Dell OptiPlex GX270 - multiple bad cap failures

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

I dunno about anyone over there other than RS.
Best ask someone on your side of the pond.

Yes FM are much better grade than FC. (In so far as ESR/Ripple ratings.)

[Choosing 10x20mm caps - Ripple/ESR]


MCZ - 2770/.011 <- Close match to Nichicon HN
MBZ, FJ, WG - 2550/.013 <- Close match to Nichicon HM, [FJ =Panasonic, WG = Sanyo.]
FM - 2180/.019
ZLH - 1960/.020 [ ZLH are Rubycon.]
ZL - 1820/.023 [ ZL are Rubycon.]
FC - 1220/.052

Can size is the best way to compare series' because, in the good brands, over a range of uF values, and within a series, all of the the same can size will have the same Ripple/ESR...
But for a specific uF/volts cap the can size may not be the same from one to another.
Example: a 1000uF 10v in brand/series "A" may be 10x20mm but in brand/series "B" it's 10x15mm.
"B" may be a better grade overall but for that specific cap "A" is better.
Happens, but not often unless the grades are pretty close in the first place.
.
MBZ are more or less the typical [standard] grade you find in P4 VRM's.
ZL would be typical in P3 VRM's.
I've seen FC as original in VRM on older P2 boards.

Boards that use more caps can usually get by with lower grades.
.

[woden]

Re: Dell OptiPlex GX270 - multiple bad cap failures

This might be the classic newb seeing something complicated where it's obviously simple to those that know, but is can size an electronics term or does it literally just mean the size of the coke can shaped body of a cap?

But if I've understood you right does that mean this Nichicon cap that seems to fit the specs wouldn't be as good as one of the Panasonics as its can size is 10x13 - just like the Nichicons on my board - compared with the Panasonic's 10x20 dimensions? Is three times the price for the Nichicons just a brand name premium?

Oh, are newer Nichicons a wise purchase given the companies history with Dell GXs? Or is that batch considered a freak one-off?

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Yes >> literally just means the size of the coke can shaped body of a cap? <<
- Normally given as Diameter x Length in MM.

On this one:
>>>
But if I've understood you right does that mean this Nichicon cap that seems to fit the specs wouldn't be as good as one of the Panasonics as its can size is 10x13 - just like the Nichicons on my board - compared with the Panasonic's 10x20 dimensions? Is three times the price for the Nichicons just a brand name premium?
<<<
I think I lost you so bad that I don't understand your question.

I will try to explain the method again but it's past my nappy time so bear with me.

WITHIN a SERIES, a physically larger cap will have better ESR and Ripple. [Just the nature of things.]
- Because WITHIN a SERIES all the caps are built to the same grade by using the same electrolyte, aluminum foils, processes, and so forth.
-
Some series may have 3 different 1000uF 6.3v caps.
Say an 8x20mm and a 10x12.5mm and a 10x20mm
They will all have different ESR and Ripple ratings.
AND there may be 10x20mm CANs in 1000, 1200, 1500, 1800 uF in that series.
Within that small range those will all have the same ESR and Ripple.
[Some exceptions but very rare in quality brands.]

~~~ ESR and Ripple are functions of:
- the can size
- the electrolyte
- the other materials
- the processes used

Now
- IF you want to compare -that- SERIES to another SERIES.
- THEN you need to find caps with the same can size in both series' and compare those.
- Disgregard uF/volts because if you go by uF/volts to compare and the can sizes are different then you just wasted your time.
UF/volts have no direct relation to ESR/Ripple.

~ And AAAALLLLLLLLLL of that ONLY tells you which series is higher or lower grade than another.
- It tells you which series' are most likely to have an equivalent or better replacement for the old cap.

ZZZZzzzzZZZzzzZZZZ Oops, dozed off there.

~~~~~~~

Now lets say you have a cap to replace.
It's 1000uF 16v and 8x20mm and HAS to be 8mm diameter to fit.
You know from looking it up [and all of that above] that FM is a good match to it's Grade.
But FM doesn't have an 8mm diameter 1000uF 16v.
[You know from all of that above] not to bother looking for FC or ZL because they are lower grade.
[You know from all of that above] that MBZ, FJ, WG, or MCZ might have a good replacement because they are better grade.

The other thing I tried to get across:
So you find a 1000uF 16v FJ but it's 8x12mm
STILL check ESR/Ripple in FJ data table.. Opps, Ripple too low, ESR too high.
[Because can is so small.]
So you look in the other series until you find one.

~~~~~~~

I know it seems complicated but it's really a lot more complicated to explain it that to do it.

Just pick a range in the middle [say 820uF to 2200uF]
Pick a common can size in that range [say 8x20mm]
And compare the [in uF range] 8x20mm caps between the two series.
Then you know which one is better than the other and by how far.

You will find that you can pick -any- can size within that uF range to compare the series' with and the results will be the same. Thing is odd-ball/unusual can sizes may not exist in both series' so it best to use a common size.

The uF range is important in that, while which grade is better/worse won't change, the properties do vary towards the upper and lower ends of the tables compared to those in the middle.
Like:
Checking same cans in 3300uF-4700uF range to 3300uF-4700uF range in the other series would be the same as 820uF-2200uF to 820uF-2200uF.
But checking the same cans in a 820uF to one in the 3300uF-4700uF range in the other series might not be.
UF corresponds to the surface area of the foil in the cap and that does affect ESR, but only a tiny bit. When the difference is surface area is that huge it adds up enough to throw things off.
.

So now are you more lost, less lost, or just bored silly?
ZZZZzzzzZZZZzzzzZZZZZzzzzz
.

[SaintAtilla]

Re: Dell OptiPlex GX270 - multiple bad cap failures

I epoxied the retainer down to the motherboard (those push clips just weren't holding tightly enough)..

Does anyone know of a tool, technique, or a "how-to" to remove the cpu heatsink plastic-stand assembly thing from the gx270 motherboard so as to avoid breaking the plastic retainer push-pins. The metal backing plate is preventing me from accessing the solder connections, and I don't see a way to remove it without breaking these retainer pins.

TIA

[jacobnbr1]

Re: Dell OptiPlex GX270 - multiple bad cap failures

It seems that these sx270/80 models have a bad GA'HUGE problem with the caps but what i have observed is that most condone the unit standing on edge and i gotta tell you i think that is a bad idea for these paticular units.

These units (when stood on edge) are starving for air supply and the machine can't get air with the vent covered.

Even if you google the reviews or look anywhere you see a picture of one, they profile it standing on edge but that bottom edge is where the fans and vent is and it will be blocked or restricted thus causing the overheat issue maybe causing the caps to fry.

If somone has one of these you will see what i'm talking about.

Just some FYI or if somone wants to discuss the issue we can.

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Those built through mid 2004 had defective caps. That wasn't Dell's fault.
All of them have bad air flow. That was Dells' fault.
GX270/280 same same
SX/GX620 are now showing up with heat blown caps.

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Does anyone know of a tool, technique, or a "how-to" to remove the cpu heatsink plastic-stand assembly thing from the gx270 motherboard so as to avoid breaking the plastic retainer push-pins. The metal backing plate is preventing me from accessing the solder connections, and I don't see a way to remove it without breaking these retainer pins.

TIA

I push the pins back through with the end of an appropriate sized allen wrench.
You have to hold your teeth right while you do it.

.

[jacobnbr1]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Those built through mid 2004 had defective caps. That wasn't Dell's fault.
All of them have bad air flow. That was Dells' fault.
GX270/280 same same
SX/GX620 are now showing up with heat blown caps.

And there is alot of em out there too. LOL

[woden]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Hi, PCBONEZ

Just back to this after a few weeks of being away and then preoccupied with another crisis...

What you've posted up there really is way beyond my meagre understanding. Not for your want of trying but I still really haven't a notion; am I still miles off with the conclusion that first, find the best series and then go for the biggest can size.... Ah! mwah! bwaaah! gaaaaaaaaah!... my brain has just dribbled out through both my ears. *runs for mop and bucket*

Could you - or anyone for that matter - put me out of my misery and tell me if which of these two caps - Nichicon or Panasonic - would be preferable for replacing my domed 1500uF 6.3v Nichicons?

[Toasty]

Re: Dell OptiPlex GX270 - multiple bad cap failures

What is the Nichicon series of the badcaps? HN? HM? I see you mentioned the height is 13mm, but no diameter. If they're either of those series, then they're 10mm diameter.

Given the choices you have available, I'd go with the Nichicon LG Polymer you picked. It's the same size can as what was there so it won't interfere with anything. It has a better ESR/ripple current rating (10 mOhm /5560mA) than the Panasonic FM (19 mOhm /2180mA).

Toast

[woden]

Re: Dell OptiPlex GX270 - multiple bad cap failures

They've got HN(M) printed on the side of them so what that means I'm not quite sure; hedging their bets and not admitting they're either, I dunno.

From your advice and the choices others have made I think I'll go with Nichicons. Though it's probably not a wise thing to rely on, the Nichicons are three times the price of the Panasonics; I wonder if the difference is indicative of superior performance on the part of Nichicons?

I'd originally been looking for the Rubycons but I can't find a supplier over here.

I take it Nichicon have resolved the problems that led to their earlier caps going bad and that there's now no more risk in using their caps as opposed to other brands?

Woden

[Toasty]

Re: Dell OptiPlex GX270 - multiple bad cap failures

(M) means ±20% tolerance to capacitive value.

If you don't think my advice is a wise thing to rely on, then don't. I do not care.

You came here seeking advice and counsel from the members here. I, nor anyone else, sought you out.

The Nichicon's are more money because they are a Polymer electrolytic and polymers are typically a higher cost industry wide when compared to the same value in the standard electrolytics.

There are risks to using anyone's caps or any components for that matter. Nichicon had a problem with 2 series out of 20 and have rectified the issue. They as well as many other(reputable) manufacturers got caught up in the bad electrolyte problem. Furthermore, Nichicon bought Fujitsu's polymer business:
https://www.badcaps.net/forum/showthread.php?t=6524

If you are still in a quandary over where to buy the Rubycons, buy them here. Topcat has them and just ask him for a quote. Hell, if you'd ordered them a month ago when you started poking about, you'd have them by now.

Here's what he stocks: https://www.badcaps.net/pages.php?vid=22

Here's the request form: https://www.badcaps.net/pages.php?vid=23

Toast

[woden]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Whoa there!

Toasty, I think you've misread my last post because there's no hostility to you or anyone else on my part. Maybe I've been unclear in what I've written and if so, I apologise for what seems like a misunderstanding.

From reading it, I'm just not sure which bit could be misconstrued as a dig at you.

From your advice and the choices others have made I think I'll go with Nichicons. Though it's probably not a wise thing to rely on,...

If it's to do with the above line, the second sentence goes on to comment on the reliability of the price gap, not your advice. Advice that I appreciate greatly.

So, apologies for any confusion.

I think I'll still go with the Nichicons; the company seem to have left their problems in the past and I can get them pretty quickly. I doubt it's worth the bother of shipping Rubycons across the Pond, they can't be that much better.

Woden

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

I've been out of town on a job.
Toasty has given good advice.

I'm not clear on -what- Nichicons you are buying as replacements.
The HN(M) mark is HN series followed by the tolerance mark. (As Toasty said.)

HN with date codes 2005 or after should be fine.

If you are going with a polymer series you need to be careful with voltages.
You can go down to 2.5v for the 9 caps in Vcore-out. [Row of 8mm next to CPU.]
I haven't traced out the other caps actual voltages on one of those boards (KC8 or TC may have) so you should check actual voltages or stick with voltages equal to or more than the originals.
.

[woden]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Hi again, PCBONEZ.

Yeah, as this is mostly way over my head, I'm just going to stick with the voltage - and capacitance for that matter - of the caps I'm getting replaced: 6.3v and 1500uF. I think if I tried to get clever with these values it'd just end in tears. So, the replacement Nichicons I'm going for are these. I presume these won't be affected by the date code threshold of 2005?!

Woden

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Hi again, PCBONEZ.

Yeah, as this is mostly way over my head, ~~~
I presume these won't be affected by the date code threshold of 2005?!

Woden

Only Nichicon HM and HN series had the problems and it's only those two you have to worry about date codes on.

-

I have 7 or 8 of those boards here now so I took some voltage readings and I modded your drawing to show what I found.
[ Was going to next week anyway for my own info. ]
Voltages taken with system was sitting at BIOS screen, all RAM slots filled, but with nothing in any add-in slots so condition was unloaded.


Note: The original cap sizes vary by board version (tower vs desktop vs SFF vs USFF) as well as by make/series of cap from board to board within those versions so I'll relate what I usually see. I absolutely have NOT seen all variations so other combinations are very likely.

Letters reference to drawing.

A: These are two of your swollen caps - [2] 1500uF 6.3v - 10mm.
[Other board versions use [2] 1800uF 6.3v 8mm here.]
These are filtering +5v either to whole board or to RAM regulators.
Usually see Rubycon MCZ // Panasonic FL,FJ // Nichicon HN
Stay with 6.3v or more.
>> Your replacements look fine for these.

B: These are your other swollen caps - [3] 1500uF 6.3v - 10mm.
These are to RAM. [Measured voltage was slightly high because RAM wasn't doing anything even though both slots were filled. Normal would be around 2.5-2.6v.]
- Usually see Rubycon MCZ // Panasonic FL,FJ // Nichicon HN
>> Your replacements are fine for these but 4v caps should work too.

C: General caps to slots and IC chips on board. - [4] 820uF 6.3v - 8mm.
- Usually see Rubycon MBZ or Panasonic FJ or Nichicon HM
These are on +1.5v so 2.5v caps should be fine.
These blow sometimes but not so often as the others.

D: The "I dunno cap" that is often blown - [1] 680uF 10v - 8mm.
The original caps seem to have problems in that spot and I suspect there is more ripple there than they can handle.
- Usually see Rubycon MBZ or Panasonic FJ or Nichicon HM
I think using a 680uF - 4v[up] solid polymer here is a REAL GOOD idea.
- I would replace this cap with a poly just because it blows so often.
I think it's only 10v to get lower ESR and keep it an 8mm 680uF so presumably the 680uF value is important here.

E: VRM-output caps. - [9] 1800uF 6.3v 8mm
- Usually see Rubycon MCZ or Panasonic FL or Nichicon HN
Down to 2.5v is okay. - Nine 820uF-2.5v solid polymer should be fine.
This is a common mod and it works fine.
[ESR is more important than uF here but you do need -some- uF.]

F: VRM-input caps. - [3] 1800uF 16v - 10mm.
- Usually see Rubycon MCZ or Panasonic FL or Nichicon HN
Must be 16v. - Three 330uF to 470uF at 16v solid polymer should be fine.
Many moddin' folks use [3] 330uF here with no problems on P4 socket 478.
I favor [3] 470uF like I see used in Xeon server boards. - Because I'm chicken.
[ESR is more important than uF here but you do need -some- uF.]

.

[kc8adu]

Re: Dell OptiPlex GX270 - multiple bad cap failures

follow this mod thread to the letter and you will be fine.
i have done this to hundreds of the 270/280 boards and no returns.
https://www.badcaps.net/forum/showthread.php?t=4969

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

KC,
Do you know what that 680uF 10v in on? (Does?)
I've found them bloated in HM (obviously) and MBZ.

[Toasty]

Re: Dell OptiPlex GX270 - multiple bad cap failures

@woden
Okay. I misinterpreted your post. Must be my personality, because I just read it again and felt the same as I did the day I replied. grrrr...

No need to apologize. Some days, I'm abrasive at best, and everything gets stuck in my craw.

We now return you to your regularly scheduled drinking...

[PCBONEZ]

Re: Dell OptiPlex GX270 - multiple bad cap failures

Cats with machine guns don't have craws.