Print

[TorZidan]

Hello Lisa gurus,

a Lisa 2/10 here is asking for your help:

Once I boot into MacWorksXL3.0 and try typing anything (e.g. when renaming a file, or in the "keycaps" test program), the cursor blinks much faster than usual, and each keystroke on the keyboard produces 2 to 4 symbols instead of one. In addition, double-clicking with the mouse anywhere on the screen (e.g. on the "hard drive" icon) does nothing. Everything else works great.

Tried replacing the I/O board with another one, and the issue goes away, meaning that the I/O board is bad.
Then, on the bad I/O board, tried replacing the COPS chip, the Eprom chip and the 6522 chip, but the issue remains.
Also, noticed that the COPS oscilator frequency can be adjusted by turning the adjustable capacitor C36 (see it on the attached diagram, above the COPS chip). So I tried turning it while the Lisa is on, but there is no change (the cursor continues to blink fast).
Any idea what the problem is? Where is the cursor speed being set? Why is C36 adjustable, if adjusting it has no visible effect on the computer's operation?

Thanks!

[sigma7]

MacWorksXL3.0 and try typing anything (e.g. when renaming a file, or in the "keycaps" test program), the cursor blinks much faster than usual, and each keystroke on the keyboard produces 2 to 4 symbols instead of one. In addition, double-clicking with the mouse anywhere on the screen (e.g. on the "hard drive" icon) does nothing.

IIRC, the blink speed, key repeat delay, and double-click speed are adjustable via a control panel. I suspect they currently have invalid values and setting them to one of the available radio-button choices will result in reasonable behaviour.

the COPS oscilator frequency can be adjusted by turning the adjustable capacitor C36... Why is C36 adjustable, if adjusting it has no visible effect on the computer's operation?

C36 has a very minor effect on the COPS frequency; it is to adjust the precision of the real-time clock (to minimize the number of seconds per day error of the clock). Assuming it works as intended, there must be a procedure for setting it accurately, but I've not seen it. Since the COPS stops when there is no power and so the time must be reset often, I suspect no-one is concerned about the error.

edit: I missed the bit about changing I/O boards affecting the problem... the settings in question are stored in the "PRAM" which is static memory on the I/O Board, some of which is shared with the floppy disk controller. The explanation may be that different memory chips on different I/O boards can default to different values when power is lost, or it may be there is a problem accessing the PRAM on one of the I/O Boards.

To test this portion of the I/O Board, try starting up with the I/O board that displays the problem, and see if setting the values in the control panel results in acceptable behaviour. Then shut down (but do not remove line power) and start up again and see if the values remain as you had set them.

[TorZidan]

IIRC, the blink speed, key repeat delay, and double-click speed are adjustable via a control panel. I suspect they currently have invalid values and setting them to one of the available radio-button choices will result in reasonable behaviour.

Wow, this answer is solid gold. Thank you!

Update:
Adjusting settings in the control panel fixes all issues.
Restarting the Lisa and keeping it plugged in the power outlet keeps the settings I chose.
But if I unplug the power and then boot it up again, we are back to "bad" settings (e/g/ no cursor speed is selected, hence the cursor blinks rapidly, the lisa prints multiple symbols per keystroke, double-click does not work).

I did the same test with the good I/O board:
If I unplug the power and then boot it up again, it forgets the last settings I chose, but it reverts to "good" settings (cursor blink speed of "2").

So why this different behavior on the good and bad I/O board?
Is there some Prom chip on the I/O board that keeps "good default settings"?
At positions 1B and 2B I see ICs "SCM21C14E" (1K x 4 CMOS Static Low Power Ram). I assume one of them is the PRAM that stores the control panel settings until power is removed?

If I can't fix the I/O board , perhaps using a MacWorks Plus II board would fix it, as it has "256 bytes of non-volatile EEPROM that retains Macintosh PRAM settings while Lisa is unplugged".

[sigma7]

if I unplug the power and then boot it up again, we are back to "bad" settings (e/g/ no cursor speed is selected, hence the cursor blinks rapidly, the lisa prints multiple symbols per keystroke, double-click does not work).

So why this different behavior on the good and bad I/O board?
Is there some Prom chip on the I/O board that keeps "good default settings"?
At positions 1B and 2B I see ICs "SCM21C14E" (1K x 4 CMOS Static Low Power Ram). I assume one of them is the PRAM that stores the control panel settings until power is removed?

If I can't fix the I/O board , perhaps using a MacWorks Plus II board would fix it, as it has "256 bytes of non-volatile EEPROM that retains Macintosh PRAM settings while Lisa is unplugged".

Those SRAM chips are 4 bits wide - each one is half of each byte. 64 bytes of the total are reserved for use as PRAM (the early Macs have 256 bytes, so there are a some settings the stock Lisa doesn't preserve), the rest is designated for the FDC.

IIRC, at startup the MacWorks software checks the PRAM for validity, and initializes it to some reasonable values iif it seems corrupted/invalid.

Since the test you performed implies that the board can store the values correctly, the fact that when power is lost the PRAM of one of the boards defaults to values that are considered valid but are unusable is not really a fault of the hardware -- the software should be better at recognizing this condition and then setting reasonable values.

I suggest a work-around may be as simple as re-arranging/swapping RAM chips to find a combination that is recognized as having invalid power-up values.

You may find that exchanging the positions of the two SRAM chips is sufficient, or you could swap one chip between two boards.

Note that the static RAM chips are static sensitive (whereas most chips in the Lisa are not), so take care when handling them.

Yes, MW+II is likely better at keeping PRAM settings.

The revisions to the 1/5 I/O Board included some changes to the static RAM circuitry that appears to have been intended to address corruption of the data when the Lisa shut down. I think these revisions are unlikely to affect the results when power is lost completely and may not apply to the 2/10, but comparing the revisions of the I/O Boards that behave differently may reveal something further.

[TorZidan]

Thanks for the valuable suggestions.

I reached out to John at VintageMicros to buy a Lisa PFG board. Hopefully it will sold my issue.

I chose to not mess up with the  sram chips for now, as they are soldered onto the i/o board. I doubt that I can remove them without having to cut some of the pins, which will render them useless...

[sigma7]

I reached out to John at VintageMicros to buy a Lisa PFG board.

To be clear, to use the PFG features, you'll need to use MacWorks Plus II -- the manual and disk images are available here: MacWorks Plus II

MW+II is intended/tested for use with Systems 6.0.3 to 6.0.8, and varieties of System 7 (although some System 7 configurations require more than 2MB of RAM to be useable).

MacWorks XL 3.0 on the other hand, is only compatible with the earliest Mac software, System 1.0 to System 3 IIRC.