286 board +5volt on AT power connector to ground.

[Robin4]

I have got 4 Mycomp PAT-286 mothboards in my possesion. I did a check on them to see if they were not shorted.
So i could hook them up on my power supply. (computer test bench)

One of them seems to have a litte battery corrosion on the traces near where the battery was. (yes this is still fixble. (call it for now board number 2)
I measured all the 4 boards on the AT power connection and board number 2 doesnt seems to have any shorts on the power connector (only have to clean op and make it tidy again)

The other 3 boards have a problem on the 5 volt power rails, the other rails seems to be fine... I measured al the 4 5volt connectors the connector have. (3 on the left and one on the right, on the right connector (can be power good signal or something)
When i do a continuity test, i got still long strong beeps meaning 5 volt is stick to the ground. So when i measure on ground on a IC its still 5 volts connection too. (which didnt happen on board 2 which is fine

I have a digital meter, but probably iam better with a soldering station then trying using a DMM. Because i can de-solder a lot of chips to take them out of circuit and see of the 5volt to ground problem will went away.

What i tried so far.

1. Kept the Dmm on 5volt and ground on the AT power connecter, let it beep continously and remove the jumper shunt one for one (maybe someone could have set a shunt wrong due no manual (you never know) but didnt helped.
2. Replace a 1K ohm resistor (in pink on the picture) which had a dis-colored leg, and i trying to avoid that could be a problem, took a another good same resistor from a donor board.
3. Removed all CPU, CHIPS chipset chips and measured again also checked the sockets for faults, and used some deoxit d5 to make the contact better, when there was corrosion_
4. Re-done some bad solder practice from the older owner to make at look better.
5. Removed almost all ICs one by one, took them out of circuit and measured again, still nothing, still 5 volt to ground.
6 Also before everything above cheched the condition of the pcb and components, nothing to strange to see.
7 After number 6 also removed all the tantalums (was not much to remove, limited quantity)) from the board.. And did replaced the most on the board with newer ones, to rule out the 5volt to ground problem. (have to add to 22uf / 16 volt where now is nothing yet)
8 Some 74 LS logic have been tested in my mini pro programmer, and checked all fine / good.
9. Bios eproms removed also did not solve the problem.




Legenda:

Black = 22uf 16v tantaal cap (have to add) 2x
Light green = components i have socket for the sake of it (didnt wanted to heat the component to much) It if it could getting bad its easier to remove the next time)
Yellow= Seems to be a PAL chip of some sort with silkscreen removed (have a board that i could see on it that the IC was a PAL, because that one had a silk screen on it)
( I took it out of circuit on the board of the photo, but still no avail.. Still 5 volt to ground. (so think that PAL should be fine)
Also removed the memory chips, these will be all checked with a memory tester i have) But with memory remove, still the same, 5volt to ground is still there.
Its also sure that the CHIPS chips should be OK, they where already marked OK with a pencil when i got this board. So probably the board had maybe already had a run when
there was no 5volt to ground on the board. But something have happend due long period of storage.

Red = Mainly components i removed (still 5volt to ground)


Iam almost done with the ICs. It could be one of the resistor arrays also, so if i make a guess, near the cpu socket probably due wrong insertion of the cpu maybe?)
It also can be the resistor array between the ISA slots.. that now it working as a shunt from 5volt to ground and internally is bad.

Is there an easier way to find the problem without damaging the board further.. (and no i dont have a bench power supply)

Last attempt would be:

Removing resistor arrays. (starting by the ones near the cpu socket)
checking those 2 3 legged things on the right side of the board. (probably doing this first)
Replacing those PLCC sockets. (maybe short underneath the socket if possible)

 

[Kelly Gray]

Using a DMM in continuity mode might be giving you a misleading reading. Typically the DMM considers any resistance below about 200 ohms to be a contact. The 5V rail on the computer typically draws several amps, which translates to an effective resistance of no more than an ohm or two. In fact, any load of more than about 25mA or so might be enough to give you a false reading. Now the meter typically uses a test voltage of under two volts, but three or four standard TTL chips could well be enough for a false reading even at that low voltage.

 

[Robin4]

I think i found something. The clue was. I was looking to an other component on the board. When i saw that at least 2 boards had been fiddeling with. On one, a different color resistor netwerk array was installed. (so that stood really out to the older black ones)
Second one had non / empty spots with open through holes.
So i thought, stop with pulling out these ICs.. And go look further what happend with those resistor network array.
To make some easy accesable work room i removed the 3 latest ISA slots so i have enough space to work there..

And behold, something is strange here..



This guy that was before me, didnt know what he was doing looks like.




These resistor array networks seems to be installed 180 degrees backwards.. Knowing the dot in the resistor network has to be pin 1.
But pin one is not installed in that tiny square.. (they seems to be flipped around wrong)



Iam going to try to flip them around and see of these resistors arrays are stil in life or might be burned internally already.

 

[Chuck(G)]

The DIPs that you've outlined in yellow are also resistor networks. 33 ohm, if I don't miss my guess. Used as part of the address-line drive in DRAM arrays.

 

[Robin4]

The DIPs that you've outlined in yellow are also resistor networks. 33 ohm, if I don't miss my guess. Used as part of the address-line drive in DRAM arrays.

Actually that color is light green as well.. If you compare it with the pal which is in yellow.
Yeah i knew those where resistor network too, only in old fashion.
I have put them in sockets, because didnt want them to re-heat them to much with the soldering iron. IF they every going bad, would be an easier replacement.
IC sockets got from a donor board.. So actually it was a free add-on.


Do you perhaps know if these red terminator resistor arrays are dual terminated? (before i buy the wrong part)

 

[Chuck(G)]

I see now--the single IC outlined in yellow has no visible markings on it, so it's difficult to speculate on its contents.

Without knowing the circuit or the part numbers, that would be pretty difficult to state one way or the other with the red resistor SIPs.