SuperMac PDQ, Spectrum 8, 24, Thunder ROMs

[EIMS]

I have this card, which appears to be a Spectrum 24 IV, but looks very similar to higher end cards. I was wondering if it might just be a Spectrum 24 IV ROM in a higher end card, but it also has an ASSY sticker with 0008235-0001.



Here is a closeup on the ROM, and it also has a 40MHz crystal.



R32, R33, and C5 are not populated.



I looked up the data sheet for the ICS1494 that those components connect to, and it is an Enhanced Video Dot Clock Generator.



Those components don't make a whole lot of sense for an ICS1494, maybe R32 connecting to XTALOUT to VSS could be for EMI suppression? C5 and R33 are connecting to pins 13 and 14, which are NC on the ICS1494, maybe those components are intended for a different clock generator chip with a similar pinout that uses pins 13 and 14?

ROM attached. I didn't take the label off, but my TL866 programmer got an ID from the EPROM, which looks like the ID for a TMS27C512.

If I was to burn a ROM for a PDQ+ or Thunder, and put it in my card, how would I tell if it is making a difference? Should the performance be different? Any other functional changes like supported resolutions or depths?

 

[MacOSMonkey]

I would love to obtain a v3 ROM for the PDQ, Series III, or Series IV, because that version allows resolution changing via the Monitors control panel. If you have one of the ROMs, please upload it. If you don't have a reader, DM me and I would be happy to read and return your EPROM.

This Spectrum 8 Series III EPROM is a ST M27256.

Good work on everything.

The latest Spectrum/8 Series III ROM is v1.3 -- it added a number of new video configs, etc. Commented elsewhere.
The latest Spectrum/24 Series III ROM is v1.1 (no 3.x update). It fixed many problems that were present in the v1.0 ROM. Commented elsewhere.
The latest Spectrum/24 PDQ ROM is v1.27 (no 3.x update). The original version from mid-1990 was v1.0 (but had bugs...acceleration problems and a filesharing issue that was System 7 reg trashing-related). v1.11 was from late '90 and fixed many of the aforementioned 1.0 problems. v1.27 was from early-mid '91 and fixed the Quark XPress bug and other remaining issues. There may also be some original PDQ boards out in the wild that have beta 1.2x ROMS on them, since there was a beta program for a while to fix Quark XPress (since it was a major sales issue). I commented on these changes in another thread...but by that time, newer designs were already underway -- Thunder, etc. So, no 3.x ROM for the original Spectrum/24 PDQ.

And now that I am thinking about it, there were also timing changes/improvements to the Xilinx code for the Spectrum/24 PDQ board sometime between v1.0 and v1.27, but the 1.27 ROM should still work in earlier rev boards.

In terms of board revs for the Spectrum/24 PDQ, my v1.11 ROM board is Rev C and my v1.27 ROM board is Rev J. But, I think the rev changes were mostly programmatic (Xilinx programming, firmware, etc.). However, there were also probably some 2nd source changes and maybe Xilinx device changes. Also, SuperMac may have originally supplied a 64.000 MHz oscillator in the optional board socket to support the original 19" 60Hz 1024x768 monitors, but by 1991, that support would have certainly ended, since everything was either 19" 75Hz 1024x768 (Hitachi/Sony) or 21" 100Hz 1152x870 (Hitachi). However, 64MHz clocks were likely still available upon request from Tech Support.

 

[MacOSMonkey]

@EIMS -- The board you have is basically a Thunder/24 card. It was originally designed/released in late 1991 with spins in 1992 as the next generation after the original Spectrum/24 PDQ as the board that switched to a newer graphics controller (SMT02), bit-shift registers (BSR03) and a custom acceleration ASIC (Squid 01, or SQD01). It was later rebranded as Spectrum/24 IV. Your board is a newer fab rev and slightly different assembly from 1992 (vs. 1991 for the original Thunder fab). For example, I think the original boards used BT DACs. Yours has an Analog Devices DAC and there was no rework on U10.

As in other comments I have posted, SuperMac's initial 24-bit board evolution was: Spectrum/24 (no acceleration) -> Spectrum/24 Series III (PAL/GAL daughterboard accelerator) -> Spectrum/24 PDQ (Xilinx-based accelerator) -> Thunder/24 (custom ASIC accelerator) -- and later re-branded as Spectrum/24 IV (as the distant successor to Spectrum/24 Series III) -- as in the board you have (but it's the same basic design as a Thunder/24).

Over time, there were various ROM upgrades. For example, your v1.7.2 ROM comes from 1993 and probably has some bug fixes, OS compatibility changes, minor improvements and additional video configuration support vs. the original Thunder/24 ROM from 1991 -- v1.15/1.16. There were also later v3.x ROMs from 1994 that included PowerPC support and additional SuperVideo 3.0 support. The ROMs were usually branded with whatever the original board marketing designation happened to be (to maintain the customer perception of a totally unique design) -- like Thunder/24 v3.0.

In some cases, there is not much substantive difference in the boards. For example, the original PDQ+ board is just a ROM-gimped version of Thunder/24 without the GWorld RAM slots. Importantly, PDQ+ boards will work with faster Thunder/24 ROMs that have full graphics acceleration performance. So, if you have a PDQ+, put a Thunder/24 ROM in it and hotrod that puppy. It was just marketing smoke and mirrors. Wait...I mean "strategy." The PDQ+ board looked different enough without the RAM slots/connectors that it was just a Thunder/24 hiding in plain sight.

SuperMac wanted to have a price and performance marketing spread across the 24-bit high-end video market -- for example:

"SLOWER/CHEAPER"
Spectrum/24 Series III -- 1989 -- no accelerator board (to replace Spectrum/24) (Codename: Easy)
Spectrum/24 Series III -- 1989 -- with accelerator (Codename: Easy Rider, "Rider" was the daughterboard)
Spectrum/24 PDQ -- 1990 -- (Codename: Shockwave)
Spectrum/24 PDQ+ -- 1991 -- (Codename: Poptart, a ROM gimped version of Thunder/24 without GWorld RAM or DSP board connectors)
Thunder/24 -- 1991 -- (Codename: Pop, full performance with GWorld acceleration RAM slots and DSP board connectors)
Spectrum/24 IV (aka Thunder/24)
Spectrum/24 V (but the same basic architecture)
etc.
"FASTER/MORE EXPENSIVE"

(For a short time, before EoL of the older products, all of the above boards were available in the market concurrently, with older boards dropping in price over time.) ...and later, SuperMac rolled in the Storm board as a plug-in DSP-based accelerator -- then on to GX, etc.

These SKUs made it hard for others to compete at any given price/performance point and ensured that SuperMac would win every benchmark or value comparison.

There was a breakfast theme for the Thunder-era board codenames. For example, Pop was Pop because of:
Snap: Spectrum 8•24 PDQ (an accelerated upgrade for the Spectrum/8 Series III board that also had 24-bit support on smaller monitors and 8-bit support on larger monitors, like 21" 1152x870)
Crackle: Thunder/8 (an 8-bit card that offered GWorld support and optional DSP daughtercard accelerator)
Pop: Thunder/24 -- the full package
...and then PopTart, the Spectrum/24 PDQ+ as a gimped Thunder/24 without GWorld slots. And "PopTart," because SuperMac needed another codename that was related to Pop and also a tasty breakfast item.

The SuperMac 8-bit board space had similar evolution over time:
Original Spectrum/8 -- 1987 -- SuperMac's first 8-bit color video card. It had a genlock sync input and worked with Julian Systems equipment, etc. The early versions of SuperVideo also had a chromakey slurping tool that would let you choose a drop-out (ie. green screen) color. TI graphics controller. I think it used to make a slurping sound when you used it to pick up a color from the background.
Spectrum/8 Series II, or "Spectrum C" -- 1988 -- a revised version of the original Spectrum/8 board with fixes, TI graphics controller. These boards are easy to spot because the ROMs say "SPEC C" on them.
Spectrum/8 Series III (Codename: George) -- 1989 -- the first 8-bit board to use an SMT01 controller. This board had a long life and was an unaccelerated, eventually "low-end" bread-and-butter 8-bit solution for SuperMac.
Spectrum 8•24 PDQ (Codename: Snap) -- 1991 as above
Thunder/8 (Codename: Crackle) -- as above
(...and as the codenames imply, the Snap/Crackle/Pop/PopTart boards were developed in the same timeframe in 1991 - busy, busy, busy)

And, of course, there were boards for smaller screen configs, custom machines/PDS, monochrome, etc., but that is a tale for another day, Bilbo.

 

[MJ313]

Snap: Spectrum 8•24 PDQ (an accelerated upgrade for the Spectrum/8 Series III board that also had 24-bit support on smaller monitors and 8-bit support on larger monitors, like 21" 1152x870)

Do you happen to know if there was a v3.0 ROM for the 8*24 PDQ? I am going to guess no... I think the latest I've seen mention of is v1.7.2 from Gamba's old site.

 

[jmacz]

I'm thinking of picking up the GQ programmer to upgrade the ROM from 1.0 to 1.1 on a Spectrum 24/III. Although I'm not sure if I'll ever use it again afterwards since I don't believe I have a real need to program ROMs outside of trying to do this upgrade trying to convince myself to do it by thinking up some other common use cases where I would need it... haha.

But that aside, is it pretty easy to program these ROMs or is there a high chance of bricking it? Just put the chip in the slots, read and backup the current version, write the new version, and that's it? Do I need to peel aside the label and erase it first or? And I assume the chip pops out easily? If something goes wrong, am I s__t out of luck or is it still possible to find a replacement chip to program?

 

[jmacz]

Nevermind on the above. Figured it out.

 

[MacOSMonkey]

I'm thinking of picking up the GQ programmer to upgrade the ROM from 1.0 to 1.1 on a Spectrum 24/III.

You should update to v1.1 if you can manage it. The 1.1 ROM was much better, more stable and had additional video configs. The only major functional loss was lock menubar, traded for quick panning that allowed you to pan the screen from anywhere. Lock menubar had many bugs and compatability issues associated with it. Also, it was A/UX compatible.

 

[MacOSMonkey]

Also, for those of you who have PDQ+ boards, just put a Thunder ROM on them (as noted in my post above). It should work and will have faster acceleration performance. They are the same board. The only difference is that the PDQ+ doesn't have the GWorld RAM slots that the original Thunder had...but who cares?

 

[jmacz]

You should update to v1.1 if you can manage it. The 1.1 ROM was much better, more stable and had additional video configs. The only major functional loss was lock menubar, traded for quick panning that allowed you to pan the screen from anywhere. Lock menubar had many bugs and compatability issues associated with it. Also, it was A/UX compatible.

Yup, I saw your note earlier in this thread and I'm running ROM 1.1 on my 24/III. Thanks!

 

[MacOSMonkey]

Yup, I saw your note earlier in this thread and I'm running ROM 1.1 on my 24/III. Thanks!

Great! v1.1 will definitely work better for you. It still has some bugs, but you would probably have to work to find them. :-D

 

[jmacz]

The latest Spectrum/24 PDQ ROM is v1.27 (no 3.x update). The original version from mid-1990 was v1.0 (but had bugs...acceleration problems and a filesharing issue that was System 7 reg trashing-related). v1.11 was from late '90 and fixed many of the aforementioned 1.0 problems. v1.27 was from early-mid '91 and fixed the Quark XPress bug and other remaining issues.

And now that I am thinking about it, there were also timing changes/improvements to the Xilinx code for the Spectrum/24 PDQ board sometime between v1.0 and v1.27, but the 1.27 ROM should still work in earlier rev boards.

In terms of board revs for the Spectrum/24 PDQ, my v1.11 ROM board is Rev C and my v1.27 ROM board is Rev J.

Anyone have a link to the v1.27 ROM for the Spectrum/24 PDQ? I've got a board incoming from @joshc and it's a Rev C board and currently looks to have the v1.11 ROM. Thanks!

 

[David Cook]

Anyone have a link to the v1.27 ROM for the Spectrum/24 PDQ? I've got a board incoming from @joshc and it's a Rev C board and currently looks to have the v1.11 ROM. Thanks!

Attached is a v1.27 file that came from somewhere. It wasn't one that I got from a board, so I can't vouch for the quality or that it really belongs to the 24 PDQ. It appears that the bits are inverted, given that the empty space at the end of the file is $FF rather than $00. Anyone have a quick utility to flip the contents?

 

[zigzagjoe]

Attached is a v1.27 file that came from somewhere. It wasn't one that I got from a board, so I can't vouch for the quality or that it really belongs to the 24 PDQ. It appears that the bits are inverted, given that the empty space at the end of the file is $FF rather than $00. Anyone have a quick utility to flip the contents?

Here you go. Nothing a little powershell can't handle. Checksum bytes verified successfully - 6d81ade5. The string "SuperMac Spectrum PDQ v1.27" appears near the top, so it seems legit?

 

[David Cook]

Here you go. Nothing a little powershell can't handle. Checksum bytes verified successfully - 6d81ade5. The string "SuperMac Spectrum PDQ v1.27" appears near the top, so it seems legit?

That was fast! Thanks @zigzagjoe

 

[jmacz]

Thanks!!!

 

[MacOSMonkey]

The v1.27 ROM should be backwards compatible with earlier Spectrum24 PDQ boards that were originally released as v1.11. The Xilinx parts were different on earlier boards. I think the later devices had slightly faster timing and there were some Xilinx code changes to accommodate them. SuperMac tested backwards compatibility, especially because of the Quark XPress issue and the need for field upgrades. However, the ROM changes should be transparent relative to the Xilinx parts. There may have also been some load time improvements to reduce the black screen interval during bitdepth changes where the acceleration code had to be reloaded (8-bit to 24-bit, etc.). Otherwise, see above for other notes. I have a v1.27 ROM if you need a master, but it sounds like yours is OK.

 

[jmacz]

The earlier ROM provided in this thread didn't work for me. The card isn't detected after burning that ROM image onto a 27C512. I verified after writing just to be sure. Can anyone verify it's a working image?

Also, should I be seeing parts on this Spectrum 24 PDQ board labeled Xilinx? I don't see any. I see a SuperMac SMT01, four SuperMac 9002328 chips, two SuperMac 9002329 chips, three BT485 RAMDAC chips, and three SuperMac BSR-02 chips.



And of course, I must have pissed off the SuperMac gods in some other life since I'm having a slight problem with this one too.. lol.

 

[MacOSMonkey]

For reference, on the original Spectrum/24 PDQ (Codename: Shockwave): SMT01 is the custom graphics controller (became SMT02 in later SuperMac board designs) the 3 BSR02s are the custom bit-shift registers (became BSR03s in later SuperMac board designs), the Bt453s are the DACS (of course -- and became a single Bt473/AD473 in later SuperMac board designs), the Bt438 is the related DAC clock, and the other large, square parts are the Xilinx parts (...36D) and their respective configuration ROMs (...37D). Use of the Xilinx parts was a next-gen speed increase over the GAL daughtercard accelerator while SuperMac worked on the Squid chip (SQD-01) graphics accelerator. I think there was a 10-pin JTAG header on the development boards, but I guess pulled for production -- the silk outline in the lower left corner with no pads is a clue. As with the gen 1 GAL accelerator on Spectrum/24 Series III, there was intentional obfuscation and, in this case, the Xilinx parts were rebranded as SuperMac ASICs to slow down competitors. But, you don't need to worry about the Xilinx parts. There was a timing difference (or issue) in the Xilinx parts on newer vs. older boards that required a configuration ROM change (and I think that is probably reflected in the ID numbering on the Xilinx parts on the respective rev boards -- but I would have to look).

The v1.27 ROM should work in earlier v1.11 boards (IIRC -- and was tested based on the possible need for Quark bugfix field updates, as before. There were also various other bugs in v1.11 -- but Quark was the worst one) and v1.11 was not a complete fix. I should be able to verify on my boards. I have v1.27 and v1.11 Boards/ROMS, can save the 1.27 master ROM image, compare it to the above bin file and burn and swap test it in the 1.11 board. I think I have blank devices.

But, before I do that -- please confirm that you burned the non-inverted version, correct? The bit-inverted one from zigzagjoe is just for readability/verification. The original file from David Cook is the one you should burn.

 

[jmacz]

But, before I do that -- please confirm that you burned the non-inverted version, correct? The bit-inverted one from zigzagjoe is just for readability/verification. The original file from David Cook is the one you should burn.

Oooh… totally misunderstood the previous comments. I burned the second inverted one. Will try the original. Thanks!

 

[MacOSMonkey]

OK - that fixed the ROM issue then.