• Please review our updated Terms and Rules here

LSI ADM-31 PSU Caps

NeXT

Veteran Member
Joined
Oct 22, 2008
Messages
8,144
Location
Kamloops, BC, Canada
I've been tasked with going over a power supply that has been recapped once but still has issues. I need to verify the values of the capacitors that have already been swapped are correct but there is no schematic or service manual online for the ADM-31 or the Boschert 49-1001 switcher this uses. Would someone be able to open their terminal and take a peek at what the value of all the lytics are?
 
Doubled-over the recap job of the owner and noted the following:

C4, C5: 100mfd200V
C16, C20: 2200mfd25V
C15, C19, C19, C22: 330mfd 63V
C17: 220mfd 63V
C21: 100mfd 100v
C9: 47mfd 10V (Subbed with a 47mfd 16V)
All caps new or original read excellent ESR

R2: Overheated and reading 132.4 ohms. Remaining color bands read "Brown, Orange, unreadable and a gold band (13something ohms, 5%)
TR1: browned to the point the markings are not readable. Reads 50 ohms cold and decreases if heated (should be good)

Need verification on the value of the resistor. There is an MCT2E which I don't have a spare on hand but a 4N33 should be a compatible substitute?
 
Last edited:
Doubled-over the recap job of the owner and noted the following:

C4, C5: 100mfd200V
C16, C20: 2200mfd25V
C15, C19, C19, C22: 330mfd 63V
C17: 220mfd 63V
C21: 100mfd 100v
C9: 47mfd 10V (Subbed with a 47mfd 16V)
All caps new or original read excellent ESR

R2: Overheated and reading 132.4 ohms. Remaining color bands read "Brown, Orange, unreadable and a gold band (13something ohms, 5%)
TR1: browned to the point the markings are not readable. Reads 50 ohms cold and decreases if heated (should be good)

Need verification on the value of the resistor. There is an MCT2E which I don't have a spare on hand but a 4N33 should be a compatible substitute?

I know this thread is really old, but I just got a dead AMD-31 myself. Very frustrating there are STILL no schematics online for the main board or the Boschert power supply. I went over your list and the only real issue I see is C21, which you show at 100mfd @100v. I have 220mfd @63v, exactly like C17. The only other differences are that some of your new caps have higher voltage ratings, which should be fine of course.

If since last May you've managed to find a schematic or even the proper voltage out on the connector, please let me know!
 
I was not able to find schematics. Ultimately the owner of the terminal in the end to replace the power supply entirely with a newer switchmode design from Meanwell. Other than a Horizontal Output Transistor that failed for unrelated reasons he's had no complaints since.
 
Why don't you just copy the schematic out from the board ?

I have done this many times, for the entire 5155 power supply, entire VDU's. It is not that difficult.

In the end, a lot of vintage computer stuff was not well documented, so we have to do it ourselves.

Firstly, copy out the schematics and make neat error free copies. Then study the design and write up the operating theory. Create the topographical component layout with component designators. Then take scope recordings of the important waveforms and put together a "service manual" for the item and put it online. This way it benefits everybody.

We either value these historical items enough to put in the work, or we don't, or they go in the rubbish to be replaced by a modern supply, but I'd hardly call that "vintage computing".

It needs to be done to a good standard, here is the 5155 SMPS manual I created as one example, only about 21 pages required:

www.worldphaco.com/uploads/The_IBM_5155_POWER_SUPPLY.pdf

Unfortunately I don't have or own an ADM supply like this (or I probably would have already created a manual in the absence of one).

I could do it on a special request as a paid job for some particular SMPS, but it would have to be shipped to me in Australia. Ideally though I would like to see others putting in the effort when there is no manufacturer data. It would probably end up as about a 20 page manual.

Another example of boards where there is no data in many cases is motor control boards , the control board in my Sieg C1 mini-lathe was like this, so I completely reverse engineered it and wrote a paper on its interesting operating theory (largely done with OP amps) and it turned into a 46 page manual, but with it, the board now documented, can always be diagnosed and easily repaired. Before this people were making stab in the dark repairs (the usual global re-capping) and mostly failing:

www.worldphaco.com/uploads/THE%20ORIGINAL%20SIEG%20C1%20MINI.pdf

And other recently, the Taylor-Wilson PET printer interface which contained over 25 IC's and again no manufacturer manual or data, requiring complete reverse engineering of the pcb, the schematic from that and creating the operating theory which turned into a 37 page manual:


It all really depends on how much you value and like the item, to put in the required work.

For many people looking for a quick fix, making the manual may not be the answer. But not having it can be equally, if not more, frustrating. A wise individual said to me once: "Never get into bed with an item of equipment without the manual".
 
The bigger issue here is it's not my terminal. I'd do this for my ADDS Consul terminal which itself is its own beast of a PSU, but this one I kinda lack motivation to keep working on this unless someone REALLY needs board photos.
 
Why don't you just copy the schematic out from the board ?

I have done this many times, for the entire 5155 power supply, entire VDU's. It is not that difficult.

In the end, a lot of vintage computer stuff was not well documented, so we have to do it ourselves.

Firstly, copy out the schematics and make neat error free copies. Then study the design and write up the operating theory. Create the topographical component layout with component designators. Then take scope recordings of the important waveforms and put together a "service manual" for the item and put it online. This way it benefits everybody.

We either value these historical items enough to put in the work, or we don't, or they go in the rubbish to be replaced by a modern supply, but I'd hardly call that "vintage computing".

It needs to be done to a good standard, here is the 5155 SMPS manual I created as one example, only about 21 pages required:

www.worldphaco.com/uploads/The_IBM_5155_POWER_SUPPLY.pdf

Unfortunately I don't have or own an ADM supply like this (or I probably would have already created a manual in the absence of one).

I could do it on a special request as a paid job for some particular SMPS, but it would have to be shipped to me in Australia. Ideally though I would like to see others putting in the effort when there is no manufacturer data. It would probably end up as about a 20 page manual.

Another example of boards where there is no data in many cases is motor control boards , the control board in my Sieg C1 mini-lathe was like this, so I completely reverse engineered it and wrote a paper on its interesting operating theory (largely done with OP amps) and it turned into a 46 page manual, but with it, the board now documented, can always be diagnosed and easily repaired. Before this people were making stab in the dark repairs (the usual global re-capping) and mostly failing:

www.worldphaco.com/uploads/THE%20ORIGINAL%20SIEG%20C1%20MINI.pdf

And other recently, the Taylor-Wilson PET printer interface which contained over 25 IC's and again no manufacturer manual or data, requiring complete reverse engineering of the pcb, the schematic from that and creating the operating theory which turned into a 37 page manual:


It all really depends on how much you value and like the item, to put in the required work.

For many people looking for a quick fix, making the manual may not be the answer. But not having it can be equally, if not more, frustrating. A wise individual said to me once: "Never get into bed with an item of equipment without the manual".

I think you're confusing Mechanics with Engineers. I'd suspect most of us have the ability to FOLLOW a schematic in order to test/replace parts, like a mechanic following a service manual. I'd suspect most of us CANNOT turn a board layout into a schematic, much less write up the operating theory. ESPECIALLY when our only source board isn't working, thus making any voltage or waveform readings completely suspect.
 
Last edited:
I was not able to find schematics. Ultimately the owner of the terminal in the end to replace the power supply entirely with a newer switchmode design from Meanwell. Other than a Horizontal Output Transistor that failed for unrelated reasons he's had no complaints since.

Ultimately, I want the original power supply repaired, but in the meantime I'd also like to check the state of the mainboard and CRT. Do you know what Meanwell product he used, or even what the definitive output voltages are of the original PS?
 
I think you're confusing Mechanics with Engineers. I'd suspect most of us have the ability to FOLLOW a schematic in order to test/replace parts, like a mechanic following a service manual. I'd suspect most of us CANNOT turn a board layout into a schematic, much less write up the operating theory. ESPECIALLY when our only source board isn't working, thus making any voltage or waveform readings completely suspect.
It doesn't have to be working to draw out the schematic from the board layout. Indentify the components and follow the tracks. On PSU boards, unlike many multi-layer computer pcb's, its usually fairly easy to see where all the components connect and how they are connected to each other and you can verify that with a meter, its not Rocket science. Once you have the schematic in most cases it is fairly obvious the sort of waveforms you expect to see in certain places when it is working, and if its not working, you simply record the waveforms after it is fixed.

The main point is this:

Without the schematic, it is not possible to logically fault find the circuit, because you are lost in the wilderness.

You might get lucky and come across a damaged part or some component that fails a passive test. Mostly though, especially in SMPS supplies, that don't run, powered tests with the aid of the schematic are required to formulate a diagnosis, check that theory with targeted parts replacements.

The notion of being able to repair something like an SMPS, without its documented design, is too haphazard to be effective, most of the time, unless lady luck is on your side. But you are better to make your own luck.
 
Last edited:
Ultimately, I want the original power supply repaired, but in the meantime I'd also like to check the state of the mainboard and CRT. Do you know what Meanwell product he used, or even what the definitive output voltages are of the original PS?
@z-ddt Would you be able to answer this as you were the one who built the arrangement for the substitute power supply? c:
 
Back
Top