TEARDOWN & REPAIR (we hope) of an OCZ MODXSTREAM-PRO 600W ATX PS

[SgtRock]

Greetings EEVBees:

TEARDOWN & REPAIR (we hope) of an OCZ MODXSTREAM-PRO 600W ATX PS - Mainboard marked 2009/16/16. This PS was bought used for $25. After a few weeks the computer shut down suddenly, and I noticed that the case was hot. I immediately shut down the power at the back of the case. After opening it up, inspecting, and allowing it to cool off, I turned it on, and it sparks came out of the fan, and I shut it down again. Removing the PS and taking the cover and fan off I noticed that there was a big blob of white adhesive holding several components together. This white stuff had attracted a rather large dust bunny, all of which was scorched black. So I unsoldered the board, and removed all three heatsinks with MOSFETs, diodes and what all still attached. I then (see pictures below) unsoldered a transformer, and was able to see a burnt resistor and a Tiny-Switch Off-Line Switcher Chip with a burnt off leg. I took over 50 pictures to make sure I could get this one back together. I have included a few below. I also downloaded the data sheets for nearly all of the chips. Build quality was generally good except for the cheap caps and the dust loving white adhesive.

 --Having discovered I could get a replacement chip from Mouser, I proceeded to check the large Teapo 390F, 400V Cap. Lo and behold it had lost a third of its rated capacitance and now read 258uF. All of the other e caps are good, though a few are on their last. I checked the silicon with a DMM, and made sure that the IC chips had no shorts. One of the chips (see below) was plainly marked but I could not find it anywhere on the Net, but it had no shorts so we will hope it never goes. So I am ordering all of the Blown Parts from Mouser and all electrolytic caps, as I just do not trust the Teapo caps.
 
 --I could not have done this job without my Hakko 470 desoldering system. I would so love to see a video of someone trying to do this with a bang stick and solder wick. Good luck! When replacing caps, I leave them soldered in place, unless I have to remove them to read values, and then replace them one at a time to prevent mistakes. I also mark each C101 or whatever when I remove them as well as making an exact list (see below). Also I take a lot of  macro photos. This procedure saved the day when I had to replace 100 odd caps on a Vizio 37" CCFL TV. I will post that job later.
 
 --I am afraid I cannot recommend this OCZ PS. Some of the caps are dated 9/11 which agrees with the mainboard date of 2009/16/16 and indicates this supply lasted at most 4 1/2 years, and maybe much less. But I do not think the large filter cap should have failed. I fault OCZ for allowing WHL (Wun Hung Lo) to assemble this unit using low grade caps. I notice that there are 3300uF Caps the same size and color but with voltages of 16 and 10 volts, and one slightly shorter of 6.3 volts. No chance of a mistake there. I will replace them all with 16 volts, as the smaller foot print 25Vs from Mouser are very expensive. At the bottom you will see the contemporaneous notes I made while deciding whether to repair or toss. I have decided to order the $30 worth from Mouser. Mouser is more expensive than DigiKey, generally, but they usually have better availability for obsolete chips and hard to find capacitors. If this repair fails I will save all the parts and the new capacitors, as I have several of the modular OCZ PSs. If the repair succeeds it should last al least another 10 years. I could by a new one but I am afraid it will blow after the 1 year warranty is up. I will post here the results of the repair attempt, but it may be a while as I have three other teardowns, that I now have the parts for, that have to be assembled first.

--If anyone has a blown Antec high wattage PS they would like to sell me cheap, I would like to do a teardown and possible repair. PM me.
 
DEVICE   COMMENTS
R505      2.18 Ohms Red, Red, Gold, Gold 2.2 Ohms 1 or 2 Watts 10% Burnt. Replace with 3 watt wire wound.
IC501      TNY278PN, TNY274-280 Tiny-Switch III Family. Has leg blown off.   Replace!
T501      Primary 0.00 Ohms, 2nd 1.25 Ohms and 3rd 0.36 Ohms, Probably Good.
U501      L0944 817B X Lite-On Optoisolator is Good.
D101      Good.
D501      FRI07 A08, Cannot identify. Switches on at 0.511 V. FR = 286.3 K Ohms, RR = 0.L, Probably Good.
ZD501      LT9250 P6KE200 = Little Fuse Transient Voltage Suppression Diode 600 W, 171-190V Breakdown. Tests Good.
HS1         GBU1005 = Taiwan Semiconductor Single Phase 10A Glass Passivated Bridge Rectifier. Tests Good
Q601      21N50C3   Good.
Q701      20N60C3   Good. All MOSFETs, Transistors and Diodes seem to be OK.
U201      PK377BN FAN4800IN Fairchild PFC CCM/DCM 64KHz 16DIP. No shorts. Modern replacements are available.
U901      J09A29A1F PS224 16DIP Cannot identify. No shorts. Hope for the best.

ELECTROLYTIC CAPACITORS
Name     Brand   Temp  Value  Volts   ESRin  ESRout   Cond.  ESR of New   Comment
C103        Teapo   105    390       400            Bad            258 uF
C300        Teapo   105   3300         10   0.01
C305        Teapo   105   3300         16   0.08
C306        Teapo   105   3300         16   0.01     0.01
C307        Teapo   105   3300        6.3   0.09    0.02
C311        Teapo   105   3300        6.3   0.01
C313        Teapo   105      2.2         50     2.3
C315        Teapo   105       10         50   0.14                     
C316        Teapo   105       10         50   1.20
C317        Teapo   105      470      25   0.03
C318      Teapo   105      220      16   0.08     0.08
C319        Teapo   105        10      50   0.90     0.87                                                
C320      Teapo   105        10      50   1.0
C?          Teapo   105    3300      10   0.05     0.03          Next to C307 w/Red Dot
C503      Teapo   105        10      50   0.60
C507      Teapo   105    1000      10   0.03
C508      Teapo   105    1000      10   0.05
C515      Teapo   105        22      50   0.35
C703      Teapo   105        47     50   0.29

"Once in his life, every man is entitled to fall madly in love with a gorgeous redhead."
Lucille Désirée Ball
 1911 - 1989

Best Regards
Clear Ether

[SgtRock]

Greetings EEVBees:

--I finally got around to finishing job. The burnt up parts, and all capacitors were replaced. The caps were all over sized and over voltage. Tested in a working computer, this PS is now for testing and my emergency replacement unit.
See pics below.

“If people do not believe that mathematics is simple, it is only because they do not realize how complicated life is.”
John von Neumann 1903 -1957

Best Regards
Clear Ether

[lpc32]

So it works?

What's "ESRin" vs "ESRout"?

[mariush]

I expected to see the tiny-switch failure, it's the 5v standby which runs 24/7.

I fault OCZ for allowing WHL (Wun Hung Lo) to assemble this unit using low grade caps. I notice that there are 3300uF Caps the same size and color but with voltages of 16 and 10 volts, and one slightly shorter of 6.3 volts. No chance of a mistake there. I will replace them all with 16 volts, as the smaller foot print 25Vs from Mouser are very expensive.

OCZ always cut corners somewhere, hence why they went bankrupt. But I wouldn't call Teapo capacitors low grade capacitors, I'd classify them a bit below brands like Rubycon or Nichicon or Panasonic or UCC. It could have been much worse, they could have used noname Chinese capacitors. 
Don't understand why you're surprised about those 3300uF capacitors, they're using 6.3v rated caps for 3.3v rail, 10v rated caps for 5v rail and 16v caps for the 12v rail.  No sense to use only 16v caps, they saved a few pennies when buying caps in bulk and maybe some pcb space (due to smaller diameters).

As for the primary cap... 285 uF is awfully close to 270uF. It could be that the oem manufacturer ordered those caps from a "gray" distributor who printed fake labels. Lots of Chinese power supplies use primary caps that have fake labels with values higher than the real ones.

[SgtRock]

Dear Marius:

As for the primary cap... 285 uF is awfully close to 270uF. It could be that the oem manufacturer ordered those caps from a "gray" distributor who printed fake labels. Lots of Chinese power supplies use primary caps that have fake labels with values higher than the real ones.

I proceeded to check the large Teapo 390F, 400V Cap. Lo and behold it had lost a third of its rated capacitance and now read 258uF.

--Good comments.  In this case I stood the oversized caps taller and was able to fan them out a bit. It runs just barely warm when loaded. It should go for a while. In anything other than an closely enclosed PS like this one, I would not have been so fussy.

--That switch certainly went out with a bang, and there was a lot of dust bunnies around that place where there was a high field charge. Oh well if it blows again I will get the pleasure of removing the liquid tape insulating everything.

"A newspaper is a device for making the ignorant more ignorant and the crazy crazier."
H. L. Mencken 1880 -1956

Best Regards
Clear Ether

[mariush]

In general it's not really a good idea to leave the leads so long. Those long leads can cause problems, you want them to be as short as possible... they can act as antennas, like some tiny inductors.

[SgtRock]

Dear Mariush:

--The next computer I build will probably get this PS installed with 1 SSD, and 4 HDs. It will be a hard working machine and will probably be left on at all times. That should give it a fair working test. Due to the nature of my businesses, I have multiple desktops running constantly. I must say I have very little trouble, considering what I put equipment through.

--Your advice about trace inductance is correct as far as it goes. But, I am guessing, mind you, that trace parasitics are going to be completely swamped in the RF cornucopia contained in the steel shell. In this case you are correct that I went against general circuit rules, but, in the pursuit of lower temperature and longer lasting capacitors. If I am wrong, it will not be the first time. Thanks for commenting.

We turned the switch, saw the flashes, watched for ten minutes, then switched everything off and went home. That night I knew the world was headed for sorrow.
Leo Szilard 1898 - 1964

Best Regards
Clear Ether

[sjc9]

I registered to say thank you. I've been looking for the parts for this power supply. I could not find the identifications of the blown 7 legged IC on the internet until I came across your post. I am going to replace my blown IC501 TNY278PN , thank you again.

[Mysion]

Not too surprising an OCZ died.   OCZ psu's have a reputation..... I have one I got on sale for 30$ witch happened to have very high failure rate. Mine still works but I've just gotten lucky so far. I hope your fix lasts and the caps weren't the only thing they cut corners on. 

[ddcc]

I also encountered a similar failure recently; your post was really helpful in identifying IC501 as another bad part that I had missed. I really wish that I have better desoldering equipment, but so far it's just solder wick and a manual solder sucker for me.

I noticed that my board is slightly different than yours; it is revision A6 from 2010/07/02, which has a 10A 250V fuse (F1) on the input, presumably to prevent the shorting that you and others have experienced. I didn't notice anything strange except the power supply dying overnight, and IC501 even appears normal on the board. This would almost suggest that OCZ (or their OEM) discovered this failure mode in early manufacturing runs, and silently "fixed" the problem with later revisions, without notifying anyone of the problem.

I've also reverse engineered part of the schematic of the board up to the TNY278, excluding the secondary, the PFC controller, and the drive circuitry for Q601 and Q602. It's a bit messy, but hopefully it's readable.

[bithead9]

Not sure if you are in the US, but a decent de-soldering iron is sold by Radio Shack.  I use mine all the time.  See here: http://www.radioshack.com/radioshack-45-watt-desoldering-iron/6402060.html  For $12 US you cannot beat it for ease of use.  I also use flux on the area first to get a nice flow off the board 

[GreyWoolfe]

Not sure if you are in the US, but a decent de-soldering iron is sold by Radio Shack.  I use mine all the time.  See here: http://www.radioshack.com/radioshack-45-watt-desoldering-iron/6402060.html  For $12 US you cannot beat it for ease of use.  I also use flux on the area first to get a nice flow off the board 

I had one of those.  I hated it but used it for several years because I couldn't afford better. Two years ago, SWMBO asked what I wanted for Christmas and I went out on a limb and asked for a ZD-985 based of Dave's review.  Right after the first time I used it, I threw the RS POS away and haven't looked back.

[casinada]

I have the same power supply, I had it on one of my computers and one day stopped working (a little bit over 3 years after I built the system). I replaced it with a spare Antec I had. I put the power supply on the pile of non working things. Today I found it on the pile of junk, connected it to test and all the voltages are present but it doesn't work when I try to power a system. The 390uf capacitor measures 330uf and the ESR is ok. no signs of bad components or smoke signal. I think it is going back to the pile of junk 

[casinada]

After 3 years in the junk pile I was able to repair the power supply thanks to the information found here and the badcaps.net forum. The only thing wrong with my power supply were the Diode  D308 (1N4148) and R315 100ohm. 

[dejan_bin_laden]

Hi guys, i have similiar psu with the almost same desgin as this, it has burned the TNY278PN so i replaced it, then i took of R506 resistor it was showing 37kohms instead of 2ohms so i replaced him also, still no love, then i replaced the big cap 390uF 400V and still nothing, i cheked everything else didnt find anything suspicious, any hints before it goes to garbage ?

Thank you

[jsh]

success story here for me. thanks to finding ddcc's website I was able to fix an old modstream power supply that was in my closet after failing 3-4 years go. the only thing wrong with mine was the big 400v cap(has almost 0 capacitance) and the TNY278(it blew up but didn't seem to damage anything else)