Question What influences a fan's life expectancy more: Start/stop cycles, or total number of spins?

[monsieurpooh]

With my custom fan curve I have the option of either having it totally stopped below a certain temperature, or having always run at least 25% speed (silent). I noticed with the fan stop option, it sometimes straddles the temperature and thus the fans will turn on and off a few times per minute -- changing the borderline temp wouldn't solve much because it would just do the same at a different game (e.g. changing from 55 to 50, it will do the same thing for a lower-load game).

The reason for my concern, is that in every single build I've made the fans are 100% the first things to start failing/rattling after a few years. So, if my interest is solely in the longevity of the fans before they start rattling/grinding: Shall I allow it to fluctuate between start/stop in certain games, which guarantees the fan is always stopped at idle? Or should I set minimum speed to 25, which guarantees no abrupt start/stop?

I am also interested to see if there is scientific/empirical evidence rather than just theoretical.

btw, the stock manufacturer setting has more advanced logic which is to turn on at 60-something and keep running until it drops below 50-something, but this is not possible with MSI afterburner fan curve; moreover the default manufacturer setting is really bad because every time it ramps up the fan it goes "WEEEUSH" really loudly no matter what.

UPDATE: After looking more closely at the MSI Afterburner settings I discovered I can have the best of both worlds. The answer is in the option "Temperature Hysteresis". Setting this to a non-zero value will cause the fan to behave in a similar way to the stock MSI settings (keep blowing a while even after temperature crossed low threshold). This prevents start/stop thrashing while still allowing for fan stop.

 

[rubix_1011]

Neither of those, I would say, or at least after the items below.

1) cleanliness - fans fail faster from dust buildup in and around the spindle and bearing, causing faster wear and drying of bearing oils. Also can cause fan blades to be slightly off-balance, causing bearing wear.

2) restriction of airflow - fans trying to move air through restrictive openings due to dust buildup or in cases with poor design cause air to be buffeted back into the blades causing bearing wear. Also, radiator or heatsink fans in which the heat exchanger is clogged with debris causing airflow restriction.

 

[Darkbreeze]

Bearings are the #1 priority, and you only get good bearing in good fans. You only get good fans if you pay for good fans. Good fans don't come cheap. I have Noctua, Thermalright (NOT Thermaltake), Corsair and Fractal design fans that have been used for 5-10 years constantly in various builds and none of them, not one, has failed yet. In fact, I've NEVER had, or seen, any Noctua fan fail in the last ten years or so. That doesn't mean they can't, but they are so well designed and use bearing of very high quality, so it is incredibly uncommon or simply the fan is so old, when it does.

The reason is, cheap bearings tend to start getting hot eventually. Hot bearings cause bearing failures because the lubricant dries out AND hot bearing ALSO result in motor failures because that heat is transferred to the motor. Fans that use expensive bearings tend to not have this problem for much longer than those using cheaper type bearings.

It's really not a matter of start/stop vs hours running, it's just a matter of RPM and bearing quality. Fan motor quality is a concern as well, because obviously a cheap motor won't last as long AND a cheap motor that is getting hot, can, reversely, cause bearing failures as well as the other way around.

But a quality fan, and it will last.

Companies usually sell both low and high quality models, so doing the necessary research to ensure you are not getting a cheap fan should be compulsory. If you buy a Noctua or Thermalright fan, you can pretty much guarantee you will get a high quality fan. Corsair has some very good quality, and some real turds. Antec, Cooler master, Thermalright and many others, have a lot of garbage and MAYBE one or two half decent models. They are mainly budget oriented, rather than designed for superior operation.

As rubix has mentioned, cleanliness is also a big factor. Running very high positive pressure configurations can be a killer on fan motors and bearings as well if they are not high static pressure models because a fan without a high static pressure capability will seriously struggle against the resistance and that creates additional heat. Heat is the enemy of any fan, and of most components in general. Anything you can do to reduce heat such as keeping clean and dust free, configuring for equal or negative pressure (Or at least not stupidly high positive pressure) and using quality components to start with are all helpful.

 

[monsieurpooh]

Neither of those, I would say, or at least after the items below.

1) cleanliness - fans fail faster from dust buildup in and around the spindle and bearing, causing faster wear and drying of bearing oils. Also can cause fan blades to be slightly off-balance, causing bearing wear.

2) restriction of airflow - fans trying to move air through restrictive openings due to dust buildup or in cases with poor design cause air to be buffeted back into the blades causing bearing wear. Also, radiator or heatsink fans in which the heat exchanger is clogged with debris causing airflow restriction.

But I have to set a fan curve and go one way or the other eventually. Based on your reasons 1 and 2, would you say it would probably be better to have the fan stop at idle, even if it means sometimes turning on and off often during a low-load game? Because having it spin less overall will result in less dust build up.

 

[monsieurpooh]

Bearings are the #1 priority, and you only get good bearing in good fans. You only get good fans if you pay for good fans. Good fans don't come cheap. I have Noctua, Thermalright (NOT Thermaltake), Corsair and Fractal design fans that have been used for 5-10 years constantly in various builds and none of them, not one, has failed yet. In fact, I've NEVER had, or seen, any Noctua fan fail in the last ten years or so. That doesn't mean they can't, but they are so well designed and use bearing of very high quality, so it is incredibly uncommon or simply the fan is so old, when it does.

The reason is, cheap bearings tend to start getting hot eventually. Hot bearings cause bearing failures because the lubricant dries out AND hot bearing ALSO result in motor failures because that heat is transferred to the motor. Fans that use expensive bearings tend to not have this problem for much longer than those using cheaper type bearings.

It's really not a matter of start/stop vs hours running, it's just a matter of RPM and bearing quality. Fan motor quality is a concern as well, because obviously a cheap motor won't last as long AND a cheap motor that is getting hot, can, reversely, cause bearing failures as well as the other way around.

But a quality fan, and it will last.

Companies usually sell both low and high quality models, so doing the necessary research to ensure you are not getting a cheap fan should be compulsory. If you buy a Noctua or Thermalright fan, you can pretty much guarantee you will get a high quality fan. Corsair has some very good quality, and some real turds. Antec, Cooler master, Thermalright and many others, have a lot of garbage and MAYBE one or two half decent models. They are mainly budget oriented, rather than designed for superior operation.

As rubix has mentioned, cleanliness is also a big factor. Running very high positive pressure configurations can be a killer on fan motors and bearings as well if they are not high static pressure models because a fan without a high static pressure capability will seriously struggle against the resistance and that creates additional heat. Heat is the enemy of any fan, and of most components in general. Anything you can do to reduce heat such as keeping clean and dust free, configuring for equal or negative pressure (Or at least not stupidly high positive pressure) and using quality components to start with are all helpful.

I too have come to realize and agree with the superiority of Noctua fans, and I have fit my system with almost 100% noctua. However, this is the GPU heatsink fan I'm talking about, and it's MSI Gaming X brand which is reputable but doesn't last forever in the same way Noctua does (on my last build, my Twin Frozr on my GTX 760 failed after only 3-5 years). And, I don't think there is an easy way to outfit my RTX 2060 Super with Noctua fans, short of some very involved project which will almost certainly void warranty, so I'll need to stick with MSI's fans. And I have to choose between having a cut-off temperature for fan-stop (which might result in thrashing between on and off), versus no fan stop.

 

[Darkbreeze]

But I have to set a fan curve and go one way or the other eventually. Based on your reasons 1 and 2, would you say it would probably be better to have the fan stop at idle, even if it means sometimes turning on and off often during a low-load game? Because having it spin less overall will result in less dust build up.

No. It's better to have very low speed operation because frequent and constant stalling is bad for the fan motor UNLESS it is a motor that was specifically designed for that sort of behavior. Some of them are, and if they are, then it's fine. Personally I'd still prefer to have no passive operation. I want at least minimal RPM on my fans at all times because you can't hear them anyhow when they are at low RPM, plus it eliminates any unnecessary additional stress on the motor by having to frequenty start up from a dead stop, plus keeping components at a more steady temperature rather than wildly fluctuating swings from low to high to low, are, I believe, more conducive to them lasting longer. Abrupt and constant changes in thermal state (As well as extended periods of very high thermal state) tend to wear down components in some cases. That part is simply an opinion based on observations and not a fact, although I'm sure there is probably some supporting science behind it somewhere.

For GPU fans, I agree, you are pretty much limited to either aftermarket cooling solutions, the same fans that the unit comes with or some kind of chinese knockoff replacement fan, because none of the cooling fan manufacturers that I'm aware of are making replacement fans for graphics card coolers. It would be nice if they did, but honestly, most people don't have issues. Of the thirty or so graphics cards that I've personally used in my own systems over the years, at least, I can't honestly remember any of the fans dying on any of them except my R9 290x, and that happened within the first two months so it was just bad hardware rather than any sort of wearing out issue.

 

[monsieurpooh]

No. It's better to have very low speed operation because frequent and constant stalling is bad for the fan motor UNLESS it is a motor that was specifically designed for that sort of behavior. Some of them are, and if they are, then it's fine. Personally I'd still prefer to have no passive operation. I want at least minimal RPM on my fans at all times because you can't hear them anyhow when they are at low RPM, plus it eliminates any unnecessary additional stress on the motor by having to frequenty start up from a dead stop, plus keeping components at a more steady temperature rather than wildly fluctuating swings from low to high to low, are, I believe, more conducive to them lasting longer. Abrupt and constant changes in thermal state (As well as extended periods of very high thermal state) tend to wear down components in some cases. That part is simply an opinion based on observations and not a fact, although I'm sure there is probably some supporting science behind it somewhere.

For GPU fans, I agree, you are pretty much limited to either aftermarket cooling solutions, the same fans that the unit comes with or some kind of chinese knockoff replacement fan, because none of the cooling fan manufacturers that I'm aware of are making replacement fans for graphics card coolers. It would be nice if they did, but honestly, most people don't have issues. Of the thirty or so graphics cards that I've personally used in my own systems over the years, at least, I can't honestly remember any of the fans dying on any of them except my R9 290x, and that happened within the first two months so it was just bad hardware rather than any sort of wearing out issue.

Interesting that you also have my same instinct that it might be better to avoid stress on the motor, but do you have any evidence or sources for this? I have read many other conflicting opinions on the internet, for example in this thread where they claim start/stop won't wear down a fan at all: View: https://www.reddit.com/r/buildapc/comments/dsk8vp/shouldnt_devices_that_stop_the_fan_at_low_temps/

 

[Darkbreeze]

Starting StressesStarting stresses a motor by:

• Applying higher than rated full-load torque to the shaft during acceleration

• Applying high magnetic forces to the rotor cage and winding end turns

• Heating the stator winding and the rotor cage. Frequent torque shocks from starting could shorten shaft life through metal fatigue. However, most shaft failures are attributed to bearing failures, excessive belt tension, misapplication, or creep during storage (large motors). Overheating the stator winding and the rotor cage occurs if the hourly number of starts exceeds the NEMA recommendations or the duration of rest time between starts is less than the NEMAallowable value. Heat from exceeding these limits can degrade winding insulation and cause thermal stressing of the rotor cage, leading to cracks and failed end-ring connections.

2 – Short Cycling
Short cycling is the process of repeatedly stopping a motor that is already heated to operating temperature, and then starting it again before it has a chance to cool. The result of underestimating the starting frequency of an application may lead to more starts per hour than the motor is rated for. Most electric motor manufacturers specify the maximum number, or frequency, of starts for a given motor type.

Motor bearings
Motor bearings within an electric motor can emerge from improper handling and storage, improper installation, misalignment, improper lubrication, start/stop or short cycling processes, contamination, overhung loads and motor fan imbalance.


Contamination is one of the biggest reasons for bearing failure modes. This occurs when foreign contaminants or moisture enter the bearings, usually during the lubrication process. You can take steps to prevent contamination during the regreasing process to ensure that they are kept out.


It is also important that your motor is properly outfitted for the task for which it was selected. This means using the right bearings for its application.

Motor insulation and windings
When it comes to motor insulation and windings, there are a number of potential issues. Contamination and moisture can lead to winding failures. Often times, this is because they are not stored in ambient areas. Overheating is another issue that can cause a motor failure. Insulation breakdown, cycling and flexing, along with AC drive stress, round out the possible failure modes for this category.


The life of the insulation in a standard electric motor is based on the temperature at which the motor operates. This means for an electric motor that is operating at a particularly high temperature, you could be cutting back on its lifespan. In fact, for every 18 to 20 degrees Fahrenheit, the insulation life is cut in half. While better insulation can extend the lifespan, temperature is easily one of the biggest factors in this instance. This means bringing in cooler outside air.


Insulation breakdown can be a big problem, as it will cause windings to short out. These problems can be detected through MCE testing and thermography. Winding shorts from turn to turn can crop up from contaminants abrasion, vibration or voltage surges.


Cycling and flexing is another problem that typically occurs from frequent start and stop operations from the motor. This kind of an operation cycle can lead to the frequent heating and cooling of windings and insulation, which can lead to wear and tear, such as holes, ultimately leading the motor to short and fail.

 

[boju]

Another aspect regarding looking after the fans when blowing out the machine is avoid spinning the fans if air source has enough pressure to do so. Like with an air compressor, i hold the fan blades stationery with something narrow like a pencil or paint brush handle if i can't get in there with fingers. Whipping fans around isn't good but it sounds good though lol.

 

[rubix_1011]

A lot of work and debate being put into one of the lower-cost components of a PC build...my recommendation would be to set the fans where you wish for noise and good temperatures and go about life.

$15-$20 fan vs. say, $300 CPU or $500 GPU, it seems petty to spend this much time arguing or debating the life expectancy of a fan when the MTBF rating is posted on nearly every fan and this will likely extend beyond the lifetime of most components in the PC before they are replaced.

It is better for the lifetime of a fan to keep it running at lower speeds than to continually start/stop a fan.

 

[Darkbreeze]

I 100% agree with all those statements.

 

[monsieurpooh]

A lot of work and debate being put into one of the lower-cost components of a PC build...my recommendation would be to set the fans where you wish for noise and good temperatures and go about life.

$15-$20 fan vs. say, $300 CPU or $500 GPU, it seems petty to spend this much time arguing or debating the life expectancy of a fan when the MTBF rating is posted on nearly every fan and this will likely extend beyond the lifetime of most components in the PC before they are replaced.

It is better for the lifetime of a fan to keep it running at lower speeds than to continually start/stop a fan.

Everyone keeps telling me this but I don't understand why. I have never, in my entire life, had a CPU or GPU fail in any way. Literally 100% of the time, the first thing to fail in my PC is always one of the fans. Last time, it was an MSI GPU fan and then a Cooler Master CPU cooler fan, both of which failed within 3-5 years. And yes, it only costs $20-$50 to replace but can take some time and risk screwing with the heatsink; moreover this is still money down the drain. It doesn't make sense to bring up the cost discrepancy with the CPU/GPU component itself, seeing that the CPU/GPU itself has virtually zero chance of failure over a whole decade or even longer.

 

[rubix_1011]

Everyone keeps telling me this but I don't understand why. It is just false.

I disagree.

I have never, in my entire life, had a CPU or GPU fail in any way.

Which as it should be, price should indicate lifespan.

Literally 100% of the time, the first thing to fail in my PC is always one of the fans.

Join the club, there are literally millions of us.

It is irrational to bring up the cost discrepancy with the CPU/GPU component itself, seeing that the CPU/GPU itself has virtually zero chance of failure over a whole decade or even longer!

I've been on plenty of forums and seen many instances where hard drives, RAM, CPUs and GPUs all fail. But, it should seem that more expensive components should last longer, especially those without moving parts....like a CPU.



So, your argument is that you are upset that you have to replace a CPU or case fan once every few years? Which is almost normal maintenance for a PC anyway?

Also, you are generally not required to remove the CPU cooler to replace the fans...nearly every CPU cooler I test allows for the fans to be changed without removing the cooler.

This seems to be a huge argument about something that is rather simple to replace as well as keep clean. I don't really understand the outrage or disbelief.

 

[Darkbreeze]

Fans have moving parts and motors. Graphics cards and CPUs do not. There is no mystery as to why one fails and the other does not. Anything with moving parts WILL fail sooner or later. There is nothing you can do about that. What you CAN do is ensure that it is later rather than sooner by making sure to purchase products which have better quality. Obviously, a 450 dollar RTX 2060 Super is probably going to come with a better cooler and higher quality fans than a 375 dollar version of the same card. Clearly, a 25 dollar Noctua fan is going to last a lot longer than a 12 dollar Arctic fan.

Keeping the dust blown out of your system and blown off the fan blades and housing, will go a long way towards helping to extend the life of any fan. Not allowing any dust filters to become too clogged and not allowing parts to live in a condition where thermal damage from not managing these factors OR from overclocking past safe levels OR from simply not buying hardware that is capable of performance at the level which you push the hardware to, all are likely to reduce your complaint.

If a fan which is under constant use lasts five years or more, then I'd say it has lived up to expectations. Also, keep in mind, leaving the system running 24/7 is going to drastically reduce the life expectancy of most, if not ALL, components. While you don't want fan motors having to constantly and regularly short cycle off and on, that does not mean that putting the system to sleep or shutting down, when it is not going to be used for 15 minutes or longer, is a bad idea. In fact, it's a good idea to do so. If I am walking away from my system and I plan to be gone more than 10 minutes, it is either put to sleep or shut down depending on how long I plan to be gone. That kind of periodic cold start is not going to be detrimental to your fan motors or bearings. It will extend their life. Short cycling every minute or every five minutes, is a different story.

 

[Darkbreeze]

This seems to be a huge argument about something that is completely normal and a non-issue.

Fixed it for you.

 

[rubix_1011]

Fair enough, also true.

 

[monsieurpooh]

I disagree.



Which as it should be, price should indicate lifespan.



Join the club, there are literally millions of us.



I've been on plenty of forums and seen many instances where hard drives, RAM, CPUs and GPUs all fail. But, it should seem that more expensive components should last longer, especially those without moving parts....like a CPU.



So, your argument is that you are upset that you have to replace a CPU or case fan once every few years? Which is almost normal maintenance for a PC anyway?

Also, you are generally not required to remove the CPU cooler to replace the fans...nearly every CPU cooler I test allows for the fans to be changed without removing the cooler.

This seems to be a huge argument about something that is rather simple to replace as well as keep clean. I don't really understand the outrage or disbelief.

Where did you get the idea that I'm "outraged" or in "disbelief" that fans fail? This is a total strawman and changing the subject. The only vaguely upsetting or outrageous thing is your suggestion that it doesn't make sense or is weird to want to prolong fan life as long as possible. If that were true, then Noctua would be out of business, right?

Also, I know CPU cooler fan is super easy to replace, but GPU fan is usually more of a hassle and more costly, in my opinion.

 

[monsieurpooh]

Thanks for the citations Darkbreeze. However, I noticed that they are all related to motors, but I read that fan bearings are the reasons that fans fail, and these may not be related to the fan motor (correct me if I'm wrong)? Also I googled your links for their original sources and I noticed your first article is actually encouraging people to not worry about start/stop on the motor and saying there is little known correlation between that and motor life.

 

[Darkbreeze]

Having to change the GPU fans every three to five years, doesn't seem to be excessive to me. If a card even lasts five years with heavy use, it has probably outlasted it's warranty and expected lifespan anyhow. Spending a half hour of your life once every three to five years and twenty bucks for new fans, does not qualify as unreasonably excessive in my opinion. And again, I personally haven't experienced the need to do that. Perhaps you are choosing the least expensive card models from a given GPU series and getting exactly what you have paid for.

 

[rubix_1011]

You seemed to be rather agitated by this, indicated by the repeated disagreements of content being discussed and replies we've provided. You can agree or disagree with whatever you'd like, that's fine. My perception of your responses caused me to think this continual disagreement was leading you to be more upset on the subject.

I just don't often hear of many people making fan life span their priority over simply replacing fans on an interval over several years. Even so, I have still had very few fans actually die in any of the systems I have owned. There have been a few, and notably, they were always 'cheap' fans.

Many GPUs are covered by warranty as long as you register the GPU at time of purchase. Also, by the time the GPU fans fail, they are typically a few life cycles off of primary market focus.

 

[monsieurpooh]

Having to change the GPU fans every three to five years, doesn't seem to be excessive to me. If a card even lasts five years with heavy use, it has probably outlasted it's warranty and expected lifespan anyhow. Spending a half hour of your life once every three to five years and twenty bucks for new fans, does not qualify as unreasonably excessive in my opinion. And again, I personally haven't experienced the need to do that. Perhaps you are choosing the least expensive card models from a given GPU series and getting exactly what you have paid for.

Yes I understand, but the whole point of my original question is, if there is a simple setting I can set to extend the life expectancy by a couple years with zero downside, I would like to do that, because there is no downside as both have silent noise levels. I am not saying it is unreasonable to have to change the fan; I am just asking what is the best way to extend the fan's life expectancy. Btw, the models I generally choose are MSI's which generally have a good reputation for decent fans.

 

[rubix_1011]

Lower the fan speed to one you are comfortable with in terms of decibel levels and temperatures.

Clean dust regularly from your system, including all fans and heatsinks.

That's it.

 

[monsieurpooh]

You seemed to be rather agitated by this, indicated by the repeated disagreements of content being discussed and replies we've provided. You can agree or disagree with whatever you'd like, that's fine. My perception of your responses caused me to think this continual disagreement was leading you to be more upset on the subject.

I just don't often hear of many people making fan life span their priority over simply replacing fans on an interval over several years. Even so, I have still had very few fans actually die in any of the systems I have owned. There have been a few, and notably, they were always 'cheap' fans.

Many GPUs are covered by warranty as long as you register the GPU at time of purchase. Also, by the time the GPU fans fail, they are typically a few life cycles off of primary market focus.

As clarified earlier, I was only agitated at the suggestion that it's unreasonable to want to eek out a bit more lifespan on the fan at zero extra cost to me; I wasn't upset/outraged at the fact that fans do eventually fail.

Yes it is true I upgrade very infrequently

Thank you for your help and suggestions

P.S.: Regarding "repeated disagreements of content being discussed and replies we've provided", I am legitimately asking for evidence, not just disagreeing for the sake of it. To verify that I do not have bias, please observe the reddit thread I linked to earlier, in which I was torn apart for taking precisely the opposite stance (there, I am torn apart for saying start/stop might decrease fan life; here, I am torn apart for saying it might not). No one seems to agree on this and I am just trying to find out the truth.

 

[Darkbreeze]

And put the system to sleep or turn it off when not in use for extended periods. There is no point in the fans having to run for hours on end when you are not actually using the system for anything. Removing those hours of operation over the course of three to five years will absolutely increase the life expectancy of the fan because you will reduce the actual hours of operation to be counted against the MTBF.

 

[rubix_1011]

^ Yep, agreed.

Turn the PC off for extended periods of non-use. Only turn on when you intend to use it.