33
\$\begingroup\$

I'm curious as to the reliability and durability of the Arduino Uno.

Does anybody have experience in "killing" one due to excessive use?

If so, how long did it take for the board to fail?

\$\endgroup\$
  • 2
    \$\begingroup\$ Good question, I started to build a small tracker with an Ethernet Shield which should run 24/7 and after some hours of testing it is getting already warm. \$\endgroup\$ – powtac Apr 9 '13 at 18:44
  • 3
    \$\begingroup\$ For cooling an Arduino see arduino.stackexchange.com/q/36/25 \$\endgroup\$ – powtac Apr 9 '13 at 21:48
  • 1
    \$\begingroup\$ I have had an arduino running, using an IR sensor and switching a few high power LEDs, for nearly a year. It finally failed due to oxide formation on some of the header pins, and thus increased resistance. A quick scrubbing of the pins, and it is up again. \$\endgroup\$ – Anindo Ghosh Apr 10 '13 at 8:12
  • 1
    \$\begingroup\$ With adequate ventilation it should run forever, or until it hits MTBF. I have a Mega with Ethernet shield and a dozen sensors running for months now, constantly monitoring network and sensors and XBee and driving a LCD display. Just had to heatsink the 5v regulator and cut out ventilation holes in the box. They're sturdy, contrary to claims to the opposite from engineering elitists. Its got electronic components, the same as any device built in a snooty ivy league college lab, and should have the same reliability :) \$\endgroup\$ – Ron J. Oct 17 '13 at 13:45

10 Answers 10

18
\$\begingroup\$

I have powered the board for days at a time. The code that was running was very simple, but there was absolutely no damage. It is worth it to note that it was being powered by a pre-regulated 5v source so the on-board regulators were not burning up.

I doubt that with anything lower than 9v there could be any sort of hardware damage, but with larger voltages the on board regulators might start to get very hot.

\$\endgroup\$
  • \$\begingroup\$ Any idea at what temperatures the hardware may be damaged at? \$\endgroup\$ – nathangiesbrecht Apr 9 '13 at 20:38
  • 3
    \$\begingroup\$ @nathangiesbrecht - Most ICs (i.e. the actual CPU on the arduino) are rated to handle at least 85°C. \$\endgroup\$ – Connor Wolf Apr 9 '13 at 22:03
  • 1
    \$\begingroup\$ Generally the ICs are designed to be reliable for at least 10 years, the usual problem will be exceeding some limit (too hot, too much current) or letting the environment get into the system (oxidation, corrosion, etc) \$\endgroup\$ – naven87 Apr 13 '13 at 7:09
  • 1
    \$\begingroup\$ Anecdotally, the most common mechanism of failure-over-time on circuit boards from all sources for the past decade or so has been failure of electrolytic capacitors, often accelerated by self-heating of a circuit inside an enclosure. So minimizing the heat produced (by the regulator) and allowing for air circulation should help. \$\endgroup\$ – Chris Stratton Apr 25 '13 at 22:10
9
\$\begingroup\$

I've had one operating a simple weather station which sat online for a couple of months without any problems - I don't see anything that would cause it to break down beyond that time either.

The only real killer (aside from external factors) would be heat, so I'd advise trialling it for your usage application and seeing how it does. If you're worried it's getting too hot, then adding a heatsink shouldn't be too hard an addition to stop temps becoming a concern.

\$\endgroup\$
8
\$\begingroup\$

Keep in mind that the Arduino is meant to be used as a prototyping device. This means there has been very limited endurance testing of the board.

Once the board reaches a steady state temperature, there is nothing in the design that would cause it to go into thermal runaway on its own.

How you program it and what you hook it up to, however, may be a different story.

\$\endgroup\$
6
\$\begingroup\$

I've done it while playing around with one (I kept it on for two days, with a simple program). Nothing happens, though it gets warm.

I would suggest that you ensure that it stays cool, especially if your code is rather heavy on the processor. A heat sink ought to do the trick, or you can attach a small fan.

Aside from that, make sure that all input voltages (power, input pins) are not prone to fluctuations. While there are enough safety resistors in the Arduino to make it less affected by fluctuations than, say, the Raspberry Pi, you can still burn one up if the input voltage is too much.

\$\endgroup\$
  • \$\begingroup\$ Do you know at what voltages would I be taking an excessive risk? \$\endgroup\$ – nathangiesbrecht Apr 9 '13 at 20:40
  • \$\begingroup\$ @nathangiesbrecht: Yes. Page 303 of the ATmega328P datasheet indicates you are taking an excessive risk if you let anything drive the input voltage above Vcc+0.5 V (i.e., +5.5 V on "5 V" Arduinos) or if you let anything pull the input voltage below negative 0.5 V. \$\endgroup\$ – davidcary Jun 26 '13 at 3:04
3
\$\begingroup\$

Excessive heat would be the only long term threat. It works like this: through the power supply you continuously add energy, most of it in the form of heat. On the other hand the Arduino will also lose heat to the environment: the higher the temperature, the more it will give off.
When the Arduino has been running for half an hour or so equilibrium will have been reached: it has reached a temperature at which the released energy matches the absorbed energy. If the temperature is OK then (less than 85 degrees centigrade) it will be OK forever. So make sure it doesn't get hot quickly. An Arduino without enclosure will reach equilibrium within minutes and the temperature will be OK. In an enclosure you'll have to provide cooling vents, or for a metal case you can mount it on another metal structure which functions as a heatsink.

All in all, if your device doesn't get hot after an hour it's probably safe to operate it 24/7.

\$\endgroup\$
3
\$\begingroup\$

I run a fleet of what I call Piduinos - paired Rpi3 and Arduino Uno for data collection and remote control.

They run 24/7 in all kinds of environmental conditions.

You can see real-time data produced by a half-dozen of them at https://www.SDsolarBlog.com/montage

Never have had an Uno completely fail after the first few days (the old bathtub curve)

Mind you, this is at a desert southwest USA solar power plant, where the outside portion experiences wild daily temperature swings.

As for what makes them fail there is always one single thing: dust. It gets into the pin header sockets. For digital devices like DHT22 temperature sensors you know it has happened because the readings simply stop. For analog voltage inputs it is obvious it has happened because voltage divider readings either begin to get too high (meaning resistance in a ground lead) or too low (meaning resistance in the sensing lead).

If you go to the montage link above it is obvious that the outside battery voltage monitor ground lead is ailing. A new board has been constructed and will be installed soon. But for now, the normal settling voltage overnight is showing way above the 12.7 the voltmeter shows at the batteries.

Thus, the term "fail" is relative. Total failure has proven to be caused by the poor quality control of suppliers. But degradation happens much more frequently and is gradual.

\$\endgroup\$
2
\$\begingroup\$

Of course it will, I usually make my own boards, I leave them for months turned on without any problem. Sometimes, like 3 or 4 times, I had to turn it off then on so that it continue working.

\$\endgroup\$
  • \$\begingroup\$ So what you're saying is that it failed 3 or 4 times within several months? \$\endgroup\$ – nathangiesbrecht Apr 9 '13 at 21:50
  • \$\begingroup\$ I would bet that the issues you had were related to strange edge-cases in your software, not the hardware itself. \$\endgroup\$ – Connor Wolf Apr 9 '13 at 22:04
  • 1
    \$\begingroup\$ Yes, it failed 3-4 times within the several months (to be exact, since July 2012) It is used daily to open the office door using a keypad. Not sure if it's a software or hardware problem, but it suddenly turn off so we have to open the door manually with the key.(everything turns off, all components and sensors connected to the board it self) \$\endgroup\$ – moenad Apr 9 '13 at 23:11
1
\$\begingroup\$

The Arduino was designed for prototyping but gets regular use in art installations and other 24/7 applications. There really isn't anything to wear out under normal conditions even running for years.

Even if one's code was very heavy on the CPU, it would be the regulator that would get warm, not the MCU, and it would be just fine.

The only issue I can see is you attach several watts of load to it and overload the regulator to just below its automatic shutdown level. It would probably still be just fine.

If an Arduino is going to die, it's most likely going to be the flash wearing out, a short circuit/overvoltage, static electricity, or connector failure/other mechanical issues/sledgehammer attack

\$\endgroup\$
1
\$\begingroup\$

Mine has been running since May 2014. Since it's solar powered, it is always ON everyday and OFF automatically at nights (so not really 24/7).

http://epxhilon.blogspot.com.au/2015/04/cheapest-commuting-challenge.html

\$\endgroup\$
0
\$\begingroup\$

I've created an incubator for growing cells(LA-4, MCF-7, etc) in a laboratory, where I work. It is powering 2 relays, 1 bipolar transistor, reading 4 sensors and displaying values on LCD screen since May 2017. It was only shut down twice, when the inside of the incubator was cleaned, then turned back on. I'm powering it with 12VDC from a very steady power supply, which has low ripple output (<5mV).

Fun fact: The sensors are constantly on rh=95-100%.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.