IBM 5160 - Troubleshooting power issues

[BakaOsaka]

Hi all,

I have an IBM 5160 (256K) with dual floppy drives. I attempted to power it on today, the PSU powered up for a split second, there was a small pop and the magic smoke was released and the PSU fan stopped spinning. Now whenever I try to power it on nothing happens - The PSU fan does not spin.

I've been following the troubleshooting guide at http://www.minuszerodegrees.net/5150/problems/5150_known_problems_issues.htm and have removed the tantalum capacitors C56, C58 and C68 (C56 and C58 were clearly blown). I've just removed them and have left their slots empty without anything soldered in to their place - I assume this is correct?

I've tested for shorts on the motherboard's P8 and P9 rails and as far as I can tell they all look fine - No obvious shorts. This is leading me to believe that the PSU may be to blame. Is there a way to test the PSU standalone? This machine is dual floppy so I don't have an MFM drive to hand to act as load.

Thanks in advance!

 

[BakaOsaka]

So I've tried connecting an MFM drive that I took out of a Compaq Deskpro to the PSU in isolation and it still itsn't working - I'm assuming the PSU has issues. I'm waiting for the driver bit so that I can disassemble the PSU - Does anyone have a list of capacitors required for the stock 130W PSU?

 

[modem7]

It sounds like:
1. You navigated through the [procedure] and ended up at the page at [here].
2. Then, per there, you removed C56 and C58.
3. Then, you reattached P8 from the PSU to the motherboard.
4. Then you discovered that the PSU still does not fire up.

Is that case?

I assume this is correct?.

Per the web page pointed to earlier, C56 and C58 are not critical to the operation of this motherboard. Normally, if they are faulty, they can be removed, and the IBM 5160 will start. But there can be exceptions.

Is the PSU the stock IBM one, or is it a clone ? For example, if a clone PSU, maybe that PSU requires a small load on one (or both) of the 12V lines, and C56 and C58 were providing that small start-up load. Just a guess.

Is there a way to test the PSU standalone? This machine is dual floppy so I don't have an MFM drive to hand to act as load.

No. The 5160's stock PSU requires an adequate load. Maybe the two floppy drives are enough. As an experiment, try both of those.

 

[Stone]

You can also load with an IDE drive or two.

 

[ibmapc]

I seem to remember someone using an automotive headlamp for a load on the 12v.

 

[modem7]

Your post #2 has appeared.

So I've tried connecting an MFM drive that I took out of a Compaq Deskpro to the PSU in isolation and it still itsn't working - I'm assuming the PSU has issues.

It is easy to fall for the trap of inadequately loading the PSU. Different makes and models have different requirements. It is easy to think something like, "This big MFM hard drive draws a lot of power at start-up; that will do as a load.", but it turns out that the drive draws most of its power from the +12V line and not much from the +5V line, and the particular model of PSU requires the adequate loading on its +5V line.

Does anyone have a list of capacitors required for the stock 130W PSU?

IBM is known to have used different 'guts' in the PSU's over the years. Therefore, if someone here does have a list of capacitors, it may not match what is in your PSU.

If the source of the 'sound of small pop with release of magic smoke' was your PSU, resulting in the fan (DC powered) no longer turning, then one of many types of components could have failed (e.g. transistor). And the 'PSU is what failed' hypothesis implies that the motherboard's C56 and C58 (quote: "were clearly blown") must have failed open-circuit earlier, not affecting 5160 operation earlier.

Maybe do a visual inspection within the PSU, if not done already. To allow time for the main filter capacitors to discharge, wait, say half an hour after disconnection of the PSU from mains AC power.

 

[BakaOsaka]

I've just tried connecting the two floppy drives to the PSU with no luck - The PSU is as lifeless as ever, no fan spin or any activity.
I'll disassemble the PSU and do an inspection once I get the required tools. In the mean time, is it advisable to use a modern ATX PSU with an ATX to AT power adapter to test the motherboard in isolation? I seem to remember that modern PSUs don't provide a -5v rail, however for testing the motherboard alone I assume this won't be an issue?

 

[modem7]

In the mean time, is it advisable to use a modern ATX PSU with an ATX to AT power adapter to test the motherboard in isolation? I seem to remember that modern PSUs don't provide a -5v rail, however for testing the motherboard alone I assume this won't be an issue?

No issue, because none of the components on the IBM 5160's motherboard use -5V (the -5V from the PSU simply being routed to the expansion connectors).

Same situation for +12V and -12V.

 

[Hugo Holden]

It does sound as though there may be a failed component in your supply. Repairing these is not without it hazards.

A while back I wrote a detailed article on the IBM 5155 computer supply, similar power rating to yours. It may well be worth a read of this article to acquaint yourself with the operations of these sorts of power supplies:

http://worldphaco.com/uploads/The_IBM_5155_POWER_SUPPLY.pdf

As noted it is important to load the +5v rail with a dummy load, or the initial turn on transient at power up will get detected and disable the supply. These supplies have fairly elaborate over-voltage protection and over-current detection too.

 

[BakaOsaka]

A while back I wrote a detailed article on the IBM 5155 computer supply, similar power rating to yours. It may well be worth a read of this article to acquaint yourself with the operations of these sorts of power supplies:

Thanks for the useful info - I'll provide updates once I've been able to disassemble the PSU.

I've powered up the system board with an ATX to AT converter. The board is in isolation, with only the PSU and speaker connected.
I have the SW1 switches set as the following:

1: On
2: Off
3: On
4: On
5: Off
6: Off
7: On
8: On

Connecting power to the board via P8 and P9 gives me no response - The speaker crackles slightly at power-up and then provides no more feedback.
I've tried replacing the RAM chips in Bank 0 with those in Bank 1, however I still get no response. I've also tried reseating the two ICs at U18 and U19 (11/08/82 BIOS), as well as the 8088 at U3. The two BIOS chips did have slight corrosion on the pins, however I cleaned and reseated the contacts. I've ordered some RAM chips, although I'm not sure this is the issue.

I tested the voltages according to this page, and I got the following:

P9
P6: +4.74v
P5: +4.74v
P4: +4.74v
P3: 0.00v

P8
P4: -11.58v
P3: +13.70v
P2: +4.45v
P1: +4.45v

These all look fine to me - Is there anything else I may have missed in the interim? Is there anything that I can do in the meantime whilst waiting for the new RAM chips to arrive?
I suppose it is possible that the BIOS chips are still defective. I assume this could cause these symptoms?

 

[Malc]

......I tested the voltages according to this page, and I got the following:

P9
P6: +4.74v
P5: +4.74v
P4: +4.74v
P3: 0.00v

P8
P4: -11.58v
P3: +13.70v
P2: +4.45v
P1: +4.45v

Your +12v line looks too high at +13.70v, I'd double check that measurement, Did you measure for shorts on the board as per Modem7's site. Edit: I should read the thread properly, I see you did.

I suppose it is possible that the BIOS chips are still defective. I assume this could cause these symptoms?

Yes it's possible that the BIOS chip in U18 is faulty as well, Can you burn your own Eproms ?, As the board has the 11/08/82 BIOS you can remove the Eprom from U19 and the board will still post without it but will throw an error, You can't do that with the later BIOS's though.

 

[modem7]

Obviously, the lack of -5V is due to the fact that your ATX PSU does not generate -5V. However, as I wrote earlier, the IBM 5160 motherboard itself, only requires +5V.

So, you have an 'appears to be dead' motherboard. Many many possibilities. The failure of any of about 80% (arguably) of the motherboard's chips can cause that symptom.

Maybe the POST is starting, maybe not. If it is starting, then because you hear no beeps when you do the 'minimum diagnostic configuration' of [here], then the POST is not getting as far as step 21 in the POST sequence shown at [here].

The basics:

* Thorough visual inspection.
* Socketed chips: Look for chips that have been inserted in the wrong orientation.
* Socketed chips: Remove all socketed chips and inspect the pins, looking for pins that have been bent up under the chip. This action also effectively re-seats the chips.
* Swap out RAM bank 0 chips for a set of known good chips.
* Swap out all other socketed chips for known-good chips.
* Use a Landmark/Supersoft diagnostic ROM in place of BIOS chip U18. See [here]. Note the special video card requirements, although, speaker beeps can be informative.
* Socketed chips: Look for 'incorrect' chips inserted by a previous owner. E.g. A VCF member found that on his 5150 motherboard, the wrong type of RAM chip had been put into one of the sockets.


Beyond that, 'diving deep' is required. E.g. Using an oscilloscope to verify that clocks are being generated. Looking for the presence/absence of other activity.

 

[modem7]

However, as I wrote earlier, the IBM 5160 motherboard itself, only requires +5V.

Oops. And the POWER GOOD signal.

 

[BakaOsaka]

Hi all, thanks for your replies. I'm still troubleshooting this board, currently without much luck.
I've bought a set of new RAM chips - MN4164P-15A's - And fitted these to bank 0. The machine still failed to POST, showing no signs of life.
I also tried replacing the CPU with an NEC V20, and again had no luck. I also removed the ROM from U19 with the same result.

My next planned steps are to use the suggested diagnostic ROM to pinpoint the cause, however I will have to purchase pre-flashed ROMs which will take some time. In the mean-time, I have hooked my logic probe up to the motherboard. Is there a list of signals I can check that may help limit the cause of this issue?

 

[modem7]

My next planned steps are to use the suggested diagnostic ROM to pinpoint the cause, however I will have to purchase pre-flashed ROMs which will take some time. In the mean-time, I have hooked my logic probe up to the motherboard. Is there a list of signals I can check that may help limit the cause of this issue?

At [here], see the linked image in step 13. If your probe indicates pulses on pin 13 of the 8253 chip, it means that the CPU has started executing the POST and is getting at least as far as step 13.

If no pulses on pin 13 of the 8253 chip, then per step 3 at [here], verify that your probe indicates pulses (the 4.7727 MHz clock signal) at the clock pins of the 8088 and 8288 chips.

 

[BakaOsaka]

Thanks for the link - Very informative.

Pin 13 (OUT 1) on the 8253 is high - No pulse. Pins 9, 15 and 18 (CLK 0/1/2) are all pulsing.
Pin 8 (CLK) on the 8284A is pulsing. Pin 10 (RESET) is low.
Pin 19 (CLK) on the 8088 is pulsing. Pin 21 (RESET) is low.
Pin 2 (CLK) on the 8288 is pulsing.

So it looks like it's at least reaching step 8 at this page - Is that correct?

1980s_john was kind enough to flash a Landmark Diagnostics ROM (This image) to an L27C256D - I fitted this to U18 and still saw no response - ie. no speaker beep at all. I'm assuming this means it's failing before it can even execute the diagnostics ROM?

 

[modem7]

Thanks for the link - Very informative.

Lots of other informative 5160 information in the IBM 5160 section of minuszerodegrees.net

Pin 13 (OUT 1) on the 8253 is high - No pulse. Pins 9, 15 and 18 (CLK 0/1/2) are all pulsing.
Pin 8 (CLK) on the 8284A is pulsing. Pin 10 (RESET) is low.
Pin 19 (CLK) on the 8088 is pulsing. Pin 21 (RESET) is low.
Pin 2 (CLK) on the 8288 is pulsing.

So it looks like it's at least reaching step 8 at this page - Is that correct?

Yes.

Although, there is a possibility that something earlier may not be happening properly. Regarding "When the 8088 CPU leaves the reset state, it starts execution at address FFFF0." The data sheet for the 8088 includes in its description of its reset pin, "The signal must be active HIGH for at least four clock cycles." In the IBM 5160, four clock cycles equates to 840 ns. The oscilloscope measurement at [here], for one of my PSU's, shows that the reset-high-minimum-time requirement of 840 ns is exceeded by a looooooong shot (300 ms).

It would be good for you to verify that the reset pin of the 8088 is initially high, then goes low. If I hook my logic probe up to the reset pin of the 8088, then power on the PSU, I see the probe indicate HIGH (for less than a second) then indicate LOW.

1980s_john was kind enough to flash a Landmark Diagnostics ROM (This image) to an L27C256D ...

I confirm that that is the right image.

You have an early 5160 motherboard. L27C256 does not appear in the table at [here], and so there is a remote possibility that the L27C256 is unsuitable for the early 5160 motherboard. Maybe someone has proven it fit for the early 5160 motherboard.

... I fitted this to U18 and still saw no response - ie. no speaker beep at all. I'm assuming this means it's failing before it can even execute the diagnostics ROM?

So many possibilities. Per the image at [here], does your logic probe indicate pulses on the /CE pin of U18 ?

 

[Malc]

.... and so there is a remote possibility that the L27C256 is unsuitable for the early 5160 motherboard. Maybe someone has proven it fit for the early 5160 motherboard......

I have a bunch of TMS27C256 Eproms which have L27C256D / E printed underneath the chip, I can confirm the TMS27C256 is not compatible with all the IBM boards.

To the OP, Where on the Eprom is L27C256D printed, Top or bottom, If it's the bottom what does it say on the top ?

 

[BakaOsaka]

It would be good for you to verify that the reset pin of the 8088 is initially high, then goes low. If I hook my logic probe up to the reset pin of the 8088, then power on the PSU, I see the probe indicate HIGH (for less than a second) then indicate LOW.

I can confirm that Pin 21 (RESET) is initially high, and goes low shortly after (ie under a second).

So many possibilities. Per the image at [here], does your logic probe indicate pulses on the /CE pin of U18 ?

With the either the original BIOS ROM or the Landmark ROM fitted to U18 the state of Pin 20 seems inconsistent. Sometimes it will pulse a few times at power-on before staying high, and sometimes it will pulse consistently until I remove power. The pulsing also seems inconsistent - For example, on the cases where it pulses a few times it may pulse at a much faster rate than it does when it's pulsing consistently.

To the OP, Where on the Eprom is L27C256D printed, Top or bottom, If it's the bottom what does it say on the top ?

On the bottom - The full text is:
L27C256D/8163027
KOREA 26

And on the front it says:
TMS
27C256-25JL
LDE 8826

 

[Malc]

On the bottom - The full text is:
L27C256D/8163027
KOREA 26

And on the front it says:
TMS
27C256-25JL
LDE 8826

Do you know if your Eprom was tested and confirmed working before being sent to you.
Sometime today i will pull out an early 5160 board and refresh my memory, IIRC the TMS27C256 Eproms i have did not work in the Early 5160 board, The board appeared Dead. I also seem to remember having the same problem if i tried them in a 5170 or 5162.

Many years ago i bought several tubes of Eproms NOS and cheap, These TMS Eproms were included and the only ones i had problems with on IBM 5160 / 5162 / 5170 Boards.