Go Back   Badcaps Forums > Motherboards with Bad Capacitors > Major Manufacturers Boards
Register FAQ Calendar Search Today's Posts Mark Forums Read

Thread Tools Display Modes
Old 01-09-2015, 09:49 AM   #1
Dumb enthusiast
alecjahn's Avatar
Join Date: Feb 2012
City & State: Wichita, Kansas
My Country: United States
Line Voltage: 120VAC 60Hz
I'm a: Knowledge Seeker
Posts: 64
Default Dell Precision 380

Hello friends,

Today's project is a Precision 380 motherboard. It's a "newer" machine, at least I'd consider it to be that (even getting on at 8 years old now...) in comparison to the more commonly plagued Precisions.

Quick side note: I have a Precision 390 as well (Core 2 Duo board as opposed to the 380's Pentium D/4 board and 955 chipset) and It had four blown caps right in the center of the board. It would restart *with error beep* (note!) and was absolutely random on whether it would detect any stick of ram in any slot at any given time. Replaced those four caps and it has been running nonstop for a good month now.
When it would die and beep, it was useless until you removed power (or held the button) and tried again. This also left a log in the BIOS that a "Thermal event has occurred" (if I remember correctly) which was bogus, the thing was spotless and had fresh Arctic Silver everywhere!

This 380 has a similar problem. A machine new to me, I slapped a hard drive in it and popped in the Dell Win XP disk. It didn't even make it past the file loading sequence before a faint click click... and it rebooted.
However, there was NO error message (diagnostic LED sequence) or new item in the diag log (BIOS). Just like you'd hit the restart button (if it had one of those).
Look in the center of the motherboard (nearly identical, again only chipset appears different) and whaddya know - four perfectly-good-looking Rubycon MCZ caps. Wait... Huh?
I have attempted to run Memtest86 (and +) and other memory tests, and it will reboot within a few seconds of the start. Never errors, just reboots.

Power supply appears fine but it is cramped and I can't see the actual cap brands.

Every cap on the board *looks* fine, but of course that might not be the case. Fortunately (or unfortunately, however you look at it) there are nine Chemicon KZG caps, two are located near the 4-pin 12v connector, four are below the CPU but obviously not quite in direct contact with it, and the other three are hiding between some of the PCI/PCI-e slots.

I hear these KZGs are bad. Heck, the post thread below me at this moment is for an XPS with similar cap choices. I also found a blog post a bit ago (and now lost it...) where the writer replaces four KZGs (with nice shiny flat tops, lookin' good) to repair a motherboard (however these were right next to the CPU).
EDIT: I just found one of the bad/replaced caps from the 390 I spoke of, and it is a Chemicon KZJ.

I'm going to look for replacements today and see how that goes. I definitely want to get all of those out of there if they're -that- bad.
There's some other small Chemicons scattered around the board, but they're all KY. I'll leave those be for now.
EDIT: Dur... The two black Chemicons in the top right are KZJ. For some reason it didn't click earlier. Looks like they gotta go, too!

Pictures attached.
Attached Images
File Type: jpg IMG_0196.jpg (1,009.0 KB, 47 views)
File Type: jpg IMG_0198.jpg (719.6 KB, 47 views)
File Type: jpg IMG_0197.jpg (698.4 KB, 45 views)

Last edited by alecjahn; 01-09-2015 at 10:20 AM..
alecjahn is offline   Reply With Quote
Old 01-09-2015, 10:20 AM   #2
Great Sage 齊天大聖
stj's Avatar
Join Date: Dec 2009
City & State: Europe
My Country: some shithole run by Israeli agents
I'm a: Professional Tech
Posts: 20,739
Default Re: Dell Precision 380

did you also look in the psu?
stj is offline   Reply With Quote
Old 01-09-2015, 10:31 AM   #3
Dumb enthusiast
alecjahn's Avatar
Join Date: Feb 2012
City & State: Wichita, Kansas
My Country: United States
Line Voltage: 120VAC 60Hz
I'm a: Knowledge Seeker
Posts: 64
Default Re: Dell Precision 380

Yeah. It looks fine... but hard to really see much in it (packed tight). I need to check the pinouts and see if a generic could be swapped-in for testing (but I seem to recall Dell doing some funky stuff at times).

I also put in a wimpy video card and removed the hard drive/CD-Rom altogether, and booted Memtest from a usb drive, just to try the most limited-power setup, but the same results. Figured that might have been a good enough test for now.

Last edited by alecjahn; 01-09-2015 at 10:33 AM..
alecjahn is offline   Reply With Quote
Old 01-09-2015, 12:04 PM   #4
Send Doge Memes
shovenose's Avatar
Join Date: Aug 2010
City & State: Fairfax, California
My Country: USA
I'm a: Forum Junkie
Posts: 6,127
Default Re: Dell Precision 380

That computer should take a normal PSU the Dell one is just physically larger. But the ATX will screw right in.
shovenose is offline   Reply With Quote
Old 01-09-2015, 05:44 PM   #5
Dumb enthusiast
alecjahn's Avatar
Join Date: Feb 2012
City & State: Wichita, Kansas
My Country: United States
Line Voltage: 120VAC 60Hz
I'm a: Knowledge Seeker
Posts: 64
Default Re: Dell Precision 380

Little update. I did a lot of digging and came up with some cap combinations that were suitable for temporary/test replacement. However, they're not the quality I would want to use permanently.
I swapped them onto the board and was able to install Windows in full on this machine without a problem. Even tossed a thirsty video card and it seems to be solid (remember, earlier, it would die before the Windows DVD was even finished loading to memory). I'm going to call it a day for now, but this has given me enough of a boost to order proper caps. Which I just did from you fine people. Thanks!

I'll chime-in again when they arrive and are installed!
alecjahn is offline   Reply With Quote
Old 03-16-2018, 10:57 PM   #6
Badcaps Veteran
Join Date: Apr 2009
City & State: Indiana
My Country: USA
I'm a: Knowledge Seeker
Posts: 330
Default Re: Dell Precision 380

Sorry to revive this old thread, but I have a similar story of a Precision 390 that I had to swap some caps on.

I have an old Precision 390 that I got pretty cheap as part of a corporate auction. It's a pretty decent machine, Core 2 Duo E6600, 4GB of ECC DDR2, etc. I have it set up on my mess of a workbench down in the basement, where I mostly use it for datasheet lookups, quick calculations, etc. It also takes part in primegrid challenges, and serves as the VNC client for a headless machine that also takes part in the aforementioned primegrid challenges. Every February, primegrid runs a month long prime finding challenge.

About halfway through February, I found that this machine was no longer grabbing work from and reporting results to the server. I went downstairs to find that it had bluescreened with a not-really-that-helpful "hardware failure" message. I restarted it, and it came back up OK and began working right where it left off. I blamed it on a transient power event or something similar.

Two days later, I found it had once again stopped contacting the server. This time I opened the side up, and found a cap on the motherboard puking its guts out. The cap is visible just south of center in this picture:

I then proceeded to disassemble the machine, to get the motherboard out. This cap is a UCC KZJ 6.3V 1800uF, and from what I can tell is on the high side of the 5VSB Southbridge VRM. There are three more just like it; one additional one on the high side of this VRM, and two on the low side. Here's another picture from after I got the board out, that shows them nicely:

I looked up the KZJ specs, and found them to be equivalent to Rubycon MCZ. I looked through my drawers of caps and found a handful of exact replacements - MCZ 6.3V 1800uF. I cranked up my soldering iron, and swapped them out. Here's an after picture:

It occurred to me that since this PC remains on all of the time, there shouldn't have been much load on this particular VRM (it being the standby VRM after all - although all of the heat from the CPU heatsink does blow over these caps - nice job with the BTX design, guys), so I checked on the caps in the power supply. Unfortunately, due the huge tangle of wires, and the inability of me to get the board completely out of the case, there are no pictures. Anyway, I found a slightly bulged OST on the +12V rail, and a mixture of OK looking OST and Ltec caps on the rest of the rails. Back to my cap drawers, I found some suitable replacements, all Panasonic FC. Here are the exact replacements:

1500uF @ 10v OST RLP + 470uF @ 10v Ltec LZP -> 2200uF 10v Panasonic FC + 1000uF 10v Panasonic FC

3300uF @ 6.3v OST RLP + 2200uF @ 6.3v OST RLP -> 2x 2700uF @ 6.3v Panasonic FC

2x 2200uF @ 10v OST RLP -> 2x 2200uF @ 10v Panasonic FC

2200uF @ 16v OST RLP (bulged) + 120uF @ 16v Ltec LZP (yes, that's 120uF, not 1200uF) -> 3300uF @ 16v Panasonic FC + 180uF @ 16v Panasonic FC

680uF @ 16v Ltec LZP + 120uF @ 16v OST RLP -> 680uF @ 16v Panasonic FC + 180uF @ 16v Panasonic FC

You might think that that 12v arrangement is a little strange. I agree. I don't recall the exact configuration, but I believe the two rails did have separate diode packs on the heatsink, and different filters, but were then connected by a shunt. Both rails are rated for 18A. Not sure exactly what they were thinking when the designed that... the power supply model number is L375P-00, and was built by Lite-On. It's seems like a pretty solid PSU, except for the not so good cap choices. To be fair, though, the datecodes on these caps indicate early 2007, so it lasted a good long time before there were any issues.

If I find myself back in the power supply for any reason, I'll take some pics and figure out the exact layout of the 12v rails.

Anyway, I did notice a couple OK looking UCC KZG caps on the motherboard as well, and so once February ended, I pulled the system apart again and replaced them too. I didn't take any before shots, but the two groups are highlighted in alecjahn's first picture: the three near the expansion slots in the south-west corner, and the group of 4 shorties in the north by north-west area. On my board, the two groups in the north-east corner are Rubycon MBZ (CPU VRM high) and Sanyo WG (RAM VRM high). I decided to leave these, and only replace the KZGs.

And, since those shortie caps are shorties because the CPU heatsink wind tunnel gets in the way, I replaced them with polymers. UCC LG series, to be exact. 16v again, but I actually was able to increase the capacitance from 680uF each to 820uF each.

The three by the expansion slots got replaced with polymer caps too, UCC PSC series this time, 470uF @ 16v (same as the originals).

Here's an after picture (you can see the Rubies and Sanyos in the SW corner, new CPU polies SSE, and new PCIe polies in the NE area):

And here's a closeup of the new CPU caps:

This machine has been idling with no issues since, and another primegrid challenge started today, so I'll get a solid week of full blast stress testing. I don't imagine I'll have any trouble in the future, but if I do I'll be sure to report back.
Attached Images
File Type: jpg IMG_5173_small.jpg (1.23 MB, 49 views)
File Type: jpg IMG_5174_small.jpg (1,002.2 KB, 47 views)
File Type: jpg IMG_5176_small.jpg (931.7 KB, 46 views)
File Type: jpg IMG_5205_small.jpg (1.73 MB, 46 views)
File Type: jpg IMG_5207_crop.jpg (942.2 KB, 44 views)
etnietering is offline   Reply With Quote

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump Technical Forums 2003 - 2019
Powered by vBulletin ®
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
All times are GMT -6. The time now is 02:50 PM.
Did you find this forum helpful?