MicroMac ThunderCache Pro find

[nottomhanks]

Hey all, a guy not far from me had a large collection of vintage computers fully restored: Coleco Adam, TI 99 4/A, Amiga, etc. i bought two LCs from him and one of them had this card:

Apparently it’s quite rare. Has @Bolle or anyone else reverse engineered this yet? There’s no FPU in it, but it has four ram slots.
This one is the 50 Mhz version.

Also, can this run in the colour classic, or is that one slightly different?

Let me know.

 

[Bolle]

Unfortunately there's no way to recover the contents of the AMD CPLDs they used on that one so that's a big no for reverse engineering it.

 

[Melkhior]

Still a very interesting device; according to this thread that mentions a very similar name, it could have the ability to expand the memory of the system using those SIMM slots.

Does it indeed enable you to break the 10 MiB limit using those SIMMs? If so, I'm very interested in how the software side achieve this - such as a dump of an onboard ROM, and/or whichever piece of software is needed to achieve this.

I'm trying to expand the memory of my IIsi by patching the ROM to enable some additional memory from the IIsiFPGA, but so far no luck, so I'm looking for existing solutions and those MicroMac devices are the only one I know of.

 

[Trash80toHP_Mini]

Unfortunately there's no way to recover the contents of the AMD CPLDs they used on that one so that's a big no for reverse engineering it.

HRMMM. I've got the non-pro version with no ASIC on board. I'd meant to ask you if you had that one in the works yet. It's just six GALs. Curiously, it has provision for a clock on board?

Nice find! Really curious about those SIMM(?) slots. Can you post a good pic of the component side?

 

[Phipli]

Does it indeed enable you to break the 10 MiB limit using those SIMMs? If so, I'm very interested in how the software side achieve this - such as a dump of an onboard ROM, and/or whichever piece of software is needed to achieve this.

I believe it adds a new memory controller on the card. I imagine it does the double boot thing too, like was discussed in another context.

 

[Trash80toHP_Mini]

Bet it's a RAMdisk using Virtual setup like using Compact Virtual on SE Accelerators? Taking the manual to work to read up on the installation for both types. I think it's the usual full line manual we've seen from MicroMac.

 

[Realitystorm]

Unfortunately there's no way to recover the contents of the AMD CPLDs they used on that one so that's a big no for reverse engineering it.

Before I try, has anyone reached out to old MicroMac employees via LinkedIn to see if they have back-ups of the contents of the CPLDs?

 

[Trash80toHP_Mini]

YIKES! REMOVE THAT MAXELL BOMB! Hope you have already, those batteries are ruinous.

Yep, looks like the Memory Extension "MM TCP" required for Cache function on the PRO models must take care of addressing the SIMMs.

10MB required on System Board and a pair of 1MB SIMMs in Bank A on the card for it to work. Takes up to four 16MB SIMMs for a full loadout of 70MB system memory. Nice board indeed!

@olePigeon you were in touch with the MicroMac gang, no?

 

[Trash80toHP_Mini]

OOPSIE! That would have added up to 74MB if my brain hadn't been offline at lunch time.

 

[Bolle]

I've got the non-pro version with no ASIC on board. I'd meant to ask you if you had that one in the works yet. It's just six GALs. Curiously, it has provision for a clock on board?

That one has been done already. It’s the same design as the Extreme Systems/Total Systems/whatever-other-name-they-branded-their-stuff-Systems variant of that card.
The clock on the Micromac version only lets you change the speed of the FPU, the CPU will always run at two times the host bus.

Total Systems Enterprise (Macintosh LC accelerator) -

wiki.pldarchive.co.uk

 

[Trash80toHP_Mini]

Nice color coding! Were you doing that before I posted pics of my Performer?

 

[olePigeon]

@Bolle I feel like maybe I've already asked this, but have you already tried contacting Hans Stupart and see if he has a copy of the code?

 

[iJol]

I finally managed to get a ThunderCache Pro too, but mine seems to be broken..
The Color Classic chimes, the white screen comes on and then it stuck..

Sad for me, don't think there's much to troubleshoot on it. Changed the SIMMs in many different ways but nothing, so the search continues..

 

[iJol]

Okay, solved the problem. The empty PRAM battery was the fault - replaced it with a CR2032, turned the 32bit memory mode on, rebooted it, turned it off, installed the accelerator and it booted fine

 

[KennyPowers]

I recently got one of these working in a Color Classic (and LCII):



(after recapping...a couple had started to leak)



However, I'm experiencing some instability that seems FPU related. It crashes reliably when running something like Speedometer's FPU tests, but runs fine otherwise:



I temporarily swapped in the FPU from another 50Mhz accelerator and it was "more" stable. FPU benchmarks would complete more often than not, but it still crashed a couple times (vs every time with the stock FPU). I fiddled with it for most of a day, and the only thing that made any noticeable difference was swapping in another FPU, so that seems to suggest that the stock FPU is having trouble running at 50Mhz?

I tried simply removing the FPU, but it wouldn't boot without it. @nottomhanks, you said yours didn't have an FPU. Do you have a picture of the front of yours? I think yours is a different board layout than mine, but seeing the front still might be helpful.

I also read that the 68882 could run at a different frequency than the CPU. That got me looking at a number of resistor footprints on the PCB next to the FPU:



The left sides of R334 and R337 connect to the FPU's clock pin. From the factory, R337 is closed with a zero ohm resistor as shown. I removed the resistor and beeped out all of those solder pads. I came up with something like this (excuse my beginner-level diagramming skills):



This is starting to get over my head, but R334 and R337 seem to select between sending the logic board's clock signal (16Mhz in this case) and the accelerator's 50Mhz clock to the FPU. Maybe someone can tell me what R333, R335, R336, and R338 are for? To my untrained eye, they look like they could pull the FPU's clock signal up or down depending on which one you populate? Or maybe they're involved if you want to run without an FPU? Anyways, I moved the zero ohm resistor to R334 and confirmed with my scope that the FPU was now underclocked to 16Mhz. I was hopeful...but, insta-"error type 11" with either FPU the instant I started the FPU tests. So, I put everything back the way it came...FPU clock at 50Mhz with original FPU installed. It seems to run very well unless I do something FPU intensive.

Anyone know where I can get a PGA 68882 that's likely to work at 50Mhz and wasn't re-marked in China?

Also, can this run in the colour classic, or is that one slightly different?

The README from MicroMac that comes with the driver says:

Now the ThunderCachePro works with the Color Classic. Chip on U34 (MACH120) must have rev. 'G' or later.

Mine has a handwritten 'G' on that chip. Also, I think you need the version with the SIMM slots on the back in order to not have the SIMMs physically interfere with the Color Classic's logic board.

 

[mg.man]

FWIW, I have a similar one [SIMM sockets are on the same side as the CPU/FPU] that had damage to the "gold labels" on the CPU and FPU - so I pulled them off. Guess what - neither were 50Mhz parts. Shame on you MicroMac!

 

[KennyPowers]

FWIW, I have a similar one [SIMM sockets are on the same side as the CPU/FPU] that had damage to the "gold labels" on the CPU and FPU - so I pulled them off. Guess what - neither were 50Mhz parts. Shame on you MicroMac!

I long figured that the real purpose of those fancy gold stickers was to hide factory overclocking. I peeled back the sticker on this one's FPU and it read MC68882RC50A. Does that mean it's a 50Mhz part? Do you use your accelerator, and is it stable? The only modifications/repairs I made to mine were replacing the original oscillator because it was rusty and I had a 50Mhz one on-hand, and replacing the electrolytic caps because I saw the beginning of leakage under a couple. Also, it didn't work until after those changes I confirmed that the new oscillator is putting out exactly 50Mhz and the caps were replaced with these:

https://www.mouser.com/ProductDetail/KEMET/T520C226M016ATE080?qs=wiJzLaOtgy5kANxiBxPp6g%3D%3D&countryCode=US&currencyCode=USD

Here's one of the original caps:



That's a 22uF 6.3V right?

 

[mg.man]

I peeled back the sticker on this one's FPU and it read MC68882RC50A. Does that mean it's a 50Mhz part?

Yes, that's my understanding.

I couldn't remember what exactly was on mine, so pulled it out for a pic :

As you can see, I too have a 50Mhz FPU, but only a 40Mhz CPU! As for stability, I haven't run it in anger, but I do wonder if heat would be an issue since the CPU / FPU would be face down - "facing" the LC/CC logic board.

 

[zigzagjoe]

I recently got one of these working in a Color Classic (and LCII):

View attachment 70016 View attachment 70017

(after recapping...a couple had started to leak)
View attachment 70018 View attachment 70019 View attachment 70020 View attachment 70022


However, I'm experiencing some instability that seems FPU related. It crashes reliably when running something like Speedometer's FPU tests, but runs fine otherwise:

View attachment 70023

I temporarily swapped in the FPU from another 50Mhz accelerator and it was "more" stable. FPU benchmarks would complete more often than not, but it still crashed a couple times (vs every time with the stock FPU). I fiddled with it for most of a day, and the only thing that made any noticeable difference was swapping in another FPU, so that seems to suggest that the stock FPU is having trouble running at 50Mhz?

I tried simply removing the FPU, but it wouldn't boot without it. @nottomhanks, you said yours didn't have an FPU. Do you have a picture of the front of yours? I think yours is a different board layout than mine, but seeing the front still might be helpful.

I also read that the 68882 could run at a different frequency than the CPU. That got me looking at a number of resistor footprints on the PCB next to the FPU:

View attachment 70021

The left sides of R334 and R337 connect to the FPU's clock pin. From the factory, R337 is closed with a zero ohm resistor as shown. I removed the resistor and beeped out all of those solder pads. I came up with something like this (excuse my beginner-level diagramming skills):

View attachment 70024

This is starting to get over my head, but R334 and R337 seem to select between sending the logic board's clock signal (16Mhz in this case) and the accelerator's 50Mhz clock to the FPU. Maybe someone can tell me what R333, R335, R336, and R338 are for? To my untrained eye, they look like they could pull the FPU's clock signal up or down depending on which one you populate? Or maybe they're involved if you want to run without an FPU? Anyways, I moved the zero ohm resistor to R334 and confirmed with my scope that the FPU was now underclocked to 16Mhz. I was hopeful...but, insta-"error type 11" with either FPU the instant I started the FPU tests. So, I put everything back the way it came...FPU clock at 50Mhz with original FPU installed. It seems to run very well unless I do something FPU intensive.

Anyone know where I can get a PGA 68882 that's likely to work at 50Mhz and wasn't re-marked in China?



The README from MicroMac that comes with the driver says:

Mine has a handwritten 'G' on that chip. Also, I think you need the version with the SIMM slots on the back in order to not have the SIMMs physically interfere with the Color Classic's logic board.

Funny you should ask, I just put together some data on that in another thread: https://68kmla.org/bb/index.php?threads/sourcing-legit-mc68882-fpus-data.46717/

Do you have pictures of your two FPUs?

 

[KennyPowers]

Yes, that's my understanding.

I couldn't remember what exactly was on mine, so pulled it out for a pic :
View attachment 70049
As you can see, I too have a 50Mhz FPU, but only a 40Mhz CPU! As for stability, I haven't run it in anger, but I do wonder if heat would be an issue since the CPU / FPU would be face down - "facing" the LC/CC logic board.

I just pulled it out to take a picture of the FPU for @zigzagjoe and decided to peel back the CPU sticker too. At least MicroMac gave you a 40Mhz CPU...look what they gave me!



I wonder if that has anything to do with my stability problems? Still, it runs fine until I do something obviously FPU related (math benchmark, etc). I've had zero crashes in games and other apps I've tried. Still...maybe I should try to find a better CPU...