SE/30 with DayStar Turbo 040 Struggles

[jasonfortheworld]

Howdy y'all. I feel like I'm beating my head against a wall, but maybe I'm missing a crucial step.

I'm trying to get an SE/30 up and running with a DayStar Turbo 040 hanging off Bolles combo PDS card (using the correct accelerator slot and not the pass through), but try as I might I just get stuck on what I believe is a RAM test screen with no boot chime at all. I've attached a photo of the screen just in case I'm wrong.

I'm using 128mb of RAM, a bluescsi v2 with System 7.1 and the Quad control panel installed, and the Garrett's Workshop GW4402B ROM set in stock IIsi ROM mode (the custom ROMs on it are not compatible with 68040 CPUs). All these components have been tested in a Mac IIci and work great with the accelerator. No issues at all.

Any chance I'm missing something obvious?

The analog board and logic board have both been recapped and I recently upgraded to this power supply from TTdesigns so I doubt it's a power issue:

4XDC Apple Macintosh SE/30 & SE Power Supply (PSU) by TT Design on Tindie

4XDC New Apple Macintosh SE/30 and SE AC-DC Power Supply Board

www.tindie.com

 

[jmacz]

Did you try testing in pieces? Like what happens if you just remove the turbo 040 and leave the other components in place? You should have plenty of power with @tt ’s psu but note if you haven’t upgraded the wiring between the analog board and the logic board, you may want to (upgrade = use thicker gauge wiring … 16 or 18 awg).

 

[jasonfortheworld]

I've gone one item at a time and everything was working great until the accelerator was added. I haven't upgraded the wiring yet, but it is on my list to do. Could that be the problem and maybe cause the voltage to sag?

 

[s_pupp]

Same problem here (description along with troubleshooting so far in detail below).

I wonder if there is residual corrosion with effects that are seen only when stressed with the requirements of an 040 .

Can crystals can go out of whack over time? Is it worth replacing those to see if that solves the problem?

Description:
I have the same problem with two Daystar Turbos tested in four SE/30’s (stock ROM’s), both with Twinspark adapters and 90% of the time with Bolle’s.

The SE/30’s work flawlessly with a Daystar 50MHz P33 030 in the same adapters and power supplies, and all worked fine with the Turbo 040 in 2012.

I’m using 16ga cables, and have tried multiple power supplies (Artmix, Seasonic, Sony). Voltages read OK.

Apple diagnostics and a TechStep report no problems with the logic boards.

The SE/30’s and analog boards were recapped in 2012. The fact that one (and only one) of the SE/30’s will boot 10% of the time with the Turbo 040 in Bolle’s adapter leads me to conclude that the accelerator cards are OK, and the problem lies with deterioration of the SE/30 logic boards since 2012.

The fact that my other four SE/30’s that worked in 2012 after a recap now have either simasimac pattern (3) or horizontal stripes (1), even with the PDS slot unoccupied, is further evidence of deterioration over time.

I plan to remove the caps and any IC’s with corroded legs and do an US cleaning followed by a microscopic view of the traces under caps and corroded IC’s.

It is definitely worthwhile to try the 16ga cables and check the voltages under load - I look forward to hearing if it helps in your case.

 

[dougg3]

Garrett's Workshop GW4402B ROM set in stock IIsi ROM mode (the custom ROMs on it are not compatible with 68040 CPUs)

Have you tried a stock SE/30 ROM just to rule that out? I guess alternatively, have you confirmed that the Turbo 040 by itself works with a IIsi ROM in the SE/30? Edit: The part I crossed out is a dumb idea that could fry your board, don't listen to me, see below.

The reason I ask is the Daystar ROM detects which computer it's running in based on the ROM checksum. I wouldn't be surprised if it gets confused by a IIsi ROM in an SE/30. I could definitely be wrong, but I would recommend ruling it out as a possibility.

 

[jasonfortheworld]

Have you tried a stock SE/30 ROM just to rule that out? I guess alternatively, have you confirmed that the Turbo 040 by itself works with a IIsi ROM in the SE/30?

The reason I ask is the Daystar ROM detects which computer it's running in based on the ROM checksum. I wouldn't be surprised if it gets confused by a IIsi ROM in an SE/30. I could definitely be wrong, but I would recommend ruling it out as a possibility.

I tried with a stock SE/30 ROM and got the same results as the IIsi mode on the GW ROM.

But can I plug the 040 turbo directly into the PDS slot on the SE/30 logic board? I was concerned that I could fry it if I did that without an adapter like Bolles.

 

[s_pupp]

I tried with a stock SE/30 ROM and got the same results as the IIsi mode on the GW ROM.

But can I plug the 040 turbo directly into the PDS slot on the SE/30 logic board? I was concerned that I could fry it if I did that without an adapter like Bolles.

The Turbo 040 cannot be plugged into the PDS slot, and does require an adapter such as the Twinspark or Bolle’s. Bolle’s is more compatible with the Turbo 040. As you say, you could fry it without the adapter. The cache slot it fits into is the one that places the card horizontally over the hard drive.

 

[dougg3]

I tried with a stock SE/30 ROM and got the same results as the IIsi mode on the GW ROM.

But can I plug the 040 turbo directly into the PDS slot on the SE/30 logic board? I was concerned that I could fry it if I did that without an adapter like Bolles.

Gotcha. Sorry about that silly suggestion, I am quite unfamiliar with the SE/30 and accelerators so definitely ignore me on that.

I just know that I've helped people with Turbo 040 issues with custom ROMs and would definitely recommend keeping the custom ROM out of the equation until you've narrowed it down further. (Even if the IIsi ROM isn't "custom" it's still not a stock SE/30 ROM)

 

[jasonfortheworld]

The Turbo 040 cannot be plugged into the PDS slot, and does require an adapter such as the Twinspark or Bolle’s. Bolle’s is more compatible with the Turbo 040. As you say, you could fry it without the adapter. The cache slot it fits into is the one that places the card horizontally over the hard drive.

Gotcha. That was my understanding too.


I did a sanity check on the voltages via the HDD molex port with the new power supply and they seem fine. I'm getting 11.88v and 4.94v on boot.

 

[jasonfortheworld]

Disregard my last voltage check.

I tested via the J12 connection on the logic board and here is what I got:

Pin 6: -4.94
Pin 7: -11.92
Pin 12: 4.75
Pin 13: 4.75
Pin 14: 11.87

So pin 12 and 13 are out of spec and should be higher. My guess was this was due to my atx extension cable not being very good for this use case.

I put the logic board back into the case and removed the extension and tested again.

Now Pin 12 and 13 read 4.88 or 4.87 which is better, but still technically out of spec. My guess is my issues booting with the accelerator are due to voltage sag from using the original harness cable to connect the logic and analog board together.

I'll need to order some parts for a new cable before I can test this any further though.

 

[zigzagjoe]

You may be able to adjust the PSU to bouy that 5v a little. Also see here for harness discussion: https://68kmla.org/bb/index.php?threads/combating-se-se30-voltage-drop.46781/

Also note the revision of the turbo040 card matters when used with Bolle's riser. The early one which is covered with GALs and has a separate cache card AFAIK needs the newer version of his board or special modifications on the older version. The one with the large ASIC and onboard cache should work with either.

 

[jasonfortheworld]

You may be able to adjust the PSU to bouy that 5v a little. Also see here for harness discussion: https://68kmla.org/bb/index.php?threads/combating-se-se30-voltage-drop.46781/

Also note the revision of the turbo040 card matters when used with Bolle's riser. The early one which is covered with GALs and has a separate cache card AFAIK needs the newer version of his board or special modifications on the older version. The one with the large ASIC and onboard cache should work with either.

I'm not to sure if I can easily adjust the trim on the TTDesign PSU. At least I don't see anywhere to do that.

But thankfully my Turbo 040 is the later revision with 4.11 firmware.

 

[zigzagjoe]

I'm not to sure if I can easily adjust the trim on the TTDesign PSU. At least I don't see anywhere to do that.

But thankfully my Turbo 040 is the later revision with 4.11 firmware.

Oh, no, the meanwell based units are not adjustable. Sorry. There are redundant connectors on the tt board which could be eliminated if all other sources of voltage drop have been dealt with. Stock LB harness is the major weak point though and should be first priority.

 

[obsolete]

And from J12, you still have the voltage drop across the logic board, riser board, and two sets of DIN 41612 connectors. Might as well just solder a 4-pin Mate-n-Lok pigtail to the Turbo040 and connect it straight to the HDD power

 

[jmacz]

Gotcha. Sorry about that silly suggestion, I am quite unfamiliar with the SE/30 and accelerators so definitely ignore me on that.

I just know that I've helped people with Turbo 040 issues with custom ROMs and would definitely recommend keeping the custom ROM out of the equation until you've narrowed it down further. (Even if the IIsi ROM isn't "custom" it's still not a stock SE/30 ROM)

Given you tried with the stock rom, this is moot. But as @dougg3 mentioned (and @Bolle as well on another thread), a non stock checksum for the ROM will mess with the Turbo040’s behavior and potentially cause problems.

I had made a stock ROM that was patched to both remove the memory check and retain the stock checksum here:

Macintosh SE/30 ROM Modifications

Sorry, third thread on this (one for each of the IIfx, IIci, and now SE/30). Again, most people will opt for the Rominator or similar custom ROM. But a few folks were looking for a stock SE/30 ROM but with the memory test disabled (and stock checksum preserved) in order to boot faster and...

68kmla.org

Leaving it here for after you figure out the issue and only if you want to remove the slow memory check.

 

[jasonfortheworld]

Given you tried with the stock rom, this is moot. But as @dougg3 mentioned (and @Bolle as well on another thread), a non stock checksum for the ROM will mess with the Turbo040’s behavior and potentially cause problems.

I had made a stock ROM that was patched to both remove the memory check and retain the stock checksum here:

Macintosh SE/30 ROM Modifications

Sorry, third thread on this (one for each of the IIfx, IIci, and now SE/30). Again, most people will opt for the Rominator or similar custom ROM. But a few folks were looking for a stock SE/30 ROM but with the memory test disabled (and stock checksum preserved) in order to boot faster and...

68kmla.org

Leaving it here for after you figure out the issue and only if you want to remove the slow memory check.

Much appreciated! If I ever get a rom flashing tool, I'll gladly add this to the GW ROM as another option.

 

[jasonfortheworld]

So I made a new harness that also uses the AUX power from the 4XDC TT design board and did some voltage tests.

When reading the voltage directly off the 040 card at C37 of the PDS connector I was getting a solid 4.94v which would mean that the power is now totally fine. Unfortunately the issue of the computer locking up at the RAM test pattern still persists.

Just for a sanity check I tossed on a Radius SE Color Pivot video card in addition to the accelerator and bolle card and tested boot and voltage again. I was getting a rock solid 4.89v on C37 of the 040 card, but still a freeze on the RAM screen.

Lastly, I did power stress test. I kept the bolle card and video card, then added back the floppy drive, and then connected an internal DVD drive, noctua fan, and rasberry pi zero 2 w running MacProxy Plus to the original molex HDD power connector from the analog board. Everything booted just fine with zero issues and TattleTech reported everything as it should be.

The Pivot video card is also recognized by the machine and works perfectly with the bolle adapter.

I then rechecked that the GW ROM, Bluescsi image, and accelerator still work in my IIci, and sure enough it's fine.

I then ran a stress test with everything in the IIci. Both Snooper and Seedometer reported no errors or issues with the hardware.

I also ran a stress test on the SE/30 with just the bolle adapter and video card. Once again, no errors and everything was in spec according to snooper and speedometer.

At this point I feel like I've exhausted all the options I personally know about. Maybe it's an issue with the bolle card itself? Or maybe there some kind of logic issue on the accelerator or SE/30 board that is causing this? Either way it's beyond me now. I'm not sure what else to do.

 

[Bolle]

Have you tried what happens with a minimum amount of RAM installed? Like going for just 4x1MB SIMMs.
I have had trouble getting Turbo040s to work with certain 16MB SIMMs (just like some SIMMs can be troublesome with Carrera040s)

Also I gather your logicboard is an original one that has just been recapped? No other parts have been swapped out? I'm asking because replacing the PALs can lead to problems when using accelerators because timing differences between the original parts and "modern" replacements.

 

[jasonfortheworld]

Have you tried what happens with a minimum amount of RAM installed? Like going for just 4x1MB SIMMs.
I have had trouble getting Turbo040s to work with certain 16MB SIMMs (just like some SIMMs can be troublesome with Carrera040s)

Also I gather your logicboard is an original one that has just been recapped? No other parts have been swapped out? I'm asking because replacing the PALs can lead to problems when using accelerators because timing differences between the original parts and "modern" replacements.

I haven't tried a minimal amount of RAM yet, but I'll be sure to give that a go. I have tried two different sets of 128mb though and both sets worked in the IIci with the accelerator, but not in the SE/30.

As for the logic board, it's only been recapped. I have a second board I just got and needs some cleaning and a recap too, so when I get a chance I'll be sure to try everything with that board.

 

[jasonfortheworld]

So I tried with 4x1mb of RAM with no luck either. Seems to be a pretty stubborn issue.

Also, I recapped my extra SE/30 board, but it looks like that has some other issues at hand. The boot chime is slow and distorted and it seems like it refuses to recognize any scsi devices. I'm guessing there's some corrosion in some of my LB chips, so that will be another issue to solve another day.