GA-990FXA-UD7 No 3.3v

[taxxin]

Dear members,

I have a nice looking Gigabyte FM3+ 990fx board here.

Symptoms: Plug the board in and no response from power switch.

Initial inspection: Everything seemed fine apart from

1. Low resistance at the memory 1.5v inductor coil (10 ohms)
2. Low resistance at the PCH power supply near the bottom center of the board (less than 5 ohm)

So anyways when turning on the power at the psu, i found that the 3.3v near the pch area was not present. measuring at the battery positive terminal there there was no voltage.

Checked around and burned my finger on the LDO voltage regulator which is part (L1117LG) 1A Fixed and Adjustable Low Dropout Linear Regulator (LDO). it powers the PCH hub (southbridge chip).

So I thought it was shorted so I removed the part. I checked the resistance of the output and it was reading less than 4 or 5 ohms. I powered up the output of the circuit with 1.2v through voltage injection and it seems to power up without a short drawing about 400 500 ma. nothing got hot apart from the chip. it also went to the battery circuit because the positive terminal of the battery also read 1.2v. I didnt think to increase the voltage because at the time i wasn't exactly sure which power rail it was.

I replaced the LDO reg with another one from an old motherboard i had and decided to fire it up. as soon as i flick the PSU switch the LDO began to heat up rapidly. Checked at the battery and it read 1.0-1.1 volts instead of the expected 3.3v also the Output of the LDO also read 1.1v. so I thought something must be shorting?

I iced up the entire area with some camping gas to see which component might be shorting and only 2 components heated up.

1. was the LDO reg, it heats up first
2. the crystal heats up second.

Theres a short video here https://photos.app.goo.gl/JxPx3moFDUv56YRZ7

Nothing else seemed to heat up to indicate any shorted capacitors?

So could it be a short inside the PCH chip? it sure gets hot right away only on stand by power, even before the main power is switched on.

If it was where would be the best place to find a replacement? is it even worth replacing?

[piernov]

Southbridge warming up when injecting on 3.3V rail = dead southbridge.

[taxxin]

Originally Posted by piernov

Southbridge warming up when injecting on 3.3V rail = dead southbridge.

My suspicion exactly was hoping someone who has experienced the same could confirm this. not looking good

[re-atari]

I don't know if it's related to the fault you're experiencing with this mainboard, but photo 4 shows cap EC26 is damaged. It's the solid alu cap situated between the memory slots and the southbridge. It looks like someone squeezed it with a pair of pliers, resulting in dents on 2 sides. Capacitance and voltage are hard to read due to a fuzzy photo, looks like it's a 560uF 6,3V cap.
It might be worth a bit of time and effort to check this cap, if only to exclude it as a possible cause.

[taxxin]

Originally Posted by re-atari

I don't know if it's related to the fault you're experiencing with this mainboard, but photo 4 shows cap EC26 is damaged. It's the solid alu cap situated between the memory slots and the southbridge. It looks like someone squeezed it with a pair of pliers, resulting in dents on 2 sides. Capacitance and voltage are hard to read due to a fuzzy photo, looks like it's a 560uF 6,3V cap.
It might be worth a bit of time and effort to check this cap, if only to exclude it as a possible cause.

Right i see it, that could be causing a short on the memory phase. I will check that out and see if it helps, but i think the problem may lie in the southbridge?

[re-atari]

I don't know which power rail the the damaged cap filters, but if it shorts the power rail to the south bridge to GND, who knows...

BTW: it makes you wonder WTH this cap is damaged this particular way, and fear for the collateral damage this treatment may have lead to.

[taxxin]

Originally Posted by re-atari

I don't know which power rail the the damaged cap filters, but if it shorts the power rail to the south bridge to GND, who knows...

BTW: it makes you wonder WTH this cap is damaged this particular way, and fear for the collateral damage this treatment may have lead to.

I think i know why!,

2 of the RAM clips directly above this cap were removed, my guess is the barbarian who removed them must have use pliers and levered against the poor capacitors.

Im pretty sure people wouldn't start scavenging parts off a board that was working fine. so im guessing that the cap was injured post-mortem. maybe its time to order a southbridge?

[re-atari]

The clips are indeed MIA, I completely didn't notice it. As you say, it may very well be the reason why and how the cap got damaged in the first place.

I'd start by replacing the cap before ordering 'exotic' parts like a southbridge IC. Replacing the southbridge isn't a job for the fainthearted, you need the necessary equipment and experience to do a succesful swap. Resembles BGA reballing.

If you're fixing mainboards on a regular basis, a POST card like a TL460, TL611 or (newest) TL631 might come in handy. You can order the TL460 and TL611 for about $ 30 on AliExpress and the like, the TL631 is a bit more expensive.

[taxxin]

Originally Posted by re-atari

The clips are indeed MIA, I completely didn't notice it. As you say, it may very well be the reason why and how the cap got damaged in the first place.

I'd start by replacing the cap before ordering 'exotic' parts like a southbridge IC. Replacing the southbridge isn't a job for the fainthearted, you need the necessary equipment and experience to do a succesful swap. Resembles BGA reballing.

If you're fixing mainboards on a regular basis, a POST card like a TL460, TL611 or (newest) TL631 might come in handy. You can order the TL460 and TL611 for about $ 30 on AliExpress and the like, the TL631 is a bit more expensive.

Funny you say that, ive just splurged out on one of these cards.

https://www.lazada.co.th/products/i2...s.item_title.1



so far hasnt been useful i guess the boards ive been trying to fix dont even boot up or complete power sequence thats what ive been trying to learn. yet its difficult to diagnose when it keeps shutting off.

Thats why ive been wondering if there is a service mode or something where i can get the power on and test every power rail.

[Repdata]

HI

here is the circuit diagram

[re-atari]

That's a TL460, should come in handy in this repair job. You do need to be aware that these POST cards have their peculiarities. I received my TL611Pro last weekend, tinkered a bit with it since then, mostly just to be able to decypher the Chinenglish user manual.
At first I feared my card was defective, as it didn't do anything when inserted in the PCIe slots on my (fully functional) MSI X370 mainboard. The display just showed '00' with decimal points lit alternating. It turned out on MSI boards the POST card only works properly when inserted in the TPM or LPC Debug connector. On Gigabyte boards, on the other hand, you must insert it in one of the PCIe slots, the card will not work on an TPM or LPC Debug connector.
It took a bit of my imagination to conclude that this was what the 'user manual' meant to say.

I have attached a photo of various TMP & LPC Debug connectors, hope it helps.

@repdata: thanx for the circuit diagram, much appreciated!

[taxxin]

Originally Posted by Repdata

HI

here is the circuit diagram

THANK YOU SO SO MUCH!!



In future I dont mind to pay you for board views and schematics for other boards if you have access to them? let me know if you have anything for MSI B85-43 GAMING, GA-Z77X-UD3H Rev 1.0. Dankeschon amigo

[dicky96]

@re-atari
I also noticed on mine (TL460s) that if you connect it to the TPM the wrong way round you also get 00 and alternating dots. But it does not seem to do any harm connecting it the wrong way on the TPM

@repdata
Yes also big thanks for the schematic

[taxxin]

Originally Posted by re-atari

That's a TL460, should come in handy in this repair job. You do need to be aware that these POST cards have their peculiarities. I received my TL611Pro last weekend, tinkered a bit with it since then, mostly just to be able to decypher the Chinenglish user manual.
At first I feared my card was defective, as it didn't do anything when inserted in the PCIe slots on my (fully functional) MSI X370 mainboard. The display just showed '00' with decimal points lit alternating. It turned out on MSI boards the POST card only works properly when inserted in the TPM or LPC Debug connector. On Gigabyte boards, on the other hand, you must insert it in one of the PCIe slots, the card will not work on an TPM or LPC Debug connector.
It took a bit of my imagination to conclude that this was what the 'user manual' meant to say.

I have attached a photo of various TMP & LPC Debug connectors, hope it helps.

@repdata: thanx for the circuit diagram, much appreciated!

Thank you for the tips!

Yes I am getting 00 as well, i also noticed that you get different responses on GIGABYTE boards if you plug it into the PCIE x1 and PCIE 16x, you dont get much from the 16x socket but in the 1x socket you get codes, i think its because of some cpus dependancy on GPU.

[piernov]

Originally Posted by taxxin

In future I dont mind to pay you for board views and schematics for other boards if you have access to them? let me know if you have anything for MSI B85-43 GAMING, GA-Z77X-UD3H Rev 1.0. Dankeschon amigo

Sale of documentation/information/software is not allowed here. This is a free help forum only. Additionally schematic requests go in the appropriate subforum.

[dicky96]

@taxxin
I have GA-Z77X-UD3H schematic, about 10 different revisions

[taxxin]

Originally Posted by dicky96

@taxxin
I have GA-Z77X-UD3H schematic, about 10 different revisions

RICH!, Muchos gracias hermano!!

[taxxin]

I failed.....


So I got the chip in the mail today, I was prepping to get it done.

So my BGA setup here is a $3 SMD heater element with 300w worth of chinese fury. For the top i have a reflow station here 858D which is 700w of the same kind of chinese fury.

So anyways I thought i might be smart and try it out with another board, it was a 970 asus with blown mosfets and burnt board. so ive placed my element underneath the board (not touching just close by) It got the surface temperature to 100Deg g (checked with my IR thermometer) cranked up my hot air to 500deg (which i think is actually F) covered the thing in flux and started on it. took quite a while but managed to reach 200deg c according to my thermometer. i started pulling up on the chip at the corner with my tweezers ever so slightly and it started to come off!. unfortunately for me some of the traces up in the corner also came off the board.

I read the chip and low and behold, it was the exact same chip i ordered!.

so anyways i thought to myself lets try it on the real thing. So i did everything the same again, this time the board only heated up to 85deg (its a gaming board) so I already had a bad feeling. I got impatient i started gently pulling on the corners of the chip, i also noticed that the chip wasnt sitting flat anyways, one of the corners were higher than the others which probably indicated that the board was probably slightly warped. the chip was coming off nicely, i would have to move the hot air to a different spot until it started to lift then proceed to the next spot and so on, right on the last corner i heard it tear, it was 3 of the trace pads lifted right up off the MB. FUCCKKKKKK!!!

so anyways i knew it was gone at that point, but I thought hey it would be good practice anyways, so i decided to go ahead knowing that it would be fail. I followed Dicky's instructions on his youtube videos about cleaning up the pads and solder wicking it all up and so on.

proceeded to start with the process. heated up the board from the underside again and covered it in flux, then popped the chip up ontop, waited to for the temp to get up and started blasting with hot air again, flux started bubbling and soon after i hear a bacon frying sound. the damn chip POP CORNED ON ME. I knew those chinese bastards were just too humid for their own good. anyways i kept going, it seemed like the balls were fully melted, but the corner was still higher than the rest, i thought id be smart and try to push down on it, when i did sodder squirted out from under the chip. i tried my best to heat it up more and push the chip around to try and break any potential sodder bridges underneath.

anyways it turns out that the new chips come with leaded sodder balls? interesting, i had no idea.


So lessons here are:

1. use proper bga tools dont be cheap.
   
2. Always heat up the chips and boards moderately first to try and remove any moisture from inside the pcb even if its brand new. All 3 of the chips that i have all popcorned, perhaps the hot air might be too hot?
   
3. dont pull up on BGA chips.

Would appreciate any tips regarding this exercise. now i just have to sell the board for scrap i dont think it can be repaired anymore.