Checker-board screen with 68030 Performer Accel in Mac SE

[DrGonzo]

Greetings all!

Wondering if anyone has built one of these and what their experience might be:

GitHub - TheRealBolle/Performer-SE-PL-CL: 16MHz 68030 accelerator for Macintosh Plus, SE and Classic

16MHz 68030 accelerator for Macintosh Plus, SE and Classic - GitHub - TheRealBolle/Performer-SE-PL-CL: 16MHz 68030 accelerator for Macintosh Plus, SE and Classic

github.com

I got a stack of boards from JLC and put one together on the quick with parts I had mostly on hand to see how it worked before building a few nicer ones (with real sockets, 68882, proper card connector vs. pin headers) for me & a friend. Unfortunately when I put it in my working SE I get a checkerboard screen and nothing else. Pull the card and all is again well. Still troubleshooting a bit but hoping someone else can confirm success with building one. I did have a couple things I was a bit unclear on when building:

- The instructions are a bit vague on burning the GALs. To pass verification I had to uncheck encrypt (obviously). Do they need to be encrypted to work? I can pull them and redo, but hoping to confirm before wasting the time pulling/resoldering.

- The picture of the card on the site shows caps and such that don't match the BOM. i.e. Tantalums are listed as 22uF on the BOM, but the pic shows 47uF. I went with what was in the BOM, so hoping that's right?

Appreciate any input! Once I get these sorted I'd like to build out a few of the Tokamac 68040 cards for my LCIIs.

I attached a pic of the one I threw together.

 

[JT737]

Burning the GALS should be pretty strait-forward; myself, I use a TL866II and have always had good luck with it. As long as you've matched up the chip properly in you chip programmer's interface, it should work. Just make sure that you have the right image!

 

[DrGonzo]

Yeah, definitely seemed like the flash went fine on my TL866Plus, and the Lattice 16v8a was listed in XGPro. I pulled the images right from the site above and kept track of the images for U3-U7 prior to mounting so should be correct as well. I just wasn't sure if the board/SE were being grumpy due to the GALs needing to be encrypted (locked?) or something. I'm used to doing 27cXXX type eproms where that's not a thing. I can certainly erase/reflash, but really wanting to get some guidance on the best approach before removing them for reprogramming. I really would like to have put sockets on the board, but there isn't enough clearance between neighboring components. Hence trying to avoid trial and error on them. heheh

Thanks for the input!

~Dr. G

 

[Nixontheknight]

iirc, there's a jumper on the board for selecting which machine you're using, maybe you have it in the wrong position?

 

[DrGonzo]

That brings up an interesting point as well. The pictured board on the site does indeed have a jumper to choose between Plus/Classic and SE. However, the board as built from Gerbers does not. It does however mention in the build notes that the GAL at U7 is only required for use in a Plus/Classic and to close SJ2 if 'not' installing U7. For my testing I'd only intended to use it in an SE so I left U7 off and put a solder blob across SJ2, assuming that was the replacement for the jumper and would select SE. Hopefully that all makes sense.

It's easy enough to wick off the solder across SJ2 so may give that a try when I get home just to see if the behavior changes though.

Thanks!
~Dr. G

 

[Garrett B]

You need GALs with active pullups. 16V8A does not have internal pullups. Rev B and onward should have them built in...

 

[DrGonzo]

Doh, well that would have been good to know. Glad I didn't buy 50 of them.

Looks like there are some 16v8D-15 versions on eBay I can get for not too much more than I paid for the A ones I have. Will give that a go and hopefully that does the trick. As above, do you happen to know whether encryption/locking makes any difference when it comes time to flashing the replacements? Would rather get it right the first time and just solder them down once.

Thanks much!
~Dr. G

 

[Garrett B]

D should work just fine. Encryption won't affect the operation of the chip. I don't encrypt because then it allows me to "read" the chip and make sure it did actually program correctly.

 

[DrGonzo]

Yup, exactly why I hadn't, as the there's a level of comfort with verification succeeding. Thanks again and will report back once I get some B+ GALs in hand.

Thanks again,
~Dr. G

 

[DrGonzo]

Success! Welll, sort of... So new GALs came in yesterday, got them programmed and swapped onto the board, and was greeted with the familiar SE beep. After some uncomfortable pauses during boot I did eventually get to the desktop. So far, so good. Doesn't seem much peppier with the 030 so checked TattleTech which seems to think I have a 68020 at 2.95MHz? More troubleshooting to come, starting with a fresh install of 7.0.1. Hoping to report back with a fully working 030 at ~16MHz soon.

Thanks again @Garret B for the tip on the GALS (rev D). Never would have guess that one!

Regards,
~Dr. G

 

[Garrett B]

Don't forget about the extension for the Performer! It can be downloaded here:

https://macintoshgarden.org/apps/micromac-accelerator-software

Things should show up correctly after that.

 

[DrGonzo]

Well Thank You yet again! For some reason I had it in my head the Performer was plug/play. I'm sure that will sort things on the quick. Will report back ASAP. Looking forward to building some cleaner ones with proper 030 sockets, 68882, and board connectors.

Then onward to the Tokamac 040s! I'm sure I'll have a few questions on those as well when the time comes. heheh

~Dr. G

 

[DrGonzo]

Success! No, really...

Extension installed and now TattleTech is showing a 68030 @16MHz with MMU. Speed test has me at 3.3 times stock with slightly faster graphics and math. Next/final builds will have the 68882 @25MHz as well because... reasons. Heheh.

I do need to bench one of my SE/30s for comparison. Looked like on LowEndMac it should be around 4.25 times stock, so not terrible at 3.3x considering only 16-bit path to memory.

Anyway, thanks for all the help/advice!

~Dr. G

 

[DrGonzo]

The final build in case anyone wanted to see the finished product.





Thanks to Bolle for taking the time to re-engineer these boards and putting them out there for all to share and enjoy, and thanks to the gang for helping through the hurdles I ran into.

Regards,
~Dr. G

 

[DrGonzo]

Well, final except for the 68030 which is still on the slow boat from China anyway.

~Dr. G

 

[Garrett B]

Beautiful work! Very impressed you added a socket for the 68030. I took a chance and soldered mine directly to the PCB...

 

[JC8080]

Does anyone know if there is a modern version of a killy clip that would let this clip over the 68000 on a Plus? I see it is designed to fit a plus, but I'm not sure how it would attach.

 

[DrGonzo]

I was looking around for one to throw on the bottom of the Performer just to add the capability, but didn't find anything. Now that I know what they're called I'll have to look some more.

~Dr. G

 

[JC8080]

Here's what I had a guy do for a Total Systems Plus accelerator I have, the accelerator has a socket that these pins fit into. It originally had a killy clip, but the plastic became brittle and failed. Unfortunately the accelerator doesn't work at the moment. I did also replace the SIMMs with the low-profile horizontal chip type to add clearance.

I'm not sure where the guy came up with this, he said he found it in parts he had laying around. He said he tried pin headers, but they were too flexible and wouldn't hold the board on.

 

[DrGonzo]

Interesting solution for the CPU. Low profile SIMMs look good too.

I wondered if the processor in a plus or 512k could be socketed and accellerators like these could replace the CPU rather than sitting on top? That way it would make for an easy restore to stock. Regarding the issue with it staying on the board with headers, I do plan to create a 3d printed mount for the performer to add some stability. With the PDS connector it's not specifically needed, but can't hut to support the board. Same could be done on the plus/512k which might be an option for further supporting the TS Plus?

Regards,
~Dr. G