VOGONS


First post, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

Im fixing a 486 board that had barrel battery acid leak. From no response on boot I am now getting 10 beeps which is a cmos error.

I cleaned this board, fixed corrosion, removed alot of components resistors capacitors etc as there was too much corrosion on the vias. I even found a few cracked traces on the back which I fixed now.

Anyway I am stuck on the cmos error, I actually have a identical but working motherboard so I checked all the components around the cmos which is integrated into the opti f82c206q chipset on this board. comparing all resistance readings ground to component pins the readings are same on both boards, no shorts or disconnected pins. The cmos chipset is near the barrel battery so it is definitely related however apart from changing the actual cmos/chipset I am out of guesses. Yes ofcourse I tried resetting the cmos chip, nothing. I checked resistance readings on the cmos chip compared to a working one and its the same, leads me to believe the chip is not internally shorted. So I am thinking some connection between bios and cmos is lost, I have not been able to trace it visually so I fear a multi layer trace might be cracked.

Its a very long shot but anybody ever solve a similar issue?

the motherboard is a chicony ch-486
here is the picture of the cmos/chipset

IMG-20200410-153046.jpg

still some flux and cleaning residue around, but I checked most things here for continuity/shorts

Reply 2 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

Yes I actually tried the physical bios chip from the good board, still 10 beeps but higher pitch, maybye they are different bios revisions? I also tried the keyboard controller chip don't know why because no logic behind it but since socketed why not, same results

Reply 3 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

This kind of gives me a idea, I can try to map which bios pins connect to which cmos pins on the working motherboard, pretty time expensive exercise but what are hobbies for

Reply 4 of 11, by Horun

User metadata
Rank l33t++
Rank
l33t++

Check the lead off L2, appears to not be soldered very well but that is probably not the issue. Also check around RP7 and Rp8, those are multi pull down or pull up resistors iirc. Have a 486 with same Opti 82C206 and it has issues also. Hope you figure yours out as it may help me too.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 5 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

So I compared resistance readings on the bios pins on good and bad motherboard, resistance is exact on each pin with ground. So it seems each pin from bios is connected to the same exact path on each bios, there goes my theory about a connection from the bios to cmos being broken.

Thanx for the good spotting, L2 I actually dug off some corrosion off that leg and it ate away some solder which I haven't added new solder yet, I made sure the Leg on L2 still has continuity which it does. I will do this later.

I did have a brief look at RP8 didn't see any shorts, however this is a good tip for further investigation. I will take readings ftom the good board RP7 and RP8 and compare. If no reply it means no success 🙁.

I will try to source a replacement cmos chipset maybye the cmos memory went corrupt? I guess the chipset wouldn't show as shorted if only the internal cmos memory went bad

Reply 6 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

So I was just poking around comparing resistance readings and I found something.
One of the main lines one the faulty board gives me no reading to ground. On the working board it does give me a reading that slightly changes, this is ok because there is several tantalum caps on this main line.

I have drawn the line in blue on the picture, its the 3rd pin on the at power supply and goes directly through 1 tantalum cap. the line proceeds all the way to the last isa slot, on the way it branches off to 2 more tantalum caps as I can see and powers each isa slot and most likely dissapears to other locations.

On the broken board the line has continuity all the way from the start to end, this is not the issue.

When I set my multimeter in diode mode/resistance and place 1 probe on earth of the board like a screw holder and other probe on the power line I get no reading.

When I do this on the working board I get a reading which fluctuates but get a reading the whole time, it fluctuates because the tantalum caps on the line are charging/discharging.

Now the way I see it the line is a positive line so it doesn't go to earth/ground direct but some component on this line must so have to get some resistance to ground, as the working board does.

Faulty board there is nothing to ground so I understand some component on this line is not grounded (the link is broken) naturally first thing I think is the first cap from the power plug, I checked with multimeter and this cap charges and discharges about the same as the good board. Should I swap the cap from the good board? I think the cap is ok if its holding charge and discharging

image-2020-04-11-18-29.jpg

Reply 7 of 11, by Horun

User metadata
Rank l33t++
Rank
l33t++

That should be the +12v line and is required. If you set your DVM to Ohms and Auto-ranging it should read near the Meg Ohm range (on most 486 standard AT boards) with Plus probe on it and Minus probe on ground if no adapter cards are installed.

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 8 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

I finely got a cmos replacement chip, cliped off the old and installed a new one and no change still 10 beeps on boot.

I went as far as comparing resistance readings on the bios chip and cmos chip from a working version of this board to the faulty one and the readings are the same.
This tells me the bios chip and cmos chip are all connected correctly on each pin to their designations.

Leaves me in a hard situation both bios and cmos connected and working properly then what could be causing a cmos error.

Just thinking loud here, motherboard posts bios checks for cmos, finds error gives 10 beeps

Bios and cmos both working and all connections right.... only issue would be if there is a indirect connection from bios to cmos going through a 3rd component that is faulty... this will be the next avenue of investigation

Reply 9 of 11, by kalohimal

User metadata
Rank Member
Rank
Member

I would first spend some time checking the connectivity of the traces where the battery leak had occurred, especially through the vias and under the keyboard connector. Then I would verify that the opti 82c206 is getting 5v on its VCC pins.

Slow down your CPU with CPUSPD for DOS retro gaming.

Reply 10 of 11, by treeman

User metadata
Rank Oldbie
Rank
Oldbie

yeah I was thiking to myself to take off the keyboard connector again, im testing it without the keyboard anyway since it doesn't even post anything on the screen, think I was a bit quick to reattach it after fixing the damage.

I thing I did check the voltages on vcc or at least checked the vcc pins show resistance to positive power rails, however good idea I need to go back to this and verify to be 100% sure and might as well check vcc on north and south bridge

Reply 11 of 11, by kalohimal

User metadata
Rank Member
Rank
Member

I'd measure it's Vcc pins for 5v with power on rather than just measuring resistance. Resistance will show up via the internal of the chip and external components connected which doesn't mean anything. But I suspect the problem is most probably due to broken traces or shorts caused by battery fluid.

Slow down your CPU with CPUSPD for DOS retro gaming.