Help Needed - 386 Power Supply Issue

[Brad Cancian]

Hi all - apologies if there's an easy answer somewhere, I am not very knowledgeable on these old computers!

I have my childhood 386 from December 1992 that up until a recent move of home was working fine. The adventure of moving house saw me have the computer in the back of my car over a 2,100km journey (approx 1,300 miles) here in Australia (well packed to as to avoid too much movement, noting that some bumps and vibrations would have been inevitable).

When putting the computer back together to get it working, the computer would not start. I think I have narrowed the issue down to the power supply. The power supply turns on the monitor, but the fan does not spin and there appears to be no power to any other part of the computer (including the hard drive, and thus presumably, the rest of the boards).

I can hear a slight electrical humming from the power supply, and I can see the fan move slightly when I try to turn the power on and back off, so it looks like it is trying to spin and turn over but is not able to do so.

Is the power supply kaput? Or is there something I can do to further fault find including pulling the power supply out and apart? Would something else be causing this? i.e. could this be a power supply that needs load to operate properly, and without this, it won't? I am curious if this might be the case given the power supply is moving the fan slightly on power up and also powering on the monitor?

This is the original power supply from December 1992.

Attached are some quick pictures of the power supply itself, so you can see the basic type (I can't see a specific make or model, so unsure of the details on what this actually is... again, apologies, I am certainly no expert here!).

If I do need to replace the power supply, where might I find an appropriate power supply these days?

Thanks all for your help!

Kindest regards,

Brad

 

[Hugo Holden]

Probably the supply needs recapping by now. It is possible that it could also have other failed components.

These SMPS's were better made back then than some new ones I have seen.

If it was my computer I would want to keep this original supply, so I would re-cap it and repair it if anything else was wrong. But to do these sorts of repairs a technician has to understand the operating principles of switchmode supplies, have good test gear (I use an isolated scope the TEK 222ps for the primary side) and be mindful of the danger of the line voltage side of it, and prepare a dummy load, because the supply won't start & run properly without one. It is not a wonderful idea to use the computer as a dummy load during fault finding, because there could be an over-voltage accidents, but mostly that does not happen, still not ideal.

So repairing SMPS's is not everybody's cup of tea and not all have experience and know how to do it. Many people grab for a generic supply as a replacement.

Somebody will suggest a make and model, ideally it would have the same array of connectors and be a size that will fit well and rated to the same or higher power output.

 

[Brad Cancian]

Many thanks Hugo for the reply!

As it was working perfectly well before the 2,100km move, I do wonder if some simple vibration has done something in terms of breaking some physical components somewhere. I checked the power connectors to the boards / drives and swapped a couple around but same result. As this was my childhood machine (and took some playing to get working properly again a couple of years ago after finding it in my grandparent's shed, where it had sat unused under a blanket for 20 years), I am more than a little bummed.

I would also like to keep the original power supply (assuming it's the power supply that is indeed the issue), but if it's a lost cause then i'd rather find a replacement.

Like yourself, I am also in Australia (Brisbane); not sure who I could turn to locally to have a look at things...?

Alas I have very little (i.e. no) experience with repairing this age and type of electronics.

Thanks again!

Brad C

 

[Hugo Holden]

When you say the supply turns on the monitor,what do you mean that the line voltage simply loops through the connector on the back panel of the power supply ? of course to do that nothing in the supply needs to be working.

The other thing is that if the supply is overloaded say a shorted Tant capacitor on the mobo or in a disk drive, that can also prevent the supply from starting up and with no load on the supply it will automatically shut down due to voltage overshoot.

How exactly did you determine the supply is at fault ?

It probably is the supply at fault, most likely.

One way to conclusively prove it would be to remove it from the computer connect a dummy load to it. Usually it is only necessary to load the +5v output, to get the supply to start normally.

You can get a 12V 21/5 watt bulb from Repco, use that as a dummy load. You can just connect the 5 & 21W filaments together making a 12V 30W bulb. It works well (on 5V) because the cold resistance of the bulb is less than an Ohm, so it applies at least a transient 30W load to the 5V rail at start up, then the current falls after that. If the supply is basically working the bulb will light up just under half brightness. To activate the supply you need to ground a control line which is usually the green wire on the connector, but best to look it up on the connector diagram. Of course to fully test the supply would require at least a 250W load.

I'm in M'dore. I don't normally offer a repair service, but it might be possible to persuade me to do it.

There is a place in Brisbane that repairs SMPS's which could be a good option, after you determine the supply is definitely the issue:

Power Supply Repairs

TechFix specialise in the repair of Power Supplies from a wide range of equipment. Often a Power Supply can be repaired for a fraction of the cost of buying a new one. This is particularly relevant…

tech-fix.com.au

 

[Brad Cancian]

Hi Hugo - indeed, the monitor turns on when plugged into the back of the power supply, so it very much is likely as you say, a through and through from the mains power. Bummer!

Alas I'm not electrically inclined - I'm a mechanical engineer by trade... all I really know about electronics is that you need to keep the smoke inside the wires My only other experience that vaguely resembles electrical know-how is collecting 1980s era casio calculator watches and getting them back into running condition through component replacements / mixing and matching (I am an unashamed nerd in that regard; I blame Marty McFly's calculator watch from Back to the Future for that obsession... but I digress!)

To be honest the only reason I thought power supply was because it was the bit that didn't seem to be working properly, i.e. the fan wouldn't spin (though it looked like it was trying to as it moves slightly at the moment mains power is applied, but wouldn't 'kick over'). It could be a busted power supply as a result of days of vibration while travelling, it could be something broken somewhere else which is not enabling the power supply to do its thing. Everything 'looks' ok to this highly untrained eye when I gaze around the inside of the box... so in reality, I am very much unsure!

I did try and turn the computer on with the drives (two disk drives and the hard drive) disconnected from the power cables from the power supply but this didn't seem to have an effect. I did not try and only disconnect the power to the mother board and leave all of the drives connected to power, however. Would this be worth trying to see if it might be an issue with the mother board tripping the power supply / not allowing the power supply to apply the required load? Or would this set up still not provide load for the power supply to kick in as it normally would?

I certainly very much appreciate the steer to test the power supply - chances are my electrical hamfistedness would most likely result in me getting something wrong with the suggested test, unless there's a tutorial or something that could step me through this?

I suspect that I may be at the point where I need someone who knows what they are doing with 30+ year old electronics to eyeball and play with the gear in hand...?

M'dore is a nice place - I hope you've been staying dry in the weather today! I am west of Brisbane, out around Ipswich way. It's a bit of a journey I know between where you are and where I am, and I certainly wouldn't have any expectations, but if I could twist your arm to come and play with 30+ year old electronics, and perhaps teach this mechanical guy a thing or two about awesome old tech, I would certainly make sure you were well compensated! Name your price and currency

Oh, and here's a pic of the set up when it was working before I moved house

Thanks again!

Brad

 

[Hugo Holden]

. It's a bit of a journey I know between where you are and where I am, and I certainly wouldn't have any expectations, but if I could twist your arm to come and play with 30+ year old electronics, and perhaps teach this mechanical guy a thing or two about awesome old tech, I would certainly make sure you were well compensated! Name your price and currency

Thanks again!

Brad

Its been over 20 years since I did a House Call (and that was for an unwell post OP patient). I have never done one for an Electronics problem.

It is quite nice system you have, when it was working, I can see why you love it.

You would have to bring the whole thing to my home/workshop in Maroochydore, and leave it with me for a few days to assess it. Then I could advise how much it would cost to fix it. Currency ?? In American dollars of course, the exchange rate is so poor currently that AU dollars can only buy a few packets of Wine Gums.

 

[BBent]

If your fan starts for a small moment to spin, it means the fan get some power from the psu (so the psu gives power, but only short). But then the start up routine sees a problem and cut of the psu. And the fan stops cause there is no more power. So are you sure you checked all connections, chips and so on, sometimes a chip or module / card can be popped up just a little bit and is making not a good contact, but you just don't see that. I had this one time after a moving, that a computer did not started... I took every connection cable in and out, stil not working, took the cards out, still not working. Took out the ram modules, put 2 back in, working... then put the other ram modules back in, and still working. So one of the ram modules just had a small bump and got a bit loose, but it was not to see by eye.

 

[Brad Cancian]

Thanks Hugo! I will definitely consider your offer of taking the computer on another road trip... though I am conscious that a road trip was probably the thing that caused the issue in the first place!

BBent - thanks for the suggestions; yes the fan moves slightly in one short rotation (2-3mm) when power button on the front of the machine is pushed, then immediately stops. The power supply keeps on making a slight humming noise. The fan does the same again when I disconnect the power by pushing the power button. The front three lights on the front of the box also so very briefly flash when the power button is pressed to turn off the power. As the lights all appear to be connected to the mother board, I can only presume that the mother board is getting power?

I tried some more fault finding today... well, I mean, taking things out, blowing on them, and reinserting them, haha....!

First I tried to remove each of the four RAM chips and see if any one of these was causing the problem. One came out at a time, and I tried to turn the computer on with each one removed, then again when reinserted… no joy there.

I then removed what I think is the card for graphics (it is the card that has the port for the monitor). It all looked fine. Reinserted it, no good. Here’s the card itself when removed:

It went into the card slot third from left; I also tried it in other slots; no good.

Here you can also see slots J6, J7, J8 and J9 (bottom of photograph). These appear to be connected to the switches and LEDs on the front of the box –

As mentioned, when the power is turned off, the three lights at the front of the machine flash briefly. So I would say the motherboard is receiving power, at least.

The next card to come out was the card to which the hard drive and disk drive are connected to. This card has the slots for the printer (if I still had one) and the mouse. It also has this little red and white lead that plugs into slot J13, whatever that is. The other end of that little cable connects to the ‘turbo’ button on the front of the machine. Never knew what that button did (still don’t!).

Here's the card with the cable from the disk drive disconnected:

Here’s the card with the cable from the disk drive and the hard drive disconnected:

Here’s my little friend, J13; he appears to be connected to the "turbo" button on the front of the machine –

Here’s the card completely removed:

Again I see nothing obviously wrong… reconnected everything, tried again… no good. Tried the board in another slot on the motherboard, same result.

Here’s the section of the motherboard where the capacitor sits, presumably for the BIOS settings. There is some corrosion here, but that was there when I retrieved the machine from my grandparent’s shed back in 2018.

I don’t suspect that this is the issue, but unsure.

Here’s another two images of this area.

For the life of me, I couldn’t get the cables from the power supply to the board to disconnect from the board, even with a little bit of force. Not sure what the magic trick here is to get these cables off to examine this area; any suggestions…?

This little fellow is also sitting homeless from the power supply. I don’t think it was ever connected to anything, but I welcome thoughts.

So, after an hour or so of mucking a about, I am still no closer to understanding what is going wrong. This has me perplexed given it was working perfectly fine before the move and travel.

Any other thoughts, observations, or things I can try?

Thanks again!

Brad C

 

[BBent]

I just used some basic, if not then else or

For example:
-If bumped when moving something get loose then start up routine fail
- if fan spins there is some power else it wont spin
- if lights on, there is some power else it wont light

- If psu stops after a few seconds then start up routine fail after a few seconds and shut down the psu else psu give wrong voltage so start up routine wont pass

- if psu gives short time voltage then start up routine fail

- if not fan spins, power or fan failure
- if not lights burn, power or board failure

It’s like a needle in a hay stack, but if using basic, maybe you can find the solution

 

[BBent]

P.s. i see some Goldstar on your images… so this seems to be a clone… Goldstar changed name… currently its known as as LG

 

[Hugo Holden]

While you were in there, you could check all of the blue Tant capacitors for shorts, they are notorious for it, spontaneously without any warning.

I don't think moving the computer had anything to do with the failure, not unless say it was dropped from a height to concrete. It is like trying to link potatoes with car accidents because 90% of people who have an accident have eaten a potato in the last few days.

 

[Brad Cancian]

Thanks Hugo; I like your potato analogy... I feel a bit like a spud here myself...! I do agree, it could be a coincidence... I have had no other issues with the hardware over its lifetime so the only corresponding thing that has happened was the move.

Alas I have zero know how or gear to check the system out to the component level on the cards. Are there any outward signs of issues with the capacitors...?

Short of studying for an electrical engineering degree, I think I am just about at the point of having to take the machine physically to someone to have it checked out

Thanks again,

Brad

 

[Hugo Holden]

Thanks Hugo; I like your potato analogy... I feel a bit like a spud here myself...! I do agree, it could be a coincidence... I have had no other issues with the hardware over its lifetime so the only corresponding thing that has happened was the move.

Alas I have zero know how or gear to check the system out to the component level on the cards. Are there any outward signs of issues with the capacitors...?

Short of studying for an electrical engineering degree, I think I am just about at the point of having to take the machine physically to someone to have it checked out

Thanks again,

Brad

Do you have a standard Ohm Meter (now days called a DVM or digital volt meter, that are multi functional, Ohms Range, AC & DC voltage ranges, current ranges etc). Good for checking if any of those blue Tant caps are shorted out.

Are you familiar with using a meter for these sorts of tasks , measuring resistance, voltages currents etc ?

( you don't need an electrical engineering degree, just knowing that pressure = flow x resistance, either in fluidic situations, or electrical ones as Voltage = electrical current x resistance )

 

[Brad Cancian]

Hi all! Some more progress with the fault finding at least. I finally manged to get the pesky power cord from the PSU to the motherboard off, which allowed me to do more checks at the component level.

The good news - PSU and the drives (HDD and disc drives) all power up when I disconnect power to the motherboard. The PSU fan ticks over fine as per normal, and I can hear the HDD ticking over as it usually would on start up... yay!! So, it's not an obvious fault with the PSU itself, or the HDD or disk drives.

I pulled everything off the mother board - cards, RAM, connectors to the lights, etc, to isolate the problem as much as possible.... here she is, naked as a babe...

I plugged the power cords back into the motherboard, and press the power button.... and the PSU won't start.

So, I am speculating that the problem is the motherboard somewhere (or the cables from the PSU to the motherboard... not sure how I could test the cables?). Probably one of those pesky tant capacitors...

I couldn't get the motherboard out of the case though to check the underside... I unscrewed the two screws but there are these plastic arrow retainers that I couldn't squeeze enough to get the board itself out of the case... any tips...?

So, right now, both joy and sadness... joy in that I have further narrowed the issue to (most likely) the motherboard... sadness as I now most likely need to delve deeper into the world of electronics to narrow down the exact fault on the board...!

Hugo - yep I do remember V=IR from back in the day but not too much else. Is there a good '101' tutorial somewhere in regards to how to use a multimeter and do the required checks that you'd recommend?

Also, as I side note, I have had a recommendation that I replace the CMOS NiCad battery... I am guessing these are both rechargable and unreliable...? How might I best do that? What do you recommend as a replacement? I also see the board has an option for an external battery...?

Thanks all for your perseverance with me on this one!

Cheers,

Brad C

 

[Hugo Holden]

One of the blue Tant caps on the Mobo will be shorted out most likely. That will stop the PSU from starting.