Electrolytic capacitors bad news for anything on full power for long-ish time

[ocset]

If you wish, Type “Buy Brushless electronics speed controller” into google.

Notice the large number that have their electrolytic capacitors mounted outside of the enclosure.
The device "use time" is typically only about an hour at a time, yet they go to these “external mounting” extents.

This is proof of the general unworthiness of electrolytic capacitors for mounting internally to Power Supply or Drive enclosures, would you agree?

Brushless Electronic speed controoller
https://www.amazon.co.uk/dp/B07D3QG55B/ref=sspa_dk_hqp_detail_aax_0?psc=1&spLa=ZW5jcnlwdGVkUXVhbGlmaWVyPUExSEY5WENLMk1VQlNUJmVuY3J5cHRlZElkPUEwMjg3NzgyMklOU1VEUFFaQlFZQSZlbmNyeXB0ZWRBZElkPUEwMzU1NjkzSU5XR1RZNTlZVEZWJndpZGdldE5hbWU9c3BfaHFwX3NoYXJlZCZhY3Rpb249Y2xpY2tSZWRpcmVjdCZkb05vdExvZ0NsaWNrPXRydWU=

[Ian.M]

Nope.   They'd have to go to more trouble to put the capacitors inside the enclosure as the values they need aren't available in SMD, and the extra board area to put the caps on the same side as the SMD parts of the rest of the circuit, and the deeper cover that would be needed to cover the caps would both significantly increase costs.

Unless the ambient temperature and/or reliability (MTBF) requirements exceed the capabilities of *ANY* electrolytic capacitor, its just a matter of specifying them correctly, positioning them in the enclosure so they aren't heated excessively by other components and buying reputable brands that actually meet their published specs.  Of course that costs money, probably at least an order of magnitude more than a far-east 'off brand' part . . . . . .

[AndyC_772]

No.

The lifetime of electrolytic capacitors, and the effects of temperature on said lifetime, are well understood and documented by all reputable manufacturers.

A design choice made by an engineer on one random throwaway consumer product is not "proof" of anything - except, perhaps, that mounting them in that location, on that product, was cheaper.

[Ice-Tea]

No, I would not agree. See all of the above.

[ocset]

The enclosure has had to be breached to get the connections of those caps back through to  the cct board. Extra expense then needed to keep the IP rating.  It surely could be cheaper to lie them flat in an enclosure. Its also an extra  pain for the assemblers.
As you know, you can get long thin radial caps  and lay them flat.

Not all devices have such "outside enclosure" placement of large components...if it was cheaper then surely we would  see it more often?

Most Electrolytic caps  have pretty poor lifetimes.....just a couple thousand hours at 105 degC...inside a plastic enclosure for a drive/power supply can easily get to 105degC

[ocset]

Thanks yes, 12000hrs (500 days) at 105degc is good going at that price.
I wonder if the reason they have the caps out of the enclosure is because they are overdriving them, and then need the extra cooling of external placement.

Because i bet most of those motor controllers are only in use for "Playtime"...ie, a few hours a week at most.

[Ice-Tea]

Most Electrolytic caps  have pretty poor lifetimes.....just a couple thousand hours at 105 degC...inside a plastic enclosure for a drive/power supply can easily get to 105degC

In which case the designer should be flogged and the poor elco should not receive any blame.

Rules of thumb for elcos (even though I can't imagine you don't know this as you have been doing... something with power electronics for a long time):

For every 10C temperature lowers from its full rating, lifetime doubles. So, even a 2k@105C cap will last 64k hours@55C. Or a 10k@105C cap will last 480k hours. With about 10k hours in a year, that's 50 years.

Don't blame the elco.

[bdunham7]

Most Electrolytic caps  have pretty poor lifetimes.....just a couple thousand hours at 105 degC...inside a plastic enclosure for a drive/power supply can easily get to 105degC

Really?  That seems like an inadequate design practice unless there  is some very specific design requirement for tight space and no cooling.

[ocset]

The Application in mind cannot even have fans, due to their lifetime shortcomings, and the fact they need vents which affect IP rating.

[bdunham7]

Then the design is being driven by the need to be either light or cheap, or perhaps both.  There are plenty of ways to cool things either passively or actively without a vent, and fans can be had with lifetimes of whatever you need--especially in an totally enclosed variant where dust won't be factor.

If it absolutely needs to be small and light and run at 105C,  you can just buy electrolytic caps that are up to the job. 

https://www.mouser.com/datasheet/2/88/HHT-1112213.pdf

[Gyro]

Treez, the stuff you linked in your OP is for RC models. Capacitor life isn't a consideration - they're packaged for minimum size and convenience. Preferably something that can be stuffed in a piece of heatshrink.

If you have a serious industrial application, then you're looking at the wrong stuff. 

[ocset]

fans can be had with lifetimes of whatever you need--especially in an totally enclosed variant where dust won't be factor.

Thanks,  we were looking for fans for use in such environments, and how effective a fan is when it is inside a totally enclosed power supply enclosure, plastic or metal.
Our worry would be the fan grease (round the bearings) melting and degrading with temperature, due to being in a hot enclosure.

[vk6zgo]

Most Electrolytic caps  have pretty poor lifetimes.....just a couple thousand hours at 105 degC...inside a plastic enclosure for a drive/power supply can easily get to 105degC

In which case the designer should be flogged and the poor elco should not receive any blame.

Rules of thumb for elcos (even though I can't imagine you don't know this as you have been doing... something with power electronics for a long time):

For every 10C temperature lowers from its full rating, lifetime doubles. So, even a 2k@105C cap will last 64k hours@55C. Or a 10k@105C cap will last 480k hours. With about 10k hours in a year, that's 50 years.

Don't blame the elco.

Yes, even in real-life situations, electrolytics commonly lasted 30 years in comms & broadcast electronic equipment in service 24/7/365, & would have probably been good for the extra 20 years, if not for the equipment they are part of being removed from service due to updating.

[ocset]

Thanks, it would be massively interesting, to hear whether  people believe there is in fact any application anywhere  which would mandate the NON-use of electrolytic capacitors?

[Kjelt]

Thanks, it would be massively interesting, to hear whether  people believe there is in fact any application anywhere  which would mandate the NON-use of electrolytic capacitors?

Yes where your product can not be serviced anymore after building it in or the lifetime and/or reliability needed would exceed multiple times the lifetime of electrolytes.
So military and space applications do come in mind, even as deep sea sensors.
I saw opened expensive broadcast and military electronics to find only tantalum capacitors used in them.

Other critical applications such as bridge and tunnel lighting, the hardware is replaced every ten years as preventive measures and the design requirements are 15 years. So that means testing and overdesigning the capacitors, 105C , higher voltage and larger casings then actually needed.

[AndyC_772]

Thanks, it would be massively interesting, to hear whether  people believe there is in fact any application anywhere  which would mandate the NON-use of electrolytic capacitors?

I design a lot of automotive products for which an operating ambient temperature well in excess of 100C is completely normal and unavoidable.

Given the spikes and surges which can appear on automotive supplies, tantalum caps are potentially a poor choice too. Instead I tend to favour high temperature ceramics.

[coppice]

Not all electrolytics are created equal. In the 1980s dry aluminium ones drove most other types of electrolytics out of long term equipment, like telecoms infrastructure, for all applications except the big caps in the power supplies. The long term reliability results were excellent.

[CatalinaWOW]

Five hundred hours of life for a model airplane controller, particularly a low end one, is effectively infinite.  It is something over 2000 typical flights, probably more like 5000.  Even top fliers seldom if ever go that many flights without a rapid deconstruction event or two.  And that number of flights assumes the temperature is at 105C over the entire flight.  A more typical profile would have the cap at some temperature near ambient (though perhaps solar heated to temps up to 40-60C), rising over the flight time to a temperature that might approach 105C.  For a flier with a huge investment in batteries you might see a subsequent flight before temps had returned to ambient, but there will still be some relief, extending the predicted life.   All of which suggests why I haven't seen a field failure of these capacitors.  Not saying it doesn't happen, but it is rare or non-existent at the flying fields I have frequented.

These controllers are pushed hard for light weight and high power.  My experience is that FET failure is far more frequent than failure of the electrolytic capacitors.  I have had that failure personally on a few controllers.

[ocset]

Thanks, i appreciate all those apps.
I believe you would agree this is another one for NO electrolytics...
https://www.blumelabs.com/

[Kjelt]

Thanks, i appreciate all those apps.
I believe you would agree this is another one for NO electrolytics...
https://www.blumelabs.com/

It is stupid to make a life lasting led bulb that is powered on 230VAC.
They should sell an external transformer rectifier with 24V or something and do the dc-dc conversion in the bulb.
I have Philips Hue light bulbs and small GU10 bulbs , after 20 minutes they are over 60^oC , I am really annoyed with that.
When does the DC grid finally arrive for these applications

[nali]

Ha - gotta love those dodgy startups with world beating wonder products and marketing heavy websites.

Company status - dissolved
Web contact info - mobile phone number

[Connecteur]

It's hard to build planned obsolescence into solid state components. One way is to use capacitors with short lifespans.  Keeps the manufacturers and the technicians happy.

[sahko123]

I think the capacitors are mounted on the outside,
A. For better cooling rather than slowly boiling inside if the temps get high.
B. Because the required capacitors for one reason or the other weren't available in smaller sizes.

[ocset]

I think the capacitors are mounted on the outside,
A. For better cooling rather than slowly boiling inside if the temps get high.

Thanks, i agree with you, but the concensus here does not.

The concensus here is that the caps are on the outside purely because it makes the enclosure able to be smaller.

[sahko123]

Which is what i meant by the capacitors are on the outside because the required ones would be to big for the small enclosure