on the databus, as the cache isnt using nubus space. its more memory mapped