DC fan life expectancy

Discussion in 'General Electronics Chat' started by tracecom, Jan 13, 2011.

  1. tracecom

    Thread Starter AAC Fanatic!

    Apr 16, 2010
    3,871
    1,394
    I have a 12vdc computer fan. It is being turned on for about 5 seconds every 25 seconds, so the duty cycle is about 17%. Is the fan's life expectancy longer or shorter than if it ran continuously?

    Thanks.
     
  2. timrobbins

    Active Member

    Aug 29, 2009
    318
    16
    Modern fans usually have electronic drives so it is not just a mechanical mtbf issue. There is a fair bit of reliability data and reports around on tubeaxial fan bearings and life - as this has been the part that has received huge focus over the last 2-3 decades wrt modern high-density power supplies. Usually that data is for continuous rated and 'lower' operating speed conditions. Milspec HDBK-217 is usually referred to for reliability issues and may have some recommended mtbf factors for duty-cycle operation.

    Ciao, Tim
     
  3. tyblu

    Member

    Nov 29, 2010
    199
    16
    Although MTTF hours are reported for continuous use and one would assume that running it non-continuously would increase its lifespan, turning it on and off may be detrimental to its health. I believe the OP is asking which has more effect: lifespan gain from lower duty cycle (5X more) or lifespan loss from in-rush current (unknown). I'm not sure. I would guess that inrush current is negligible in comparison to flyback suppression, conductor reliability at higher power, and mechanical reliability, such that lifespan is increased overall.
     
  4. timrobbins

    Active Member

    Aug 29, 2009
    318
    16
    Yes it's a tricky call, and given the high reliability of good quality fans (assuming we are not talking about low quality stuff) then I doubt there is a definitive answer - which is why I suggested the OP look for any substantiating data.
     
  5. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Although anecdote is not particularly useful, my oscilloscopes' fan has been in use for almost 17 years - it's not on all the time, of course.

    One of the most common failures I've seen with fans is them getting clogged with dust, followed by stalling and overheating, usually over a long period of time. On the computer side, I've also had more catastrophic failures. A small screw or something fell into a case fan. This promptly caused the fan to destroy itself, losing two blades. It still worked, but it was very noisy and vibrated a lot, because it was out of balance. It's a catastrophic failure, but not likely to happen. And it's very obvious when it does happen.
     
  6. tyblu

    Member

    Nov 29, 2010
    199
    16
    Takes a heavy duty fan to be able to lose a blade! Wouldn't stick my finger in that one...
     
  7. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Any fan going at 2000 RPM doesn't like a piece of metal stopping it almost instantly. If the blade doesn't break entirely, it will probably crack. (I used to experiment with old computer fans. They go really fast on 30V.)
     
  8. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Most solid state fans (brushless) are constant speed. Varying voltage does not speed them up or down. Older models, the ones that use brushes, can be varied.

    It is basically the same reason solid state fans burn out more or less instantly if spun up with compressed air. The counter EMF zaps the electronics.
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Sorry - this isn't the case.

    Brushless DC computer fans can and do change speed with the voltage applied across them, and the load of stuff on their blades.

    I made a temperature-regulated fan speed control using a thermistor in the feedback loop of an LT1171 wired as a flyback regulator. It works quite nicely; a 12v fan that is noisy as the dickens (85 CFM) when powered by 12v alone is nice and quiet when its' full capacity is not required to keep the computer guts cool.
     
  10. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    It was designed for it then. Ever take one apart? A chip (8 pin SMT DIP), resistor, capacitor, and 6 coils. Most RC oscillators don't very a lot with power supply variations.
     
  11. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Most are designed to vary with applied voltage; the computer uses PWM, which effectively modulates the voltage to the fan.
     
  12. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    CPU fans nowdays yes, but not standard units like what is in a power supply. PWM will not work on a fan that uses the arrangement I just described (the RC oscillator). The chip sequences between the 6 coils.
     
  13. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Generally, 3 wire fans are PWM capable, while some 2 wire small fans aren't good at variable speed or reversal.

    Dust is the biggest killer, followed by heat, which causes the oil in the sintered bearings to seep out too fast.

    If you have a ball bearing fan, and not in a dusty environment, they can run forever if kept oiled. Sintered bearings (oil "smashed in" with ceramics for a sleeve bearing) are catching up and even surpassing ball bearings in computer sized fans, though.
     
  14. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Bill, I'm sorry, but I have to disagree with your assertion that an RC oscillator keeps a stable frequency over voltage. That feature is unique to 555's and some forms of relaxation oscillators built around comparators. It is likely just using a Schmitt inverter internal to the chip to generate the oscillations; and the frequency increases as voltage increases. I have found 90% of modern power supplies include a fan regulator which runs the fan slowly under low load and fast under high load. It is done using PWM.
     
  15. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Bill,
    Yes, I've dissected more than a few fans.

    The 6-pole fan you've taken apart seems to be a somewhat unique case. Most computer fans nowadays have four coils and two transistors, driven by a Hall-effect sensor, providing the motive force.

    I can only suggest that each fan needs to have its' performance evaluated on a case-by-case basis; as no single solution will work in every case.
     
  16. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    As Wookie just mentioned, there are other methods. As for RC circuits being stable, look again. The charge curves are in percentages. Where the variations come in are from fixed voltages from junctions. Stable is relative. You will not get a 50% change from a simple RC circuit.

    Can't argue with thatoneguy either, I've blown a lot of dust out of equipment over the years.
     
  17. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    But it's not just ratios - for most simple oscillators, the on and off thresholds are not ratiometric to the supply voltage. So they may be 2.8V and 2.0V at 5V, but 1.5V and 1.2V at 3.3V. Try it - you'll find that a Schmitt oscillator will vary considerably. In my case, from 2V to 6V, I got it to go from a few kHz to about 40 kHz - that's a massive variation.
     
  18. tracecom

    Thread Starter AAC Fanatic!

    Apr 16, 2010
    3,871
    1,394
    So, after all this, is there an answer to my question? :)
     
  19. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Two sides to the issue.

    First: Electronics HATE being switched on and off. Example: Hard Disk Drives left running will work for 7 years or more, while those turned off each day die within 3-4 years. Incandescent light bulbs will burn out sooner due to the rapid temperature changes.

    Second: "Mean Time Between Failures" is rated at continual duty. each time the fan turns, a tiny amount of the shaft and oil are "used". The fewer revolutions, the less wear.

    The answer is a combination of the two. Which failure mode is more likely? I'm sure you've had fans that didn't run, but if you "helped" them get spinning, they would run fine until powered off again.

    Is the main goal extending the lifetime of the fan, saving power, or reducing noise?

    Longest lifetime is a combination of running cool and at a lower RPM.

    Saving power is means waiting until the fan needs to move hotter air, and adds to the electronic stress of the off/on temp cycles.

    Reducing noise means lowering the RPM or a better blade design.

    Typically, "Power Saving" kills electronics faster than leaving them run continually, IF they are run under the maximum limits. Heat stress/shock damages both mechanical and electrical devices. There is no answer without contacting the manufacturer, as some fans are designed for optimal lifetime in one "mode" of operation, and using that fan in a different mode will shorten the lifetime a good deal.
     
  20. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    I've had fans die just from handling them. They were super cheap units, but the bearings can wear out and still work until the hub of the fan is moved.

    It's like a lot of things, you spend the money, you get the quality.
     
Loading...