Adventure in fixing a Lisa DataPower PSU

[berskyboy]

Excellent. Before you install them, you should check them with the diode test to make sure they look plausibly like transistors. Measure all 6 permutations of red & black to the 3 connections.

I just got the Q1 and Q2 transistors. 

Measurements as follows (I bought 4, 2 spares)

DIODE TEST           
Red to Black            1, 2                  3, 4
Lead 1 to Lead 2      0.537, 0.536   0.530, 0.540
Lead 2 to Lead 1      OL, OL           OL, OL
Lead 1 to case          0.467, 0.471  0.471, 0.462
case to Lead 1          OL, OL          OL, OL
Lead 2 to case          OL, OL          OL, OL
case to Lead 2          OL, OL          OL, OL

I will have time to put them in the PSU tomorrow.

[sigma7]

I just got the Q1 and Q2 transistors. 

Measurements as follows (I bought 4, 2 spares)
...

Awesome, they look like NPN transistors!

[stepleton]

If you have a sensitive scale, then before you install the transistors, you might consider weighing the ones that you have, old and new.

Counterfeit parts are more and more of a problem these days, and I've seen a post or two on other forums complaining of replacement power transistors that weren't what was advertised. When one poster cut open their new transistors' TO-3 cans, what they found was a substantially smaller and flimsier chunk of silicon inside. Now, as manufacturing processes improve, you might expect the silicon to shrink somewhat over a component's production lifetime, but I assume that for a power transistor this is only to a point. A radical die shrink probably isn't so realistic, as I assume you must have some amount of silicon to conduct current.

For the purposes of this project, you might use this information as follows:
- If the PSU still doesn't work (or soon breaks again) and the new transistors were a lot lighter, you might consider the possibility that you got phony parts and try sourcing new replacements from somewhere else.
- Otherwise, if there are problems and the new transistors were roughly the same weight, you might suspect that there is still another fault elsewhere in the PSU.

Specialised equipment like a curve tracer could tell you more about whether the new transistors were likely to behave like the older ones, but I suspect you don't have one of those. (I wish I did!) Lacking that, weight is one of the few comparative measurements that's easy to take.

[berskyboy]

If you have a sensitive scale, then before you install the transistors, you might consider weighing the ones that you have, old and new.

thanks for the tip, yup it's 10.8g same as the original.  I weighted all of them.  So this distributor is reputable.

[berskyboy]

Excellent. Before you install them, you should check them with the diode test to make sure they look plausibly like transistors. Measure all 6 permutations of red & black to the 3 connections.

I've tested Q1 and Q2 on the board

{7-8, 10-11}   0.00   Q2, Q1 e-b (in parallel with pulse transformer)   Ov, 0v. [0, 0 ]
{8-7, 11-10}   0.00   Q2, Q1 b-e (in parallel with pulse transformer)   0v, 0v. [0, 0  ]
{7-9, 10-12}   0.46   Q2, Q1 e-c [ 0.468, 0.455]
{9-7, 12-10}   inf   Q2, Q1 c-e [OL,OL]
{8-9, 11-12}   0.46   Q2, Q1 b-c [0.468, 0.463]
{9-8, 12-11}   inf   Q2, Q1 c-b [OL, OL]

[berskyboy]

So I put the testing edge connector and measured around to 5V getting 5V then drops slowly (not as quick as before all the Q1 and Q2 issue).   The 12V rail 11.5v then dropping slowly.

Here's the two videos

5V Rail
https://www.youtube.com/shorts/F0PeQeJKY8M

12V Rail
https://www.youtube.com/shorts/ScXZ-lT0yqA

Just for a kick this is what Gemini AI said.  (I know you want to keep AI honest)

Your videos provide excellent context, showing the voltage initially jumping up to 5.4V on the 5V rail (0:07) and around 11.5V–11.8V on the 12V rail, followed by a slow, gradual decline (0:07).

The fact that the voltage is dropping slowly (and much slower than it did before you resolved the Q1 and Q2 transistor issues) is a massive clue. This behavior points directly to how the Apple Lisa power supply unit (PSU) interacts with a computer that is stripped down on a test bench.

🔍 Why the Voltages are Slowly Dropping
This slow drop is typically caused by two specific hardware behaviors:
1. Lack of a "Minimum Load" (The Most Likely Culprit)
The Lisa's power supply is a switching power supply.
• Switching power supplies require a minimum electrical current draw (a "load") to stay active and regulate their output.
• If you are testing with only the video board connected—or with the main logic board, card cage, and mechanical Twiggy drives removed—the PSU does not sense enough current draw.
• The Result: The PSU triggers its internal over-voltage or under-load protection and shuts down. The "slow drop" you see on your DMM is actually the large filter capacitors inside the power supply slowly draining their stored energy because nothing is consuming the power (0:07).

[sigma7]

Just for a kick this is what Gemini AI said.  (I know you want to keep AI honest)

Code Select Expand

filter capacitors inside the power supply slowly draining their stored energy because nothing is consuming the power (0:07).

This might be a useful point. See if the voltage measured at the load resistors behaves in the same way (test with red and black at the two ends of one resistor, then again with the other resistor).

[berskyboy]

This might be a useful point. See if the voltage measured at the load resistors behaves in the same way (test with red and black at the two ends of one resistor, then again with the other resistor).
 

I re-seated the edge connector, and put alcohol and tested again.  Success 5.37V and 12.49V and it's NOT dropping!

https://www.youtube.com/shorts/Ptwl7Cb_I84

[sigma7]

Success 5.5V and 12.5V and it's NOT dropping!

Yes, it looks like it is now working... congrats.

To finish:
Adjust R11 so that +5 is between 4.9 and 5.1. It may need to be adjusted again once you measure it in the fully assembled Lisa.
Check the +33, -12 and -5V outputs.
Consider adding a small low speed fan to enhance longevity.

[berskyboy]

To finish:
Adjust R11 so that +5 is between 4.9 and 5.1. It may need to be adjusted again once you measure it in the fully assembled Lisa.
Check the +33, -12 and -5V outputs.

Sounds good will adjust R11, do I need to touch R26? 

And how do I check +33V output, would that be while placed in the Lisa?

[sigma7]

do I need to touch R26? 

And how do I check +33V output

I don't think you need to adjust R26.

If you're going to use a real Widget (the only device which monitors the Power Fail signal AFAIK), then you might need to adjust R26. Perhaps someone else already has a procedure for setting it without a variac?

Check +33, -12 and -5 (those are negative signs) with the same test jig setup you just used to check +12 and +5... ie. don't change anything, just move the red probe.

[berskyboy]

heck +33, -12 and -5 (those are negative signs) with the same test jig setup you just used to check +12 and +5... ie. don't change anything, just move the red probe.

it measures -5V when red is on ground on the 5V rail and -11.62V on the 12V rail.  Where is the 33+V test?

[sigma7]

Where is the 33+V test?

Pin "D" on the card edge connector. If you can't get to the bottom for some reason, you could measure something closely related at L5 or the top lead of CR22.

With the card edge connector test jig, pin "D" is readily accessible.

[sigma7]

heck +33, -12 and -5 (those are negative signs) with the same test jig setup you just used to check +12 and +5... ie. don't change anything, just move the red probe.

it measures -5V when red is on ground on the 5V rail and -11.62V on the 12V rail.  Where is the 33+V test?

With black on ground, and red to the -5 rail the meter should read -5V, and with red to the -12 rail, it should read around -12V.

[berskyboy]

71.2V, that and measured at CR22 also

https://www.youtube.com/shorts/aSk2IGHACGA