Dell GX620

[HCT]

Hello,

I have a GX620 USFF, was working 100% OK then next attempt it wouldn't power on at all, light goes green on the power brick but the box is dead. All the caps as far as I can see look OK not weeping, not bulging. Any suggestions as to most likely cause of failure? Are any particular caps worth trying first? The board has a mixture of K topped and Y topped caps.

Cheers

H

[PCBONEZ]

Read the whole thread. The small ones dry out and don't usually bulge.

[bluto]

It seems that this thread has been dormant for a while... Time to bring it back.

I note that most of the GX620s shown here are the ultra small (external power supply) version. Has anyone had issues with failing caps on the slightly larger SFF chassis version of the 620?

I have some GX 620 SFF units in an enclosed kiosk application and on a whim I decided to crack a few open and check the caps. Oh and these babies were quite warm as others have stated! 2 of the 3 machines that I looked at each had a single bulged and leaking cap. The first one had a bulged Rubycon MCZ 1800uf 6.3v 105c (PET date code?). This specimen also had some Panasonic 2200uf 6.3v 105c caps a bunch of little Chemi-con 220uf 16v SMG 85c caps.

The second machine had Chemi-con caps in place of the Panasonic and one of them was bulged.

These machines were manufactured mid 2006.

I apologize for the poor camera phone pictures.

So far these machines still run fine and there are no symptoms of capacitor failure evident. Also there have been no problem reports from the field either.

What is the prognosis for these machines? Will they go for a while or is death imminent? I note that topcat now has a pre-made kit available for the 620 SFF, so that can't be a good sign right there.

Unfortunately, there is no good way for these machines to be recapped. They would likely end up getting replaced. I am just trying to get an idea on how soon that would have to be.

[PCTechNY23]

I have seen them start to blow the 16v 220uFs and PSU fans. So far all under warranty but they are about to run out.

[PCBONEZ]

PET is the material used in the sleeve.
Forgot exactly what it stands for. [Some kind of plastic.]
It's there for some ecology reason.

[Player2]

Hello all,

I've got a few GX620 boards that are really giving me a headache, hopefully someone has a suggestion. I recently received eight of these boards for repair; four worked fine after recapping, the other four refuse to do anything. When plugged in, I've got an amber LED lit (not the power button LED, but near that on the board) which leads me to believe that the boards are all receiving power. When the power button is pressed I get nothing; no momentary spinning of the fans, no diagnostic LEDs on the back, nothing.

I have replaced ALL the caps, all the way down to the 220 uF's and 22 uF's. I have also reset the CMOS and replaced the batteries. All of the components I am testing with (PSU, CPU, RAM, HDD) are good, they worked in the four good boards as well as an SX280 board I had laying around. I found a few results via Google indicating that shorted USB ports can cause this, checked them all and they're fine. I have even checked the power buttons just to be sure. Also tried running one of the boards without the tray, just in case something was shorting out, but no luck.

Has anyone seen this before? Any ideas on what else I should be checking? I don't see GX620 boards too often, but I work on SX280's all the time (very similar board) and haven't seen anything like this. Hopefully it is something simple I am just overlooking.

Thanks in advance for any advice!

[s2ndsight]

I want to post this information here, since this forum helped me figure out what was wrong with my GX620 Mini-tower. I also purchased all of my replacement caps from this site as well. This info, I just finished posting on the Dell community forum. I hope it is helpful.

The other posts in this thread are concerning the GX620 ultra small form factor machine, which I believe has a history of worse capacitor problems due to the heat buildup in the little case. My problem was somewhat different, but it appears that replacing the caps was still what fixed it!!

This Dell Optiplex GX620 Mini-tower appeared to work perfectly in all respects, except that if it lost power or was unplugged, I would have to continually reset the bios date/time when it booted back up. These CMOS batteries have a very long life, and it is not uncommon for them to last 10+ years, especially if the machine stays plugged in for most of that time. Replacing the battery was the easy try, but that did not fix it for me.

I replaced the following capacitors around the 12v connector on the motherboard which is labeled HH807: C575, C555, C566, C545. Each of these caps appeared to be a no-name brand and were 1800mf, 6.3v. There were also absolutely no visible signs that these capacitors were bad. Also, C559, which was a Rubycon and 1000mf, 16v, also, no visible sign that it was bad. The 1000mf capacitor had blue ink on the top, like it had been deliberately marked, and I have a sneaking suspicion that THIS was the only capacitor that needed to be replaced. But i've read (on badcaps.net) that the no-names are sometimes very prone to failure. I also replaced C413, which is closer to the center of the board. It was also an 1800mf cap. No visible sign of failure.

So, I can't specifically narrow it down to which one of those 6 capacitors was causing the problem, but I am fairly sure it was one or more of those. A few days earlier, I had replaced all of 27 smaller caps, and these were all of the 220mf, 16v caps. They looked slightly bulged to me and were only rated for 80 degrees C, as opposed to the other caps which had a 105 degree C rating. In retrospect, it appears that none of these smaller caps were bad or needed to be replaced and I probably wasted my time doing so. They were also a bit harder to cope with than the larger caps, so if I had found a posting like this one before I started work, it would have saved me a significant amount of time.

I also noticed something else when I was finished. Before I replaced the 6 caps I mentioned, when I would plug the machine back up after sitting idle for a bit, I would see nothing of the power diagnostics on the front of the machine... I think... or possibly I might have not noticed before. I doubt it. I believe that plugging the machine in before the repair resulted in nothing visible except for the amber "aux power" light on the motherboard. AFTER replacing the 6 caps, whenever I plugged the machine in, it would appear to run a quick power diagnostic, which was visible from the indicators on the front of the machine. I push the power button, and it boots up now without begging to know what time of day it is and my bios settings are preserved.

Thanks to badcaps.net for the assist. I hope this info makes very easy work for the next person to have this problem!

[PCBONEZ]

Those boards are built for Dell on contract with Intel and they don't use no-name caps as original.
If they are truly no-name caps then board was previously re-capped by some twit.

The 85C [it's not 80C, it's 85C] do fail often enough due to heat dry-out. It usually takes them longer 'to go' because they don't handle so much ripple as the larger caps and so are working further from their absolute limits.
.

[PCBONEZ]

Oh and.... CONGRATS!

[s2ndsight]

Yes, 85C. Writing from memory. It didn't look like this motherboard had been touched by anybody. What I'm calling a "no-name" capacitor looked just like the Rubycon, essentially with the same coloring, but without the name printed on it. I thought that I had seen in another post somewhere some reference to these being a knock-off of some type, or perhaps it was just a brand that doesn't print the name on the capacitors.

[PCBONEZ]

Probably Panasonic FL or FJ.
Was the vent shaped like a 'T' with a curved top bar?
.

[s2ndsight]

The vent is a Y, except all three divisions are equal size. They have other markings that I don't know the significance of, other than the temp rating. (M)105C, 6(7), 5 I.

[PCBONEZ]

Depending on what it looks like the (M) may be the uF tolerance [+/-20%] or it may be Matsushita's trademark.
Panasonic = Matsushita

Can you upload some photos of these mystery caps?
.

[s2ndsight]

I could. And I will, but I don't think they are mystery caps. I am almost sure I read in another post somewhere on this site about someone else who described the Rubycon lookalikes in some detail. I will find that link and post it with a couple of pics asap.

[370forlife]

Y with all three divisions the same size...UCC?

[s2ndsight]

I looked around a bit and found a thread titled "A Look at Caps for Indonesia Market" and found some great photos of Chemicon caps that look exactly like these that I suspect I've had problems with. It wasn't the discussion I thought I had seen on "fake Rubycons". The only difference is in the coloring. These caps look very similar to the Rubycon just in the fact that they have the black and gold coloring. In each photo, the Rubycon cap is on the left for a comparison.

[IMG]P1010195.jpg[/IMG]
[IMG]P1010196.jpg[/IMG]

[PCBONEZ]

Chemicon KZJ
Susupected of the same heat issues as Chemicon KZG.

[Chemicon = UCC]

[speg_head]

I just got my first ESR meter (micro 3.1) and I have some 220uf's from a gx620 that rate at 0.04-0.07. Are these ok? Also have some 2200uf that rate 0.00 ohms. The board I have has no physical sign of bad caps so I'm taking them off one by one.

[PCBONEZ]

First - you need to look them up in the data sheets so you know what you are expecting.
That is because ESR varies -A LOT- from one make/series to another.

On the 2200uF.
Higher grades of low ESR caps can have ESR less than 0.01 and those won't always show an ESR on the meter.
When you run into that also check them with a regular ohm meter.
If they read 0.00 [or anything at all low] on that meter too then they are shorted.
If they have a high resistance [after the cap charges and the meter steadies out] they are probably good caps with a very low ESR.

[0 on ESR meter] + [high on ohm meter] => good cap 99.999% of the time.

On the 220uF.
Without knowing the make/series/volts/can size it's hard to say if they are in spec.
0.04-0.07 is probably an okay range for 220uF -in general- for higher grade variety of 220uF but if they are all the same make/series the warm fuzzy would be ?firmer? if they didn't vary by almost 100%.
-- 0.07 is almost twice 0.04.
They may all be good but it is possible some are good and some just beginning to degrade.

- For reference:
Assuming they are 25v 220uF caps,, The ESR for a Panasonic FM is 0.056 and for a Nichicon LXZ it is 0.120.
Most motherboards use something like LXZ grade [or worse] for caps under 470uF but this particular Dell board is a different animal.

.

[speg_head]

Most of the 2200s, 820s, and 220 are rubycons. I will check the series and compare the data sheets. Would it be safe to say that 0.00 is a good reading if the capacitance reading on those caps are accurate? The esr 3.1 thankfully also shows cap value and they are pretty accurate at that. Thanks pcbonez this will help a lot.