MicroMac Power Workstation

[KennyPowers]

I think this 2-slot riser card may be as simple as a passive adapter, although I think I'd want to squint really hard at one of these adapters to make sure I'm not missing a buffer or inverter or something tricky that's helping prevent collisions.

Looking at the blurry image of the top of the 2-slot card from the manual, it looks like there *might* be an IC or two next to the slots. I think I see some legs on the edge of the image that aren't part of the LC board:



Here's the full page from the manual: https://archive.org/details/micro-mac-power-workstation-manual/page/n15/mode/2up?view=theater

 

[KennyPowers]

I was talking to some much more knowledgeable people than myself (zigzagjoe and @max1zzz) about this dual-slot PDS adapter, and they both suggested simulating a completely passive adapter by installing a passthrough on a cheap LC PDS ethernet card and stacking an accelerator on top of it. In fact, I believe @max1zzz did exactly that with an Interware Booster clone while developing his accelerator/ethernet combo card. So, I installed a passthrough on this Farallon card I had laying around:



After modification, the ethernet card still worked and passed all diagnostics without an accelerator installed, and all of my soldering looked good under magnification. These are the two LCII accelerators I own:



With either of these accelerators installed on top of the ethernet card, the accelerator works, but the machine doesn't see the ethernet card at all. With the first accelerator installed, Tattletech shows this in slot $0E (same as it shows with only that accelerator installed):



So that accelerator presents a DeclROM apparently. With the MicroMac accelerator installed, slot $0E shows as empty:



With only the ethernet card installed:



So maybe it has to do with the relative complexity of these accelerators compared to some of the simpler ones that just add a faster 030 and an FPU. I also just now noticed this blurb on MicroMac's site for the Power Workstation:

...includes a specially modified MicroMac ThunderCache accelerator with a dual-slot PDS adapter...

I wonder if those "special" modifications are what allow that accelerator to recognize an additional PDS card with a purely passive dual-slot PDS adapter? If so, then finding one of the adapters wouldn't be much use without a specific accelerator that works with it. Actually analyzing the bus activity of these two accelerators to try to figure out what's preventing them from seeing a PDS card is probably beyond me, and even if it weren't, I'm probably not interested in REing/modifying either of them to work in this configuration. I would be interested in plugging a simpler accelerator like a Booster 30 clone or a MMXL99 into this modified ethernet card to see if they work with it, but I don't have either of those. That would just be a curiosity though, as I don't want to sacrifice the extra RAM these accelerators provide.

Of course, none of this is very practical since I could just put a wifi BlueSCSI in this thing instead of a physical ethernet card, but then using 30-40yr old computers is already a bit impractical. That's what makes it fun

 

[KennyPowers]

Noodled around with an oscilloscope with the stacked LCII accelerator/NIC. Seems like this accelerator (and likely others) assume an expansion card won't ever be present. With the accelerator installed, the /PDS.AS line is never asserted. If I'm understanding things correctly, the LCII logic board asserts that line whenever a valid PDS slot address is on the bus (so $FE00 0000–$FEFF FFFF for the LCII in 32bit mode). I'm guessing this accelerator doesn't have logic to assert /PDS.AS when addressing that range. I'm also guessing that simpler accelerators than this one that don't add RAM/cache/etc may just work with an expansion card by coincidence due to their relative simplicity? I think I'm done messing with it, but it was interesting seeing the different signals with and without the accelerator installed.

 

[KennyPowers]

It took me 3 years, but I believe I've finally found the dual-slot PDS adapter that was an option with the Power Workstation:



There's no programmable logic on it...just a single hex inverter, and near as I can tell, none of the inverted signals are actually used. Only two of the inverters in the 74F04 are used, with both taking the signal from the PDS pin circled in red in the first picture as input (I couldn't find an LC PDS pinout). One of the inverted signals goes through R1B and then appears to go nowhere. The other inverter's output goes to the unconnected side of R2B. I wonder what all that stuff on the right side of the board was for? Aside from that, it appears to be a completely passive adapter. I didn't check all 96 PDS pins, but I checked several of them, and each pin I tested on the male slot was connected directly to the same pin on each of the female slots.

I haven't tried using it yet. Micromac's site specifically mentions using the adapter with a "specially modified MicroMac ThunderCache accelerator". Two ThunderCache accelerators were included in the lot with this adapter, and everything was said to have been "pulled from working installs". I couldn't get any further details about the machine this was pulled from though.

 

[LeadedSolder]

Nice find! Are you going to make a clone of this riser? It looks very straightforward but the dimensions can be tricky.

 

[KennyPowers]

Nice find! Are you going to make a clone of this riser? It looks very straightforward but the dimensions can be tricky.

I have zero PCB layout/design experience, and you wouldn't be able to use this in a regular LC pizzabox...at least not with the lid on. I also suspect it will only work with either very simple, or very specific accelerators. Still, there's nothing programmable on it, so figuring out the schematic should be relatively easy.

near as I can tell, none of the inverted signals are actually used

I've poked at it a bit more, and I was wrong about this (shocking). The clock signal coming in to the adapter goes to two separate inverters in the 74F004. R1B and R2B control whether the adjacent slot receives the unmodified clock signal from the main logic board, or if it receives an inverted clock signal passed through a ~22ohm damping resistor, which to my simple brain means the clock signal for that slot will be delayed by one half cycle? In my pictures, the top slot has a resistor installed, and the clock pin of that slot gets the inverted logic board clock passed through that resistor. However, R2B next to the lower slot has the bottom pads bridged, meaning the lower slot gets the unmodified clock signal directly from the logic board.

So basically, it looks like that stuff on the right end of the board allows you to independently delay/invert the clock signal going to each PDS slot. The Power Workstation manual says the accelerator should go in the lower slot, so it seems to be delaying the clock for any expansion card installed in the top slot. I suppose if you wanted to install those the other way around, you could simply move the resistor. Also, maybe this would allow tweaking timings for specific cards