Capacitor identification help & replacement help

[NYEngineer]

Hi all,

I have a 1990 vintage computer CPU board with about 10 of these capacitors on it. Can anyone please help me identify what kind they are, and if you're really bored or know offhand, a reasonable replacement from Mouser, Digikey, or other US vendor?

They're axial, look like electrolytic (or maybe styrene?), and are marked:

22 UF
15 DC
150D±10%

I am supposing they are electrolytic 22 microF 15V 10% tolerance, but really have no idea what the 150D might mean.

I wonder if replacing them with axial tantalum capacitors would be okay or not.

Pictures attached. Any ideas please? Thanks in advance!

[Gyro]

Those are Tantalum, dry ones. You shouldn't need to touch them unless you've found a shorted one (unlikely with those, it's the bead ones that fail).

P.S. Data here... http://www.vishay.com/capacitors/list/product-40015/ Hope it helps.

[NYEngineer]

Those are Tantalum, dry ones. You shouldn't need to touch them unless you've found a shorted one (unlikely with those, it's the bead ones that fail).

P.S. Data here... http://www.vishay.com/capacitors/list/product-40015/ Hope it helps.

Totally awesome and fast, thank you so much. Is there a way I can test these in circuit? Presumably shorted ones would be an easy continuity or resistance test.

My guess is that these are spread out around the board to handle current spikes and they probably all go from positive to negative 5V planes on the 4-layer circuit board. But, I don't really know. If that's the case, an LCR on the leads while the cap is in place won't tell me anything.

[Gyro]

You're welcome. 

Yes, the caps are almost certainly bulk decouplers between +5V and ground. They're decent quality hermetic ones without any specific failure modes unlike Aluminum electrolytics, so personally I wouldn't go poking them with a hot soldering iron for testing. There's probably a lot of redundancy there too.

The easiest thing is to just bring the board up on a current limited supply, if you can, to check that there are no 5V to ground shorts.

[NYEngineer]

The easiest thing is to just bring the board up on a current limited supply, if you can, to check that there are no 5V to ground shorts.

I don't really have this ability. The power supply has four outputs: +5V (running about 5.12V), -5.2V (running about -5.25V), and +/- 12V. I honestly have no idea how to do what you suggest. I have no idea what the -5.2V is used for, for that matter.

That said, wouldn't a resistance test with my Fluke show if any of these are shorted?

Thanks!

[Cubdriver]

Considering that the supply seems to be happy, it doesn't appear that any of them are shorted.

-Pat

[Gyro]

Agreed, if the supply is happy then they are fine. Yes a resistance test should catch any gross shorts, you'll see leakage current from other circuitry and the 'resistance' will rise as the caps charge.

[NYEngineer]

I measured the resistance of over 30 of these across the two (VMEbus-like) boards of the vintage computer. They ranged from 5 ohms for a few, to 8 ohms for most of them, to dozens of kilohms for the rest.

The computer is unstable and will crash after about 20 minutes from a cold power on. The 5V and 12V rails seem stable, according to a Fluke 289 logging multimeter logged on 1-second intervals. No "events" when it starts misbehaving.

I think it wouldn't hurt to replace a few of these, especially the ones with 5 ohms resistance. However, I have never worked on a 4-layer PCB. Does anyone have any thoughts on how to replace through-hole tant caps on a 4-layer board?

I'm not sure how to apply "Gyro's" post about the resistance rising. Is that something I should look at?

Thanks!

[Cubdriver]

I measured the resistance of over 30 of these across the two (VMEbus-like) boards of the vintage computer. They ranged from 5 ohms for a few, to 8 ohms for most of them, to dozens of kilohms for the rest.

I'm not sure how much I'd trust those 5 and 8 ohm readings - a cap leaking to the point that it has 5 ohms of resistance put across a 5V bus will draw 1 amp, and I'd think things would be getting unnaturally hot especially with several of them there!

The computer is unstable and will crash after about 20 minutes from a cold power on. The 5V and 12V rails seem stable, according to a Fluke 289 logging multimeter logged on 1-second intervals. No "events" when it starts misbehaving.

I know enough about computers to be extremely dangerous with them, so I'm not going to touch this other than to say that this may not prove anything, as a lot could happen on the rails in between the readings - the sort of power glitch that could cause a computer to gork is more in scope territory than DMM.  I'll defer to someone more knowledgeable on that one.

I think it wouldn't hurt to replace a few of these, especially the ones with 5 ohms resistance. However, I have never worked on a 4-layer PCB. Does anyone have any thoughts on how to replace through-hole tant caps on a 4-layer board?

You'd want a solder sucker to clear the holes, and a decent iron for component removal.  Since you'd be pulling them as duds and don't care about the part being removed, I'd cut the longer of the two leads off as close to the cap as possible, then heat and remove the cut lead and cap/remaining lead one at a time by heating and pulling them out with pliers.  Use the solder sucker (after adding some fresh solder) to clear our the hole.  All that being said, I wouldn't do anything to the caps unless I was quite certain they were bad, and I don't think you're there just yet.

I'm not sure how to apply "Gyro's" post about the resistance rising. Is that something I should look at?

Thanks!

He's referring to how a cap behaves when connected to an ohmmeter.  When first connected (assuming it's discharged, and relatively large - more than a few microfarads), it will appear electrically as a near short and draw as much charging current as the source (in this case the ohmmeter) and wiring can supply.  As it begins to charge, its apparent 'resistance' will increase as the capacitor voltage rises and the charging current decreases.  Ultimately, the cap will effectively reach the same voltage as that supplied by the ohmmeter and stop charging.  In theory at this point it should be an open circuit, with infinite resistance.  It's easiest to see this on an analog meter - when first connected, the needle will quickly deflect towards 0, then move back towards infinity as the cap charges.  The larger the cap, the more time it will take for the needle to go back to the high end of the meter scale.  The following video found on a quick search of youtube shows it:


You can also search on RC time constant for a better explanation of caps charging and discharging.

-Pat

[NYEngineer]

...

First off, Pat, thank you for your detailed response!

Now: as to the Fluke 289 and its logging, it also counts "events." I'm not an expert on this complex instrument but I those "events" are things that happen below the resolution of the capture interval. It shows a count of these things, but to see the details you have to download the data to a computer (which I didn't do). However, I did not notice the "event count" increase, so maybe it still didn't see anything. I have a Rigol DS1000E so I could also scope it. (I did that in the past to measure the stability of the power supply; it doesn't have more than 80 mV of ripple on the 5V or 12V lines at around 10-20 ms period. Pretty noisy looking on the scope until you use the averaging feature.) Since the computer starts acting weird after 15-20 minutes I doubt I could record enough data in enough resolution to see anything, anyway, without having hugely fast reflexes (and it's super boring to watch something to see when it crashes).

All that being said, I wouldn't do anything to the caps unless I was quite certain they were bad, and I don't think you're there just yet.

I welcome thoughts as to what else to do. Here are my thoughts:

1. Next time it crashes, quickly disassemble the computer and look at it at with a FLIRone camera. This essentially entails pulling off a backplane connector, pulling out two VMEbus-style boards, and examining them and the internal backplane. They will probably cool down quickly during this so I am not sure it will be of any value.

2. Replace or recap the power supply, see if it helps. It has 4 feeds: 5V, 12V, -12V and -5.2V (no idea what that last one is for).

3. Pull all socketed chips, spray contact cleaner in the chips and sockets, and reseat. There are probably 60+ normal sized DIP chips, plus maybe a dozen large DIP chips, three huge square chips (one being a Weitek FPU) and one that is a chip with pins on the side that goes into a square slot with pins on the side that I have no idea how to remove. I fear doing this because I haven't removed chips in decades and breaking any of them spells the permanent end of this irreplaceable computer.

4. Measure current on the 5V and 12V rails. I believe this power supply puts out 35A on 5V though, so I'd need quite some big connectors for the Fluke. (Todd MAX-354-1252) A watt meter connected to the 110V line shows it drawing about 230W in stable state operation. See how stable that current is. I'd have to rig something up to get the current through the meter, or get one of those probes that can go around the wire (but those are ridiculously expensive when I last looked).

Other ideas?

He's referring to how a cap behaves when connected to an ohmmeter.

I understand how caps behave in general; my question was more about how to use the resistance function of the meter to discern whether a cap may be bad in-circuit. I don't know what sort of voltage or current the Fluke 289 puts out when measuring resistance, but I can't imagine it's much. For the caps that measure in the kilohms, you definitely see the resistance change while measuring them. For the ones with 5 and 8 ohms, the resistances stay pretty stable.

Frankly, I don't know how to test the caps in-circuit or if there is even a reasonable way to do it. If these are bypass caps as I expect they are, then measuring one is probably like measuring them all (and makes me wonder why there would be different resistances at all in the first place).

...instructions on replacing the caps...

So, it sounds to me that replacing the caps on a 4 layer PCB is no different than doing it on a regular PCB, from your general description. I was concerned that there might be connections, somehow, to the internal layers. The solder exists on both the top and bottom of the board for each cap, either by design or pulled through by capillary action.

Thanks!

Doug

[SeanB]

Wet slug tantalum capacitors, almost invariably the most reliable component ever, there are  literally millions of them used on space craft, and they are the exact ones you have here n the metal case with a glass seal. You are more likely to have a loose connection in a socket, or a cracked socket pin, especially if it is a rolled pin socket and not a turned pin one, they are known to snap inside the body and go intermittent.  Remove each socketed chip and check the pins are straight then reinsert, you will find that fixes most issues, though the broken pin is hard to find, you can only look for the single stamped pins that look slightly different in alignment than the others, and then check continuity on all of the socket pins while the IC is plugged into it and applying pressure to the IC.

[Gyro]

I think you're probably barking up the wrong tree with the caps on the board too (In fact I'm be more likely to suspect the PSU in that respect). As Sean says, there are plenty of other potential candidates for your problem. I certainly wouldn't go unsoldering them from a multilayer board at this stage, you'll most likely cause more damage than good. I know that 'replace all of the capacitors' is a common theme on this forum, but that's cheap little consumer grade things not like what your computer is using.

Instead of scatter gunning this problem I think you need to get a bit more methodical. Do you have any diagnostic software, can you write a simple memory test for instance, it could just as easilly be a memory related issue if you see what I mean.

BTW, when I answered your cap question yesterday I didn't realise you already have an active thread running on this. Running two threads in parallel is never a good thing - you want to concentrate all the ideas an input in one thread so everyone knows what everyone else is suggesting rather than duplicating.

[NYEngineer]

BTW, when I answered your cap question yesterday I didn't realise you already have an active thread running on this. Running two threads in parallel is never a good thing - you want to concentrate all the ideas an input in one thread so everyone knows what everyone else is suggesting rather than duplicating.

Yes, my intention was not to have two threads about this. One was about caps, the other was to find vendors, but they have morphed.

[NYEngineer]

Instead of scatter gunning this problem I think you need to get a bit more methodical. Do you have any diagnostic software, can you write a simple memory test for instance, it could just as easilly be a memory related issue if you see what I mean.

I have a diagnostic program that runs a test of all RAM. It's ECC RAM. It says it tests all devices, but doesn't seem to really do that.

I also have two sets of RAM for the machine. The computer crashes in the same way regardless of which RAM boards I have installed. So, I don't think it's the RAM.

I guess I will have to remove and clean each chip. That will probably take quite some time given the huge count. I don't know about removing the PLCC (there's one) and the three large PGAs.

I also found a place which will refurbish the power supply (for a not inconsiderable amount of money).

Also, the reason I suspect these tantalum caps is that the person who sold me this computer said he had three of them fail on another one of his. So, it's not a strong reason.

Thanks!

[Gyro]

At the risk of extending your second thread (sorry)...

All useful information, you have clearly covered a lot of the bases in terms of faultfinding, RAM etc. 

Yes, you may well be in for a long haul in terms of checking chips, connectors etc. You would probably do well to find someone with a scope to actually check the level of ripple on the supply rails.

In terms of the capacitors, I understand now your reasons for suspecting them. Two possibilities if so:

1. As Cubdriver said, if any of them are leaking (electrically) enough to make a difference then they will heat up significantly and so should be easy to identify.
2. Second option, one or more has lost enough capacitance to increase supply noise. In that case, soldering a low ESR electrolytic cap in parallel will cause an improvement. It shouldn't cause any harm and might be a good diagnostic step. Obviously be careful of polarity.

@SeanB: These are solid dry Tantalums (Sprague, now Vishay, Tantalex 150D), not Wet Tants. Still reliable but not as rugged in low impedance situations.

[helius]

Now that picture looks familiar!

I have no idea what the -5.2V is used for, for that matter.

It is used for powering ECL circuits.
https://en.wikipedia.org/wiki/Emitter-coupled_logic

Does the computer boot? Does it boot if you leave it powered on for 20 minutes before booting? Better exclude software issues or problems with some peripheral that are triggered by software.

You may want to contact computer museums that have experience restoring these machines. I know one in Rhode Island has done so.

[NYEngineer]

Hello again all,

Given the comments here in this thread and the arrival of the FLIRone camera in the mail I decided to go through and see if I could see anything on the FLIR "heating up."

In using the FLIR on the power supply while waiting for the 20 minute crash, I was surprised to find a capacitor that seems to be leaking in the computer power supply. (All without the FLIR.) I know for certain that this was not obviously bad when I first received the computer as I did a very detailed visual inspection of all caps. However - see for yourself.

This capacitor is a large pink axial 2.2uF, 250V, 10%. The markings are hard to read but there's another seemingly identical one in the computer's console/CRT power supply which is easier to read. 522/255-1, 2.2uF, +/- 10%, 250V - W3, -55ºC / +100ºC. There are two more lines that are hard to read, one which seems to say "EAU" or "EAO" on an angle in a rectangular box followed by MK _1_ where "_" means I can't make it out. Finally a barely legible line seems to say "2 2/250". It is about 13mm in diameter and 30mm long as measured by caliper.

Not sure what kind it is, but googling some of the numbers implies "metalized polyester film capacitor" which Google's pictures don't actually look like this one. My totally uneducated guess is still that it's some sort of a film capacitor though.

It's from a Todd MAX-354-1252. It looks like the power supply here: http://www.slpower.com/data/collateral/MAX-250_DD.pdf - indeed, you can see the pink cap between the two large ones in the picture.

Any thoughts on this? What's it's purpose? It's nestled between two large 1,200uF 200V caps. I'll put another picture or two in the next post.

Thanks!

Doug

[NYEngineer]

Here are two pics of the seemingly same cap from the console power supply.

What is the replacement please?

Thanks,

Doug

[NYEngineer]

It is used for powering ECL circuits.
https://en.wikipedia.org/wiki/Emitter-coupled_logic

Thanks!

Does the computer boot? Does it boot if you leave it powered on for 20 minutes before booting?

The computer boots and runs fine for about 20 minutes. It starts getting flakey (spurious errors in programs) and then crashes to the point where it won't boot past it's equivalent of loading the MBR in a PC. I could stop the boot process and leave it sitting for 20-30 minutes at the PROM prompt to try your test. However, given the issue above, I think I'm done with powering this thing on until I replace that cap or the whole PS.

You may want to contact computer museums that have experience restoring these machines. I know one in Rhode Island has done so.

(Parenthetically, I want to visit that one, one weekend. Which is good, cause it's only open on Saturdays with appointments in advance.)

Thanks,

Doug

[Cubdriver]

What is its normal orientation?  That doesn't so much look like the capacitor itself leaked (film caps don't really have anything inside them to leak, unlike electrolytics) as it does that something like rosin or something similar in color that has dripped on the end of the cap in question.

-Pat

[helius]

Any thoughts on this? What's it's purpose? It's nestled between two large 1,200uF 200V caps. I'll put another picture or two in the next post.

It's a conducted EMI/RFI suppression capacitor, to prevent noise on the power line from entering the machine.
The brand is ERO, which was folded into Roederstein under the Vishay conglomerate.

The original brand: http://www.electronicsurplus.com/ero-f1772-522-2000-capacitor-2-2uf-250vac
The currently manufactured replacement: http://www.mouser.com/ProductDetail/Vishay-Roederstein/F17735222000/

I agree that this does not look like leakage from the cap, but something dripped onto it from overhead. In the first picture above, the yellow drop cap on the left has some sort of dust or particles on it. Is it just dust, or fragments ejected by an exploded component?

[helius]

I think what I would try to do, if careful inspection of the boards didn't reveal any visible damage, is to use a digital oscilloscope on each power rail, set to trigger on any glitch (if your scope doesn't have a glitch trigger, you can just use a slope trigger instead). Then when the rail is stable it should not trigger, but if something happens to disrupt power you will capture a trace. From what you know so far, you can't conclude that the power simply goes bad after it warms up: even a momentary disruption in power will cause a crash, and you won't be able to see it if your eyes are elsewhere. Whether the glitch happens on all rails or just one will point directly to the area causing the problem.

If the problem is caused by heating, it may not necessarily be in the power supply. Both cold solder joints and pin sockets are affected by thermal expansion and can manifest as this kind of problem. This would not appear as an obvious hot spot on a thermal image, since the same heating is happening all over the board but it just happens to break a weak joint.

[helius]

You could also test the power supply in isolation if you can put together a load capable of dissipating enough power.
It's not a real great idea to operate a switching PSU with no load, because often they require a load to stay stable.
But a load on each rail that's between 25-50% of the rated current should work. You can build a passive load from wirewound resistors, incandescent bulbs, or loops of wire submerged in water:



The idea is that needing to repair the PSU could be expensive, but at least all the components are available (some may need to be substituted). Whereas the logic boards contain chips that are unobtainium with no substitute. One thing to keep in mind is that your loads should connect securely to the PSU's outputs, whether they use ring terminals or pin contacts. You don't want them to slip or get hot from being loose. Once the loads are securely connected you can activate the enable pin on the PSU (usually a 1k or 0r to ground)

[NYEngineer]

if careful inspection of the boards didn't reveal any visible damage

I can go over the boards with a loupe and check more closely. I'm on the cusp of needing reading glasses (actually, I'm past the cusp but in denial) so when I was having problems with the keyboard I didn't think anything was wrong until I noticed bad solder joints on several keys after looking at the pictures I took with a macro lens.

use a digital oscilloscope on each power rail, set to trigger on any glitch (if your scope doesn't have a glitch trigger, you can just use a slope trigger instead).

I have only the most basic oscilloscope in the world, a Rigol DS1000E 2-channel digital. (I actually have a Tek 2225 too, but it's way out of calibration.) I've only used it to teach my kids about Arduino and its analog and digital outputs a few years ago, so I'm not remotely an expert in its use. I did try to characterize the 5V and 12V ripple using it (I'll attach pics) but I would have to do a lot of learning to figure out how to set a slope or glitch trigger.

I can also order a Rigol DS1074Z-S Plus from Amazon for Sunday delivery. Maybe that has glitch triggering? I'd prefer the bigger screen too, I'm sure. With 4 channels, I could also have it watch all four power rails concurrently. One interesting thing about the power supply is that it has three different ground points. One for 5V, one for +/- 12V, and one for -5.2V. That last one seems to be also tied to chassis and the "ground pin" of the power plug. I noticed no voltage between the various ground pins so not sure why there are different.

...but something dripped onto it from overhead. In the first picture above, the yellow drop cap on the left has some sort of dust or particles on it. Is it just dust, or fragments ejected by an exploded component?

I didn't notice any obvious problems, and the power supply is quite dusty (even after being air-can blown out). So, it's probably dust, but I will double check later this weekend.

Finally, here are some pictures of the power on the 5V and 12V rails I took earlier this week showing the ripple. This was, however, soon after power-up. I also did data logging with a Fluke 289 and it didn't see anything on those rails. (It took me a while to remember I could switch to AC coupling to see the ripple much more easily!)

Thanks,

Doug

[NYEngineer]

What is its normal orientation?  That doesn't so much look like the capacitor itself leaked (film caps don't really have anything inside them to leak, unlike electrolytics) as it does that something like rosin or something similar in color that has dripped on the end of the cap in question.

Just back from a weekend out. Just checked, and it does indeed seem this orange stuff is external to the cap.

The cap is normally horizontally oriented (the axis is parallel to the ground).

I should have a 4-channel Rigol to test for any glitches later today, so I'll let you know how that goes, assuming I can figure out how to work with it.

Cheers,

Doug