HM HN, Dates for the bad boys?

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

2009 is fine.
Bad ones were 2001 thru 2004.
-
The Hillman-Helmold study doesn't apply to the HM/HN problem.
It DOES explain why caps made with CH and TW sourced materials are crap and that's why you hear 'use Japanese caps' so much.
-
The study was done by CALCE out of the University of Maryland.
There is all sorts of interesting stuff on their site.
http://www.calce.umd.edu/
.

[japlytic]

Re: HM HN, Dates for the bad boys?

On a Palit 8400GS Super video card, I saw a 1500uF 6.3V Nichicon HN series on it, made on the 13th week of 2009. Should that unit be OK?

[Per Hansson]

Re: HM HN, Dates for the bad boys?

Thank you kc8adu, that was a very interesting read!

So afterall my readings where very much in line with what to expect for HM/HN, very interesting!

[kc8adu]

Re: HM HN, Dates for the bad boys?

tried to attach the .pdf explaining this.its a few kb too big.
google 2004_Electrolyte_Hillman-Helmold.pdf

kc8adu; yea, I posted about it here; https://www.badcaps.net/forum/showth...light=nichicon

PCBONEZ; maybe they ran out of the batch they had on this mobo so they had to change in the middle, I dunno...

Anyway, I bought a decent multimeter so I can also measure capacitance, a Fluke 113
What was interesting was that those two caps above that have not bulged and have good ESR indeed also do have good capacitance at ca 1800uF (they are rated at 1500uF)
Quite an achievement since all the others had failed...

Anyway, what I found really strange was that the caps that had failed had very high capacitance, many near 3000uF and some way over that too
What's up with that? I thought leaking cap=poor ESR and poor capacitance?

I did make sure to measure some brand new Samxon, Rubycon and Sanyo caps to verify the meter works and indeed it does, measured values very close to expected...

As a second note I measured the 1x failed cap in this router; https://www.badcaps.net/forum/showthread.php?t=7475
The 1000uF rated failed cap had only 370uF left, and it's ESR was 3.0...
The rest (not bulging) had both good ESR and capacitance

Has the overfilling of the electrolyte in the HM / HN caps given them much higher capacitance or what is the deal?

[Per Hansson]

Re: HM HN, Dates for the bad boys?

kc8adu; yea, I posted about it here; https://www.badcaps.net/forum/showth...light=nichicon

PCBONEZ; maybe they ran out of the batch they had on this mobo so they had to change in the middle, I dunno...

Anyway, I bought a decent multimeter so I can also measure capacitance, a Fluke 113
What was interesting was that those two caps above that have not bulged and have good ESR indeed also do have good capacitance at ca 1800uF (they are rated at 1500uF)
Quite an achievement since all the others had failed...

Anyway, what I found really strange was that the caps that had failed had very high capacitance, many near 3000uF and some way over that too
What's up with that? I thought leaking cap=poor ESR and poor capacitance?

I did make sure to measure some brand new Samxon, Rubycon and Sanyo caps to verify the meter works and indeed it does, measured values very close to expected...

As a second note I measured the 1x failed cap in this router; https://www.badcaps.net/forum/showthread.php?t=7475
The 1000uF rated failed cap had only 370uF left, and it's ESR was 3.0...
The rest (not bulging) had both good ESR and capacitance

Has the overfilling of the electrolyte in the HM / HN caps given them much higher capacitance or what is the deal?

[kc8adu]

Re: HM HN, Dates for the bad boys?

seimens industrial mobo?
just had one with 04 dates on hm.

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

Also interesting that caps made 22 weeks apart ended up on the same board.

Nichicon's announcement about the problem suggested that only some batches were affected but they didn't elaborate so there is no way for us to know.
[The assumption is only one/some shifts or one/some facilities made the mistake.]

Possibly their markings don't give any way to know which batches without cross-referencing with factory records [Like: "On xx day crew zz made the HM's" - or something similar.] so putting out more info would have been useless to us anyway.

.

[Per Hansson]

Re: HM HN, Dates for the bad boys?

Here is a few more dates, all caps are Nichicon HM 1500uF 10v 10x16mm, datecodes start with the letter "H"

0251; 13x bad capacitors on industrial PC
0143; 11x bad capacitors on industrial PC

What was interesting with the second system was that two capacitors where still ok, see picture
Those two had a datecode of 0213 and not 0143 like the rest!

There was also another industrial PC where all capacitors I could see where ok, they had a datecode of 0213
These three systems run 24/7 and are over 5 years old

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

That's the first bloated "A" date code I can remember.

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

A0332 x1 680uF 10v HM 8mm

Detailed:
GX270 w/ "Vesuvius" motherboard.
Single Bulged HM - 680uF 10v 8mm - A0332 - next to the 2x2 pin "P4 Plug".
There are other 8mm HM on the board w/~same date but none bloated.
Has MCZ in Vcore with date codes one week later so this one isn't a rebuilt unit.
Has Panny FJ for the 16v caps.

Have another ~4 mos newer GX270 with same board and cap arrangement except in the exact position where the bulged HM is there is an MBZ (not bloated).
There is only one MBZ in it and where the non-bloated 8mm HM are in the first system are also HM in the second.
Date codes suggest the MBZ is original.
.

Should have 3 more of these to look at next week.

[zandrax]

Re: HM HN, Dates for the bad boys?

A GX270 owner reported me this date for a brunch of bulged 1500uF 6.3 V HNs: H0325, so 25th week of 2003.

Zandrax

[gdement]

Re: HM HN, Dates for the bad boys?

Nothing unexpected here, but some more bad Nichicon HN's (taken from Dell GX270):

H0343 x1 1500uF 6.3v
H0351 x3 1800uF 16v

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

Thank you gdement.

Including your input my bean counting for 2001 has thus far found bad:
HM-H0135
HM-H0140
HM-H0142
HM-H0146
HM-H0147
I've seen no bad HN with 2001 codes.

.

[gdement]

Re: HM HN, Dates for the bad boys?

I'm looking at a bad Nichicon HM cap I just found lying around. I don't remember what board it came from, but the date code is only H0142, so that's older than the 2002-2004 range normally talked about as defective.
It's a 3300uF 6.3v, bulged with crusty orange leakage all over the top of it.

I also have a few Dell GX260 boards with good Nichicons all over them. I'll have to look at their dates sometime.

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

http://support.gateway.com/s/MOTHERB...102175nv.shtml

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

Thank you for helping!

It's an Intel board made for OEMs.
D915GAG or D915GAGL - (I think the L means onboard LAN.)
Gateway(for sure) and eMachines(I think) used them.

http://www.everestcomputers.ro/produ...oducts_id=1817

--

I think the A and H are codes for the fab facility they were made at but that's totally a guess on my part.
Possibly the code indicates the facility and which shift.

.

[gdement]

Re: HM HN, Dates for the bad boys?

0429 HN:

For what it's worth, I have an Intel board with 3x 1000uF 10V 10mm Nichicon HN(M) caps on it. 1 of the 3 has exploded. They have 0429 date codes on them, but 1 of the good ones I can't actually read.

The one that exploded is next to the top PCI-ex slot (where a video card would be). I haven't actually confirmed what it's connected to though.

I was recapping this board today, and can give more accurate information:

All 3 caps are HN A0429 1000uF 10v 10mm. There were actually 2 of them failing. The one that exploded had high ESR, but another of the caps that I previously called good (near the RAM) was actually bulging. It still reads good ESR. The 3rd cap seems fine both in ESR and visually.

The one that exploded could have been buried under a video card, but the board has onboard video so most likely there wasn't a card there. Owner only had 2x256MB RAM so I don't think he was an upgrader.

I'm pretty sure the motherboard is an Intel D915GAG - but the board model isn't actually printed on it. It has the name "Augsburg" on the BIOS sticker. I found data on the board at Gateway's site, and that fits my memory of what computer the guy said it came from. If only it would POST I'd know for sure.

My date code and stevo's start with different letters, which might be significant. My bad HN's are A0429, his good HM's are H0422. I wonder if that letter is the plant code.

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

Thanks Stevo!
Yes 2004 week 22 is correct.
.

[stevo1210]

Re: HM HN, Dates for the bad boys?

stevo1210
2004 is a critical year. That's the year Nichicon supposedly fixed the problem.
Next time you are in that case anyway for maintenance or whatever please post the whole date code for those. - Thanks!

These are photos from a few weeks ago, last time that I opened it up to clean out the dust bunnies and changed the CPU cooler. (Sorry for the bad image).... If anyone has noticed, I have edited out my motherboard's serial number.




As you can see, the date code is H0422, which I think means made in 2004 in week 22??
So far I have seen no bulging or leaking. My computer is as stable as rock even when overclocked from 3GHz (Prescott) to 3.4GHz.... I'm just happy that my Delta PSU can keep up to all of this.

And to add to that, this was my brothers old PC in which he used to play games (such as WoW) everyday. The previous owner before my brother (my uncle) used this PC for video editing and had it on almost 24 hours a day.

The idle temperature of the PC is about 55-60c (Before I replaced the cooler), now with the cooler replaced, it idles at 43c.

Thanks.

[PCBONEZ]

Re: HM HN, Dates for the bad boys?

Nah,, Maybe, but I don't think so:
The VRM out caps are 1500uF 6.3v. KZG and MBZ are rated the same in that size.
I think they just source caps where ever they can that day. (So to speak.)
Seen similar from other MoBo manufacturers and PSU's too for that matter.
I have two identical MSI boards with different VRM caps.
(In fact they are KZG vs MBZ also.)
Pretty often I see posts by someone that has some other board I have (or recorded) and their's has different caps.
Is not really unusual.

I'm guessing this:
I think the first 4 digits of a Gigabyte Serial Number are the date code.
- On these two boards the dates on the caps make that seem possible.
---- Can anyone confirm or disprove that?
.