3

I'm researching 5V regulators for a project and came across one that looks promising, but I noticed that its datasheet specifically calls for a tantalum capacitor on its output. Normally I wouldn't mind, but I remembered advice that if a datasheet is calling for very specific components, I should be skeptical of it as it may still be in the testing phase and is only known to perform well with those very specific components.

Granted, that advice was given in context of a regulator not just asking for "10uH 2A inductor", but rather a very specific part with a number from a specific vendor, whereas the regulator I'm looking at now is just saying a tantalum capacitor, but I was hoping for a second opinion

E.HP.S
  • 165
  • 1
  • 10
  • Usually when I see a recommendation for outdated components like tantalum caps I take that as a sign that I should look for a slightly more modern part. – user1850479 Aug 17 '23 at 17:23
  • @user1850479 I didn't notice until others pointed it out, but aluminum electrolytic also works, according to the datasheet. That said, this is good to know for the future. Thanks! – E.HP.S Aug 17 '23 at 17:51
  • If anything, if I see a recommendation for a specific part number I tend to "be happy" as I don't need to pick it. As for being skeptical or not, just use a reputable supplier - which Microchip is. Regardless of the type of cap they recommend, this datasheet is from 2016... Older components can be great, cheap & reliable, but you might want something more modern as other suggested. – Vladimir Cravero Aug 18 '23 at 10:13

2 Answers2

12

No, you should not be skeptical. At least in these instances, you should believe the datasheet, even if it does not tell you why it is important.

And if you do read the datasheet, the section 4.2 Capacitor requirements tells you that aluminium electrolytics are adequate, and tantalum capacitors are better (which is true for this application) and capacitors with extremely low ESR may cause instability.

This is true for many other LDO type regultators of the era, from many other manufacturers as well, so this is nothing new and not just not a quirk of this regulator.

For the other regulator device which suggested a specific inductor, it is just tested or validated with that specific inductor, and they want you to know what inductor it is, so you can either simply choose that exact inductor and get the expected performance, or you can at least read the datasheet of the inductor to learn about the parameters it has, to help you choose another inductor which will hopefully work. Simply saying "10uH 2A inductor" is not enough for selecting an inductor for a switch mode regulator, as it is not even defined what the "2A" means, is it the RMS current rating which is basically defined based on the inductor not overheating, or the peak current rating which is what you need to handle so that the switching currents don't cause the core to saturate and the inductance dropping too much. There is also ESR which affects efficiency and some coil may not be even intended for the switching frequency you need.

Justme
  • 127,425
  • 3
  • 97
  • 261
  • Thanks. I was more thinking in the sense that if a datasheet isn't just telling you the specs for the part you need for the regulator circuit to work (e.g. 10uH 2 ampere tolerant inductor), but specifically "get this very specific inductor made by [INDUCTOR MAKER INCORPORATED] with [PART NUMBER FOR PART FROM INDUCTOR MAKER INCORPORATED]", that might be a regulator best avoided. I mainly just wanted to make sure that a datasheet specifying a specific capacitor chemistry wasn't a similar red flag. Thanks for the advice! – E.HP.S Aug 16 '23 at 19:03
  • 2
    Neither is a red flag. The biggest red flag is to use old regulators that have requirements for old caps. Tantalums are, in multiple ways, problematic, and they can be easily avoided. Preferred option is to use a modern regulator which can work with modern ceramic caps. – Justme Aug 16 '23 at 19:17
10

With regulators (linear or switcher), the output capacitor can interact with the regulator’s feedback behavior and cause it to be unstable. This is mainly an issue when ultra-low ESR ceramics are used in place of electrolytic or tantalum. Otherwise it’s ok to use an aluminum electrolytic or polymer type in place of tantalum, they're similar enough.

Read further into the datasheet - it doesn’t require a tantalum. An aluminum electrolytic is adequate. They recommend tantalum only if very large load transients are present.

Which leads me to, why is tantalum in their reference design? Habit, mostly. They probably used tantalum to show the very best transient response for their part. Likewise with your inductor example: a reference design used it to best show off the part. Either way, for your design you can make different trade offs for cost, space, efficiency, etc.

That said, tantalum caps have some advantages, especially in density, stability and low leakage. They also have downsides:

  • Tantalum is a conflict mineral and should be avoided.
  • Tantalum caps can catch fire when overstressed (overvoltage, reverse voltage, temperature.)
  • Tantalum caps are kind of expensive compared to other options.

For these reasons I haven’t used a tantalum in a design since the 1980s.

What about those 'other options'? Aluminum polymer electrolytic caps offer similar transient-handing benefits to tantalum without its drawbacks. With care, some designs can even migrate to MLCC.

More cap selection stuff here: https://resources.altium.com/p/which-type-capacitor-should-you-use

For their part, IC manufacturers have also responded and have reduced the need for tantalum caps. Some voltage regulators are stable with MLCC (ceramic) output caps, or can be made stable if designed carefully (such as adding a small series resistance to the cap to model the ESR needed for stability.)

Meanwhile, MLCCs have become more dense, offering superior high-frequency bypass performance to tantalum which allows downsizing bulk caps from what would be required if they were tantalum or electrolytic. This is especially the case for high-frequency (1MHz or so) switchers which are often designed to work with ceramic caps.

More about migrating to MLCC here: https://product.tdk.com/en/techlibrary/solutionguide/mlcc_replace-guide.html

hacktastical
  • 49,832
  • 2
  • 47
  • 138
  • Good catch. I'd missed that. – E.HP.S Aug 16 '23 at 18:40
  • Hmm, you say polymers would be OK for this regulator. But 20 years back in the era of these regulators, in year 2000, ceramics and polymers were really not viable options for LDOs. I would think polymers would have lower ESR than tantalums, but the ESR requirement is not stated in this regulator data sheet. – Justme Aug 16 '23 at 19:14
  • @Justme I'll keep an eye on that. I'm still picking through other datasheets – E.HP.S Aug 16 '23 at 19:24
  • SMPS circuits’ performance is quite sensitive to the physical layout, so the device manufacturer will have gone to some trouble to optimise the layout of the reference design. It’s quite possible that they’ve picked an inductor and stuck with it (for whatever reason or none) and specified it in the datasheet. That shouldn’t be taken as meaning that no other inductor will work, but that no other inductor has been tested extensively. – Frog Aug 16 '23 at 20:25