In my experience it's not a major problem, I have some Socket A motherboards with bad caps, and they have a resistance of 15-20 ohms to ground, they still run fine. But I also have this one board that begins smoking when turned on, and a resistor unsolders itself and moves back and forth in the liquid solder. Quite spectacular. I still have to investigate what exactly causes this, but I remember measuring an extremely low resistance there. Wouldn't that mean that the maximum amperage from the power supply goes through the circuit?

snufkin wrote on 2023-09-25, 09:14:

Just wondering, but is 55 ohms actually a problem? That's only going to be a few 10s of mA, which might just be leakage currents. It doesn't sound like a metal-metal short. Anyone know what would be an expected value would be on this board? I've got a vague memory of measuring resistances on a working Athlon motherboard and getting low resistance (10s of ohms) on at least one supply rail.

I may have made a storm in a teacup, but I must be sure.

kinetix wrote on 2023-09-25, 10:37:

snufkin wrote on 2023-09-25, 09:14:

Just wondering, but is 55 ohms actually a problem? That's only going to be a few 10s of mA, which might just be leakage currents. It doesn't sound like a metal-metal short. Anyone know what would be an expected value would be on this board? I've got a vague memory of measuring resistances on a working Athlon motherboard and getting low resistance (10s of ohms) on at least one supply rail.

I may have made a storm in a teacup, but I must be sure.

If I remember correctly, the “so-called resistance” (i.e., the voltage drop across the semiconductor when current passes. The magnitude of the current depends on the range of measured resistances) for the Radeon 9600 core is something like 6 “so-called Ohms”
And so what?

I was checking my GA-5AA rev 3.2 and schematics is quite different. But maybe it will help you (TC40 is on the left side from the bottom):

Round soldering point is connected to ground (as suggested) and there is no measurable conductivity between capacitor pins.

I think your TC39 is my TC40.
Can you check that one, is the square point connected to the inductor?

My real problem is
What veins touch the square? Top? Back? Both?

Right now I think, almost sure, are both. Both veins are connected. The inductor solder points are connected in both sides. So it does not matter which side the capacitor leg is soldered

I will risk and test in a few hours , at home, and report back

Well, I'm pretty screwed.

I soldered the capacitors that I had removed and energized the board. The monitor briefly went from "no signal" to wanting to show some signal, but I power off immediately when I remembered that I had to check something. I didn't realize if it had shown anything on the screen. When I turned it on again, there was no response, but I thought I heard a very faint creak on the board a couple of times. I tried a couple of times without result.

I removed the capacitors again (arrrgh) and checked the soldering point. Once again I had forgotten to verify the welding points. Now the rounded point, which I assumed had no contact with the upper line, is giving me continuity with it. It seemed to have an almost invisible solder joint. I dedicated myself to cleaning the contact, even using an injection needle to scrape and make sure that the solder and the metal of the pass-through did not touch the vein.... but it continues to give me continuity. dam!

I really need certain information about which veins contact the legs of the capacitors, which veins in wich side touch each hole.
do the round point (1) contact the back vein (A)? and the front side?
do the square (2) contact the veins (B) in both sides?
I leave the reference images.
remember it is a rev 2.2

From your photo it does not look as if there is a connection between "1" and "A" on the top side. Switch your multimeter to resistance mode and measure the resistance from the solder hole to the exposed metal in the top plane "A". If there is a short-circuit, you will get a very low resistance. If there is a resistance of 10-50 ohms, then it is just leakage current that occurs across the other capacitors that are still in the circuit.

If there is a connection, the easiest way to clean it is to apply flux to the area, then gently go across the area with a clean iron tip to suck it up. Keep dwelling time short and the temperature below 380 to prevent further damage to the solder mask.

And I forgot to mention: If "A" is a ground plane, there will always be a connection between "1" and "A", regardless of a possible solder bridge, because "1" is also connected to the ground plane on the bottom side.

If the issue is if 1 makes a connection to A, and 2 to B,

take a sharp tool, scratch some solder mask from A and B (until you see the copper) and test it for continuity and resistance. Simple as that.
Flux + solder braid and clean the area.

kinetix wrote on 2023-09-25, 16:41:

My real problem is
What veins touch the square?

https://www.youtube.com/watch?v=dip6LPpZ5Uk
From 4:00 or 4:40 minutes.
Or use the factory SOLDER BRAID to clean the soldering area - it will be immediately visible.

Hello. I asked you as I saw in a different thread that you have this board, so maybe you can take a look or a picture to the connections.
And Measure the upper side connections ( no need to unsolder the cap).
Right now I have no Flux nor braid, just the soldering tool and tin (with internal flux), I ll buy some this week or next weekend

right now I m not even sure 1 was originally connected to A in the backside

kinetix wrote on 2023-09-25, 16:06:

I think your TC39 is my TC40.
Can you check that one, is the square point connected to the inductor?

I checked TC39 and it's as you wrote, square point is connected to the inductor.

elcrys wrote on 2023-09-27, 19:46:

I checked TC39 and it's as you wrote, square point is connected to the inductor.

thank you very much
now I need to be sure how the round one is connected. hope some with rev2.2 can tell me

2 is connected to B, to coil L10 and mosfet Q8.
1 to A.

Nexxen wrote on 2023-09-27, 21:33:

2 is connected to B, to coil L10 and mosfet Q8.
1 to A.

1 to both sides As?

kinetix wrote on 2023-09-27, 23:21:

Nexxen wrote on 2023-09-27, 21:33:

2 is connected to B, to coil L10 and mosfet Q8.
1 to A.

1 to both sides As?

Dude, 1 is connected to ground... 😀
The + just next to the cap circles indicates polarity and means "Voltage in" and is connected to C42 to 46 (the capacitors next to the cpu socket)

Q18 and 17 are connected to that huge plane and are GND. For that lead of the capacitor to be connected to ground is correct, otherwise you'd have a huge kaboom!

Relax and be done with this. It's all OK.

Again, could you upload high definition pictures of your motherboard.
Top and bottom, thanks.

Nexxen wrote on 2023-09-28, 00:50:

Again, could you upload high definition pictures of your motherboard.
Top and bottom, thanks.

here you go.
I had to limit resolution, quality and filesize to what Vogons allows.
I horizontally flipped the image of the backside so that it coincided with the topside and was easier to analyze.
The green capacitors are the ones I changed, in addition to TC39 and TC40
It is interesting that the C116 capacitor appears to have never been included on the board. In the images that I have found this capacitor exists, but apparently it was not included in this board when it was manufactured. The solder points have tin but the "drop" and the holes are too regular to have been removed latter.

mmm, maybe I'll send to "theretroweb" some better images than the ones they have of this motherboard

kinetix wrote on 2023-09-28, 16:53:

Nexxen wrote on 2023-09-28, 00:50:

Again, could you upload high definition pictures of your motherboard.
Top and bottom, thanks.

mmm, maybe I'll send to "theretroweb" some better images than the ones they have of this motherboard

Yes, yours are way better. They'll be happy to oblige.
I can't see anything obviously busted, your board looks close to mint new.

well, it boot!!!
right now!
At the beginning, after putting in the last capacitor it didn't start, no image.
I checked the configuration switches (according to the stencil on the board) and tested at 3.3 and 3.4 volts. nothing.
tried another Pentium 133. nothing.
I changed the memory stick, et voila!!
The bad thing is that it is only 16MB (16MB PC66). The others I have and tested with are PC100 and PC133, in theory compatible, but either they are bad, or the board doesn't "like" them. I thought the higher speed ones would work at a lower speed. Maybe the pentium133 is the culprit and they do work with a CPU with a higher FSB (ie the K6-2/450 I got with the board). anyway, I have to test them on another PC (a PIII Compaq Deskpro EN that I have here and recently restored). everything else I have available are 72 pin FPM and EDO sticks.
I'll look for some additional compatible ram sticks.

the bios is the 1.3 version, so maybe, if in the end it works correctly, maybe I will update the bios to its latest stable version.

Now I´ll continue working on it because it is stuck in "Checking NVRAM". maybe is that ramstick, or the bios. keep working

mmm, it doesn't have a jumper to clear the CMOS