Macintosh Classic stalls booting when anything is plugged into the SCSI bus

[Verault]

I have a MAC classic board which will boot and work fine from a floppy drive but the second I connect a hard drive via internal ribbon or external connector the system just sits forever without progressing to even the floppy drive. The mouse cursor still works, its just perpetually waiting it seems forever. I am looking into sourcing another SCSI chip could it be something else? I did try the ROM from another Classic board and it made no difference.

 

[dochilli]

Are you sure that the HDD is funcional? Old HDDs often fail after 30 years.

 

[Verault]

Two working hard drives, 6 working blue SCSI devices, and two working external SCSI2SD devices that all work fine on other system. Its definitely not the devices or the cables because they work on other machines.

And I guess the more information the better right. This system was already recapped and it did not suffer battery leakage or severe cap leakage damage. The board was pretty clean prior to the recap.

It boots fine if only a Floppy drive is installed and IT can boot to the onboard ROM drive as well. Put anything on the SCSI bus and it just hangs forever but the mouse cursor will still move.

 

[ymk]

Could be something wrong with the /ACK or /REQ line.

Once you're booted from a floppy measure the data and control lines at the external connector.

They should idle at around 2.8V.

SCSI External D-Sub (PC/Amiga/Apple Macintosh) pinout and wiring @ old.pinouts.ru

 

[Verault]

Thanks YMK I will test it tomorrow and report back what I find.

 

[ymk]

I forgot to add, since the Classic doesn't have board termination, you'll need a hard drive plugged in (with its termination resistors). Disregard the part about booting from floppy.

Also, look for ~4.5v at the TRMPWR pin.

 

[Verault]

Yes I am aware about the termination and there was termination set on the drives. I will post back with the pin voltages.

 

[Verault]

Ok here are the voltages. First I took them with no SCSI devices in the machines. PIN1 /REQ gave 0.11 VDC Pin 5 /ACK gave 0.18 VCD and pin 25 TRMPWR gave 4.5 VDC

Next I tried an internal drive connected to the internal SCSI connector. With pin 1 /REQ I got 2.19 VDC with pin 5 /ACK I got 2.2 VDC and with pin 25 TRMPWR I got 4.03 VDC.

 

[ymk]

Those readings look good. Voltage on the rest of the control and data lines should be similar to /REQ and /ACK.

 

[Verault]

Well I dont know what it could possibly be. Like I said this thing will just wait forever and thats not normal behavior

 

[ymk]

By my comment, I meant measure the rest of the lines. They should also be around 2.2V.

The state of the control lines may also reveal what stage of the transaction the machine is stuck in.

 

[Verault]

Ok Ill check the rest as well. Does it matter if I have a device connected or not?

 

[ymk]

Check first with the hard drive attached. The lines should behave similarly to /REQ and /ACK; around 2.2V with the drive and 0V without.

 

[Verault]

Just to be thorough here are all the voltages without and with a scsi device:

 

[ymk]

Those voltages look OK and show the bus is idle.

Check the pins of the controller for electrolyte and corrosion. The nearby rail caps often leak on to it.

Swapping or reflowing the SCSI controller is what I'd try next.

 

[Verault]

ok ill clean the contacts with some deoxit. i have a new scsi chip on order. the board was already clean when it was recspped but i can try cleabing the scsi connector better.

 

[Verault]

Just curious if there are any other components besides the SCSI chip which could be involved in this?

 

[ymk]

I'll check the schematics later. What do your +5V and +12V rails measure with that hard drive attached?

 

[Verault]

So I am getting 4.87 without a load and 4.83 with a load on the 5v rail and 11.66 without a load and 11.61 with a load on the 12v rail.

 

[ymk]

Those are low, but I doubt they're causing a problem.

Here are the lines on the SCSI controller that reach other ICs. The ones in cyan go to the CPU. I couldn't find any other reference to /IRQ in the schematic. I did trace it to pin 14 of UI6 (which I also couldn't find in there) on a spare board.