So, my Antec Neopower 480 has been out of action for a while. Jump-starting it with a staple/paperclip on the green/black pins works, but it just wouldn't boot a machine. So it didn't work basically.

Also for a while, my MSI K8N Neo2 Platinum had lost the ability to overclock my CPU or RAM, even though the 3500+ Venice is capable of a solid 2.92 GHz 24+ Prime95 and the 2x512 G.Skill LE is capable of 308 MHz @ 2.5-4-4-8-1T for 24+ Prime95... both CPU and RAM are capable of higher clocks for benching... that is on my DFI NF4 Ultra-D however...

Anyway, so PSU didn't work, and Mobo was reducing the CPU and RAM to stock speeds! Oh the agony... I couldn't live with myself running a machine at stock and having a once-great PSU out of commission...

Soooo, I looked at the capacitors on the motherboard and the capacitors on the PSU to see if either had any notoriously bad brand caps... and they both did... G-Luxon, Teapo, Fuhjyyu - all really bad brands... and then there were some questionable types: United Chemicon KZG, Panasonic FL

I decided to replace them all with Rubycon and Panasonic capacitors

Here is the listing of the original and the replacement caps for reference purposes:

Antec Neopower 480

C46: Fuhjyyu TMR 1000uF 10V --> Rubycon ZL 1000uF 10V
C45: Fuhjyyu TMR 1000uF 10V --> Rubycon ZL 1000uF 10V
C22: Fuhjyyu TNR 470uF 25V --> Panasonic FM 470uF 25V
C27: Fuhjyyu TMR 2200uF 16V --> Panasonic FM 2200uF 16V
C28: Fuhjyyu TMR 2200uF 16V --> Panasonic FM 2200uF 16V
C32: Fuhjyyu TMR 3300uF 10V --> Panasonic FM 3300uF 10V
C34: Fuhjyyu TMR 3300uF 10V --> Panasonic FM 3300uF 10V
C37: Fuhjyyu TMR 3300uF 10V --> Panasonic FM 3300uF 10V
C39: Fuhjyyu TMR 3300uF 10V --> Panasonic FM 3300uF 10V


MSI K8N Neo2 Platinum

EC37: G-Luxon 100uF 6.3V --> Rubycon ZA 100uF 6.3V
EC39: G-Luxon 100uF 6.3V --> Rubycon ZA 100uF 6.3V
EC45: G-Luxon 100uF 6.3V --> Rubycon ZA 100uF 6.3V
EC56: G-Luxon 100uF 6.3V --> Rubycon ZA 100uF 6.3V
EC35: United Chemicon KZE 100uF 25V --> no change
EC55: United Chemicon KZE 100uF 25V --> no change
EC27: G-Luxon 470uF 6.3V --> Rubycon ZA 470uF 6.3V
EC3: G-Luxon 470uF 6.3V --> Rubycon ZA 470uF 6.3V
EC42: G-Luxon 470uF 6.3V --> Rubycon ZA 470uF 6.3V
EC5: G-Luxon 470uF 6.3V --> Rubycon ZA 470uF 6.3V
EC13: Rubycon ZL 470uF 10V --> no change
EC19: Rubycon ZL 470uF 10V --> no change
EC36: Rubycon ZL 470uF 10V --> no change
EC38: Rubycon ZL 470uF 10V --> no change
EC40: Rubycon ZL 470uF 10V --> no change
EC41: Rubycon ZL 470uF 10V --> no change
EC50: Rubycon ZL 470uF 10V --> no change
EC51: Rubycon ZL 470uF 10V --> no change
EC57: Rubycon ZL 470uF 10V --> no change
EC2: Teapo 470uF 10V --> Rubycon ZL 470uF 10V
EC15: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC31: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC32: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC4: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC8: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC9: Teapo 1000uF 6.3V --> Panasonic FC 1000uF 6.3V
EC12: United Chemicon KZG 1500uF 16V --> Rubycon ZL 1500uF 16V
EC18: United Chemicon KZG 1500uF 16V --> Rubycon ZL 1500uF 16V
EC23: United Chemicon KZG 1500uF 16V --> Rubycon ZL 1500uF 16V
EC6: United Chemicon KZG 1500uF 16V --> Rubycon ZL 1500uF 16V
EC11: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC14: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC17: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC20: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC22: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC24: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC25: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V
EC29: Panasonic FL 1800uF 6.3V --> Rubycon ZLH 2200uF 6.3V


So, here are the results:

Antec Neopower 480

Before: Could not even boot up machine
After: Boots up machine fine


MSI K8N Neo2 Platinum

Before: Reboots during overclocked load from 2.4 GHz upwards at any voltage/setting
After: Passed SuperPi 32M and OCCT (30 min test) @ 2761 MHz


I'm happy

Took 2.5 hours to recap the motherboard, and maybe 1 hour to recap the PSU...

I bought a load of caps at the same time (to meet the minimum order of Farnell.com) so I have caps to do the Tt Purepower 480... that psu works fine, and has done ever since I bought it around 8 years ago, but it can't hurt to swap out the Jenpo and Teapo caps to Panasonic and Nichicon


The Tt Purepower recap looks like it will be a lot easier than the Neopower, because there is hardly any space in the Neopower to begin with, and also the replacement caps were almost all bigger than the rubbish Fuhjyyu caps, so it was a VERY tight fit lol...

Anyway... here's the Tt Purepower listing:

Teapo 3300uF 10V --> Panasonic FM 3300uF 10V
Teapo 3300uF 10V --> Panasonic FM 3300uF 10V
Jenpo 3300uF 6.3V --> Panasonic FC 3900uF 6.3V
Jenpo 3300uF 6.3V --> Panasonic FC 2200uF 16V
Jenpo 2200uF 16V --> Panasonic FM 2200uF 16V
Teapo 1000uF 16V --> Nichicon HD 1000uF 16V


Hopefully this information will be helpful for other people

Good that you got it working again.

I would just like to point out that the UCC KZG's are actually marginally "better", at least in terms of ripple current and ESR, than the Rubycon ZL's you replaced them with! Just something for anyone doing the same thing to keep in mind. For you, I doubt that putting them back would make any noticeable difference to anything

I'll know what's probably wrong should my neo2 get a bit flaky.
I'll check over it the next time I have it out of the case to see if any are looking a bit dodgy.

it was a tough choice to bother replacing the KZGs at all... but the listing of Good caps at Badcaps.net mentions that the KZGs are no longer recommended due to the number of reported failures...

as I don't have a component analyser, I decided to replace every single capacitor that was not on the recommended list

I'm gonna bin the G-Luxon, Fuhjyyu, Teapo etc... all the really bad brands... but the Panasonic and United Chemicon caps I have I'll probably keep until I get the chance to use a component analyser or DMM with capacitance test, to see what's what

Congrats on successfully recapping the mobo and PSU. I'm just a little too scared to try something. I'm scared to even get within an inch of the round things with the copper coiled around them on my mobo...Been shocked before and I definitely didn't like not feeling anything in my hand for a week...

I wouldn't be surprised if they had only actually seen a few bad ones and decided to label the whole series as bad, to be honest I mean, even good caps can fail. But maybe they really are a bad range. Don't suppose you know how many failures they had actually seen?



Fair enough



You could give this a try. Costs pennies to make that and I found it worked well for me. I made one of my usual rubbish posts about it here on OCC some time ago actually!

Shocked on DC voltage? I thought you wouldn't feel that... I can press my hand to the back of a running board...don't feel even a tingle... a PSU I would feel...

Sure you can.

What does that actually do?

I've been on a late shift and my head hurts lol

Aaah sorry. Have a **giant man hug**

It can measure resistance, inductance and capacitance via your sound card with some software. See here

I have a 1 farad DC cap in my car... I bet you'd feel that And in case you don't know, 1F is about the size of a liter bottle of soda. A full discharge could probably take an average-sized person to the ground

The whole point of a capacitor is to hold a voltage for quick discharge, which is why bigger ones can actually be pretty dangerous if not discharged properly first.


This is a cool thread though. My neo2 recently "let loose the ghost" and I'm not sure why. Maybe I'll do a closer inspection of the caps. I don't have any particular need for it at the moment (replaced with a nf4 Infinity), but I'm always looking for a reason to fire up the soldering iron and multimeter

hmm, well here's a lil update...

having issues with extended testing of stability, 2761 MHz failed after about an hour of SP2004, 2750 MHz failed after about 3 hours... gritted my teeth and reduced the speed to ~2700 MHz and it seems to be ok now...

oh well, better than the NO overclock before lol... and certainly ample for the use of the computer (music recording/production)

I also replaced the last 3 opamps of the Audigy2 ZS (all 4 output and 1 input opamps upgraded to LM4562 ) so it's a bit of a retro upgrade all round for that rig really hehe

You really amaze me sometimes... I have a hard time just trying to solder LEDs.

Where'd you get your caps from?

Holy lots of work, Batman! I have no patience for that stuff anymore. I did too much solder, PCB, and component repair in the US Marine Corps.

Farnell

they are one of the few decent component suppliers in the UK

Man, where are the photos? I want to see photos! I'm a photo junkie, what can I say?

Thanks Hardnrg, I guess I'll try this instead of RMA'ing my neopower 480 which died from a bad cap. Make this a sticky as I would rather not use the search tool to find this, and post pics

well, in the Neopower, you could *barely* see the difference because everything is rammed so tight in that psu... and for the K8N Neo2 Platinum yeah I guess you could see the difference, but I didn't think they would be all that exciting as photos... also, after that much desoldering and soldering, I just wasn't in the mood to take pictures lol

I forbid you to make such excuses!

Yes Mistress!

After a couple 48+ hour runs of SP2004, I might take the whole rig apart to take yer bloody pictures...

...if you're lucky!

Don't sweat it tho - they are pretty redundant without before and after shots, but then I suppose it wasn't performed with the intention of being a worklog. I'm just being picky

No worklog - no post! He took the time to individually name all of the caps and the replacements, but can't take pics, Bah-Humbug! Action pics would have been good, though. Some people would actually like to learn this stuff, I'm sure.

Hey nrg, I'll help you out. You take the new pics, I'll take the old My neo2 is sitting on a shelf, so no de-installation necessary! I get the feeling it's in a photogenic mood, too

Apart from having a Zalman chipset cooler, it should suit the "before" pictures nicely

I've never removed a soldered cap, and I'm kind of curious about your process too. Is it hard to maintain enough heat on both pins simultaneously so that it'll come off the board? It seems like if you heat one, and then the other, the first will cool before the second one's hot. Do you have to kind of "shimmy" the pins out little by little, one pin at a time? Details, please!

Once you've de-soldered one leg it should be free from the board. When you de-solder the other leg it should come out pretty easily.

The way I did mine was just to heat one pin and with the other hand gently push the cap, so you're pushing it on the heated pin side and towards the other. This slightly lifts the heated pin out of the hole. Then repeat the action on the other pin. I just did this over a couple of times until one of the legs was fully out of the hole, then you can heat the other pin and just gently pull upwards. Because the leads are short it doesn't really take long at all. I think that's what you mean by shimmy?

As much as people seem to hate the solder suckers, I seem to have had good luck with mine and the majority of the time it will remove all of the solder in the hole so the cap just falls out. I've used a vacuum cleaner once and improvised too... it worked, but it was really quite tricky and is probably not worth the effort

shimmy shimmy is the key yes lol... I figure out which caps I need to desolder, and then look to see which solder points they correspond to on the underside of the motherboard, then I hold onto the capacitor and push it to one side, in line with both legs, so that one leg is effectively being pulled away from the motherboard, then I use a 50W variable soldering iron on maximum and dab it on the solder point until the solder melts and the capacitor leg gets pulled up a bit... then I wait for the solder to cool (this is like a second) and then do the other leg, shimmy shimmy, until it's out

yeah chavalcito, action pics would have doubled the modding time I'm sure lol... I knew it was going to be an arduous mod session anyway, so I didn't take pics

the process is fairly straightforward and is very repetitive, shimmy shimmy, out pops the cap... I did groups of caps because I'd noted down the values beforehand so I wouldn't be looking at the board thinking "arg, what went there?" lol

the replacement process is pretty easy too, I trimmed down the legs of each cap to about 7mm and lined it up to the holes... there is a circle symbol on the motherboard for each capacitor, and the white semi-circle is the negative side which corresponds to the negative stripe on the capacitor itself... if the "holes" are totally clear you can just push the cap through and splay the legs outwards to about 45°... if the holes are partially filled and there is a kinda pit in there, then you can line up the legs into the pits, then heat up the solder points on the underside and the cap will slide into place... if the holes are totally filled then it's a bit tricky and you have to hold the capacitor exactly in place and do the same solder heating from the other side to get the cap to slide into place... so in all situations you get the capacitor body touching the motherboard, and splay the legs out 45° to hold it in place, then solder it in place, trim the legs

it's a very tedious and fiddly mod, but beer and music help, as with pretty much every mod

I've always used solder suckers, I'm not sure why people hate them either.

Yep markie and nrg, that's exactly what I meant by "shimmy"

Sounds like a solder sucker would be the easiest way. Maybe I should invest in one of those someday.

I think active solder sucking heat guns are the best. You don't worry about shimmying anything. You just put the tip over the through hole solder point, count to three, and pull the trigger for a few seconds. You do that to all points and the cap just comes off. It is the same for resistors, caps, DC jacks, it doesn't matter how many contact points there are. The only downside is that you have to heat the point for the right amount of time; too little causes more work, too much burns and won't allow you to re-solder.

It's pretty easy to practise soldering really. Just grab an old bit of kit that you don't mind destroying (I'm sure everyone has old network cards or whatever lying around). Find stuff with the kind of components you want to practise on then go to town removing and reattaching them.

mini update again:

Bios: 1.C Final TCCD (1.CFT.rar) from LejaBeach

11x246=2709 MHz was passing SP2004 after 3 hours, I rebooted to change to the currently working settings, ran SP2004 again and it failed this time after a few minutes lol
11x243=2673 MHz was passing SP2004 after 15 hours, rebooted, running SP2004 again and it's still passing after 1h 24m...

Here are my currently stable settings... I'm fairly confident that these will work for me for a 48+ hour SP2004 run seeing as it's passed for over 15 hours already...

CPU

1.400v x 110%
11 x 243 = 2673
HT Multi = 4x

Ram (2x 512 G.Skill LE TCCD)

tCL = 2.5
tRCD = 4
tRP = 4
tRAS = 8
CPC = 1T
tRC = 12
tRFC = 18
tRRD = 2
tWTR = 4
Drive Strength = Weak

...

so obviously it's a smaller overclock than the 10 x 292 = 2920; ram 292 @ 2.5-4-3-7-1T-7-14-2-3...
but also it's a lot bigger than stock speed haha... I can live with 2.67 GHz for that machine... makes it faster than a 4000+ Orleans (512k L2, 2.6 GHz) so it's not as bad as I first thought

2.7ish is still pretty nice for a neo2. I was a big fan of mine when I first got it, but like you, that kind of fell off when I discovered the DFI nf4 boards.

Out of curiosity though, how much did you spend to recap the board?

I got 2.5 on my Neo2 with a 3200+ Winchester. It didn't stay stable there very long. I was definitely held back by my RAM and PSU. I didn't know different chips in RAM made a difference then, and my PSU got great initial revues, but then after tie it showed that the 3.3v and 12v rails really drooped.

£14.21

Probably not a lot, seeing as he works at Maplin (electronic store)

dah: he got there already!

He ordered from Farnell though. I love Farnell

Maplin don't sell high quality capacitors, only "general purpose" caps... not the type of capacitor you want on a motherboard, psu, audio equipment, etc

Farnell is indeed fantastic... I used to think RS Components were good til I found Farnell... they do a nice range of connectors and stuff too

I is teh n00b! -1 vision of thread

Extremely late post... but well done 'NRG!

Just curious what release/revision of K8N Neo2 platinum you have, because the K8N Neo2 I've got (MS-7025 V1.0) only the sub-470uF components are Teapo (6 x 100uF) with the remainder utilizing primarily Rubycon with UCC

I've since sold on the mobo once I confirmed it could OC the 3500+ (better than ever before by far, within a few MHz of my Ultra-D)... erm, I forgot what I was gonna say lol... here is the cap list:



Click to view attachment

uploaded the Excel file for yours and others convenience


I think I replaced all the red/orange