Just a real long shot, but you might try calling the folks at JE Capacitors.....

I believe these are the same folks previously known as JEA....Jaques Ebert and Associates.

Many years ago, they were very good at being able to supply rare and obsolete types. I don't know if they still deal with things that old or not, but they used to be very knowledgeable about older styles of caps.

Its cheaper just to replace my old scope with a new digital one. all the parts are obsolete.

Probably quite true. Unless it's the challenge or the sentimental value of the old scope.
Perhaps the link may be of use to someone at some time.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Read this thread with interest, and was wondering if anyone had a suggestion as to what is considered the "best" capacitor right now, money no object (within reason, no weird Mil spec stuff from Area 51, etc). I remember reading about Asus bragging about the new style caps that they were using which were supposed to be so much better than the previous style, not sure which type they were. Is there such a this as Best, or does it depend on it's intended application - MB, vid card, etc? Thanks for the enlightenment!

The "Japanese Solid Capacitors" seems to be the latest "thing", but as far as long time reliability I don't know.

Maybe you mean Sanyo OS-CON
http://en.wikipedia.org/wiki/OS-CON

http://www.edc.sanyo.com/english/products/capacitor/oscon/outline.html

IIRC this type of capacitor was invented by/at Sanyo ~30 years ago but sadly become famous (in bigger use) only just now (for the last few years).


"An incorrect electrolyte formula within a faulty capacitor causes the production of hydrogen gas leading to bulging or deformation of the capacitor's case, and eventual venting of the electrolyte. ...
After some normal use the bad capacitors fail predictably far sooner than normal end-of-life;
Faulty capacitors have been discovered in motherboards as old as Socket 7 and have affected equipment manufactured up to at least 2007"
http://en.wikipedia.org/wiki/Bad_capacitors

Yeah, I have a Dell at work that I rescued from a skip and rebuilt; it suffered from the rash of bad caps but by the time I worked that out it was a month or two past the cutoff date for a Dell replacement. I got a replacement kit from a chap in the US who specialises in this stuff, and reprogrammed my (un)soldering skills. As JG said, it was easier in the first instance to break the cap off its leads (by wiggling back and forth repeatedly) which then meant I could remove the leads one by one rather than having to keep two leads hot while removing the cap in toto.

Figured it was time to post a follow-up for this. Just a week shy of two months since doing the cap replacement, and I've yet to have a single issue or hiccup with anything.

Everest 'uptime' output:


I cleared all the event logs after the first boot with the new caps, and the 16-minute downtime was to pull one of the case fans and put two drops of 10w-30 motor oil in the bearing. This system is running 24/7, so anytime I have more than about 5 minutes of power-down, that's the only way to notice bad fans is when they cool off and start up again.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Blowing some dust off of this thread. Main cruncher is still doing just fine with its replaced caps. The Sempron machine started acting up again though. Same symptoms as before.. will memtest for days on end without issues, but loading Windows throws it into a random power-off.

I looked inside and saw one of the replaced caps with some ooze, so I replaced that one again, but that didn't fix anything. I've decided that since it is a 7..or 8 year old Biostar board, it has done far more than I've ever expected a $60 board to do, and it is retired. Fortunately, I have another board from an emachines carcass that I think is a s754 board, and it's a full-size ATX board instead of mATX, so I can have more RAM!

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Sounds like a reason to buy a new system.

---

Grant
Darwin NT

Blowing some dust off of this thread. Main cruncher is still doing just fine with its replaced caps. ...

Strangely enough, I've just had to replace some capacitors in a switch-mode PSU...

It was a close call as to whether to return the unit or to just fix the obvious fault. One capacitor was torn wide open, and a second was badly distorted. Curiously, the PCB outlines didn't match the capacitors that had been used. Also, there was space for a parallel bank of capacitors where only two had actually been put in place... I decided swapping the capacitors was going to be quicker than the returns route...

The values checked out ok except... The only replacements that I could find were physically much larger for the same supposed value of the originals... So, a bit of engineering ingenuity later and careful use of a soldering iron, and all has tested out fine under load for a good few hours now.


I wonder if the factory had substituted cheap 'equivalent' capacitors of the 'same value' but with a bad "ESR" not knowing that you can't do that for high currents?... If so, then I've saved myself replacing one dud PSU with another...

For that company's sake, I hope their engineers quickly sort out their ignorant purchasing people!


All good fun...

Cheers,
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

I have replaced many capacitors in my time also.
Most have been while working as a technician with a local Magnavox service center and have been in tv's, stereos, vcr's satellite systems, etc. (you wouldn't believe what a bolt of lightning can do to a power supply filter cap! hehe!)
On the tv/home electronics side a lot of times there was no outside evidence of failure, but there would be symptoms of mishandling of the signal which would lead us to suspect a capicitor failure. (An example: If a "crt" tv (one with an actual "picture tube", for you younger folk, hehe)showed a black band at the top and bottom of the picture and the picture was "squished" near the top and bottom of the visible part suggested a cap failure in the vertical oscillator/output section.) Most however have been small electrolytics, but a lot were ceramic, tantalum, and mica. On a lot of tv's we would automatically replace the caps in the high voltage output circuit whenever we would have a horizontal output transistor to fail. This was because either failure of the cap could cause the transistor to fail or failure of the transistor could cause damage to the cap leading to early failure of the replacement transistor.
Also, as one poster mentioned the old hv caps for a scope, I have an "oscilloscope" type automotive "ignition scope" which lets you view the voltage waveforms on the spark plug leads of an automobile. It has a series of (if I remember right, four) 1000v caps in a voltage quadrupler circuit to supply the anode voltage to the crt. One of these caps failed and shorted out the power transformer, which burned out. I found a replacement transformer and caps for it and replaced them, but this was years ago. One of the caps has failed again and now I can't find either a replacement hv cap or the power transformer for this unit so now it's just a piece of sentimental "junk"!
Someone suggested bridging a known good cap across a suspected bad one, and this is one way of testing, but it won't work in many cases. If the cap in question is "shorted" then it won't work. If the cap is in a circuit where the capacitance is critical (i.e. in an oscillator/frequency dependent circuit) it might or might not work, depending on the combined capacitance of the leaky/defective cap and the new one. The only time this will work with 100% accuracy is if the defective cap is "open" and the bridging cap is the same value as the original. You're welcome to try this, but be warned that it might not work, and with the newer circuits it could possibly cause more damage due to the large inrush of current required to charge the bridging capacitor (especially with larger electrolytics). (A lot of power supplies are designed to bring the power up gradually to avoid just such a large startup current demand.)

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

A lot of times a manufacturer will design one circuit board for use in a lot of different models. There will be unused spots for the parts that are not used/required in your particular model. It's cheaper to manufacture one "generic" board which can be used in many variants than it is to manufacture a special board for each model.
A good example is in the wiring harnesses for cars. If you ever look behind/under your dash you will find many connectors just hanging loose. This is because they are in the harness for other models, but not your particular one. One amusing example was with a fellow that had a station wagon with integrated side marker lights and taillights. He had his right taillight burned out. He took the panels off to try to find the problem and found an extra connector which is supposed to be used for a separate side marker light on other models. He rigged a special jumper to plug into it to jump one to the other thinking that this would fix his taillight problem. It shorted his entire park light system! I had a h*&* of a time finding the problem, but when I removed his jumper all of his problems except for one blown bulb were fixed!
So replace any defective parts ONLY as they are on your specific board. If you have other spaces/parts defined, ignore them! They are not used in your particular model and may cause problems if you try to add them.
If you are *absolutely* sure that your extra spaces are in parallel, then you might substitute, say two 220mf caps for a 440mf one, but you would have to make sure that it uses only a single layer for the interconnections. A dual or multilayer board would require you to actually determine the circuit traces for each layer to make sure they aren't used by separate sections of the circuit and are in fact in parallel (such as a dl trace jumping from one side of the board to the other).

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

A lot of times a manufacturer will design one circuit board for use in a lot of different models. There will be unused spots for the parts that are not used/required in your particular model. ...

If you are *absolutely* sure that your extra spaces are in parallel, then you might substitute, say two 220mf caps for a 440mf one...

And that is exactly my suspicion for what manufacturing tried for the sake of cost-cutting...

Although that substitution is fine for the capacitive value, you likely also suffer twice the "Effective Series Resistance" for a particular type of capacitor. For a high ripple current, the x2 voltage drop can cause enough heating to blow the capacitor...

Hence my suspicions for my example. The replacement capacitors I used are physically much larger for the same capacitive value and for a similar working voltage, suggesting a much lower (better) ESR. I also suspect the replacement capacitors are more expensive than the originals.

Capacitors are much more than just a capacitance and purchase cost. If only the accountants and purchase people realised that!


Keep searchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Right, so the sempron machine is up and running again. The board was in fact a s939 board, so I kind of had to keep using the 200MHz slower CPU that was on it. Put Arctic Silver on the heatsink instead of the OEM crap that turns into epoxy after about two years (you go to pull the HS off and the CPU comes with it.. that's always fun). Blew the tarred smoker dust out of the fan and fins, swapped the RAM, added some more to it, put it in the rig and booted up.

Took some time to find all the new hardware and controllers, did a reboot. Went to the HP website and got the drivers for two things that Windows couldn't find (donor board was from a Presario SR1617CL.. I thought it was an emachines), and now it's back in business, crunching and doing WSUS. It's great having more RAM now.. the entire SQL database can fit in RAM without swapping.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

When I need capacitors changed I do it the easy way.....














































Send it to SciManStev...

---

Official Abuser of Boinc Buttons...
And no good credit hound!

One time when I was investigating a client's computer stability issues I noticed bulging tops on the motherboard capacitors. I showed him this and he just shoved them back down again with his finger :| "sorted"

---

Brian.

One time when I was investigating a client's computer stability issues I noticed bulging tops on the motherboard capacitors. I showed him this and he just shoved them back down again with his finger :| "sorted"

LOL...that'll fix it. Denial in computerland.

---

"Time is simply the mechanism that keeps everything from happening all at once."

So I think the problem I fixed is actually fixed, but as predicted by others, some of the other caps may be on their way out as well. I haven't done any testing, but it seems semi-random, but happens when there is a heavy RAM load (games) after more than 24h of uptime. If I reboot and then play a game, I can do it for hours just fine, but after ~24 hours, it starts off fine and then just randomly does weird stuff.

Weird stuff like sound freezes and loops the same 100ms of audio indefinitely. At this time, I can still do caps lock and ctrl+alt+del and pull up task manager, but that's about all I can do.

When this machine gets "retired" and becomes my 24TB file server, I'm pretty sure I'm going to pull the second CPU out and move all the RAM to the cpu_0 bank. That will drop 130w of power consumption right there, and hopefully leave the quirky behavior of cpu_1 and its memory banks behind.

For now.. I just have to reboot to play a game. Sure it takes about 4 minutes and I have to umount all the NFS shares on the remote machines before doing so (otherwise they complain about stale handles), but I don't want to mess with anything while its the only thing I have for a daily driver. Last time I changed the CPU configuration (added second CPU), it worked for two POSTs and then went brick with no explanation (got an RMA of course).

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Oh yes, I've very recently had to replace a capacitor on a mother-board. Its a slightly sad tale involving, a very large Scythe Ninja 2 HS unit with very little room to push down the HS 'locking pins' to the S775 socket mounting on an MSI P35 board..........and a screwdriver. Yep, the screwdriver slipped, after it had taken me 25 mins to get the HS off in the first place! Managed to replace it OK, but instead of the nice 'metal can' (is that what they call 'solid', these days?) electrolytic that was there, the only thing I could find of the same ratings, was a standard item from a suspect Gigabyte board. It works fine, but I'll replace it with the proper item, when I can find one from a UK supplier.

To The Kitty-man. Blowing the mid-range cross-over caps to Space Truckin'? Respect!

---

Don't take life too seriously, as you'll never come out of it alive!

...
Managed to replace it OK, but instead of the nice 'metal can' (is that what they call 'solid', these days?) electrolytic that was there,...

The metal can is intended to keep non-solid electrolyte inside. Solid electrolyte doesn't need a metal can...

                                                                   Joe

To The Kitty-man. Blowing the mid-range cross-over caps to Space Truckin'? Respect!

LOL...thanks.
That was about 33 years ago.
And still "Space Truckin'".

Meow!

---

"Time is simply the mechanism that keeps everything from happening all at once."

Indeed, Joe. I could see there was no real difference and yet my particular board was sold as a factory repaired item using 'solid Japanese capacitors' and it was only two particular capacitors that appeared to be different to the others on the board, with the tell-tale signs of re-work on the underneath. Perhaps the word 'solid' referred to their reliability/performance! When I was more heavily involved with the replacement of components, I think the descriptions and selling of the various components was done more on the basis of quality/tolerances, rather than using 'buzz words', hype or 'suggestion'. I'm just a CGLI Elect Tech and simply replaced one broken component with another component of the same working parameters. In my case, to re-cap, I replaced a cap that was short and fat and had no plastic sleeving around it, with one that was taller, narrower, had sleeving and appeared to be of lower quality, but the MSI board still works just fine. Incidentally, I have come across large electro-caps with integral screening cans, which were used in TV transmitters.......

---

Don't take life too seriously, as you'll never come out of it alive!

The term 'solid' refers to the internal construction of the capacitors.
'Solid' types do not use a liquid electrolyte, and hence are not subject to drying out or internal gassing that eventually leads to the failure or exploding of standard electrolytic types.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Okay so I've had a few more freezes/lock-ups/crashes in the past few weeks. They've gotten progressively worse/more random, too. I can play a heavy graphics but light memory (~700mb) game for hours and hours and it's fine, but watching a 1080p movie in VLC will be fine for 30 seconds to an hour, and then it just freezes, and everything starts going downhill in a hurry. Explorer.exe doesn't respond (taskbar, etc.), then applications try to close but don't. This random crash/freeze/hang causes a handful of applications/services that are running to corrupt their .ini file, so I've started keeping "known good" backups of them.

After four of these crashes this evening, I decided to run memtest just to see if it was a bad DIMM. Since I have to go inside the case and swap the SATA cable for the optical drive from one controller to another, I noticed that the big cap next to the RAM bank that I replaced with a larger capacity one is already fairly bulged, and two more smaller ones next to the RAM are also bulged. [sarcasm] Great! [/sarcasm]

Ran memtest anyway for about 30 minutes and it only made it about 30% through the first complete pass, and I just shut it down. Moved all the RAM to the cpu_0 bank, and thankfully my board isn't stupid. BIOS shows 2 physical CPUs, 4 logical CPUs, 4096mb RAM. Windows shows 4 cores and ~3300mb of memory. While I was in BIOS, I also turned ECC on and enabled BG Scrubbing. If I see a huge performance hit, I'll turn ECC off, but I know I'm having memory problems, so the more fault-tolerant I can be with that, the better.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)