ECS K7S5A - picture story

[Rainbow]

This is how it looked



Caps removed



Caps



Thoroughly fried MOSFETs



MOSFETs are removed



Diodes' legs disconnected for testing



They tested all OK so they're soldered back



Holes cleaned, new solder added to the pads



Here is one step missing - testing and replacement of the small SMD transistors that drive the MOSFETs

New MOSFETs



The MOSFETs came from this dead Chaintech board

[Rainbow]

New MOSFETs are installed

New caps are installed

Measuring voltage without CPU

Measuring voltage with CPU

And finally it's alive

[Maxxarcade]

After seeing this I think I should recap my K7S5A before it blows the MOSFETs, those are hard to solder. How did you do that? I think when they are done at the factory it is infrared soldered.

[kc8adu]

i heat each leg and lift with a dental pick.then heat the tab.lift with tweezers/hemostats.
i see a bunch of these with bad caps.
its not if but when.

[Tom41]

There's something wrong with this thread for me... I can see the 'thumbnails' in the attachment section properly, but all the inline images are red Xs

[Shroomie]

Me too...actually, I don't even see the Xs (Firefox 1.0.7).

Great work there, Rainbow. Good to see a dead board working again.

[Rainbow]

Looks like the links are to http://badcaps.net instead of http://www.badcaps.net. If you're logged in via both addresses, the pictures will probably work.
I desolder the MOSFETs either by heating the tab first and lifting the back or lifting the legs first - it works both ways for these big ones. The small ones found on some boards can be easily cracked so the back must go first - if you try to lift a leg, it will break off.

[gonzo0815]

i am somhow suppriesed, as the K7s5a i recaped lately has measured the exact same 1,69v instead the exact 1,65 the duron should get. Strange, may be this board is desinged for badcaps, not for god rubys Mcz.
The vrm of this board is very hot, i hope my MCZ will last for it.

[Rainbow]

I measured that at diode's tab - before the inductors. And there is a voltage drop across them. This Duron should have 1.6V (at least I think). The hardware monitor in BIOS displayed 1.619V. Most of boards I've fixed with this type of VRM display exactly 1.600V.

[MD Willington]

K7 Electrical specifications

Want to try the latest 3rd party bios

http://p199.ezboard.com/fk7s5amother...cID=1118.topic

[Rainbow]

I flashed the latest official BIOS already.

[iro]

I got one board with bad caps. After replaceing, board starts but screen dont show anything.

How can i measure (test) MOSFETs and diodes.

[Fizzycapola]

<3 K7S5A I owned one and loved it, cost next to nothing and outperformed alot of the *premium* boards with room to spare, I'm glad to see them still around and being maintained so nicely.

Originally Posted by gonzo0815

i am somhow suppriesed, as the K7s5a i recaped lately has measured the exact same 1,69v instead the exact 1,65 the duron should get. Strange, may be this board is desinged for badcaps, not for god rubys Mcz.
The vrm of this board is very hot, i hope my MCZ will last for it.

Good point there gonzo, I've been wondering about this also, the makers of boards such as ECS know they're going to put caps bad caps on then it posses a question whether one should account this to avoid overvoltage.

[58pacer]

Rainbow, thanks for that great post...you inspired me to try the capacitor switch on my K7S5A. I noticed from the photos that your initial post display shows "CMOS Settings Wrong, Keyboard Error, CMOS Mmory Size Wrong."

I'm getting the same display after my re-cap job. I'm wondering if you had the keyboard connected when this picture was snapped (mine is), and how you resolved the CMOS errors.

I've removed the battery, reset CMOS at JP4, replaced battery with new, and jumpered back to default CMOS settings with no improvement.

I'm so desperate, I wrote ECS tech support in search of any clues.

Just wondered how you resolved the CMOS errors.

Thanks in advance to you or anybody else who could provide any ideas.

[gonzo0815]

Lost CMOS bug is IMHO not cap related.It is one of the more annoying problems, some of this boards may develope.
There are some solutions out there, but IMHO theyre not that bullet proof.
The Keybord issue is bios related, and it should be fixed in the latest versions.

[Rainbow]

CMOS checksum error is normal after battery replacement, BIOS upgrade or clear CMOS. Just go into SETUP and save.
The keyboard was probably not connected as it was first test.

[58pacer]

Thanks, fellas.

My keyboard was connected, but there is no recognition of its presence from the MoBo. I can't get to setup because I can't input anything from the keyboard!

Latest ECS-supplied BIOS is loaded. Had Cheepo-man's before, but went back to ECS to try to resolve the erratic post...it didn't help, so that's when I discovered this website and identified the bad caps.

Sounds like the keyboard *should* respond if everything's been done properly.
My solder joints are crappy! I'm getting a better tip, and going at it one more time.

BTW, Rainbow, I substituted a 1800uf cap for the 1500uf at C425 because I couldn't find the right size to harvest from a dead board. That's the cap right next to the AGP, but it's more than half-way across the length of the board from the PS-2 ports, so I think that machs nix! Just thought I'd mention that in case anybody knows what that cap has influence over.

Another BTW, the initial BIOS screen comes up quickly and reliably after the cap replacement...so much better than before. I know I'm almost there...just have to resolve this keyboard issue.

[58pacer]

Update...

Never ASS-u-me anything...or so the Murphy's law derivative proclaims.

The keyboard that *was* working fine before the cap replacement decided to take this opportunity to go tits up.

I replaced it with an old made-in-Thailand Dell with a bad spacebar, but everything else OK, and the keyboard problem is gone.

In the interim, I had ground down a soldering tip to give me a nice, pointy, usable working surface, and reheated the solder connections to make sure everything was nice and tight.

Onward and upward...thanks for the help!

Gunther

[58pacer]

Latest update...

Success! Booted up to Windows...everything a-OK!

Another successful cap replacement saga for the record.

Gunther

[Scenic]

sorry for posting in this old thread.. but i'm reviving this board (K7S5A V3.1) right now
the board still did POST, but won't power up at the first try, gave me BSODs and so on...
same thing as usual... almost all 6.3V 2200µF G-Luxcraps bulging

i already replaced all of them with OSTs (price matter.. it's going to be used with an old Athlon 1GHz.. and i had some here.. (OSTs.. not athlons.. well.. actually i have 5 of them.. umm... off topic.. ))

looks like pro, doesn't it?



well.. until you see the solderpoints (no photo for now.. 4:20 AM )

as soon as i get the other needed caps i'll finish this job..

any suggestions where to get good caps? i'm running out of them and the next electronics store here only has OSTs not a good choice (at least for my 3 other (more "high-end") boards waiting for a recap..)

looked on eGay/ePay ... some NCCs (Nippon ChemiCon) or Panasonics from Thailand (o.O) or G-Luxcraps, Lelons, Fulltecs (Fullcrap ), ROE (Roederstein (never heard of that before)) or other crapcaps (?) with shipping from germany
as well as some fully noname caps.. -.-