GA-BX2000 Recap: identify capacitors, voltage regulator

[rattlesnake]

When I got this motherboard it had some bulged capacitors but worked. One day, after some intensive CPU testing, it died peacefully.

Since I enjoy learning about electronics and repairing things, I decided to recap it. Got a 60W soldering iron and a few other things and I took note of the position and values of the capacitors on the motherboard:

10 1200uF 6.3v
19 330uF 25v
3 22uF 16v

But after desoldering all caps, I noticed I made a mistake, as one of the 3 smaller capacitors is 4.7 uF/50v instead of 22uF. I guess I didn't see clearly before desoldering. Now I wonder where this 4.7uF/50v is supposed to go.
My best guess based on some (not-so-clear) pictures I took before desoldering is that the 4.7uF 50v capacitor goes in the middle of the other two, i.e:



Can somebody with this motherboard please confirm this?

So if this is right, the correct count of the aluminum electrolytic capacitors is:

10 1200uF 6.3v
19 330uF 25v
2 22uF 16v
1 4.7uF 50v

I have another issue, this one apparently is the definitive issue.
I noticed the capacitors near the voltage regulator in the center (LINFINITY LX8383A - 00CP PHIL 9935C), (datasheet) were bulged, and the bottom of the PCB right under the voltage regulator looks discolored. So maybe this means the LX8383A is bad too? or this discoloration maybe was caused by the bad caps nearby?




If the voltage regulator is bad, what replacement should I buy? What are the alternatives?

I tried to understand the datasheet, I noticed this isn't a fixed voltage regulator and it needs a capacitor. But honestly I couldn't understand much.

Thank you for sharing your knowledge

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

But after desoldering all caps, I noticed I made a mistake, as one of the 3 smaller capacitors is 4.7 uF/50v instead of 22uF. I guess I didn't see clearly before desoldering. Now I wonder where this 4.7uF/50v is supposed to go.

Don't worry about it and just use the 22 uF capacitor in its place. Most of those 22 uF and 4.7 uF capacitors are just general filters for the PSU's 3.3V, 5V, and 12V rails. You could probably even use standard 16V/100 uF/5x11 mm caps everywhere in place of those.

As for the 330 uF 25V capacitors... again, you don't have to use exactly 330 uF or 25V rated capacitors. The only reason this was done on old motherboards was because low ESR capacitors were expensive and not easily available back then, so manufacturers often used general purpose or entry-level low ESR caps with a higher voltage rating. The higher voltage rating makes the capacitor bigger, and the capacitor ESR and ripple current are directly proportional to the cap size (ESR decreases and ripple current capability increases with capacitor size). So you could also use 16V/ 470 uF capacitors in place of those 25V/330 uF caps.

Before replacing anything though, you should first look at the specs of the original capacitors. The capacity (uF) and rated voltage (V) is not enough. The other two important specs of a capacitor - at least for computer gear - is the ESR (impedance) and ripple current. To find those, you need to also post what brand and series the capacitors are, as well as the dimensions (diameter and height in millimeters). This will enable us/you to look for datasheets for the caps and thus find the specs of the original capacitors, after which you can choose proper replacements.

So for example, the green capacitors close to the CPU socket are Sanyo (now Suncon). Sanyo has several popular series that are used on motherboards: SE, WX, WG, and WF. This is an older motherboard, so those are probably either SE or WX (please confirm if that's the case). Now Sanyo is a good quality Japanese brand in general, so you don't need to replace any of the Sanyo caps. But the black capacitors with the "Y" vent on top appear to be a cheap Chinese/Taiwanese brand, so you should probably replace all of them. Again, however, we need to know what brand and series they are so that we can try to find proper replacements. (Though if I had to guess, these are probably entry-level low ESR, so anything like Panasonic FC, Nichicon PW/PS/PM/PJ/HE, Chemicon LXY/LXV/LXZ/KY, and Rubycon YXJ/YXG should work as replacements for those.)

I have another issue, this one apparently is the definitive issue.
I noticed the capacitors near the voltage regulator in the center (LINFINITY LX8383A - 00CP PHIL 9935C), (datasheet) were bulged, and the bottom of the PCB right under the voltage regulator looks discolored. So maybe this means the LX8383A is bad too? or this discoloration maybe was caused by the bad caps nearby?

It could well be that the bad caps caused the linear regulator to run hotter than normal. But it's also not uncommon to see linear regulators like that to run very hot on motherboards and discolor the area around them. Most motherboards nowadays are black or dark brown, so you can't really see it. But it still happens and it doesn't mean the regulator is bad - just crappy design on the manufacturer's part. I have seen many motherboards over the years with discolored regulators like that, so I won't conclude that yours is bad just yet. Try the recap the motherboard first and see if it works. If not, then we will get to the regulator and test it and/or replace it, if necessary.

By the way, that's a great classic slot motherboard you got there! I'm glad you decided to save it. Once recapped with good Japanese caps, these older motherboards are pretty much indestructible.

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Thanks for the swift reply momaka!

I knew about the ESR, the temperature (all must be 105°), but I wasn't fully aware of the importance of the ripple current.

This is all the written data from the previous capacitors and their measurements:

1200uF 6.3V / SANYO / 105°C / S.E.97 / D X. | size 9x15mm

330uF 25V / CHOYO / XR 105°C / 3 5 A | size 7x14mm

4.7uF 50V / CHOYO / MR 105°C /2 0 A | size 4x7mmm

22uF 16V / CHOYO /MR 105°C /1 3 A | size 4x7mm

I have a capacitance meter and tested all the extracted ones to confirm how good/bad they are. Probably I could use the SANYO 1200uF 6.3V because they still have a low ESR, but most of them show these values:

1. C=1.10mF, ESR=.05 ohm
2. C=1.12mF, ESR=.03 ohm
3. C=1.11mF, ESR=.04 ohm
4. C=1.11mF, ESR=.03 ohm
5. C=1.10mF, ESR=.02 ohm
6. C=1.10mF, ESR=.00 ohm
7. C=1.11mF, ESR=.00 ohm
8. C=1.09mF, ESR=.00 ohm
9. C=1.10mF, ESR=.00 ohm
10. C=1.05mF, ESR=.02 ohm

Aren't they a little low on the capacitance?

So far I bought these ones for replacement:

1200uF 6.3V / JACKCON / -40+105°C / 02 08A LOW E.S.R | size=8x16mm , they show these values:

1. C=1.24mF, ESR=.02 ohm
2. C=1.23mF, ESR=.03 ohm
3. C=1.24mF, ESR=.03 ohm
4. C=1.23mF, ESR=.04 ohm
5. C=1.23mF, ESR=.04 ohm
6. C=1.23mF, ESR=.04 ohm
7. C=1.24mF, ESR=.04 ohm
8. C=1.26mF, ESR=.04 ohm
9. C=1.25mF, ESR=.04 ohm
10. C=1.24mF, ESR=.03 ohm

Are they appropriate?

I also tested the extracted CHOYO 330uF 25V, they show values like, C=360/365uF, ESR=.05, just two are C=345/ESR=.04 ohm and three are around C=340/ESR=.04 ohm, and of course at least two or three were completely out of spec and I already threw them away, they had different values, I remember one with a ESR=3.5 ohm.

I already bought these for replacement:

JACKCON / 330uF 25V / -40+105°C / 02 09C LHK | size = 8x11mm

I tested some with the capacitance meter:

1. C=343 ESR=.13 ohm
2. C=340 ESR=.13 ohm
3. C=341 ESR=.14 ohm
4. C=331 ESR=.13 ohm
5 C=349 ESR=.13 ohm
6. C=339 ESR=.13 ohm

These ones seem to have a higher ESR. Are they useful for this project or I should save them for a different project and buy different ones? In that case, what should I look for, 330uF or 470uF like momaka said?

About the smaller caps I removed, these are the values I measured:

C=5.15uF ESR=3.0 ohm
C=25.1uF ESR=2.1 ohm
C=24.7uF ESR=2.2 ohm

And these are the replacements:

22uF 25V / HITANO / EHR 105°C | size=4x11mm

1. C=22.2uF ESR=1.1 ohm
2. C=21.5uF ESR=1.2 ohm
3. C=21.3uF ESR=1.2 ohm

Are they appropriate/good enough to replace them?

Thanks!

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

This is all the written data from the previous capacitors and their measurements:

1200uF 6.3V / SANYO / 105°C / S.E.97 / D X. | size 9x15mm

330uF 25V / CHOYO / XR 105°C / 3 5 A | size 7x14mm

4.7uF 50V / CHOYO / MR 105°C /2 0 A | size 4x7mmm

22uF 16V / CHOYO /MR 105°C /1 3 A | size 4x7mm

Looks like your diamter values are off by 1 mm. The standard sizes are 5 mm, 6.3 mm, 8 mm, 10 mm, and 12.5 mm.
So your 9 x 15 mm caps are probably 10 x 15 mm.
The 7 x 14 should be 8 x 14 mm.
And 4 x 7 is probably 5 x 7 mm.

I have a capacitance meter and tested all the extracted ones to confirm how good/bad they are. Probably I could use the SANYO 1200uF 6.3V because they still have a low ESR, but most of them show these values:

1. C=1.10mF, ESR=.05 ohm
2. C=1.12mF, ESR=.03 ohm
3. C=1.11mF, ESR=.04 ohm
4. C=1.11mF, ESR=.03 ohm
5. C=1.10mF, ESR=.02 ohm
6. C=1.10mF, ESR=.00 ohm
7. C=1.11mF, ESR=.00 ohm
8. C=1.09mF, ESR=.00 ohm
9. C=1.10mF, ESR=.00 ohm
10. C=1.05mF, ESR=.02 ohm

Aren't they a little low on the capacitance?

No, that capacitance is fine. Electrolytic capacitors usually have a 20% tolerance for their capacity. That means a 1000 uF capacitor can read as low as 800 uF and as high as 1200 uF and still be considered in spec. Likewise, a 1200 uF cap can read as low as 960 uF and as high as 1440 uF. ESR-wise, your Sanyo capacitors appear okay as well, so I would keep them.

So far I bought these ones for replacement:

1200uF 6.3V / JACKCON / -40+105°C / 02 08A LOW E.S.R | size=8x16mm ...
Are they appropriate?

Ugh, Jackon - cheap caps. I would use them only if I had bulging/leaking capacitors that are out of spce and nothing better to replace them with.

Definitely keep the Sanyos.

I also tested the extracted CHOYO 330uF 25V, they show values like, C=360/365uF, ESR=.05, just two are C=345/ESR=.04 ohm and three are around C=340/ESR=.04 ohm, and of course at least two or three were completely out of spec and I already threw them away, they had different values, I remember one with a ESR=3.5 ohm.

I already bought these for replacement:

JACKCON / 330uF 25V / -40+105°C / 02 09C LHK | size = 8x11mm

Choyo is know crap cap brand. But replacing them with Jackcon.... I don't know. Both have equally poor chance of being a proper permanent fix.

See if you can find better capacitors in your country, like Panasonic, Rubycon, Nichicon, and United Chemicon. Sanyo is okay too, but you're very likely to get counterfeits (they are probably the most counterfeited brand).

I'm not sure what is available in Argentina in terms of online stores, but in North America, we have Digikey.com and Mouser.com. Badcaps.net also has a store that can ship caps internationally, but it's $14 USD for international shipping, if I remember correctly. Other options include Farnell/Element14 and RS Components, but I think those are more popular for Europe. Finally, there is also the seller egekecu on eBay with supposedly genuine Japanese caps, but I've never shopped from him so I can't confirm that.

These ones seem to have a higher ESR. Are they useful for this project or I should save them for a different project and buy different ones? In that case, what should I look for, 330uF or 470uF like momaka said?

If you can find better caps, then YES - save these for a different project (hopefully one where the ESR doesn't matter too much, like linear power adapter filtering). You can look for either 330 uF or 470 uF caps, whichever is easiest to find for you. There won't be any harm going to 470 uF capacity.

About the smaller caps I removed, these are the values I measured:

C=5.15uF ESR=3.0 ohm
C=25.1uF ESR=2.1 ohm
C=24.7uF ESR=2.2 ohm

And these are the replacements:

22uF 25V / HITANO / EHR 105°C | size=4x11mm

1. C=22.2uF ESR=1.1 ohm
2. C=21.5uF ESR=1.2 ohm
3. C=21.3uF ESR=1.2 ohm

Are they appropriate/good enough to replace them?

Yes, those should be okay.
Like I said, those small caps are usually general filters, so just about anything will do. In reality, you probably don't even need to replace the original ones, as they don't see to be that bad in terms of ESR. For small 5x11 mm general purpose (GP) caps, 1-5 Ohms ESR is usually within spec. But if you already have replacements, you might as well go with them.

That said, I suggest keeping the original caps from this motherboard, just in case (or at least keep the non-failed ones anyways).

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Thank you very much for all your help and guidance momaka. I really appreciate it a lot!

I was looking for better caps this afternoon, I went to three major electronics stores here, all I could find was 470uF HITANO, which aren't low ESR -the seller told me-, and in another store I found some of the brand REC that I wasn't careful enough to examine in detail. Is this brand good/better?

IF you tell me the REC are definitely better I'll go buy them. I'll also keep looking for better capacitors, maybe another option is desoldering the caps from broken motherboards? I could buy used/discarded mobos.

I was also reading about capacitors: your replies, the web, etc. So you say the smaller caps (22uF 25V) are for filtering the PSU 3.3v, 5v and 12v rails, how does this happen? and what about the other capacitors? What are they for? Stabilizing the current (or is it voltage?) for the ICs? Do you have any reading material that you can recommend me? Maybe something with useful examples I can try and have fun.

Tomorrow I'll recap.

One more thing, I have this dead ASROCK K7S41GX motherboard that has 14 1000uF 6.3V / TMZ / (M) 105°C / 3 (k) / D b | size=8x11mm capacitors. If I desolder these, and test them, and they are fine, should I use them? But I'll be short of 5, and now another question for a change: Is it a wise idea to mix brands and values? How can one tell?

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

No problems, you're welcome.

I was looking for better caps this afternoon, I went to three major electronics stores here, all I could find was 470uF HITANO, which aren't low ESR -the seller told me-, and in another store I found some of the brand REC that I wasn't careful enough to examine in detail. Is this brand good/better?

REC is just another cheap Chinese/Taiwanese brand, just like Jackcon. Hitano are not suitable for motherboards either.

Only these are the good Japanese capacitor brands worth using:
Panasonic, Rubycon, Nichicon, United Chemicon, and Sanyo/Suncon
The only Taiwanese brand that is okay is SAMXON, and that's only if you use their GS, GC, and GD series. Beware of GF (commonly found in LCD monitors and some power supplies) - they go bad quite frequently.

In general, it's unlikely that you will find good low ESR capacitors at any local stores. Most small shops haven't even heard of low ESR capacitors, let alone the difference between them and standard GP caps.

maybe another option is desoldering the caps from broken motherboards? I could buy used/discarded mobos.

Yes, that is definitely a better option if you can get bad motherboards for cheap. In fact, this may be your best chance of getting good quality *genuine* Japanese capacitors if online shopping is not an option for you.

In particular, be on the lookout for newer (Pentium 4/Athlon 64 -era) Abit, Gigabyte, and ASUS motherboards. All of these usually use good quality Japanese caps. OEM PCs, like Dell, HP, Gateway, Acer, and IBM/Lenovo also usually have motherboards with good caps (but not always!)

That said, even the Japanese brands listed above have several series that are known to have problems. So here is what to avoid from them:

1) Nichicon HM and HN series with a date code string H01xx, H02xx, H03xx, H04xx (where "xx" is another 2-digit number.)

2) United Chemicon KZG, KZJ. In particular, definitely avoid the 6.3V/3300 uF KZG (10 mm dia.), 6.3V/1800uF KZG (8 mm dia.), and 6.3V/820 uF KZG (8 mm dia.) - all of these *will* go bad regardless if you use them or not. The only KZG and KZJ capacitors that ~somewhat~ okay (at least for a temporary recap) are the 16V/1500 uF in 10 mm dia. and 16V/1000 uF in 8 mm dia.

3) Sanyo WF, especially the 6.3V/1800 uF 8 mm caps.

In addition to that, be careful with Rubycon MCZ - they are good caps but a bit sensitive to heat. If you install them in a really hot spot, they will bulge and leak.

Other than those, the Japanese brands have been pretty solid over the years.

So you say the smaller caps (22uF 25V) are for filtering the PSU 3.3v, 5v and 12v rails, how does this happen? and what about the other capacitors? What are they for? Stabilizing the current (or is it voltage?) for the ICs?

Well, the small caps may not be only used for the PSUs rails. Some of them could be filters to a 1.5V, 2.5V, or 3.3V linearly-regulated rail that's generated by the motherboard (especially if you see any SOT-232, TO-252, and TO-262/D2PAK regulators with markings like 1117, LM317, 1084, 1085, or 1086 near the caps). My main point is: if the caps are filtering a PSU or a linearly-regulated motherboard rail, then the cap ESR and capacity is not too critical to match exactly.

The caps around the CPU, Northbridge, and RAM are usually more important - especially if they are filtering a buck-regulated rail. You can tell if a power rail is buck-regulated, because there will be a coil/inductor nearby with some MOSFETs around it. On your motherboard above, there is only one: it's right above the Northbridge / next to the CPU socket. On older motherboards like yours (i.e. Pentium, Pentium II, and some Pentium 3), the CPU is the most power-hungry part, and is the only one that will use a buck-regulated rail. Just about everything else will be linearly-regulated. On motherboards made after the Pentium 3 era, you may find more buck-regulated rails (i.e. more inductors/coils).

And yes, capacitors are used to smooth the voltage on whatever rail they are filtering. So for example, that motherboard of yours probably gives the CPU a core voltage of 2.50V. If the caps are bad, the voltage will start to drop down when the CPU is under heavy load or even have very big "dips" if the caps are in very bad condition. Because digital circuits communicate with 1s and 0s (i.e. high voltage for a "1" and no voltage for "0"), if the voltage drops down too much or has sudden "dips" and "spikes", the CPU may mis-interpret a "1" for a "0" and vice-versa. That when you can start getting errors and your computer to crash or even refuse to boot.

Do you have any reading material that you can recommend me? Maybe something with useful examples I can try and have fun.

I can't think of anything formal at the moment. Most of the knowledge about motherboards and their power rails was self-taught (just measuring things with a multimeter to see how things are connected).

But I have posted a number of motherboard and video card threads where I show a break-down of the power rails on those devices. So here are just a few:
https://www.badcaps.net/forum/showthread.php?t=63041
https://www.badcaps.net/forum/showthread.php?t=63211
https://www.badcaps.net/forum/showthread.php?t=63069
https://www.badcaps.net/forum/showthread.php?t=63163
https://www.badcaps.net/forum/showthread.php?t=63164

One more thing, I have this dead ASROCK K7S41GX motherboard that has 14 1000uF 6.3V / TMZ / (M) 105°C / 3 (k) / D b | size=8x11mm capacitors. If I desolder these, and test them, and they are fine, should I use them?

Oh, I forgot to mention... TMZ and TMV is another series from United Chemicon that are known to go bad. They are not as bad as KZG, but I would *not* use them in place of the Sanyo caps around your CPU.

You probably could use them in place of the two bulged Choyo 25V/330 uF cap next to that hot regulator, but don't expect them to last too long (a year or two if you are lucky). Also note that they are only rated for 6.3V, while the original Choyo are 25V. In general, you should not replace caps with ones that have a lower voltage rating. BUT, if the voltage that the new caps are going to see is less than what they are rated for, then you should be fine. If I had to guess, that hot regulator is probably either for the Northbridge, RAM, or AGP bus, so you are unlikely to find more than 3.3V on its output and more than 5V on it's input. Hence why 6.3V caps should work in that spot.

But I'll be short of 5, and now another question for a change: Is it a wise idea to mix brands and values? How can one tell?

Yeah its fine. If the caps meet the specs of the originals and/or the circuit, then you can use whatever you got on hand.

[Wester547]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Only these are the good Japanese capacitor brands worth using:
Panasonic, Rubycon, Nichicon, United Chemicon, and Sanyo/Suncon
The only Taiwanese brand that is okay is SAMXON, and that's only if you use their GS, GC, and GD series. Beware of GF (commonly found in LCD monitors and some power supplies) - they go bad quite frequently.

Samxon is actually Chinese. Beware of the GK series too. I would add Taicon to the list of somewhat acceptable brands. BTW, “S.E” (as this thread has shown) is a factory code for Suncon, not a series. DX, AX, GX, CG, CA, and HC are several other series popular on older motherboards, along with certain “undesignated” series.

1) Nichicon HM and HN series with a date code string H01xx, H02xx, H03xx, H04xx (where "xx" is another 2-digit number.)

I would add early 2005 datecodes to that, just for peace of mind, and 2005 dated Nichicon HZ as well.

2) United Chemicon KZG, KZJ. In particular, definitely avoid the 6.3V/3300 uF KZG (10 mm dia.), 6.3V/1800uF KZG (8 mm dia.), and 6.3V/820 uF KZG (8 mm dia.) - all of these *will* go bad regardless if you use them or not. The only KZG and KZJ capacitors that ~somewhat~ okay (at least for a temporary recap) are the 16V/1500 uF in 10 mm dia. and 16V/1000 uF in 8 mm dia.

I’ve seen almost every 6.3V and 10V rated KZG and KZJ go bad on the shelf, and some 16V rated ones too. The ones that don’t bulge and leak of their own volition, tend to go out-of-spec without signs. Both series, I’d avoid no matter what.

3) Sanyo WF, especially the 6.3V/1800 uF 8 mm caps.

The 820uF 6.3V 8mm WFs are notorious for bulging too... heck, the whole series is regardless of case size or capacitance/voltage rating.

In addition to that, be careful with Rubycon MCZ - they are good caps but a bit sensitive to heat. If you install them in a really hot spot, they will bulge and leak.

Rubycon MFZ is known for that “sensitivity” too.

Other than those, the Japanese brands have been pretty solid over the years.

Minus the quaternary salt electrolyte capacitors.

Oh, I forgot to mention... TMZ and TMV is another series from United Chemicon that are known to go bad. They are not as bad as KZG, but I would *not* use them in place of the Sanyo caps around your CPU.

I would say that 680uF 4V TMV and 680uF 4V TMJ are bad (those too will vent in storage), but I’ve yet to see evidence that TMZ is a truly bad series or even ultra low impedance. They are used interchangeably with Rubycon YXG, LTEC LZP, OST RLS and RLP, and Taicon HD on older ASUS and ASRock motherboards (and IIRC, interchangeably with Teapo SC and Rubycon YXG on older Abit motherboards too), and I’ve never actually seen one bulge before, just a few either leak from the bottom seal or dry out. I think they are okay if you keep them cool with exceptions. The datecode on that 1000uF 6.3V TMZ looks to be December 28th, 2003, made in Indonesia. Being ~14 years old now, it does appear to be fairly long in the tooth and there may be better alternatives.

But yes, the first order of business is replacing the bulged Choyo capacitors on the motherboard. That will likely restore the motherboard’s stability.

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Gentlemen, it doesn't boot.

I get nothing, it's exactly like before. I use a POST PCI/ISA card and I get nothing. At least the CPU fan is spinning, but that's it.

I was extremely careful in my work, I cleaned all the caps with solder wick, re-measured them, I cleaned all the holes, the legs of the used capacitors, re-measured them, soldered, cleaned the flux, I was very meticulous, delicate, I was a craftsman. Yet it failed silently. No explosions.

I soldered the SANYO 1200uF 6.3V back, replaced the CHOYO 330uF/25V with the JACKCON 330uF/25V, and replaced the small capacitors (4.7uF 50V, 2x22uF 25V) with the HITANO 22uF 25V.

What is the next step? What is worth doing to fix this motherboard or at least to finish the diagnose? By the way, the PSU is OK, the CPU is OK (they work well with other motherboards).

Thank you for the resources and all the data about capacitors. I'll save it and consider it a bible. I'll see if I can get a good deal getting discarded motherboards. I found a guy that sells used working P4 or Socket 462 motherboards for a good price.

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Gentlemen, it doesn't boot.

I get nothing, it's exactly like before. I use a POST PCI/ISA card and I get nothing. At least the CPU fan is spinning, but that's it.

I was extremely careful in my work, I cleaned all the caps with solder wick, re-measured them, I cleaned all the holes, the legs of the used capacitors, re-measured them, soldered, cleaned the flux, I was very meticulous, delicate, I was a craftsman. Yet it failed silently. No explosions.

I soldered the SANYO 1200uF 6.3V back, replaced the CHOYO 330uF/25V with the JACKCON 330uF/25V, and replaced the small capacitors (4.7uF 50V, 2x22uF 25V) with the HITANO 22uF 25V.

What is the next step?

My assumption from this description is you tested the motherboard with all the caps installed, correct? Because if you test the motherboard without any capacitors installed, that can cause damage. Just making sure I point this out.

Anyways, with the capacitors replaced, now it's time to check all the regulators on the board, especially that overheated one (the LX8383A-00cp).

According to the LX8383A-00cp datasheet, pin 1 is Adjust/Ground, pin 2 (and also the tab) is V_out (the output voltage), and pin 3 is V_in (the input voltage coming to the regulator.) With the motherboard powered on and everything connected correctly, use a multimeter to measure the voltage between V_in and ground (any metal part in your computer case, but NOT pin 1 on the regulator). Then do the same for the V_out pin: measure the voltage between it and ground. Report back what results you get here.

There appears to be two more Linfinity regulators, so we need to check those as well. One of them can be found at the upper-right corner of the motherboard, next to the RAM slots. The other is below the CPU slot and right next to the AGP slot. Test both of them the same way you test the LX8383A-00cp mentioned above and post back what results you get. This should tell us if all of the voltages going to the various motherboard components are correct.

What is worth doing to fix this motherboard or at least to finish the diagnose?

Depends on how much you value your time and the motherboard.
If you do have the time and you enjoy doing this, then you might as well attempt everything reasonable to troubleshoot it.

I usually give up on motherboards when the problem is not power-related nor blown ethernet chip (due to a thunderstorm - somewhat common issue), of course after troubleshooting everything else (like trying known good RAM, CPU, PSU, CMOS battery, and reset CMOS).

Speaking of which, have you tried resetting the CMOS? I see this motherboard has dual BIOS. Have you tried booting from the second BIOS? What does the POST card say with the second BIOS? What about the jumper/switch settings for the CPU/frequency/host clock - tried setting those to "stock" / minimum?

Minus the quaternary salt electrolyte capacitors.

Yeah, those too.
But I think you're unlike to see those on a motherboard, so I didn't mention them.

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

My assumption from this description is you tested the motherboard with all the caps installed, correct? Because if you test the motherboard without any capacitors installed, that can cause damage. Just making sure I point this out.

Yes, I never turned it on when it was "cap-less".

Anyways, with the capacitors replaced, now it's time to check all the regulators on the board, especially that overheated one (the LX8383A-00cp).

According to the LX8383A-00cp datasheet, pin 1 is Adjust/Ground, pin 2 (and also the tab) is V_out (the output voltage), and pin 3 is V_in (the input voltage coming to the regulator.) With the motherboard powered on and everything connected correctly, use a multimeter to measure the voltage between V_in and ground (any metal part in your computer case, but NOT pin 1 on the regulator). Then do the same for the V_out pin: measure the voltage between it and ground. Report back what results you get here.

I'm using the motherboard on a table, there's no computer case. Furthermore, I think that there is no third pin in this electrical installation, that is to say, no ground (if there is something, I think it doesn't go all the way to the ground in a proper way). Anyway, I used the PSU case and it worked! but I'm really interested to know how can this be possible? And why am I to avoid using the pin 1 (GND/ADJ)?

This is what I measured:
8383A
V_out 3.316
V_in 4.99
Accidentally I also measured the voltage between pin 1 GND and the PSU case, it was 2.070v

8384 (near the RAM sockets)
V_out 3.541
V_in 4.746

8384 (near the CPU/AGP slot)
V_out 1.506
V_in 4.998

Full info about the other two voltage regulators:
LINFINITY LX8384-00CDD THAI 9935U, they are surface-mounted. This is the datasheet.

Depends on how much you value your time and the motherboard.
If you do have the time and you enjoy doing this, then you might as well attempt everything reasonable to troubleshoot it.

Yes! Let's go for it. I really enjoy this process as I'm learning a lot and it's quite fun! But it was disappointing not to see any changes after the whole recapping business.

Speaking of which, have you tried resetting the CMOS? I see this motherboard has dual BIOS. Have you tried booting from the second BIOS? What does the POST card say with the second BIOS? What about the jumper/switch settings for the CPU/frequency/host clock - tried setting those to "stock" / minimum?

I have set the CPU Bus to 100 MHz, which is the standard, and the multiplier is locked, but I also set it right. Changing the bus speed to 66 MHz didn't help.
I reset the CMOS many times, I tried it with and without the battery on.
The motherboard has two BIOS chips, from the manual: "2M bits FLASH RAM, Support Dual BIOS", and one can select to use both or a single one, but this function is optional, as I don't have pins where the jumpers JP18 and JP19 should be, so it is permanently in dual BIOS mode. Anyway I went on and closed the contacts where the pins should be with some solder, so it would use only a single BIOS. Nevertheless, it didn't yield a different result and I removed the solder.

The POST card isn't showing anything, it just powers on but the display is fixed at "----". As if there was no CPU. No error code.

I don't think the BIOS is compromised, but you never know, it could be! The question is that removing it is quite hard as it is surface mounted and has a lot of pins, I don't have the proper tools to desolder it, and I don't have a fancy external/USB BIOS flasher.

Mmm... what are your thoughts?

Thank you for the detailed answers as always and Merry Christmas!

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

There are two MOSFETs and this is what is written on them: "NEC / K3296 / 97M".

Maybe they are the culprit? How can we check them?

I think this is the datasheet, right? Is it strange that the name on these doesn't include "2S"?

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

The two MOSFETs are very unlikely to be the problem (typically, they would make your PSU shut down if there was something wrong there).

Regarding the linear regulators: all input and output voltages appear okay. Well, the one near the RAM slots appears to have an output that is a bit high at 3.5V and an input that is a bit low for a "5V" rail. So it does make me wonder a bit if there is something wrong with it (though I highly doubt it).

Unfortunately, I am doing a bit of traveling tomorrow for the next 2-3 days, so I can't give you more to test/check right now. But when I come back, I'll look into your motherboard again. So please be patient if you don't see a response from me for a few days.

[rattlesnake]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

Thanks momaka.

I look forward for more responses with things to do to troubleshoot this motherboard!

Have a good trip.

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

I'm back. Happy New Year.

I'm using the motherboard on a table, there's no computer case. Furthermore, I think that there is no third pin in this electrical installation, that is to say, no ground (if there is something, I think it doesn't go all the way to the ground in a proper way). Anyway, I used the PSU case and it worked! but I'm really interested to know how can this be possible? And why am I to avoid using the pin 1 (GND/ADJ)?

Because the Ground/Adjust pins are not always connected directly to ground (as you saw with the voltage measurements for your first regulator above).

So if you measure the input and output voltages with respect to the regulator's GND/ADJ pin, you may not get an accurate reading.

You are indeed right in noting that most *linear* voltage regulators don't necessarily have a direct connection to ground. But that's normal.

Yes! Let's go for it. I really enjoy this process as I'm learning a lot and it's quite fun! But it was disappointing not to see any changes after the whole recapping business.

Yes it can be disappointing at times when you seem to have tried everything and nothing works. But not everything is fixable, unfortunately. Still, if it's worth it to you, then you should try everything you can, indeed.

That said, looking at the voltage readings of the regulators above once more, again I don't see any issues there. So unfortunately, the bad caps near the regulator in conjunction with you running load tests on the motherboard could have killed the motherboard's Northbridge. Otherwise, I just don't see why it won't work with the new caps you put.

But before declaring the motherboard dead, there's a few more things we can try.

I have set the CPU Bus to 100 MHz, which is the standard, and the multiplier is locked, but I also set it right. Changing the bus speed to 66 MHz didn't help.
I reset the CMOS many times, I tried it with and without the battery on.

Good.

I also suggest testing with only a single RAM module, and try inserting the RAM in and out once or twice more, just in case this is a contact issue. Same goes for the VGA card and CPU.

The POST card isn't showing anything, it just powers on but the display is fixed at "----". As if there was no CPU. No error code.

Interesting.
Did you try all PCI slots? (Assuming your POST card is PCI). If I remember correctly from reading some article a long time ago, some motherboards have PCI slots that may not show anything on a POST card. So try each one, just in case. I find it very strange that the car is sitting with a blank code. Most dead motherboards I have at least display "FF" or "00".

I don't think the BIOS is compromised, but you never know, it could be! The question is that removing it is quite hard as it is surface mounted and has a lot of pins, I don't have the proper tools to desolder it, and I don't have a fancy external/USB BIOS flasher.

Mmm... what are your thoughts?

Yes, I doubt the BIOS is bad, especially considering you have dual BIOS.

But try the above few suggestions to see what we get. They may yield more information about what the problem is.

If nothing else, sometimes I go as far as testing the small transistors and resistors on the board. It's tricky and finicky work and rarely yields any results. But I did sort-of revive one motherboard that was refusing to boot at all otherwise (now it boots only when I plug in and power the PSU at the right moment and change some CMOS settings, then save CMOS).

[Mad_Wolf]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

What are those 2 caps under the agp slot that are marked.
Specs ?

[momaka]

Re: GA-BX2000 Recap: identify capacitors, voltage regulator

What are those 2 caps under the agp slot that are marked.
Specs ?

Looks like 25V, 330 uF.
My guess is they are low voltage and under 5V, so probably can use 6.3V caps.
Could also up the capacitance to 470 uF, 680 uF, or even 820 uF without issues.
But 470 uF @ 10V or 16V would be a very conservative and safe bet.