SE/30 Xceed greyscale adapter cloning thread

Yes, the manual is on VintageApple.org:

https://www.vintageapple.org/macbooks/pdf/Micron_Technology_Xceed_SE-306-48_1989.pdf

Yes, the manual is on VintageApple.org: https://www.vintageapple.org/macbooks/pdf/Micron_Technology_Xceed_SE-306-48_1989.pdf

Click to expand...

Thanks much, pcamen. I figured the user manuals were out there somewhere, but this one's a much better find. Brain's all but shut down ATM, don't remember seeing that tech doc before. I wonder if I got the Development Docs/Technical Info for the Color30 there too? Pretty sure it was somewhere else, a link to that would be good, gang. []

techknight said:

Well I was thinking along the side of driving the analog board from the video card, not sync the video card to the analog board. Seems backwards to me.

Click to expand...

Not backwards at all, if you look at it the other way around!  [] You're syncing the video card to the VSYNC output from the Logic Board connector, not the A/B. HSYNC is carried along for the ride on the harness and passed with VSYNC to the A/B from the Color30, with the HSYNC signal used by Color30.

Bare bones representation of 306-48 specific harness signals:



< Tangential Silliness >

Originally, talk was about driving the GS Neck Board from a modified (very well documented) Toby Card and then looking at doing same from my very early (first?) release programmable SuperMac Spectrum card from a NuBus Mac for shiggles and gits. Sometime in that time frame Bolle figured out we could be putting the A/B at risk.

If there's a way to sync my Spectrum or a heavily modified Macintosh II Video Card (Toby) or clone to VSYNC from any Compact Mac along an unused DA-19 pin it opens up some serious playtime opportunities. Think Macintosh Plus with a modified GS harness installed and the GS neck board on its CRT. The setup would be A-B switchable from its own motherboard's internal video output in Plus Mode to GS input from any NuBus Mac with the Plus acting as its secondary display. You're talking something on the order of six to eight lines on a cable patched through a Compact's security slot?

That'd put my SE on top of the pet IIfx in one big hurry!  [}]



In the diagram above, the Compact Mac to NuBus Mac Video Cable, the A-B switching setup and NuBus card would be represented by J4. Are Plus or SE Logic Board to A/B connectors the same (all Compact Macs?) as those on the SE/30?

< /Tangential Silliness >

edit: massive page return continuity corrections

Here is the same type doc for a card called the Xceed SE/301 0-78, also an SE/30 card.  That couldn't be the Color30 could it?

https://www.vintageapple.org/macbooks/pdf/Micron_Technology_Xceed_SE-3010-78_1989.pdf

I also found this Color 30 users guide:

https://www.vintageapple.org/pic/MicronXceedColor30Manual.pdf

That's all I can find in my stash Xceed related. 

However, I did find this on Archive.org, Xceed technical documentation. 

https://archive.org/details/MicronXCeedTechDocs/page/n3

Interestingly, there is also a file there called Micron XCeed ROM.sit.  Here are the contents:



Does that look helpful?

Actually those XCeed Tech Docs look pretty interesting.  You guys seen those before? 

Maverick Theory of Operations in those docs with ICKY, Bartlett and Goose as the cast of characters, "Top Gun" being the SE/30 PDS would be what what many have been working from/on for many years.

...hmmm, is this Micron Xceed Grayscale Kit so much special? What makes them so rare and pricy at all?

I mean...mine just layed inside a cardboard box for some years now - never ever really used it, the kit came out of my SE/30 as I built it back to stock.

I just tested/tried this Xceed Kit a few times this year but couldn't find much fun using it :/ sure, 256+ gray scales look pretty fantastic but...

My recorded video with this Micron Xceed Grayscale Kit:

https://www.youtube.com/watch?v=ZsUVtx3pP8Y 

I like my b/w SE/30 like it is ... I am open for trade offers if someone wants this 'Unicorn Kit' ...first time some of my tradeworthy stuff is kind of "goldweight" - there is also a thread already: 



cheers,

Marcus

@pcamen if your still willing to assist with some funding I think @MarNo84 Card would be the perfect candidate for reverse engineering processes.  Since his already has some issues, destroying it won’t feel as bad.  I contacted an engineer who can then modernize the card with fewer and newer chips. Besides talking to the engineer I’ve not had much time to do other things. 

I believe @Bolle is reverse engineering the 306-48 card with the intent of coming up with a new (simpler) design for an internal greyscale only variant that can be mass produced.  The idea is do do away entirely with the external video option and focus just on the internal greyscale. 

That said, if we can also try to clone the existing Xceed greyscale card, that might be interesting too.  And perhaps going down both roads gives us a better chance of success.  I'm open to chipping in for that. 

I wasn’t aware of @Bolles project. Perhaps it might be beneficial to wait on this project until we see how that develops. Don’t want to waste time overlapping work.  

MarNo84 said:

...hmmm, is this Micron Xceed Grayscale Kit so much special? What makes them so rare and pricy at all?

Click to expand...

Back to car analogies, it's a rare high performance vehicle modification akin to Shelby modified Mustangs. I don't know how many were made, but nothing like a sizeable percentage, if even a significant percentage of SE/30 production. It's not really in the unicorn class where a Shelby Cobra would be. Micron had a patent on the Xceed GS mod, so no others became available for the niche SE/30 market. The cards were inexpensive enough for the time, but use of a second, Color Screen possibly topped the list of reasons to buy any Micron card, especially so the 1024x768 capable versions.

Nearly done with the schematics and will post them after cleaning them up a little bit.

The breakthrough on the Carrera040 front got me distracted from finishing them in the first place.

With the schematics and the Micron docs including the GAL equations at hand someone can go ahead and do the maths on how to work those counters so the card gives us the desired resolution and refresh rate.

Not sure what I should think about reverse engineering the GS capable card. @maceffects how would your engineers go about the custom silicone?

I guess something along the lines of decapping it and extracting the internal layout from die photos?

The schematics for that card are already out there so not "much else" to be done other than reverse engineering that ASIC (and the few GALs) to get full insights on what the card does on a hardware level.

@Bolle I wasn’t aware we already have the schematics. That will save time.  I think there are two options, one more reliable than the other. One used some mound kind of radiation scanning, which is very reliable and not destructive. The other sounds more like the method you mentioned which is cheaper but destructive and may required a second chip. 

I fail to see the need for doing reconstruction of the original boards at all. Once the internal GS only 306-48 clone is done, we should take another route. PDS pincount is problematic as compared to NuBus, but I'd think a far more capable Dual Head card would be the way to go:





Keeping the likes of a GS screen in a clear SE/30 display next to a primary workstation should be the goal IMO. With that Dual Head HDMI output and peripheral connection Wombatted to USB peripheral inputs over a KVM hookup to a primary workstation would be killer.

The HDMI section could be entirely discrete on the GS only Clone, using a second interrupt, leaving the third available for a NIC.

Just tossing the crazy notion out here. If you've got engineering expertise on tap, maybe have them take a look at the prospects for doing something more advanced than rehashing ancient, wholly obsolete, proprietary ASIC based hardware?

@Trash80toHP_Mini I’m definitely willing to do whatever it takes to make grayscale a possibility for all to enjoy.  I really feel like that’s the only downside about the classic Macs.  If you or @Bolle think of anything I can do, just let me know.  Or if you you me to sit down and enjoy the ride while the real engineers go to it, that’s good too. I’m not an electrical engineer.  My skills are quite low.  I often quip that I’m the only one to get a Masters from Purdue University that wasn’t in engineering.    :lol: :lol: :lol:

I think my problem is I want to help, but just don’t know where to direct my attention. 

Would it be unrealistic to try for an SE/30 motherboard "interpretation" that includes a micron grayscale fpga built in? I believe someone in a 2017 post somewhere said they were close to recreate an SE/30 motherboard, which would have the potential to draw much less power and would be clean without 30 years of gunk? I already have a twinspark +040 and ethernet and it seems very crowded. I don't know if my power supply could handle another pds card.

How many people would want both the internal grayscale option and external function? I would be happy with just the internal grayscale. Would this be easier to accomplish?

Another post I saw is someone replacing the CRT with an OLED pasted to the inside of modified glass which seems like it would be better on the eyes if it could work with grayscale.

Personally, I only want internal grayscale and the only modification I'm comfortable with is one that keeps the original motherboard and CRT.

Jinnai said:

Personally, I only want internal grayscale and the only modification I'm comfortable with is one that keeps the original motherboard and CRT.

Click to expand...

I agree, I could make an adapter that would allow a LCD to fit in and look stockish but, the real appeal is the original CRT and Grayscale.  I have zero interest in external video either. 

A completely brand new SE/30 board would be wonderful, though not really financially possible.  To say nothing of the lack of parts, unless using newer stuff.

I mean, I think ants got a WORKING color LCD up in his SE/30, even made a curved cover for it to give it that factory correct look, but said he didn't have interest in making it color anyway (it would also do grayscale no problem). There's a hacking thread for that elsewhere.

I'm also of the opinion, stated further up or in a tangential thread, that I have 0 bytes of interest in external display support, and would rather have grayscale internal support, if anything. This is party to discussion further up about project feature creep, and a simplified internal-only card. Also, one that would work and play well with TwinSpark or Bolle's PDS adapter card for cache accelerators, NICs, etc. IIRC the compat problems were centered around the cache cards. but still, where concession can be made…

I think for simplicity our focus should be to provide only the internal greyscale enhancement.  That is obviously one of the most highly valued enhancements to the SE/30 that people desire, given the prices of the Xceed greyscale capable cards. 

Once that is done, the work can be leveraged into other things. 

I personally would be all for new hardware that can run the classic macOS versions, like a newly envisioned compact Mac form factor with a newly developed system board.  I am mostly for the OS experience personally and get frustrated constantly having to futz with the hardware issues.  But on the other hand I find emulators not fully reliable.  Newly developed hardware would be a good in-between point. 

A co-processor type add-one for a modern Mac would be completely awesome in my book. 

One reason I was wondering about mb modified with an onboard grayscale fpga is the Carrera040 Info /Hacking Thread:

trag said:

Is this a correct summation of the problem:

---------------------------------------------------------------------------

The SE/30 has only an 8 bit data path to the VRAM (D24 - 31).

This requires byte-wise writes to the VRAM, instead of using the 680x0's normal 32 bit data path.  This uses a feature of the 680x0 chips called Dynamic Bus sizing.

The 68030 uses a different control signal scheme to signal/setup Dynamic Bus Sizing than the 68040.

...

Seems disappointing to me that the data path to VRAM is only 8 bits wide.   Aren't there two VRAM chips?  Are they only 4 bits wide each?

Click to expand...

I wonder @Bolle are these problems with xceed not being compatible with an 040 related to the Carrera not compatible with the VRAM?

If these are related, then it seems like it would require modifications to the mb anyway to get an 040 and grayscale to both work?