PDA

View Full Version : DEAD 386 board



6885P5H
April 27th, 2017, 12:52 PM
This motherboard is DEAD. I guess it's a "UNICHIP". It says 367C REV 1.0 on it and also 386WB. When powered on nothing comes up on the screen other than a spike at power on. The power supply is good, the RAM is good, and the video card is good. I don't know about the CPU, it's soldered. What a disaster! Nothing on AM radio when the board is powered up. Well, nothing other than a soulless, eerie tone... I can switch num lock on the keyboard a couple of times, after that it stops responding. But I think this behavior might come from the keyboard itself and not any kind of activity in the motherboard. The motherboard had one of those nasty Varta barrel battery installed and it had already started to spread its corruption. I removed it, the board looks kinda funny there and 2 traces don't look too good. But they both are related to the CMOS battery. Shouldn't affect anything other than that right. RAM or no RAM, BIOS or no BIOS, video card or no video card, all I hear is always the same lifeless, eerie tone.

I am not really looking for assistance. I already know that nothing can be done to bring this board back to life, realistically. But what if, it's simply not jumpered correctly? What if for example it needs a battery to function?

Why is it dead? Most of the parts seem simple and standard. Dead CPU, or dead chipset? Would it be possible to find that out, what is broken? Could a person be able to unsolder the CPU or whatever or are the pins way too small and close to each others?

I am definitely not happy about this. I hate broken things... https://img4.hostingpics.net/pics/884519DSC04780.jpg

Stone
April 27th, 2017, 01:07 PM
... The motherboard had one of those nasty Varta barrel battery installed and it had already started to spread its corruption. I removed it, the board looks kinda funny there and 2 traces don't look too good. But they both are related to the CMOS battery. Shouldn't affect anything other than that right....Wrong.

There are hidden traces that could be bad as well. You can't see the inner board layers.

Chuck(G)
April 27th, 2017, 01:50 PM
If you don't like the QFP 80386, there's an unpopulated PGA socket position. Not that this board would be worth the trouble.

Far East manufacturers sometimes laid out their PCBs for both QFP and PGA so they could use whatever they could easily purchase. Obviously, since you don't need a socket, a QFP is the preferred low-cost move.

modem7
April 27th, 2017, 10:12 PM
Try re-seating the BIOS ROM chip if you have not already.


But what if, it's simply not jumpered correctly?
Setting appear to be at http://www.rainbow-software.org/manuals/367c.html

6885P5H
May 1st, 2017, 11:24 AM
Wrong.

There are hidden traces that could be bad as well. You can't see the inner board layers.

Inner board layers? Oh no.... I had heard about such an atrocity before. I didn't want to believe that something like a 386 board could have them... So do all 386 boards have inner layers?


If you don't like the QFP 80386, there's an unpopulated PGA socket position. Not that this board would be worth the trouble.

Far East manufacturers sometimes laid out their PCBs for both QFP and PGA so they could use whatever they could easily purchase. Obviously, since you don't need a socket, a QFP is the preferred low-cost move.

Of course I don't like it. Maybe I'd think about adding a socket if I was sure that the problem was with the 386. But then I don't know if the board is even ready to accept a PGA CPU.


Try re-seating the BIOS ROM chip if you have not already.


Setting appear to be at http://www.rainbow-software.org/manuals/367c.html

I have. By "not jumpered correctly", I meant maybe there is a quirk with the board that wouldn't be explained on the silkscreen or on TH99. I've messed with the jumpers, always the same thing. I'll try reading the ROM to see if something is off but...

I get 5.09V at the first VCC point of the PGA. No problem, that is under the tolerance range... Should I measure the other VCC points?

Because the battery was near the power connectors, maybe the internal traces it damaged were for power. If power gets to the CPU, then that would mean, that the battery hasn't damaged any power traces. But it could've damaged something else or... I don't know I'm just trying anything I can to find out if the problem is with the CPU or something else...

It's a shame because it sounds good on paper. 40MHz 386, 128KB of cache, 128MB of RAM... But the construction is awful.

I wonder why it's also called a DataExpert even though that's not written anywhere on the board.

modem7
May 1st, 2017, 02:20 PM
So do all 386 boards have inner layers?
I would be very surprised to find one that didn't.


I wonder why it's also called a DataExpert even though that's not written anywhere on the board.
The designer possibly licensed the design out to various manufacturers, with DataExpert being one of those manufacturers.


Because the battery was near the power connectors, maybe the internal traces it damaged were for power. If power gets to the CPU, then that would mean, that the battery hasn't damaged any power traces. But it could've damaged something else or...
Coincidentally, I am in the process of repairing an IBM 5150 motherboard that was damaged by a leaking battery. By where the green/blue 'stuff' was, I can tell that the battery was on a card in slot #4. Also, the pins and solder joints on five chips near the front of slot #4 have dulled considerably, whereas all other chips have bright silver pins and joints. Sure enough; I have already found two of the 'dulled' chips to be faulty. I am guessing that the green/blue 'stuff' has made its way into the chips. I am presently replacing all 'dulled' chips, because I just know that even though some may be working now, those chips may fail in time because they have been compromised.

modem7
May 1st, 2017, 02:27 PM
Connected a speaker to your motherboard yet (to hear if there are BIOS error beeps) ?

What about a POST card ?

Xacalite
May 2nd, 2017, 09:25 AM
128MB of RAM
Wow, rare thing for a 386 board, indeed.
So, probably well worthy of rescue.
And I doubt the problem is battery leakage, there's very little corrosion visible, I've seen boards with much more green stuff and still working fine.

Stone
May 2nd, 2017, 10:02 AM
And I doubt the problem is battery leakage, there's very little corrosion visible, I've seen boards with much more green stuff and still working fine.Unfortunately it's more often what you don't see than what is visible on the surface that ruins these boards. The inner layers can't be viewed but can really be source of the worst problem with regard to battery leakage.

Malc
May 2nd, 2017, 10:25 AM
Unfortunately it's more often what you don't see than what is visible on the surface that ruins these boards. The inner layers can't be viewed but can really be source of the worst problem with regard to battery leakage.

Very true, Horrible stuff, I've had a fair few boards that suffered that fate, One in particular i have is a nice Commodore 386 machine though unfortunately dead as a door nail, It's battery barfed over the mobo :-( I've cleaned, replaced sockets and IC's but as yet no luck in getting it working, One day maybe, It's too nice to dump.

GiGaBiTe
May 2nd, 2017, 01:55 PM
I would be very surprised to find one that didn't.

I'm currently trying to repair a 386SX-40 motherboard which had pretty severe battery leakage, which is only a two layer board and makes things far easier.

http://i.imgur.com/BNAo0R2h.jpg

Stone
May 2nd, 2017, 02:05 PM
...which is only a two layer board and makes things far easier.Does two layers mean no internal layers, i.e. just top and bottom surfaces used for traces?

GiGaBiTe
May 2nd, 2017, 02:14 PM
Yes, there is only a top and bottom layer. I can shine a bright light on the bottom of the board and see through the board, there are no inner layers.

6885P5H
May 3rd, 2017, 09:26 AM
Connected a speaker to your motherboard yet (to hear if there are BIOS error beeps) ?

What about a POST card ?

I don't know how to connect a speaker to it (pin 7 and 10?). Not that it matters, as like I've said before I have an AM radio so I would hear the beeps if there were any. I hear absolutely nothing other than a constant "tone". Zero activity.

I don't know how POST cards work, but my guess is that it would display nothing. I guess I could buy one.... They seem insanely cheap on ebay.


Wow, rare thing for a 386 board, indeed.
So, probably well worthy of rescue.
And I doubt the problem is battery leakage, there's very little corrosion visible, I've seen boards with much more green stuff and still working fine.

Ehm, it's more just me hating broken things, and giving everything a chance before I give up. My mentality is to try to use things until they get so worn out, that they can no longer realistically be used or fixed. Heck not too long ago when I accidentally fried the card reader of my computer, I did some research to see if it could be fixed. Yes I will even give a chance to something as insignificant as a 10$ card reader. If you're interested the answer to this was, it could theoretically be fixed by replacing the controller and verifying if parts like capacitors are still good. But realistically, no one's ever going to do that so I just bought another one. I was just curious to see if something like that could be fixed. Theoretically, it could....

There is one thing I haven't tried, I haven't verified if the cache chips are good. I'm not sure how I would do that, I have never messed with cache before. Ha, but just thinking about it makes me laugh.... As if that's the problem.... I'll just post the sounds of the board here so that you can see what I'm dealing with here....


https://www.youtube.com/watch?v=VEQu2sTuleI

6885P5H
May 3rd, 2017, 09:31 AM
Yes, there is only a top and bottom layer. I can shine a bright light on the bottom of the board and see through the board, there are no inner layers.

Well there you go modem7!

modem7
May 3rd, 2017, 11:21 PM
So do all 386 boards have inner layers?

I would be very surprised to find one that didn't.

Yes, there is only a top and bottom layer. I can shine a bright light on the bottom of the board and see through the board, there are no inner layers.

Well there you go modem7!
Well, I'm very surprised.
But maybe there is another factor at play here:

* All of the inner traces are either running underneath a surface plane, or underneath surface traces (following their exact path).

* The motherboard is actually a 286 board masquerading as a 386 board.

* The change in density between the layers is resulting in the torch light bending around the inner traces, making them invisible.

* Defective torch.

6885P5H
May 6th, 2017, 10:03 AM
Well it seems that the corruption spread further than I originally thought... I saw that a capacitor, an inductor and a resistor had been corrupted, and even 2 of the cache chips were starting to get corrupted. There is even some corruption growing on a RAM slot. I cannot remove any of the parts, I would need to get an high-end soldering iron with a microscopic point, something other than my big solder sucker to remove the solder, and, some kind of magnifying glass. To then find out that the board is too far gone to save it in anyway? When parts start getting corrupted, that can make a circuit stop working right. Ahhhhh well I hope that this motherboard isn't too rare...

I'd like to read the BIOS chip but unfortunately the label covers the information I need to know. Can I damage the EPROM if I read it with wrong settings?? What capacity is the BIOS EPROM on a 386 motherboard like this usually?

6885P5H
May 7th, 2017, 05:45 PM
The BIOS chip is a 512 kilobits EPROM. It is 100% flawless.

If you'd like to download it, look up its MD5 checksum online. E553C6666719B6B78F814F291767F830

Uhmmmm so I don't have any 386 boards then. Death by a battery, if that's what it is, it's a pretty awful death.

Does that mean I should remove the batteries in my 2 busted Sharp PC-7000s? They're broken, but they still work kinda! Wouldn't want them to get even more messed up by battery corruption.

glitch
May 7th, 2017, 05:52 PM
This past week has been barrel battery mayhem rework for me! I think I've removed 12 of them from motherboards and clock cards. This was the heaviest corrosion I found:

http://i.imgur.com/2qxoEMn.jpg

There was so much on that board that I had to jump several traces, replace the keyboard DIN connector, and replace an ISA slot since the battery gunk had wicked its way into the slot. Happy to report it's now working :) I also worked on an 8-bit ISA multifunction board (RAM/RTC/serial/parallel/game) that I'm pretty sure was left out in the rain. Many sockets had to be replaced, but it's also fully functional now. Same thing as you -- bothers me to no end if I can't fix a thing!

6885P5H
May 9th, 2017, 11:27 AM
Bothers me to no end if I can't fix a thing, but unlike you I have zero skills or knowledge... But one day I'd like to take an electronics class.

glitch
May 9th, 2017, 03:03 PM
Bothers me to no end if I can't fix a thing, but unlike you I have zero skills or knowledge... But one day I'd like to take an electronics class.

Taking on basket case projects is often a fine way to learn! I've taken very few actual electronics courses.

EpongeX
May 12th, 2017, 04:58 AM
Hi, got the same board, looks like the jumpers are right
Did u try an other video card even if this one works in an other system ? (weird things happens with theses vintage boards)

6885P5H
May 14th, 2017, 04:40 PM
Yes I have. It has already been established that it is dead, probably from battery corruption.

You have one? Cool! In page 2 I said something like, I just hope that it isn't too rare... Well, looks like it's not rare, but extremely common! That's a great thing. You should remove the battery on yours before it too gets corrupted.