Mystic Died. Looking for an MC88916DW80 chip.

[YMK]

The LEDs indicate that the Mac isn't hanging up in the middle of a SCSI transaction.

The SCSI controller is good to 25MHz, so it's within spec.

Will it boot if you turn extensions off?

Some more troubleshooting options:

• Try a shorter SCSI cable
• Try running MacSD at 33MHz
• Boot from another medium and run some SCSI tests (benchmarks, copying/unstuffing large files, etc).

Also, I've mentioned this before, but raising the 5V rail can help with stability at higher clock speeds. 5.25V should be safe.

Compare your 5V rail when using MacSD vs the spinning hard drive. Load on the 12V rail can affect 5V rail voltage.

 

[JDW]

Will it boot if you turn extensions off?

Yes, but such is nothing new. The same is true when I try to boot from BlueSCSIv2. Pressing SHIFT allows it to boot to the Desktop. But it cannot be an extensions conflict, for two reasons:

1. I can boot my spinning platter drive just fine without pressing SHIFT.
2. I've spend hours disabling most everything with Conflict Catcher in the past, and nothing other than pressing SHIFT seems to get MacSD or BSv2 to boot while the 575's CPU has a high overlock like 48MHz. Booting 7.1 & 7.5.5 work at 48MHz though.

• Try a shorter SCSI cable

Doesn't change anything. I used SCSI Rider with BlueSCSIv2, which allows me to attach it externally with no ribbon cable at all. I then used SCSI Rider with a 15cm "short" ribbon cable to attach MacSD (since MacSD cannot be directly plugged into SCSI Rider), but that changed nothing.

• Try running MacSD at 33MHz

Tried that, but no change. Still freezes at the same point, unless I press SHIFT.

• Boot from another medium and run some SCSI tests (benchmarks, copying/unstuffing large files, etc).

I must interpret your word "medium" as being "another drive." And if by that you mean, "boot from your spinner drive with MacSD attached," I've done that many times. My spinner will boot just fine. And I can run benchmarks, copying/unstuffing large files, etc. just fine in that case.

However, even without booting from another drive, if I press SHIFT, I can boot from MacSD to the Desktop, as I have said. And from there, I can run benchmarks like Speedometer 4.0 and Norton System Info, run the DayStar Power Demo app, copy large files and do pretty much anything that doesn't require a CP or Extension to be loaded. The same is true when I press SHIFT and boot to the Desktop with BlueSCSIv2.

But I must repeat, this isn't an extensions conflict because, as I've said, I can boot from my spinner drive just fine WITHOUT pressing SHIFT, and booting with the same CP's and Extensions. And I've spent a long time with Conflict Catcher trying to narrow the freeze down to some extension or control panel (when booting from MacSD or BSv2 at high CPU clock speeds), but I cannot find anything that triggers it.

And like I said, all I need to do is just reduce the CPU clock speed to 43.2MHz and then I can boot MacSD and BlueSCSIv2 just fine, with all those extensions and CP's in place! So it is absolutely positively not an extensions conflict.

Also, I've mentioned this before, but raising the 5V rail can help with stability at higher clock speeds. 5.25V should be safe.

Compare your 5V rail when using MacSD vs the spinning hard drive. Load on the 12V rail can affect 5V rail voltage.

Because I can't do such a test with the Analog Board removed, where do you recommend I measure +5V?

If I press SHIFT to kill extensions and then boot to the Desktop, if I then measure between pins 5 & 7 on MacSD's Expansion Header, I get 4.65V. But that's Termination Power, so I don't know if that's really the best location to measure, although it is easiest to access.

 

[YMK]

And like I said, all I need to do is just reduce the CPU clock speed to 43.2MHz and then I can boot MacSD and BlueSCSIv2 just fine, with all those extensions and CP's in place! So it is absolutely positively not an extensions conflict.

Ok, sorry I missed those details.

Because I can't do such a test with the Analog Board removed, where do you recommend I measure +5V?

You can get 5V at the ADB port, ideally with no peripherals attached.

If I press SHIFT to kill extensions and then boot to the Desktop, if I then measure between pins 5 & 7 on MacSD's Expansion Header, I get 4.65V. But that's Termination Power, so I don't know if that's really the best location to measure, although it is easiest to access.

It's not a good place. It goes through a diode and it's also subject to current spikes on the SCSI chain.

 

[JDW]

You can get 5V at the ADB port, ideally with no peripherals attached.

The "no peripherals attached" part is problematic because you need an ADB keyboard in order to press the PowerKey and power-on the Color Classic Mystic. And removal of ADB devices while the machine is powered is not advised. When I have disconnected an ADB device in the past by accident, it usually freezes the Mac.

 

[YMK]

And removal of ADB devices while the machine is powered is not advised.

If you don't want to disconnect a cable, you can power on the CC by shorting the ADB power pin (not 5V) to GND.

I haven't observed any lockups by pulling an ADB cable. The worst I've seen is the mouse going to its slowest setting.

 

[JDW]

you can power on the CC by shorting the ADB power pin (not 5V) to GND.

I am seeing some incorrect info posted on the internet when it comes to ADB pin-outs, but it would appear the following should be correct (I think) when looking at the back of the LC575 motherboard.



When I touched my multi-meter's Black/GND probe to one of the motherboard connector bolt holes (which should be GND), and then press the RED probe against a jumper wire inserted into "3" above (which should be +5v), I get nothing with the machine powered off, which is correct. And when I put a wire into "2" above and press that wire against the same connector bolt hole (effectively Grounding pin-2), the machine won't power on, but I then am able to measure 4.835v on "3" above. I touched "2" to the bolt-hole GND for a short duration and a long duration, but that won't power the machine on, despite the fact it theoretically should.

If I disconnect my jumper wires and then plug in my keyboard and press the PowerKey, then the machine powers on. And if I put my meter into beeper mode (to check continuity) and if I put one probe on "2" and the other probe on "4" (on the ADB cable coming from the keyboard, not the motherboard connector, mind you), I get no beep when I press the power key.

Going under the assumption that the bolt hole isn't a good enough GND, I used by jumper wire (with male pins at both ends) to connect "2" & "4". To my delight, that powers on the machine. I then had my meter's RED probe on "3" and quickly put the BLK probe on "4". I measure 4.925v initially. Then when MacSD or BlueSCSIv2 stops working (the point at which the machine freezes), the voltage measures 4.930v.

I don't have my scope here at home to detect very quick voltage dips, but I see nothing on my 121GW multimeter that indicates the voltage is unstable at all. And let's be honest here, 4.9x volts is truly good enough, especially when we consider that I am measuring this externally at the motherboard's ADB connector.

So my feeling is, this test shows that the 5-volt line's voltage level is not the cause of the freezing when the 68040 is overclocked to 48MHz.

 

[JDW]

@YMK

I've also been talking to @eric via PM, and he asked me to install Macsbug and then enter the debugger by pressing CMD & PowerKey at the time of the lockup. When I do that, I see the following (two different screenshots because I tried it multiple times):



To escape out of the debugger, I type "ES" and press RETURN. Most of the time, that triggers a total machine freeze, and the arrow pointer no longer moves when I move the mouse. But there are those rare times when I type "ES" and press RETURN, only to find the machine then boot to the Desktop. So the sheer act of jumping into and then out of the debugger can sometimes cause the machine to boot into 7.6.1.

But please don't forget that:

1. I can boot into 7.1 and 8.1 just fine, with MacSD or BlueSCSIv2, with the 040 clocked at 48MHz.
2. My spinning platter hard drive can boot into all the OS's including 7.6.1 and 8.1 even at 48MHz.

 

[YMK]

The purpose of measuring the 5V rail was to determine if having the HDD installed affected it. I see you've measured 4.930v for the solid state drives. I may have missed it, but I didn't see one for the HDD.

And let's be honest here, 4.9x volts is truly good enough, especially when we consider that I am measuring this externally at the motherboard's ADB connector.

It's good enough for operation at the stock 33MHz. Beyond that, you are promised nothing.

It's common knowledge in the PC overclocking scene (from the 486 era to present day) that running higher voltages ("overvolting") is often necessary for stable operation at higher speeds.

Conversely, the silent PC crowd that aims for slow/no fans, knows that when you reduce CPU voltage/wattage, you must often reduce clock speed to below the CPU's rating to achieve stable operation.

In both cases, there's a strong correlation between CPU voltage and maximum achievable speed.

Measuring at an unloaded ADB port should be a very good representation of CPU voltage. I don't see why it should measure any lower.

If you're pushing the envelope on clock speed with any platform, millivolts matter. The good news is since it's running on the low side of the normal range, you have quite a bit of headroom. I would still recommend increasing the 5V rail regardless of the outcome of this specific issue. It may correct other problems you've been seeing, like video noise. At 4.9V, you are most likely leaving several MHz on the table.

Don't take my word for the voltage issue. Ask around or research it. If I'm mistaken, I welcome anyone to correct me.

 

[JDW]

The purpose of measuring the 5V rail was to determine if having the HDD installed affected it. I see you've measured 4.930v for the solid state drives. I may have missed it, but I didn't see one for the HDD.

In my conversations with @eric regarding BlueSCSIv2, he recommended that I completely disconnect both power and the SCSI cable from my spinner drive so my testing could focus specifically on the SD card drive in question. So I have been testing MacSD separately but in the same way as BSv2.

Although I will try boosting the voltage later, I’m not holding my breath that will work.

All said, I hoped to swap out my spinner drive with MacSD or BSv2, but because those drives don’t offer me overclocking flexibility like my spinner drive, if raising the voltage doesn’t help, my solution will be to continue using my spinner drive and simply enjoy the OS booting compatibility it offers.

 

[YMK]

All said, I hoped to swap out my spinner drive with MacSD or BSv2, but because those drives don’t offer me overclocking flexibility like my spinner drive, if raising the voltage doesn’t help, my solution will be to continue using my spinner drive and simply enjoy the OS booting compatibility it offers.

Before boosting the voltage, try getting a 5V measurement with the spinner. It's an important data point.

MacSD and BSv2 use completely different SCSI implementations and both work perfectly well on much faster hosts than the Mystic. The fact that you're getting the same exact outcome with two very different SCSI devices suggests neither one is at fault.

Another quick test you can try is booting from MacSD with only the power cable connected to the HDD.

 

[JDW]

Before boosting the voltage, try getting a 5V measurement with the spinner. It's an important data point.

MacSD and BSv2 use completely different SCSI implementations and both work perfectly well on much faster hosts than the Mystic. The fact that you're getting the same exact outcome with two very different SCSI devices suggests neither one is at fault.

Another quick test you can try is booting from MacSD with only the power cable connected to the HDD.

I will connect the spinner drive to get that data point.

Thank you for all your kind help to date!