• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Indiana Warmech and The Last Crusade, Or, Obtaining My Final Grail Item: A Lisa 2

warmech

Well-known member
I was cruising Facebook marketplace a couple weekends ago and came across something I never thought I'd see an ad for on there: a Lisa 2. The asking price was reasonable-ish, but once I saw the description and photos of the battery area, my heart sank a bit. "Used to power on, but now no longer does..." accompanying the telltale blue-green hue of corrosive battery damage. I decided to ask about it and see where things went. No mouse or keyboard were included (bummer), but the seller was more than happy to send additional photos of the corrosion. I decided to post a price check thread on here to solicit opinions about what in the heck something in this state may go for and got a handful of extremely helpful responses - the $500-1000 range seemed fair and on-point. I asked the gentleman if he would be willing to meet up for me to take a look at it in person, so we setup a time to meet in front of the police station down the road from him.

The next Wednesday afternoon rolls around and my wife and I head across town (more like three counties - welcome to DFW) to take a look. The guy was really, really nice and we struck up a conversation about the Lisa; it belonged to his great uncle (who bought it new, if I recall) and had sat in his house for the last few decades untouched. His great uncle came across it and gave it to him to sell for money for school, figuring it was probably worth something. The guy had already sold the mouse, keyboard, and Profile on their own on eBay and made decent money off of them - unfortunate for me, but I was genuinely glad he'd managed to get good money for them. He knew the Lisa would be difficult to ship (and was dead as a doornail) so he wanted to try to hand it off locally rather than take a chance on shipping. I took a look at the board and, sure enough, it'd been hit by the usual cause of death via battery leakage. Fortunately, it looks manageable and to have only hit replaceable components. The traces are a mess and will take extensive work to repair but the COP421 looks completely clean, so my fears of that thing taking a hit were relieved. By far, the most difficult thing to replace on this is going to be the card edge connector (more on that later). We talked for a few and I threw out a figure that we negotiated on and eventually came to an agreement: $600. With that, and after over a decade of waiting, I walked away with my very own Lisa. :)

On to the state of this thing. The I/O card edge is going to need some work; thankfully, I know just the guy to re-pad it. It doesn't look like any of the connectors have completely gone, but a fiberglass pencil isn't gonna save it either. The card edge connector on the main board, however, will require replacement - its pins are absolutely ruined where there was corrosion (to the point of disintegration) and I have a replacement on the way. The resistor packs will need to be replaced and some traces cleaned up, but it otherwise looks okay. As for the I/O card itself... oof. It's gonna need some new ICs here and there (the 1488 and 1489 are absolutely toast) and a ton of traces repaired, but neither are beyond my skill or patience. The CPU and RAM cards look fine and dandy, so I'm not really worried about them. The PSU is... shrug? It had a blown fuse (which I have replaced), but I have yet to diagnose further. If the soft-power traces are toast, though, will it even power on? I have yet to attach power to it (nor do I have any intention to yet), but it no longer powers up according the the seller so we'll see if it was, in fact, just the fuse. I'll certainly make sure the transistors are tested and that there are no visible component failures, as well as replace the caps and deoxit the pots, but I need to read up on how to test the PSU. Any Lisa owners out there have any advice on how to proceed? Not just with the PSU, but with anything and everything. I'm no stranger to working on PCBs, but this is probably the only Lisa I'll ever come across and I'd rather like to not do any further damage to it.

I don't have a keyboard but, if I can get this thing to come back to life, I'm willing to drop the cash on one. Since a Mac mouse can be used, I really, really don't care to spend three-figures on a Lisa mouse. Wish me luck folks... it's gonna be a bumpy ride.

(Also, how does one remove the cards installed in the expansion sockets? I have no idea how they're supposed to come out of there, lol.)
 

Attachments

  • IMG_3954.jpg
    IMG_3954.jpg
    941.1 KB · Views: 50
  • IMG_3956.jpg
    IMG_3956.jpg
    1.4 MB · Views: 45
  • IMG_3957.jpg
    IMG_3957.jpg
    1,017.1 KB · Views: 45
  • IMG_3958.jpg
    IMG_3958.jpg
    1.1 MB · Views: 45
  • IMG_3962.jpg
    IMG_3962.jpg
    1.1 MB · Views: 48
  • IMG_3963.jpg
    IMG_3963.jpg
    1 MB · Views: 51

stepleton

Well-known member
how does one remove the cards installed in the expansion sockets

Pull on the metal tab that protrudes from the yellow thing. Pull all the way out, then twist to "unclamp" the expansion socket.

Enjoy --- one of many over-engineered attributes of the Apple Lisa.

Good luck with the repair!
 

Byrd

Well-known member
Thanks for posting your adventures - with some gentle cleaning and lots of vinegar (+ aforementioned fibreglass pencil), the damage to the PCBs might not seem as bad as thought (for example 1488 and 1499 should be salvagable). I also find working with corroded solder really difficult. The PSU should be OK presuming no corrosion once recapped.

Have fun! :)
 

LCARS

Well-known member
That is a great find and even better that after a decade of waiting you found the machine with a good story to go with it. In 2007 I found the first piece of my collection in a similar way (Performa 550) and that memory makes the whole package happier than an eBay transaction.

Yikes, these clock batteries have done more harm...
 

warmech

Well-known member
Pull on the metal tab that protrudes from the yellow thing. Pull all the way out, then twist to "unclamp" the expansion socket.

Enjoy --- one of many over-engineered attributes of the Apple Lisa.

Good luck with the repair!
Are you freakin' kidding me? Really, Apple? Lol, what an oddball approach - thank you for the explanation! The first slot's tab is broken, so I didn't want to risk the populated one without asking.

Thanks for posting your adventures - with some gentle cleaning and lots of vinegar (+ aforementioned fibreglass pencil), the damage to the PCBs might not seem as bad as thought (for example 1488 and 1499 should be salvagable). I also find working with corroded solder really difficult. The PSU should be OK presuming no corrosion once recapped.

Have fun! :)
I'm really hoping it's not quite as rough as it seems once I get everything pulled... and good to know regarding the PSU. Everything on it looks in good shape, so it may just get a reflow in a couple places and that's all it needs.

That is a great find and even better that after a decade of waiting you found the machine with a good story to go with it. In 2007 I found the first piece of my collection in a similar way (Performa 550) and that memory makes the whole package happier than an eBay transaction.

Yikes, these clock batteries have done more harm...
Agreed - most of my favorite pieces are as much because of the circumstances I got them by as for what the machines themselves are. And yeah, hard to think about how many Lisas have been thrown in a landfill due to those stupid batteries... :(
 

bibilit

Well-known member
Congrats, i have worked in a pair at some point, more or less the same kind of corrosion.
The PSU should not be an issue.
the issue here will be all the broken traces you will have to deal with.
The bottom board ( the Logic Board IIRC ) will probably be the worst area, as the acid from the battery will directly drop and remain for a long time.
wish you luck.
 

cheesestraws

Well-known member
Agreeing with all the above: it'll take patience but these machines are very repairable, being chunky enough to reasonably easily see what's going on. Even I'd attempt it, and I have the fine motor control of The Hulk.

Also, if you haven't already found it, lisalist2 is a great resource and has people on it who are very very knowledgeable about these machines.
 

warmech

Well-known member
Alrighty - update time!

First, I have a keyboard on the way, so yay!

Second, I've had a chance to clean this thing up quite a bit. the I/O board pins are fine and a fiberglass brush did indeed take care of them, much to my surprise. There are some broken traces coming off of them, but none of the pins are lifted of the board which means I will not have to re-pad them. There was a decent amount of corrosion elsewhere though, and that has been largely mitigated. I've still got a bit to go, but traces have been taken down to copper where the corrosion was the worst and I've verified that most of the traces are still intact. Some are broken, though, and will require bodges; that's more than doable, though. I'm going to have to replace most of the ICs in the area around the batteries, but I have them on-hand and ready to go into sockets when the time comes. The COP421 has been sealed up in an anti-static case, so it's safe and sound for a later time, lol.

The motherboard was in better shape than I was anticipating. I've removed the large connector and cleaned either side of the board where it was. Thankfully, only three pads came off and they appear to be easily bodge-able to their respective destinations. The EMI filters were absolutely toast, though, along with the two serial ports - they are all being replaced. The EMI filters, for those who read this, are still available at Digi-Key, but boy oh boy are they expensive now ($6.24 apiece)! Also, for those curious, the motherboard screws on a Lisa 2 are #4-40, in case some of yours were as corroded and stripped out as mine were.

The PSU is the major stumbling block at the moment, more to come on that...
 

Attachments

  • IMG_3998.jpg
    IMG_3998.jpg
    1.1 MB · Views: 32
  • IMG_3999.jpg
    IMG_3999.jpg
    947.4 KB · Views: 32
  • IMG_4001.jpg
    IMG_4001.jpg
    1.1 MB · Views: 33

warmech

Well-known member
As for the PSU, it makes the familiar chirping when attempting to power on the Lisa (the I/O board is removed for cleaning, but I wanted to verify the machine would even power on first), so I've been going over everything I can think to. Using the Lisa 2 restoration thread by rdmark over at TinkerDifferent as a guide, the six main transistors all seem to check out fine; the only issue I can think of is that Q1 returns .462v when testing base to collector and .525v when testing base to emitter. I'm not familiar enough with transistor testing to know if those are within spec, whereas all the rest of the 3904s test in the .6v range and look completely fine. Outside of that, I've located what I think to be a short, but I'm a bit confused by it (mostly due to my not being very knowledgable about transformers). In the attached image, the four transformer pins all have continuity, which is slightly alarming to me as it means that all three pins boxed in red in the top-left also all have continuity. As those three pins go to CR17, I'm pretty sure that's no good, lol. As for the lower left-hand box, those four pins also have continuity between Q1's base pin and the trace to its left. Is this a sign of a short within the transformer? With the transformer removed these continuities go away, so I know it's local to the transformer. Any thoughts or suggestions? Additionally, I'll be recapping the PSU this evening as well; the caps are all 40-ish years old and I assume surely can't be in spec anymore.
 

Attachments

  • IMG_3995.jpg
    IMG_3995.jpg
    1.3 MB · Views: 35

stepleton

Well-known member
all three pins boxed in red in the top-left also all have continuity. As those three pins go to CR17, I'm pretty sure that's no good, lol

Let's look at the schematic and see where the pins of CR17 (a weirdo "programmable unijunction transistor") go.

Pin 1: straight to primary-side ground.

Pin 2: this is the gate of the PUT; I don't really know for these devices, but for regular BJTs, can't current always flow through the gate and out the emitter? If that applies here too, then depending on how you've hooked up your continuity tester, this pin could look like a wire that connects to primary side ground.

Pin 3: up to the base of Q1, then (depending on how you've hooked up your continuity tester) through CR24, CR25, CR26, then entering the transformer through pin 27, then four turns to come out on transformer pin 22, then to primary side ground.

It may seem odd for CR17 to connect to ground on all sides, but it looks to me like it spends a lot of its life dealing with AC, and AC continuity and DC continuity are different things.
 

warmech

Well-known member
Let's look at the schematic and see where the pins of CR17 (a weirdo "programmable unijunction transistor") go.

Pin 1: straight to primary-side ground.

Pin 2: this is the gate of the PUT; I don't really know for these devices, but for regular BJTs, can't current always flow through the gate and out the emitter? If that applies here too, then depending on how you've hooked up your continuity tester, this pin could look like a wire that connects to primary side ground.

Pin 3: up to the base of Q1, then (depending on how you've hooked up your continuity tester) through CR24, CR25, CR26, then entering the transformer through pin 27, then four turns to come out on transformer pin 22, then to primary side ground.

It may seem odd for CR17 to connect to ground on all sides, but it looks to me like it spends a lot of its life dealing with AC, and AC continuity and DC continuity are different things.
Dang - thanks for explaining that! PSUs aren't exactly my strong suit, so I'm grateful to anyone who can help me understand things a bit better.

Small update of sorts as well:

I recapped the PSU and checked the six main transistors. Mostly satisfied with the results, I checked everything for glaring issues, dropped the PSU back in its cage, and fired everything up... to hear the CRT coming to life and see the power indicator light on the front button light up! For about 5-10 seconds, after which everything died again and I could smell the familiar smell of dead caps. Sure enough, the fuse had blown again and caps C18 and C21 had very puffy tops (see below); oddly, C19 is part of the same section, but looks fine. I'm looking at the schematic for this section and it looks like they're part of the +5v section; they're seated in the correct polarity, so I'm a bit unsure what would cause that.

Back to the drawing board for the moment, I suppose. Anyone have any ideas on how to approach this, lol?
 

Attachments

  • IMG_4026.jpg
    IMG_4026.jpg
    798.5 KB · Views: 20
  • Screen Shot 2022-06-06 at 7.17.55 PM.png
    Screen Shot 2022-06-06 at 7.17.55 PM.png
    550.5 KB · Views: 21

warmech

Well-known member
Minor follow-up.

Been in touch with a fellow on Facebook who's walked me through building a PSU bench test rig. I'm having Carpal Tunnel surgery tomorrow (yay, finally getting feeling back in my right hand!), but Friday or Saturday I'll get it assembled and have an external means of testing the PSU. The good news is that the rest of the unit seems to be fine; I pulled some logic chips off the CPU board and they tested fine, so overvoltage protection looks like it did its job.
 

stepleton

Well-known member
A test rig is a good idea. You can make a very basic one with power resistors as long as you have some way of dealing with the considerable amount of heat that they will have to dissipate. Behold my one-of-a-kind Apple Lisa 0 (which I only run for about a minute or so):


The PSU expects each rail to have a current demand within a certain range --- if you're out of that range, the power supply will shut down. PDF page 259 of the hardware manual describes the range of acceptable current draws for each rail: you can use Ohm's law to compute the resistor you'll need on each rail to simulate that draw. For the +5V rail, draws between 4A and 8A are acceptable, so to make the math easy, let's choose 5A.

R = V / I = 5V / 5A = 1 ohm is the resistor size you'll need. But take care of the power it requires to handle 5A:

P = I * I * R = 5A * 5A * 1 ohm = 25W.

So you need a resistor that can handle 25W. That's a big resistor! (It turns out to be the largest one you'll need for this job.) It'll get HOT! Take care.

"Electronic loads" are a more expensive option that may have longer endurance. You can spend thousands on these, or you can go the Aliexpress route and spend dozens, which I've done for other applications with some success.



The capacitors that blew up are bulk (smoothing) capacitors on the secondary side --- that is, the side of the PSU's transformer that's connected to the computer, not to the wall. As you note, they're on the +5V rail. It's a bit of a mystery to me as to why they're blowing up, but I'm not exactly a PSU expert. Testing the voltages you see on the rails when the PSU is powering a dummy load will be instructive.


One last warning: be careful around power supplies. You probably know to do this instinctively, but don't guess: please read what you can about the risks of working on these things. Know why you have to be careful about using an oscilloscope to probe around inside (even if you don't have one, it's good to know). Know where the rectified mains are in the PSU and how to avoid them --- and what makes them especially worth avoiding. Understand how the capacitors can hold onto a charge, how to test for a charge, and how to discharge them. And more... (Thanks for reading this warning!)
 
Last edited:

warmech

Well-known member
A test rig is a good idea. You can make a very basic one with power resistors as long as you have some way of dealing with the considerable amount of heat that they will have to dissipate. Behold my one-of-a-kind Apple Lisa 0 (which I only run for about a minute or so):


The PSU expects each rail to have a current demand within a certain range --- if you're out of that range, the power supply will shut down. PDF page 259 of the hardware manual describes the range of acceptable current draws for each rail: you can use Ohm's law to compute the resistor you'll need on each rail to simulate that draw. For the +5V rail, draws between 4A and 8A are acceptable, so to make the math easy, let's choose 5A.

R = V / I = 5V / 5A = 1 ohm is the resistor size you'll need. But take care of the power it requires to handle 5A:

P = I * I * R = 5A * 5A * 1 ohm = 25W.

So you need a resistor that can handle 25W. That's a big resistor! (It turns out to be the largest one you'll need for this job.) It'll get HOT! Take care.

"Electronic loads" are a more expensive option that may have longer endurance. You can spend thousands on these, or you can go the Aliexpress route and spend dozens, which I've done for other applications with some success.



The capacitors that blew up are bulk (smoothing) capacitors on the secondary side --- that is, the side of the PSU's transformer that's connected to the computer, not to the wall. As you note, they're on the +5V rail. It's a bit of a mystery to me as to why they're blowing up, but I'm not exactly a PSU expert. Testing the voltages you see on the rails when the PSU is powering a dummy load will be instructive.


One last warning: be careful around power supplies. You probably know to do this instinctively, but don't guess: please read what you can about the risks of working on these things. Know why you have to be careful about using an oscilloscope to probe around inside (even if you don't have one, it's good to know). Know where the rectified mains are in the PSU and how to avoid them --- and what makes them especially worth avoiding. Understand how the capacitors can hold onto a charge, how to test for a charge, and how to discharge them. And more... (Thanks for reading this warning!)
I cannot thank you enough for your insight - PSUs are not my specialty either, but this does a fantastic job of helping me understand more about them. I'll follow up with my math and a proposed parts list if you wouldn't mind double-checking my work. I'm recovering from hand surgery at the moment, so give me a few to sit down and figure things out (and please excuse any typos, lol). Also, your Lisa Zero is awesome! :D

Your warning is not lost on me, trust me. After working on repairing CRTs (mostly 15KHz Wells & Gardner and Electrohomes) for the last ~10-12 years, getting bit once is all it took. "Safe" and "cavalier" aren't exactly compatible mindsets, which is why I feel so dumb for powering it all up at once (personal safety aside). I don't appear to have damaged anything on the motherboard side of things by powering on the Lisa, but that PSU isn't going back into it until I've taken much more care to diagnose it.
 

warmech

Well-known member
@stepleton Okay, here's what I've come up with; chosen values reflect a combination of affordability and availability.

VoltageRange (Amps)I (Amps)R (Ohms)P (Watts)Chosen Resistor (Ohms)Chosen Resistor (Amps)
+54-85125125
+120.35-2112121215
+330.3-0.70.447514.527515
-50.0-0.20.1500.55010
-120.1-0.20.16751.927515
+5 (Standby)0.0-0.10.051000.2510010

The Mouser project BOM is available as well. Let me know if this is just off as can be or if it's actually functional. This is my first foray into bench testing a PSU, so please have mercy on my inexperience. :p
 

stepleton

Well-known member
Looks good to me; I think it will do the business! The only bug I spot is that the current range for the -12V rail is 0.01A to 0.2A. Your selection is within range regardless.

With your loads in order, interconnect is next. You've calculated that the +5V and +12V rails draw 5A and 1A respectively, so it will be worthwhile to use stout wires for those rails. Then there's the matter of the connector --- I think it's a good idea to purchase this rather than try to jury rig something that clips onto the power supply PCB's fingers. Here's what I used on mine --- was it that pricy back then? Sheesh.

Refer to the schematic to figure out how to wire it up. Note how one side of the connector uses numbers to identify the pins while the other side uses letters: refer to the circuit board itself to see which is which.

Once you're all set, you'll need to energise the ON/OFF pin (pin X) in order for the PSU to believe that it's connected to a Lisa that is powered on. Shorting that pin to the +5V Standby rail will do the trick, and that's pretty easy since it's right on the opposite side of the connector. My slightly fancier option makes use of a crocodile lead.

Finally: when you've completed the dummy load and after you've double and triple-checked everything, you'll plug in the mains and find that it doesn't work. Nothing happens. This will be because you don't have the rear panel pushing in the panel interlock microswitch on the back of the power supply. Defeat the interlock by wedging a little piece of cardboard or something in the microswitch access hole.


Once you're cooking, you'll literally be cooking because the three positive rails will collectively dissipate over 50 watts. A minute with this thing plugged in will be about as much as you're going to want to do unless you work out some kind of heat dissipation strategy, and (as suggested) you'll want to be careful not to touch the resistors for a while afterwards. It would be nice to stress the power supply for longer, but this should be enough time to take some measurements...
 
Last edited:

warmech

Well-known member
Looks good to me; I think it will do the business! The only bug I spot is that the current range for the -12V rail is 0.01A to 0.2A. Your selection is within range regardless.

With your loads in order, interconnect is next. You've calculated that the +5V and +12V rails draw 5A and 1A respectively, so it will be worthwhile to use sturdy wires for those rails. Then there's the matter of the connector --- I think it's a good idea to purchase this rather than try to jury rig something that clips onto the fingers. Here's what I used on mine --- was it that pricy back then? Sheesh.

Refer to the schematic to figure out how to wire it up. Note how one side of the connector uses numbers to identify the pins while the other side uses letters: refer to the circuit board itself to see which is which.

Once you're all set, you'll need to energise the ON/OFF pin (pin X) in order for the PSU to believe that it's connected to a Lisa that is powered on. Shorting that pin to the +5V Standby rail will do the trick, and that's pretty easy since it's right on the opposite side of the connector. My slightly fancier option makes use of a crocodile lead.

Finally: when you've completed the dummy load and after you've double and triple-checked everything, you'll plug in the mains and find that it doesn't work. Nothing happens. This will be because you don't have the rear panel pushing in the panel interlock microswitch on the back of the power supply. Defeat the interlock by wedging a little piece of cardboard or something in the microswitch access hole.
Excellent! And good catch on the -12v rail - exactly why I wanted another set of eyes looking at my work, lol.

As for the connector, I bought a collection of card edge connectors off a friend who picked up a ton while they were at a surplus component market in California a few years ago and, boy howdy, I'm sure glad I did. I think I paid less than a dollar per connector at the time. I needed a variety for building wiring harness adapters for some more uncommon arcade games and just told him to grab some common sizes; sure enough, I had one 22/44 left in the bag. I cannot believe how expensive those connectors are nowadays... It's about the same price at Mouser for the EDAC ones.

As for the wires, would you say 18ga would suffice, or would 16ga be better? I think I have both, but may just have 18ga.

Roger that on the ON/OFF - is that a static connection, or a momentary one? Also, that interlock has got me every time; thanks for reminding me well in advance, lol.
 

stepleton

Well-known member
As for the wires, would you say 18ga would suffice, or would 16ga be better? I think I have both, but may just have 18ga.
Hmm... I went bigger I think, but don't know quite what I have here --- it's just some stranded wire I had handy. You can always double up on the two 1A+ lines.

Roger that on the ON/OFF - is that a static connection, or a momentary one?
You'll need to have a constant +5V on the ON/OFF pin in order for the main (that is, non-standby) PSU to operate. See hardware manual PDF page 263, "On-off Control".
 

kkritsilas

Well-known member
I would suggest that the 5V (main) resistor be higher power rating than the 25W you are showing. Thw 25W power rating is a maximum, and running resistors at their maximum power rating is a little questionable. At the very least, it will run very hot, at worst, if the 5V is a little high (anything higher than 5.0V), and you are risking a burn up/meltdown or the resistor if you apply power for anything but a very brief amount of time. I would also suggest that you up the power rating on the +12V also be raised, and would think that the resistors for the -12V, -5V for those rails can be safely reduced to 5W, or even lower. Does the Lisa actually have a +33V rail, or is it a 3.3V rail.

I am making the assumption that the Chosen Resistor (Amps) is actually in Watts, not Amps.
 
Last edited:
Top