electrolytic caps VS. tantalum caps

[chemelec]

I also read the tantalum comes from a Gorilla habitat and that it is damaging that habitat by us taking it?

I Really Don't belive that statement.

 

[mramos1]

Gorillas are breaking in the plant at night, doing bad things to the caps and leaving?? Not sure I believe it either..

 

[Sceadwian]

Nigel, you keep stateing that tantalum's are very unreliable or have been in the past, what are the failure mode's they're known for, and specifically why do they fail? The link Gramo posted seems to praise tanlums in pretty much every deparment, except price? chemelec doesn't recommend them for use in 'brute force' filtering, but I have some from a bulk cap pack I got and was thinking about using them as motor filter caps eventually.

 

[gramo]

They are both good to a degree, and both are necessary for some circuits.

Tantalum are capable of storing much more energy than an electrolytic - But they discharge about twice as fast; hence there use in high frequency designs. Also Tantalum's have a 'infinite' life span. They are the superior product, and you pay for what you get, its just too bad that Tantalum is one of the most rarest materials.

You would not use a Tantalum in a high current circuit such as a DC motor power supply - an electrolytic would be more suited. Electrolytic have a much slower charge/discharge time, hence a lower value Tantalum can replace a higher value Electrolytic. Electrolytic’s also have a rather short life span.

Not sure if it’s an issue with Nigel, but many people don’t realize Tantalum's are polarized components as well, and incorrect use will make them very unreliable.

 

[Oznog]

I Really Don't belive that statement.

Actually true. Tantalum caps are made from coltan ore, found primarily in the politically unstable Congo. It has helped fuel a bloody war. Most of the ore on the market has probably been smuggled out of the Congo and sold in neighboring countries. The mining operations may adversely affect the threaten Eastern Lowland Gorilla population.

https://en.wikipedia.org/wiki/Coltan

 

[Nigel Goodwin]

Nigel, you keep stateing that tantalum's are very unreliable or have been in the past, what are the failure mode's they're known for, and specifically why do they fail?

They always fail S/C, usually as supply decouplers on PCB's fed from regulated supplies - but I suspect that's probably just because that where they tended to be used?.

As to why they fail?, I've no idea - as far as I remember they were usually 25V rating on 12V regulated lines? - I've always assumed it's because they are crap!

I don't know if people here remember the valve days?, but a common capacitor manufacturer was Wima (non-electrolytics) - and those were also extremely unreliable. To the extent that if you ever saw them used in a piece of equipment for repair, the first thing you did was replace them all - and that would cure 90%+ of the faults. Tantalums aren't quite that bad, but if you see any on a board the first thing you do is check them all to see if they are S/C - this finds the majority of faults (or did, back when they were used!).

 

[CanonMan75]

I don't know if people here remember the valve days?, but a common capacitor manufacturer was Wima (non-electrolytics) - and those were also extremely unreliable. To the extent that if you ever saw them used in a piece of equipment for repair, the first thing you did was replace them all - and that would cure 90%+ of the faults. Tantalums aren't quite that bad, but if you see any on a board the first thing you do is check them all to see if they are S/C - this finds the majority of faults (or did, back when they were used!).

Were Wima caps as bad as Hunts?

 

[eng1]

On regulated lines the failure rate increases because of the low output impedance of IC regulators. In those applications the operating voltage should be much less than the rated one.

 

[Nigel Goodwin]

On regulated lines the failure rate increases because of the low output impedance of IC regulators. In those applications the operating voltage should be much less than the rated one.

This was pre-IC regulators - but in any case that doesn't make much sense?.

 

[eng1]

This was pre-IC regulators - but in any case that doesn't make much sense?.

Some app notes suggest that the operating voltage should be 30% of the rated one, when tantalum caps are used with regulated power supplies. If the output voltage is 12 V or more, 36 V rated tantalum caps are expensive. And aluminum electrolytic technology has improved... they fearure low ESR as well and higher voltages.

An interesting reference: https://www.electro-tech-online.com/custompdfs/2007/01/e_infotumtal.pdf

 

[Nigel Goodwin]

Some app notes suggest that the operating voltage should be 30% of the rated one, when tantalum caps are used with regulated power supplies. If the output voltage is 12 V or more, 36 V rated tantalum caps are expensive. And aluminum electrolytic technology has improved... they fearure low ESR as well and higher voltages.

So even the manufacturers think they are crap as well! - what's the point of printing 36V on a capacitor then hiding away in the datasheet that you shouldn't use it above 12V?.

 

[eng1]

Nigel, I think that you'll never choose tantalum capacitors!?!

 

[Nigel Goodwin]

Nigel, I think that you'll never choose tantalum capacitors!?!

Except for the one case I suggested! - I'm even more against them now you've posted that Hitachi link!.

 

[eng1]

As I said, I always use them after 5 V LDO regulators, that require a low ESR.
10 uF - 16 V tant caps are cheap and I find them easily. As I don't know exactly the ESR of my aluminum caps, I prefer to avoid failures!

 

[chemelec]

Nigel Goodwin, Since your mostly opposed to Tantalums, Here is Something else you might Consider when using Electrolytics.

The Physically Bigger the cap is, Generally speaking the Lower its ESR is!

So for two caps of the Same Voltage and Capacitance, if you Need Low ESR, Use the Bigger one.

Actually, Testing my Tantalums on my "Sencore LC102", Capacitor/Inductor Analyser, this also holds true for these types. "Physically Bigger is Better in this way".
If you Have the Room.

Also, Testing a lot of NEW Electrolytics that I have just Bought, I find most of them FAIL for the ESR Test.

 

[Nigel Goodwin]

Nigel Goodwin, Since your mostly opposed to Tantalums,

I've never seen any reason to trust them?, and the lack of them now in anything I ever see seems to prove me right. Perhaps tantalums are more reliable now?, but domestic electronics manufacturers don't seem to be convinced after the previous disasters!.

I'm also somewhat dubious about using components with have a specific rating, then a warning in the datasheet not to exceed 1/3 of that rating - it seems a poor way to specify a component!.

 

[Optikon]

On regulated lines the failure rate increases because of the low output impedance of IC regulators. In those applications the operating voltage should be much less than the rated one.

Actually, this problem is not as bad when they are on the outputs of IC regulators. A much worse scenario exists when they are on the "input" side. This side _usually_ connect to an even lower impedance source and the startup surge current often does these caps in. A filtering inductor in the path helps the situation greatly. IC regulators usually have current limit built into them and I often wonder if this is responsible for output tantalums surviving repeated startup surge currents.

 

[Optikon]

I've never seen any reason to trust them?, and the lack of them now in anything I ever see seems to prove me right. Perhaps tantalums are more reliable now?, but domestic electronics manufacturers don't seem to be convinced after the previous disasters!.

I'm also somewhat dubious about using components with have a specific rating, then a warning in the datasheet not to exceed 1/3 of that rating - it seems a poor way to specify a component!.

I believe if you specify a quality manufacturer and use a trusted distributor you can get quality tantalums. They can be used reliably if you apply modest de-rating and watch out for repetetive surge situations. By the way, you can get fused tantalums for those who remain ultra-paranoid.

Also generally, de-rating components will improve the longevity no matter what the compopnent. So if the tantalum manufacturers are putting de-rating warnings into the dataheets, consider it a blessing. When was the last time you saw an aluminum electrolytic datasheet that read:

CAUTION: The electrolyte in this capacitor will dry out over time and cause the component to fail. It will likely be the most unreliable part in your design. Please de-rate generously to imporve longevity. Enjoy!

 

[eng1]

Actually, this problem is not as bad when they are on the outputs of IC regulators. A much worse scenario exists when they are on the "input" side. This side _usually_ connect to an even lower impedance source and the startup surge current often does these caps in.

Personally, I use aluminum type at the input because they're cheaper and low ESR is required at the output to avoid oscillation.

IC regulators usually have current limit built into them and I often wonder if this is responsible for output tantalums surviving repeated startup surge currents.

Good observation!

 

[Nigel Goodwin]

When was the last time you saw an aluminum electrolytic datasheet that read:

CAUTION: The electrolyte in this capacitor will dry out over time and cause the component to fail. It will likely be the most unreliable part in your design. Please de-rate generously to imporve longevity. Enjoy!

But they don't say "while we rate this capacitor at 63V, please don't attempt to use it at more than 21V"

Admittedly electrolytics are troublesome - but less so (in my experience) than tantalums - certainly far more reliable compared to where the tantalums historically failed. Small electrolytics used as supply decouplers on PCB's hardly ever fail - tantalums did. Where electrolytics fail big time is in switchmode PSU's!.