Questions durability Raspberry PI 3

[W. H. Heydt]

I fully expect any Pi I have, operated sensibly, to out live me. So what the heck?

Same here...but I'm 67, so that doesn't mean they need to last all that long to achieve that....

[Heater]

I'm not far behind you.

We have to have something to leave to our children that still works though.

[Gavinmc42]

Have some model B+'s that have been running 24/7 for 18 months.
My oldest used one is a 2012 model A but it does not run 24/7.

MTBF, worst part on the A, B is the 220uf Capacitor, 5000hrs? 200 days?
Newer models don't have electro caps.
Hey, that might explain why some of my old B's used in security cams since Aug 2014 are flaky.
Two have been running 24/7 since then, 4 since late 2014, got 7 now, newer B+.
Better solder some new caps in on the old ones.

[W. H. Heydt]

I'm not far behind you.

We have to have something to leave to our children that still works though.

I have a Pi2B (soon to be a Pi3B) set up that my grandson uses. He's 8. If those Pis outlast him, that would be impressive.

[davidcoton]

MTBF, worst part on the A, B is the 220uf Capacitor, 5000hrs? 200 days?

Common sense, and a quick web search, shows your figure is way out. Imagine a board with 20 electrolytic capacitors (used to be common). At a MTBF of 5000 hours, it would be failing every ten days or so. Not true.

MTBF turns out not to be so useful for electrolytics, and is very dependent on factors like temperature and ripple current. Manufacturers instead quote an expected life, which is of the order of 15 years (all the electrolytics fail in a pattern around that average).

We'll find out in 15 -- 20 -- 50 years how many Pis are still running.

[mikerr]

IME PSUs are the main failure point in pi setups over time,
especially as you can get away with any old phone charger or usb lead hanging off something else.

The pi3 being more fussy on power means many have to move to a "proper" PSU from the start
which is good for longevity !

[Heater]

davidcoton,

MTBF turns out not to be so useful for electrolytics, and is very dependent on factors like ....

Quite so. There are some neat online capacitor lifetime calculators that try to take all that into account. Like:
http://www.illinoiscapacitor.com/tech-c ... ators.aspx

Playing with that I'm getting a projected life spans of the order of 10,000 hours to 100,000. Depends critically on the max temp rating of the cap. So 1 year to 10 years.

I don't think Gavinmc42 implied that a MTBF of 200 days meant that they failed randomly from day one like some radio active decay process. If they did we could talk about a "half-life".

I imagine there is something more like a "bath tub" curve of failure rate.

Anyway there seems to be a lot of other failures to worry about, like the SD card going chernobyl on one of my Pi. Another
failed when I applied 12 volts to it the second time. Amazingly it survived the first time !

[gamemaniac]

what mtbf of the pi3 in hours of use?

[kusti8]

what mtbf of the pi3 in hours of use?

The point is nobody knows , it's so long. Anyway you phrase the question , it's going to be long but unknown. I'm curious why you need this. I always have my Pis on and they've never failed.

[jamesh]

The oven testing the SoC's goes through gives us a MTBF for chops run at normal temperatures.

For PI1/2 this was over 30 years, Pi3 runs a bit hotter, so ISTR is slightly reduced, but I would expect at least 25 years from it.

Certainly long enough for the majority of use cases.

[gamemaniac]

eu vou usar case de acrilico, dissipador e cooler fan no pi3 qual seria o mtbf real dele?

[davidcoton]

eu vou usar case de acrilico, dissipador e cooler fan no pi3 qual seria o mtbf real dele?

[I will use acrylic case, sink and cooler fan in PI3 what the actual MTBF of it?]

No-one can give you an MTBF for your particular combination of parts. But expect 20+ years for the Pi, unless you break it first with wrong connections to GPIO, or some bizarre accident. (The fan may not last that long, so you will probably have to replace that some time.)

[Poil]

The oven testing the SoC's goes through gives us a MTBF for chops run at normal temperatures.

For PI1/2 this was over 30 years, Pi3 runs a bit hotter, so ISTR is slightly reduced, but I would expect at least 25 years from it.

Certainly long enough for the majority of use cases.

Hi Jamesh,

It's been a couple of years since this post so I was wondering if any of this data available now? especially for the Compute Module since that is aimed at industrial markets. I have had a hunt round the internet to find this information but with no result.
It's important for industrial customers to have good confidence in the lifespan of their products.

Thanks.