VOGONS


First post, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

After recently experiencing my first tantalum explosion, I'm feeling a bit less adventurous when it comes to using 25+ year old cards and motherboards with tantalum caps. Especially ones that already have failed tantalums.

I don't really like the idea of using "new" tantalums as replacements, as its hard to know how old they are if you find some, and they could potentially be only a few years away from exploding already depending on how long they've sat in storage. Plus, they aren't cheap.

I used an electrolytic I had laying around to fix a New old stock EGA card I bought last year and it worked fine, but it was only one cap. If I'm redoing a whole board, I want to be sure the replacements are functionally similar enough to keep the device operating as designed.

Longevity is also a concern. I have read that modern multi layer ceramic caps (MLCC) can be suitable replacements for tantalums in some situations these days, but I'm not that savvy with the electrical aspects of these things to know why this would or wouldn't work on, for example, an old video card or motherboard. I like the idea of using ceramics, because they are so easy to work with, durable and reliable.

Just as an example, I'm looking to replace dozens of 10uf 16v tantalums on several different boards. I'm looking at these for possible replacements:

https://www.mouser.com/ProductDetail/TD ... Qvfayqc%3d

Would those work? What would be a suitable electrolytic alternative?

Now for some blitting from the back buffer.

Reply 1 of 12, by Evert

User metadata
Rank Member
Rank
Member

You can replace Tantalums with MLCC's, but take not of the voltage bias. They lose their capacitance as the voltage that is applied to them, gets closer to the rated value. A through-hole polymer capacitor is a better replacement in my opinion. However, you can offset the voltage bias by using MLCC's with high voltage ratings.

sigpic2689_1.gif

Reply 2 of 12, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie
Evert wrote:

You can replace Tantalums with MLCC's, but take not of the voltage bias. They lose their capacitance as the voltage that is applied to them, gets closer to the rated value. A through-hole polymer capacitor is a better replacement in my opinion. However, you can offset the voltage bias by using MLCC's with high voltage ratings.

The usual solution to this is to have MLCCs with higher voltage rating to begin with, or same capacitance and voltage rating in larger package.
The capacitance derating starts quite early, and in fact a 16V 10uF capacitor could have as little as 20% capacitance rating, (2uF) when used at 12V.

MLCCs also have much lower ESR than tantalums, so if the circuit is the output capacitor of a LDO regulator, that is a bad thing. Otherwise, lower ESR is usually a good thing.

These caps might be good for replacing generic tantalum bypass caps on a motherboard. Or then change to aluminium electrolytics, they normally have a bit higher ESR than tantalum electrolytics, but you can choose a low-ESR aluminium electrolytic or maybe a polymer electrolytic to get lower ESR.

Reply 3 of 12, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Evert wrote:

You can replace Tantalums with MLCC's, but take not of the voltage bias. They lose their capacitance as the voltage that is applied to them, gets closer to the rated value. A through-hole polymer capacitor is a better replacement in my opinion. However, you can offset the voltage bias by using MLCC's with high voltage ratings.

Would something like this be suitable?

https://www.mouser.com/ProductDetail/KE ... bDtw%3d%3d

These electrolytics are significantly cheaper... would there be any noticeable difference for most uses?

https://www.mouser.com/ProductDetail/Ni ... hPFFGt0%3d

These ones are a little larger but according to the datasheet they appear to have lower impedance... and that's a good thing when replacing tantalums, right? 😊

https://www.mouser.com/ProductDetail/Ni ... aO1xpBw%3d

Now for some blitting from the back buffer.

Reply 4 of 12, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Jepael wrote:
The usual solution to this is to have MLCCs with higher voltage rating to begin with, or same capacitance and voltage rating in […]
Show full quote
Evert wrote:

You can replace Tantalums with MLCC's, but take not of the voltage bias. They lose their capacitance as the voltage that is applied to them, gets closer to the rated value. A through-hole polymer capacitor is a better replacement in my opinion. However, you can offset the voltage bias by using MLCC's with high voltage ratings.

The usual solution to this is to have MLCCs with higher voltage rating to begin with, or same capacitance and voltage rating in larger package.
The capacitance derating starts quite early, and in fact a 16V 10uF capacitor could have as little as 20% capacitance rating, (2uF) when used at 12V.

MLCCs also have much lower ESR than tantalums, so if the circuit is the output capacitor of a LDO regulator, that is a bad thing. Otherwise, lower ESR is usually a good thing.

These caps might be good for replacing generic tantalum bypass caps on a motherboard. Or then change to aluminium electrolytics, they normally have a bit higher ESR than tantalum electrolytics, but you can choose a low-ESR aluminium electrolytic or maybe a polymer electrolytic to get lower ESR.

Thanks for the info. So, an MLCC with a significantly higher voltage rating that still physically fits and has the correct capacitance could be a decent option to replace most tantalums, aside from the situation you mentioned? Does the dielectric code matter for usage well below the maximum temperature? I see the codes like X7R, X7S and X8L, but those ratings just seem to be related to how much the capacitance changes beyond a certain temperature (which is orders of magnitude hotter than any PC environment).

The prices for these caps are very high though. In quantities of 100 they're still barely under $1 each, which is way too high for me.

Are there any situations where a good quality low-esr aluminum electrolytic would just not work in place of a tantalum? They're so much cheaper its hard not to want to go that way and order a whole pile of them.

Now for some blitting from the back buffer.

Reply 5 of 12, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Sorry, have to keep making posts because I can't edit...

How do I actually find and identify Low ESR caps that are suitable for motherboards, video cards, sound cards, etc.? Looking on Mouser, some say "Low ESR" others don't mention ESR or even Impedance anywhere. You'd think they would have a way to just see all of the caps that are low-ESR...

Now for some blitting from the back buffer.

Reply 6 of 12, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

I believe Panasonic is now recommending polymers for design replacement of tantalums.

All hail the Great Capacitor Brand Finder

Reply 7 of 12, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
gdjacobs wrote:

I believe Panasonic is now recommending polymers for design replacement of tantalums.

That's good to know. I'm going back and forth between 9 cent low-esr electrolytics and these 16 cent polymers.

I have also come across some devices that have rectangular surface mount tantalum caps of the same capacitance (10uf, 25v). Is there any good way to use the same radial lead polymer caps to replace these? Or is that a really bad idea? It sounds fiddly to me, but if people do it I'd give it a shot.

Now for some blitting from the back buffer.

Reply 8 of 12, by Unknown_K

User metadata
Rank Oldbie
Rank
Oldbie

How big of a problem are blown tantalum capacitors? I have seen a few on 90's era gear but nothing like the issues we have with aluminum SMT capacitors.

Having the black mark on aluminum capacitors be the negative side while tants are the positive side can cause some fireworks.

Collector of old computers, hardware, and software

Reply 9 of 12, by RogueTrip2012

User metadata
Rank Oldbie
Rank
Oldbie
Unknown_K wrote:

How big of a problem are blown tantalum capacitors? I have seen a few on 90's era gear but nothing like the issues we have with aluminum SMT capacitors.

Having the black mark on aluminum capacitors be the negative side while tants are the positive side can cause some fireworks.

In my experience Tants and ceramics CAN short the circuit its in causing other things in the line to blow and be a PITA to troubleshoot. Alu Electrolytic can leak electrolyte on the board that will eat copper traces and other metals but they usually open in resistance not always harming the circuit. I've seen exceptions of course.

I haven't worked around Polymers enough yet to see how they hold up.

All types can blow up with nice pop and sometimes smoke and fireworks. They alone can ruin a circuit board if damage is bad enough.

> W98SE . P3 1.4S . 512MB . Q.FX3K . SB Live! . 64GB SSD
>WXP/W8.1 . AMD 960T . 8GB . GTX285 . SB X-Fi . 128GB SSD
> Win XI . i7 12700k . 32GB . GTX1070TI . 512GB NVME

Reply 10 of 12, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

One issue with ceramics is their tendancy to generate electrical noise due to piezoelectric effects.

All hail the Great Capacitor Brand Finder

Reply 11 of 12, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie
Ozzuneoj wrote:

I have also come across some devices that have rectangular surface mount tantalum caps of the same capacitance (10uf, 25v). Is there any good way to use the same radial lead polymer caps to replace these? Or is that a really bad idea? It sounds fiddly to me, but if people do it I'd give it a shot.

I'd say if you get a surface mount polymer it should be possible. But the packages and component footprints on circuit board won't match, so you have to be creative how to select a polymer which you can solder in place of tantalum.

gdjacobs wrote:

One issue with ceramics is their tendancy to generate electrical noise due to piezoelectric effects.

True - but nowadays microphonics should not be a huge issue any more, and even if it were, it should not cause much problems when the capacitor is used only as a power supply bypass instead of some high impedance sensitive circuit. I could be totally wrong though 😀

Reply 12 of 12, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Okay, the more I look this up, the more complicated it gets.

Now, I've read some comments online from people saying that the age of the tantalum doesn't matter, its that the rest of the circuit has changed over time in a way that overloads a tantalum, causing it to explode. I've never heard this before. They say that it would be best to use tantalums rated for significantly more voltage than the ones being replaced...

I'm also seeing very few people recommending completely replacing all old tantalums. Most say that it is good enough to only replace ones that have failed (either shorted or exploded). Ones that haven't failed after many years aren't likely to explode\fail any time soon.

This is fine, except they can potentially damage a board or other components nearby, no to mention the safety hazards. My biggest problem, honestly, is the stink and smoke! Its horrible, and it can't be healthy. It is 10 degrees F outside here and won't be much warmer until maybe April. Having to open windows this time of year is a very bad thing. I guess I can just hook up all of my tantalum-equipped stuff outside while wearing eye protection, but that seems like such a haphazard way of going about things. Also, what if one tantalum pops, I replace it, and then the next one in the chain goes? Seems like a lot of risk...

Now for some blitting from the back buffer.