Hipro 250W (Dell) PSU /w Teapo and G-Luxon

[UraBahn]

Got a Dell with a Hipro 250W P-PFC power supply running really hot even under normal load, so I opened it up. Fans working fine. No badcaps on the mobo (the motherboard uses Nichicon PW or something like that) so I checked out the PSU. I was greeted by four Teapo capacitors (including two LXK input caps) and three G-Luxon capacitors. I also found a new cap brand (obviously cheap): 'Asiacon'. None looked bad, but I replaced them anyway (except for the Asiacons, which I don't have any replacements for ATM).

Original Caps / Replacements:

2x Teapo LXK 680uf 200v ---> Rubycon USR 680uf 200v + original Teapo LXK (didn't replace, couldn't get it out!)
2x Teapo SC 3300uf 16v ---> Chemi-con LXJ 3300uf 16v + Panasonic CE-HFQ 3300uf 16v
2x G-Luxon HM 2200uf 10v ---> Nichicon PL 2200uf 16v
1x G-Luxon HM 1000uf 16v ---> Chemi-con LXJ 1200uf 16v

Power supply runs cooler now and I don't hear the fan kick up to high speed as often anymore. Trouble is, I used some relatively old capacitors in the recap so I'm gonna get some new 3300uf 16v caps to replace the ones I put in there. (I think Panasonic's CE HFQ was one-upped by the FC series, but I'm not sure.)

Sorry no pics yet.

[linuxguru]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

I've recapped several of these (model nos. similar to DLP2507...). Try to replace the primary-side 47uF, 50v cap near the auxiliary (+5Vsb) flyback transformer also. I've used 100uF, 35v without any impact.

Asiacon is definitely crap - get rid of all of them, including the 4700/10s and the 3300/16. It doesn't matter if you use slightly lower-valued replacements, like 3300/10 and 2200/16.

[davmax]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

The 10V can be replaced by quality 6.3V caps. This helps keep the diameter smaller.

[kc8adu]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

The 10V can be replaced by quality 6.3V caps. This helps keep the diameter smaller.

ya beat me to it!
that definatly helps in some of these new sardine can units.

[jonnyGURU]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

Why replace Teapos that don't need to be replaced?

AFAIK, the chronic Teapo failures are limited to mobo caps.

[UraBahn]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

Thanks for the replies and suggestions everyone. Every day I learn something new on these forums .

@linuxguru: I had a gut feeling that Asiacon was bad...
I don't have any caps that'll replace the 47uf 50v ATM, I'll have to go get some when I get more caps soon.

@kc8adu and davmax: OK, but I'll have to get some more caps before I can evict the Asiacons.

@johnnyGURU: I replaced the Teapos just to be sure. I'll get some newer caps (I have my sights set on some new-old-stock Panasonics on eBay) to replace the old caps I put in there. Yeah Teapo doesn't go bad too often in PSUs, but they do go bad, and when they fail they like to hide it.

[UraBahn]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

Well well... a few minutes ago I heard a fairly loud *POP* come from either my computer(s) or my Mitsubishi DiamondPoint NXM56LCD TFT monitor. After a few comical moments of panic, I immediately shut down both computers and the ethernet switch and unplugged them.

Seeing as I didn't have the guts to open my beloved flat-panel monitor, I checked my computers, ethernet switch, and even the speakers. The caps in the switch looked fine (anyone here seen Shoei and TOSIN caps??), the computer with the recapped Jetway 993AN mobo and 200W Hipro was OK, but in the Dell's Hipro I found the soldering on one of the Chemi-Con's legs broke!

Guess I didn't solder it as well as I thought I did. How embarrassing

Fortunately nothing looks like it was toasted!

Incidentally I also found a very slightly buldged 'Scatronics' 2200uf 16v capacitor in my speaker's power supply. I replaced it with a Jamicon WG of the same value. (Jamicon are supposed to be OK for audio, right?) I'll try to get a pic of that Scatronics cap in a bit, but right now I gotta unwind.

[kc8adu]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

so we now have a crapcap brand that admits its shit right on the label!

[UraBahn]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

(sorry about the multiple post...)

I finally got my mind together and re-soldered the joint in question. I also replaced one of the Asiacons with a Chemi-con LXJ I just so happened to have lying around (taking extreme care to make sure I soldered the dang thing right this time!). The result...

...it works! Guess nothing was fried in the PSU anyway...

...now to gather enough courage to open my TFT flatpanel monitor...

P.S. I think I found the source of the pop... the 9v Duracell battery in my cheapo Jamcam popped open! badbatteries.net

[Shroomie]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

anyone here seen Shoei and TOSIN caps??

Yeah, I've seen a few Tosins in old (AT) power supplies. Never seen a failed one, but then I haven't encountered many. Watch them, see if they fail, I guess.

[yanz]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

Why replace Teapos that don't need to be replaced?

AFAIK, the chronic Teapo failures are limited to mobo caps.

i still dont' get the best explanation for that assumption. who know if his teapo in his psu getting bad or not without testing them with esr meter? then how many people care to test them?

in my opinion, crap caps in mobo just wont get better in psu. they both placed in area where low esr is required. probably caps in mobo are failling faster because they are closer to the load(?).

but i aware there are many factor that involve in these condition:
-bad caps in psu but good caps in mobo
-good caps in psu but badcaps in mobo

which is the worst..? which is the better?

why with the assupmtion that crap caps in psu are getting bad slower than the same caps in mobo? could it be true?

[linuxguru]

Re: Hipro 250W (Dell) PSU /w Teapo and G-Luxon

In a PSU, the secondary caps are subject to surge currents at a couple of 100 KHz at most. On a mobo, the VRM O/P caps are subject to surge currents at several 100 *MHz* from the CPU - which are not always fully bypassed by the SMD ceramics on the CPU PGA carrier or the mobo.

Assuming that each cycle causes a small amount of irreversible electrolytic chemical change in the capacitor, it is reasonable to assume that VHF stress on the cap is worse than LF stress, other things being equal. The idea of soldering an X7R ceramic below each electrolytic (as suggested by Trodas and others) is an excellent suggestion, both for durability and overclock potential.

I'm warming up to the idea of populating all the empty locations for SMD ceramics, both at the VRM O/P and directly below the CPU socket - I'm satisfied with the first experiments I did with an old S-370 mobo and X5R ceramics cannibalized from a Slotket. No appreciable overclock improvement, but stability seems to have improved a great deal - repeated computation of md5sums on large (multi-GB) ISO files now consistently return the same result, which is an improvement from earlier, when occasional miscalculations were encountered.