Bolles finds

[omidimo]

Oh you are right! Just look at the old catalog and noticed there were different versions. Derp me! 

 

[trag]

It’s a 66MHz part... first thing I checked.

I have a spare 100MHz chip, not sure if I should attempt the swap though. That’s some damn fine pitch right there.

Debatable.   I wouldn't do it.    And you would need to have an ICS9178 chip in order to perform the speed bump, otherwise it just runs at 66.   Guessing, I haven't actually read the datasheet carefully, the PPC601 doesn't generate its own internal frequency based on jumpers, but gets it from an external clock source.   It's easy to generate a 2X or 1/2 clock.  But generating a 3X clock is more challenging, hence the need for the ICS9178.   I could be completely wrong.  I really need to dig out the PPC601 datasheet, assuming I have one here somewhere.   But that's what I've been assuming all these years.  Otherwise, why do the clock tripled PPC601s need a special PLL/clock buffer?

I hesitate to mention this as I don't want to encourage any murders, but the ICS9178 is also on the PM8100/100 and faster.   I think I've seen it come up on Ali Babba or other Chinese suppliers as well.  No idea if those are trustworthy.   I've seen an MPC part that serves the same function, maybe on the 120MHz PPC601 CPU card for the original PowerTower (Power Computing), oh, and probably on the faster 7200s, but it is not pin compatible with the ICS9178.

Instructions for switching from 2X to 3X are on Marc Schrier's Clock Chipping Home page, which is also gone now, but preserved/mirrored.  (Is it mirroring if it's gone?   Imaged?)   However, the Turbo601 does not like to operate much outside its 33MHz bus speed.    So, you can't slow down much below 96.   And I'm not sure if faster will work.   It didn't for me, but that could have been the limits of my PPC601 chip (rated for 66 or 80, can't remember) not the limited range of bus speed of the whole card.

I know for sure it won't slow down much.  It might speed up.

Almost forgot, though you may already know this.  Some (all?) PPC601 at 100MHz and greater operate on 3.3V instead of 5V.  There was a transition there somewhere between 90MHz and 100MHz.    The PCC Power 120 has a bunch of level shifters on the I/O pins because of this.

View attachment ICS9178_02.pdf

 

[Franklinstein]

The 601 and 601v are not directly pin compatible because of voltage differences. The fastest 601 you could swap in would be 80MHz. Also the 601 may not support 1/3 bus ratio so an overclock to 100MHz on a 33MHz bus may not be possible. 

 

[Solvalou]

Yeah I agree, leave it alone. Even if it was as simple as a CPU swap, these things are hen's teeth, especially the NuBus varients.

I'm lucky I have a 100MHz PDS one for my Q650 but frankly, after trying that and a stock 6100, there's not a huge difference. Probably partly thanks to the 6100 having a little more cache.

I am stararting to wonder if I should pull the heatsink off mine and replace the paste though. It kicks out some crazy heat!. Especially when Quaking!

 

[Bolle]

I could press that reset button all day and listen to that sweet chime coming out of an SE/30.

 https://www.dropbox.com/s/jghslvq25kwot99/IMG_720.mov

Too bad there won't be happening anything else.

The card seems to be working fine in the IIsi and IIci at least in a quick test in both of them.

It lists the following info:

ROM Family: $77D
ROM Checksum: $7C4F0317
ROM Version: 2.5f2
Patch ROM Version: 1.1

So according to the unofficial Turbo 601 site this card does not have the 256 color fix for the IIvx but still I can select more thousands of colors...

 

[trag]

So according to the unofficial Turbo 601 site this card does not have the 256 color fix for the IIvx but still I can select more thousands of colors...

Interesting.   I haven't read the UT601 site in a while.  Was there some indication that hte problem affected some but not all IIvx users?

 

[trag]

The 601 and 601v are not directly pin compatible because of voltage differences. The fastest 601 you could swap in would be 80MHz. Also the 601 may not support 1/3 bus ratio so an overclock to 100MHz on a 33MHz bus may not be possible. 

The 601 definitely supports a 1/3 bus ratio.   The PM8100/100, 8100/110, 9150/120 and Power 100 and Power 120 all used it. Oh, and the PM7200/120.    Also, the Turbo601 I switched from doubling to tripling had either a 66 or 80 PPC601 on it, so it's not a 601v vs. 601 feature.

 

[Franklinstein]

To be fair the only proper 601 you mentioned in your list was on the Turbo 601 where all the others were 601v; the original 601 didn’t exceed 80MHz and Apple never used them at 3:1, only the later >100MHz 601v-based models (the only exception being the 7500/100, running 2:1 on a 50MHz bus). 

 

[trag]

Are the 601 and the 601v different enough that you would expect one to support 1:3 and not the other?

You seem better versed on this than I am.   Is it correct that the 601 (and 601v) don't synthesize their own CPU clock and must be fed both the bus clock and the internal CPU clock?  It may be a failure of my imagination, but I can't think of another reason why the 3:1 cases would need the ICS9178 chip.

 

[Franklinstein]

I was on mobile so was not inclined to try to dig up and quote the specifics (i hate touch screens), which is why i said “may not” instead of “does not.”

After looking through the original 601 UM on a proper computer, buried in chapter 8 it appears that the 601 will support operating at bus multiples other than 1:2 but for that you need to do some clock skewing with that chip you mentioned; it’s not as easy as simply changing a resistor for other chips. 

Speaking of other chips, I do know for sure that the 603 does not support half speed multipliers where the 603e does. 

 

[trag]

Thanks.  That's cool.   I'll have to give chapter 8 a read some time.

 

[Bolle]

Just picked these up from the post office:



-Total Systems Mercury 030 Plus/SE :

Already have one of those but I killed one of the GALs on mine when experimenting. Going to check that missing GAL out to repair mine, then one of them probably has to go again as I don't have enough SEs and Pluses to put stuff into.

-PAK68/2:

That one was a DIY accelerator published in the german c't magazine. It's supposed to work in any 68000 machine if you have the correct GAL setup. The code is public so I should be able to get this to work somehow I think. It's missing two GALs, not sure yet if I need those or not.

-SCSI2SD:

Not much to say about that one.

 

[trag]

Nice haul.  Good luck with the GALs.   Boy, that has a different meaning amongst non electrical types.

 

[omidimo]

Very cool. Looking forward to what you do with the PAK68/2.

 

[Bolle]

Good luck with the GALs.

Thanks  Didn't kill that GAL this time and was able to fix up my broken one with a copy of the clock generation GAL.

View attachment 28903

Now on to the PAK68/2.

Found all the GAL listings including jedec files as well as the equations.

When used in the Mac SE I need U4_13, U5_14, U6_30 and U12_31 with the 68000 removed from the logicboard.

U5_16 is needed when the 68000 stays in place. This will disable the generation of the E clock signal in one of the GALs because the 68000 will generate that signal and the bus logic on the PAK will sync to that instead.

Gonna remove the 68000 on the spare SE board I have and install a socket instead.

View attachment CT9111_314.TXT

 

[Bolle]

And here we go... the PAK works like a charm after some trial and error to find the correct GALs.

The 68000 came off the board nicely...

...and two socket strips went in.

Now I can either plug in the original 68000...

...or switch to the PAK

I ordered a set of four 27C512 EPROMs to fill those ROM sockets on the PAK and have a copy of the SE ROM sitting right on the 32bit bus of the 020.

This is supposed to give you a nice speedbump on everything toolbox-related.

I noticed that the crackling sound issue that exists with most accelerators in the SE and Plus is not as bad as with all other upgrades I have tried so far.

In some games there aren't even any crackles and sound playback is completely flawless.

Speedometer benchmarks it twice as fast as the original SE.

 

[Trash80toHP_Mini]

Amazing stuff, my friend. :approve:

 

[Bolle]

Cards, cards, cards...



That would be a Daystar cache card for the Quadra PDS. It came with a PDS adapter for the Quadra/Centris 610.

Then we have two Formac video cards from the late80s/early 90s: a ProNitron 80 and a Prisma2. Both nothing special. The ProNitron has a fixed 1024*768 output at 256 colors. The Prisma2 does not show up in TattleTech or SlotInfo. I dumped the ROM and it looks totally different from any other Formac ROM I have seen so far - it might be broken.

The card in the upper right corner is pretty interesting. It took me a while to figure out what it even is.

It says "TransMac Board V2" and is made by Hermstedt GmbH. They were popular for their "Leonardo" line of ISDN Nubus/PCI cards.

The card right here shows up as "BAM Communication Board" in TattleTech. After digging around in old magazine articles I found out that the card together with a piece of software (supposed to be called "Macsteam") can be used to connect to Transdata networks simulating a Fujitsu/Siemens DSS9750 terminal. It was supposed to be able to connect to up to 4 different BS2000 mainframes at the same time.

Pretty useless nowadays but interesting nevertheless.

 

[trag]

Cool stuff, Bolle.  So do you pretty much have all the info you need to replicate the PAK upgrade, except, maybe, the schematic for the board?  Also, I am curious where you found the GAL listings, etc.

 

[Bolle]

The schematics are available on the net and I also got the original magazines here. Got the GAL listings from someone on a1k but the basic ones are also printed in the magazines.

Not sure if it's worth it though to replicate it. I have been working on replicating the newer PAK68/3 but I am not yet sure on what features I should implement right into the design.

There was a bus buffer board (PuPla) and a RAM extension (FRAK) for the PAK3 as well and it would be nice to have them all in one board. The RAM board was not supposed to work in the Mac though so that might need some messing with the GAL code to get it to work.

Also not sure if the buffer extension works on the Mac or if it is needed at all. There are not a lot reports on using any of the PAK cards in the Mac. Most stuff you find is on using it in the Atari as it seems.