Mac Classic - will it ever run again!!

[rubber_jonnie]

Given that the original BluePill was a cheap no-name Chinese board they're all "fakes".

Ok then, a fakier fake than the original fakes :)

[olivier.jan]

So finally I took some time to open the Mac Classic and test the BlueSCSI and the results are mixed.
First, using 6.0.8 from ROM or from a floppy makes things difficult as one software is not properly working with this version, so I ended up downloading the pre-installed 7.0.1 image from the BlueSCSI wiki.
Anyway, although I manage to boot with this image, it keeps crashing with "illegal instructions" "bad address", etc...
Tried multiple things, always ended up with a crash.
I also tried booting the image which works flawlessly on my SE/30 and even with extensions OFF it finally crashed before the end of the boot.
At some point I ended up with a corrupt SD Card :(

After a while, I removed the BlueSCSI and used an external BlueSCSI (Plugged to the DB25 port):
- My SE/30 image boots without issues (with extensions OFF).
- The pre-installed 7.0.1 image from the Wiki works well and is stable.
- I managed to initialize a blank image and install a fresh copy of 7.0.1 on it without issue.

There is clearly something wrong with the internal version of the BlueSCSI.
I see two possibilities:
- We are very unlucky and we both have Internal versions with crappy STM32 and I have 2 external versions with good stm32.
- The power delivered by the internal SCSI bus on the Classic is causing issues.

I keep the Classic open and on the desk for the WE, so let me know if you need further testing.

Olivier

PS: The Classic is fully recapped, both MB and AB.

[Tomswork]

I have built 5 internal and 5 external with blue pills from 3 sources. 3 external pills work perfectly. 3 internals work perfectly. And the other 4 have been night mares to program and most run the slowest versions of the firmware and require special programming tools.

Tom

[rubber_jonnie]

So finally I took some time to open the Mac Classic and test the BlueSCSI and the results are mixed.
First, using 6.0.8 from ROM or from a floppy makes things difficult as one software is not properly working with this version, so I ended up downloading the pre-installed 7.0.1 image from the BlueSCSI wiki.
Anyway, although I manage to boot with this image, it keeps crashing with "illegal instructions" "bad address", etc...
Tried multiple things, always ended up with a crash.
I also tried booting the image which works flawlessly on my SE/30 and even with extensions OFF it finally crashed before the end of the boot.
At some point I ended up with a corrupt SD Card :(

After a while, I removed the BlueSCSI and used an external BlueSCSI (Plugged to the DB25 port):
- My SE/30 image boots without issues (with extensions OFF).
- The pre-installed 7.0.1 image from the Wiki works well and is stable.
- I managed to initialize a blank image and install a fresh copy of 7.0.1 on it without issue.

There is clearly something wrong with the internal version of the BlueSCSI.
I see two possibilities:
- We are very unlucky and we both have Internal versions with crappy STM32 and I have 2 external versions with good stm32.
- The power delivered by the internal SCSI bus on the Classic is causing issues.

I keep the Classic open and on the desk for the WE, so let me know if you need further testing.

Olivier

PS: The Classic is fully recapped, both MB and AB.

Interesting. I can boot from 6.0.8L on floppy and initialise the HDD (BlueSCSI) with the patched HDD tools. I can copy files there and it all works fine, until I reboot, and then it give me the same errors, crashing with "illegal instructions" "bad address" etc.

In fact I thought I had it cracked earlier today, I changed the SD card and after installing 6.0.8L it booted fine, but then crashed about 5 minutes later. They tend to usually crash just as it boots into desktop. I'm really thinking it's termination, so I am going to look at that further, but I have to replace the brakes on my wife's car this weekend, so time may be limited.

[rubber_jonnie]

Ok, so since I replaced the BluePills in my BlueSCSI devices, the Mac Classic has been running perfectly with 2x 2GB 'hard' drives, so I'm really happy with the result.

But, as usual, the Classic has thrown me another curveball, the PSU side of the Analog board has decided to play up again:

PXL_20230908_124647549.jpg

The 5v line is reading just over 4v at this point, 12v line is 11.1v, so the PSU is just not delivering properly. Leaving it for 5 minutes results in a chime and successful boot, and all voltages return to normal, and this is either with the original hard drive or the BlueSCSI.

I'll have to retrace my steps to remind myself of how I fixed this the first time, as I want to button up the machine with the BlueSCSI installed, and this means I can't.

Ho hum.

[stephen_usher]

That "feels" like a capacitor issue, though I s'pose it could be a dry joint somewhere.

[rubber_jonnie]

That "feels" like a capacitor issue, though I s'pose it could be a dry joint somewhere.

That's what I thought, but the PSU part has been completely recapped. I'm going to have to pull the board and stick it under the magnifying glass to see if there are any dry joints, or if there have been any subsequent cap failures.

It's very annoying.

[stephen_usher]

What does the +5V line look like under an oscilloscope? It could be that it's not 4V but an oscillating 5V with an average of 4V.

[rubber_jonnie]

What does the +5V line look like under an oscilloscope? It could be that it's not 4V but an oscillating 5V with an average of 4V.

Yes, I agree, just need to get the Spectrum+ off the workbench, which I will be doing shortly now it's fixed.

IIRC last time this happened, it was taking it's sweet time getting started, so it could well be oscillating. Hopefully have a look this afternoon.

[rubber_jonnie]

What does the +5V line look like under an oscilloscope? It could be that it's not 4V but an oscillating 5V with an average of 4V.

Yep, it's oscillating ...