I won a local auction for an Apple Lisa!!!

[berskyboy]

thanks so much.  I've put in an order at DijiKey

here's a nice website / wiki with info on each part.
https://caps.wiki/wiki/Apple_Lisa_1.8A_Power_Supply_(Datapower)

Thanks for the tips on how to test it.  High Voltages scare me, but that's a good thing, just be careful.

[berskyboy]

With RIFAs, you just can't know in my experience: visible distress does correlate with propensity to blow up, but they do just seem liable to go at some point regardless. If I chose to power up a machine with RIFAs in it, I'd just be prepared for them to blow at any time. They probably won't permanently damage anything if they do, but they'll make a stink and a mess. These days I just replace them if I find them.

Note that in the Lisa, the computer doesn't have to be on for at least some RIFAs to have a voltage across their leads. If the Lisa is just plugged in, they are at risk of popping.

There are ways to test PSUs. For the Lisa, I have a harness with a brace of power resistors that I plug into the power connector, one resistor for each of the voltage rails, plus the ability to short the pins you need to short in order to tell the main PSU to turn on. (Note that there is also a standby power supply inside the PSU: it always provides +5V whenever the Lisa is plugged in.) The values I chose for the resistors are worked out from the minimum and maximum current draws listed in the Lisa Hardware Manual, but while the "1.2 amp" Apple-made Lisa 2/5 PSU needs some current draw to operate, I don't know if this is the case for the 1.8 amp Datapower-made 2/10 PSU that you have (maybe it can just run with no load; I've never tried it). I can share the values I use if you're interested. Anyway, once the PSU is up and running on the load, I check voltages. If they look good, then I feel good about putting the PSU back in service.

Naturally you must always be careful when working with power supplies, and note that in the 2/5 PSU at least, some of the heat sinks have quite high voltages on them. One contact can give you a good zap!

(Apologies that I can't check the capacitor choices you've listed right now!)

I just replaced the RIFA Caps and the two large ones.  To test it should I proceed to do what this video said?

https://www.youtube.com/watch?v=GGUZQWQytwY

Or is there another simpler way to test it?  And what am I looking for.  For the video it looks like measuring voltages with my multimeter is the method. 

[sigma7]

To test it should I proceed to...

Yes, you will use a DC voltmeter to check the output voltages are within range.

(If there is a more subtle problem, an AC voltmeter may (or may not) reveal excess ripple on one of the DC voltages, but this is easier seen with an oscilloscope... it is unlikely you'll need to go this far, and not necessary to verify the PSU is safe to install in the Lisa)

The technique in the video looks fine, but if you can afford to buy a card-edge connector instead of soldering to the PSU circuit board, then it will be more convenient; eg. you will be able to test a PSU while still in the case etc.

These threads/posts may help you understand what is required:

https://lisalist2.com/index.php/topic,789.msg5772.html#msg5772

https://lisalist2.com/index.php/topic,504.msg3532.html#msg3532

Note in particular that:

• you must connect the +5 Sense to +5V, or the output voltages will rise uncontrolled.
• the interlock micro-switch must be activated

Even with the remote voltage sense, there is still a voltage drop in the card cage, so you may need to adjust the voltage to a bit more than 5V at the PSU to obtain 5V on the CPU and I/O Boards. The amount of memory, expansion cards, drives, etc. connected will affect the amount of voltage drop.

The normal supply voltage tolerance for TTL parts is 5%, so "+5" should be between 4.75V and 5.25V

[berskyboy]

Hi I connected up the four lights like this YouTube Video: https://www.youtube.com/watch?v=GGUZQWQytwY

1. no lights came on.  I measured the 12V and 5V rails.  They registered then drop to zero
2. I took off the lights one by one, seeing if one would light up.  Leaving the 5V light, still didn't light up
3. I took off all the lights, no sounds, the 12, 5V rails do that same.  They startup then go to zero
4. I measured the ohms at the 5V and 12V rails they start at 0 then climb up.

Here's a video of the board. 
https://youtube.com/shorts/QDEb45ZxE3s

Images
https://www.dropbox.com/scl/fi/yzhxi7d9vpqcol9al0ukk/Lisa-PSU-Jumpers1.HEIC?rlkey=8tx6ekyhw9iecmeiflzy10nr5&st=ewgjzgu5&dl=0

https://www.dropbox.com/scl/fi/nf89dl77g8qk0zqst7ffi/Lisa-PSU-Jumper2.jpeg?rlkey=r2ivchebqohl8noumxsrp0zrc&st=ogvu8e1p&dl=0

I'm assuming I need to replace the small 5 capacitors and maybe other components.  AI said check R29 but I can find that dial.

[stepleton]

One theory is that these bulbs are drawing too much current at startup and the PSU is shutting itself down in order to protect itself from what it thinks is a shorted output. I wonder if the bulbs in the video are subtly different somehow.

When you say the other rails "registered", do you mean they briefly had the correct voltages before that voltage decayed?

[berskyboy]

One theory is that these bulbs are drawing too much current at startup and the PSU is shutting itself down in order to protect itself from what it thinks is a shorted output. I wonder if the bulbs in the video are subtly different somehow.

When you say the other rails "registered", do you mean they briefly had the correct voltages before that voltage decayed?

Yes the rails had the correct voltage then down to zero.  The bulbs are 13v, I bought from a car shop, they are standard 12-Volt miniature automotive bulbs

[sigma7]

The bulbs are 13v, I bought from a car shop, they are standard 12-Volt miniature automotive bulbs

Note that at timestamp 3:16 in the european (I think) video he says the two bulbs in series are 24V each, while the other two are 12V.

If you are using all 12V bulbs, you'd want at least 3 in series for the 33V supply. However, I don't think the 33V supply absolutely has to have a load to start up the PSU, so I'd first try just disconnecting those to see if the PSU will start up.

More importantly, the video shows the PSU has one white jumper installed (on the top side) for 220V, does yours have the 3 jumpers installed for 110V?

edit: I see in your video of the top side that you do have the jumpers installed for the 110V configuration.

edit 2: at timestamp 0:40 in your video I see a loose lockwasher... remove it as you don't want it relocating and causing a short circuit as that may damage multiple components.

edit 3: are the symptoms still repeatable? If not, check if the fuse has blown

edit 4: R29 is in the video brightness circuit, this is not connected to the power supply circuitry itself so doesn't affect it powering up outside the Lisa.

Re: measuring the resistance at the output of a PSU:

• If the PSU is powered up, or has significant charge on the filter capacitors from being powered up recently, the ohmmeter may be damaged.
• It is normal for the resistance to start at 0 and rise, this is due to the current from the ohmmeter quickly then slowly charging the filter cap you have connected to.

[berskyboy]

I have removed that lock washer.  (Thanks for that)

I remove the jumper wires tested, 0V
then I re-added them and getting 5V then goes to zero, and 11.5 V then to zero. 
it makes a quiet shhhhhh sound (like when you would tune a radio), but very faint, and then off.

So do I need to start replacing components?  It seems it does produce power, but then shuts off.

This is my DijiKey List.  (Note I already replaced the RIFA caps and the two large Capacitors)

https://www.digikey.ca/short/nqwvwb17

[sigma7]

do I need to start replacing components?  It seems it does produce power, but then shuts off.

I recommend not blindly replacing components, at least not yet. It is very likely that after replacing many components it would still not work.

I'll look over your photos to see if the way it was wired has any obvious problems.

Given that it seems to almost work, it is probably possible to narrow down the problem substantially, but it will take some time/patience and back and forth measuring things.

edit: what is the wattage or current rating of the 12V bulbs you are using? (or their part number if the wattage isn't specified)

edit 2: the bottom of the board shows a couple of dark patches (scorch marks), what are the component designations that correspond to these? For example, it looks like R3 beside the white jumpers is one of the components that is getting hot (which is normal, but maybe not the others).

[berskyboy]

The bulbs are 4057LL.Bp2 LAMP Bulb

https://www.canadiantire.ca/en/pdp/4057-sylvania-long-life-mini-bulb-2-pk-0202857p.0202857.html?utm_content=shopping&gclsrc=aw.ds&gad_source=1&gad_campaignid=24075369818&gbraid=0AAAAADojZpheGGp8Mr_9WqX3o-9pEnlse&gclid=Cj0KCQjwkt_UBhDMARIsALpnOAxAAhAfH3vqS6Pp_CuVl7QJVEUvGO24I8ThIRwVkpKtLZZqC1Ge5EQaAlpCEALw_wcB#store=397

12.8V 14V, but maybe the WATTS are wrong. LOL

Also more photos of the bottom and the top

https://www.dropbox.com/scl/fi/8odk8cynrm5y4n206e9s6/Lisa-PSU1.jpeg?rlkey=e7wxzctih90zpu8tfi8b8m7vn&st=lsr47ej6&dl=0

https://www.dropbox.com/scl/fi/1pee4kbx8o6ngpqj72rrr/Lisa-PSU3.jpeg?rlkey=5tjltsf7o0nbmp3f5drsvw991&st=ch6ec0we&dl=0

https://www.dropbox.com/scl/fi/bh635l9pnbfii2qoyahym/Lisa-PUS2.jpeg?rlkey=a1zrhnlz8vfoho1xpzyvnk1yw&st=nd3ey7ft&dl=0

AI told me to go with resisters vs bulbs,
4. High-Wattage Dummy Load Resistors (To replace your bulbs)

What they do: To permanently graduate from the light bulb trap—which keeps triggering your over-current latch due to cold inrush current—you need true resistive loads.

What to order:One 2 Ohm, 25-Watt or 50-Watt aluminum-housed chassis mount resistor (for your +5V rail)One 12 Ohm, 25-Watt aluminum-housed chassis mount resistor (for your +12V rail)

[sigma7]

The bulbs are 4057
12.8V 14V, but maybe the WATTS are wrong. LOL

Those are dual filament bulbs for tail lights. They have a dim filament for the parking/running lights and a bright filament for the brake light.

The two power ratings of 27W and 6.7W correspond to the bright and dim filaments. Using the common formula P = V*V/R, we can calculate the resistance to see how much current they will draw when connected to a different voltage such as 5V.
The bright filament is 27W at 12.8V -> 6 ohms
The dim filament is 6.7W at 14V -> 29 ohms

So connecting the bright filament to 5V, you'll draw about 0.8 Amps, which is not very much for that power supply.

I'm supposing that if you have/had connected the dim filaments, that wasn't enough load for the PSU to start up. You can use your ohmmeter to check which filament is connected, or you may be able to tell by looking through the glass. The dim filament is a longer length (more coils) of thinner wire.

AI told me to go with resisters vs bulbs,
4. High-Wattage Dummy Load Resistors (To replace your bulbs)

What they do: To permanently graduate from the light bulb trap—which keeps triggering your over-current latch due to cold inrush current—you need true resistive loads.

What to order:One 2 Ohm, 25-Watt or 50-Watt aluminum-housed chassis mount resistor (for your +5V rail)One 12 Ohm, 25-Watt aluminum-housed chassis mount resistor (for your +12V rail)

Thanks for pointing out that is suggested by AI... it is important that we keep tabs on the sort of information it provides!

For this particular application, the inrush current is probably not an issue, as 1980's digital electronics will also have a high inrush current.

The recommendation of single aluminum power resistors is not cost-conscious. Those are appropriate when saving space, which isn't normally an issue for a test jig. Cost goes up rapidly with power capability, and goes down with quantity, so 5 5 watt resistors is often much cheaper than 1 25 watt resistor. In addition, by spreading the load among multiple resistors, they probably won't be as hot.

As for the values suggested, 2 ohms on a 5V supply will draw 2.5 Amps and dissipate 12.5 watts. Doubling the power capability to 25W is derating the resistor to extend its life, but for the few minutes it will be in use, that's not necessary here. Increasing further to 50W is multiplying the cost substantially.

The 12 ohm load on 12V would draw 1A and 12W.


Suggestion: double check that you have the bright filaments connected as loads to the +12 and +5 voltages, and see if it behaves any differently.

[sigma7]

I am able to get a Lisa DataPower PSU to start up with a 30 ohm load on +12 and a 9 ohm load on +5, anything less than that shuts down promptly. I only tested one unit - others may be a bit different.

So if using 6 ohm brake light bulbs, one is likely to be sufficient for each of +12 and +5; 29 ohms may work for +12, but not +5.

I suppose one dual filament bulb might be sufficient if the dim filament was connected to +12 and the bright filament connected to +5.

Note that these are light loads compared to a normally configured Lisa, so the voltages output could be outside of normal specifications, but if close-ish, eg. within 10% of nominal, I'd suppose it was safe to install in a real Lisa. (If still in doubt, remove any extra boards/drives so as to not risk more components than necessary.)

With no load on the 33V line, it would likely read somewhat higher than +10%.

Once the Lisa is running, check the +5 voltage on the I/O Board and adjust the PSU to get 5V +/- 5% (4.75 - 5.25V).

IIRC, the DataPower voltage adjustment (R11 on the PSU daughterboard) is counter-clockwise to increase.