Max safe temp for hard drives?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
A report on my drive temps, in a stock Antec SLK3000B case, rear 120mm Tricool on minimum and a Seasonic 430W psu doing far less for through-case airflow. On a hot summer's day (well, very hot for the UK, 29C in the shade ):
60Gb Seagate ST360015A - 56C
80Gb Seagate ST380011A - 52C
Adding a 7-volted Akasa Amber 120mm to the front blowing through the cage has brought drive temps down to 37C and 35C respectively. Case temp reported as 30C.
I'd say a front fan is essential in this case, but it doesn't need to do much in the way of airflow to get satisfactory results. The fact that there's some flow in the right place is enough.
On a side note, I'll agree with MikeC et al that the Antec drive grommets in this case are not all that effective for stopping low-frequency noise...
60Gb Seagate ST360015A - 56C
80Gb Seagate ST380011A - 52C
Adding a 7-volted Akasa Amber 120mm to the front blowing through the cage has brought drive temps down to 37C and 35C respectively. Case temp reported as 30C.
I'd say a front fan is essential in this case, but it doesn't need to do much in the way of airflow to get satisfactory results. The fact that there's some flow in the right place is enough.
On a side note, I'll agree with MikeC et al that the Antec drive grommets in this case are not all that effective for stopping low-frequency noise...
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
-
- Posts: 10
- Joined: Mon Oct 02, 2006 2:46 am
I put a Schyte Anti Vibration & Schyte HD cooler (attaches on the bottom) and with a little modification I have the drive hitting 36C on heavy load...and the drive is in the 5.25" bay with not much ventilation (there is no fan on the top of the case or side) but 2 80MM on the bottom for exhaust and one 80mm fan behind the CPU for exhaust.
Works a treat. No noise, AND keeps things cool...AND cheap. $13 total.
It's a 400GB Perpendicular Seagate drive (very loud when in 3.25" bay, as my case is aluminium).
If you get the Scythe cooler (black or aluminium) make sure to take off the fan and put it on the bottom so there's more ventilation.
Works a treat. No noise, AND keeps things cool...AND cheap. $13 total.
It's a 400GB Perpendicular Seagate drive (very loud when in 3.25" bay, as my case is aluminium).
If you get the Scythe cooler (black or aluminium) make sure to take off the fan and put it on the bottom so there's more ventilation.
-
- Posts: 2198
- Joined: Thu Feb 10, 2005 11:20 am
- Location: TN, USA
That is because Dell wants the HD to die after the 3 year warranty is up so you'll buy a new PC. They are even willing to have it die before the 3 year warranty is up because some people don't get the 3 year warranty and some will buy a new PC instead of dealing with the support process...Bories36 wrote:ahh my 160gb hd is running at 61c, thats way to hot, but its a dell and its been stable...i hope if fails so i can build a new comp
I opened up my dell and theres 2 fans, 1 case fan and 1 fan on the psu, and the case fan is ducted over the cpu heatsink, so theres no airflow over the hd...
that is FUD, dell just did a poor design.dhanson865 wrote:That is because Dell wants the HD to die after the 3 year warranty is up so you'll buy a new PC. They are even willing to have it die before the 3 year warranty is up because some people don't get the 3 year warranty and some will buy a new PC instead of dealing with the support process...Bories36 wrote:ahh my 160gb hd is running at 61c, thats way to hot, but its a dell and its been stable...i hope if fails so i can build a new comp
I opened up my dell and theres 2 fans, 1 case fan and 1 fan on the psu, and the case fan is ducted over the cpu heatsink, so theres no airflow over the hd...
my theory on drive cooling, keep it as cool as you are willing to listen to. once you experience a drive failure, your tolerance will go up
-
- Posts: 5
- Joined: Fri Feb 09, 2007 5:59 pm
- Location: Sweden
Google's take in a study of over one hundred thousand drives in real life:
http://216.239.37.132/papers/disk_failures.pdf
http://216.239.37.132/papers/disk_failures.pdf
Very good google article
Hey goblindojo,
Thanks for very good article links. I went thru whole article and this is really helpful.
It seems HDD temp is not as critical as we think, but it also seems keeping the HDD in between 30- 40 C is the best bet. Not below 30 or not above 40.
The conclusion is SMART error rate is more relevant than any other params.
Thanks for very good article links. I went thru whole article and this is really helpful.
It seems HDD temp is not as critical as we think, but it also seems keeping the HDD in between 30- 40 C is the best bet. Not below 30 or not above 40.
The conclusion is SMART error rate is more relevant than any other params.
I would not conclude that below 30 deg C is bad. Google's study had several drive models in it, revolving at 5400rpm and 7200rpm. As you may know only Maxtor made 5400rpm drives recently. Maxtor also had a bit of quality problems recently too. And even Google admits there's huge differences between drive models & batches. And 5400rpm do run cooler than 7200rpm.
So what's my conclusion? It's that most of these sub-30 samples were drives that had low power consumption. They were not sub-30 because they had better cooling than the rest.
Correlation and implication aren't the same thing. For example I can claim that crime rate is higher here in Finland when weather in California is hot. Is this really because of temperature in California or is it just a coincidence? Either way, it's undoubtedly true. Studies support the theory.
But we should use some intuitive thinking in evaluating the true meaning of the result. For example days are hotter than nights. And when it's day in California, it's night in Finland (because they are on the opposite sides of the globe). Crime rate tends to be higher at night. So is the weather in California really the reason that dictates crime rate in Finland? If we were to lift off the California and drag it to Eastern Atlantic, next to Norway, would hot weather still cause higher crime rates in Finland?
So what's my conclusion? It's that most of these sub-30 samples were drives that had low power consumption. They were not sub-30 because they had better cooling than the rest.
Correlation and implication aren't the same thing. For example I can claim that crime rate is higher here in Finland when weather in California is hot. Is this really because of temperature in California or is it just a coincidence? Either way, it's undoubtedly true. Studies support the theory.
But we should use some intuitive thinking in evaluating the true meaning of the result. For example days are hotter than nights. And when it's day in California, it's night in Finland (because they are on the opposite sides of the globe). Crime rate tends to be higher at night. So is the weather in California really the reason that dictates crime rate in Finland? If we were to lift off the California and drag it to Eastern Atlantic, next to Norway, would hot weather still cause higher crime rates in Finland?
-
- Posts: 32
- Joined: Wed Jan 03, 2007 5:13 pm
I have read thru all the posts on the topic. But still confused...
Anyways, after searching thru google with hot, hard drive, fan = this post on this forum. Thanks to everyone who contributed.
Just wanted to post some questions here and may be some insights.
I have a PC that hard drives are running too hot (to me.) 2 hard drives stacked on top of each other but are separated about 1/2 inch in between (about thickness of a hard drive. The cage can hold 3 hard drives.) According to Smartfan they are running 47C and 50C. It started going to 52C for the 2nd hd since today where the weather turned warm here (it it 90F or 32 C).
The hard drives have been running in about same temp since I built this PC. Its has smart PSU (with 2 fan and one does not run at all at these temps.) So I added another case fan at the rear of the case (Some cheap case fans that runs all the time.) But this issue has not improved.
Wondering if moving the case fan to the front of the hard drive case would make a difference ?
Following are the results from Speedfan "in-depth online analysis" results from the SMART tab of the program.
This is for the HD that is running at 50C.
When the weather runs hotter than today, It may kill the Harddrive!! may be!!
Anyways, after searching thru google with hot, hard drive, fan = this post on this forum. Thanks to everyone who contributed.
Just wanted to post some questions here and may be some insights.
I have a PC that hard drives are running too hot (to me.) 2 hard drives stacked on top of each other but are separated about 1/2 inch in between (about thickness of a hard drive. The cage can hold 3 hard drives.) According to Smartfan they are running 47C and 50C. It started going to 52C for the 2nd hd since today where the weather turned warm here (it it 90F or 32 C).
The hard drives have been running in about same temp since I built this PC. Its has smart PSU (with 2 fan and one does not run at all at these temps.) So I added another case fan at the rear of the case (Some cheap case fans that runs all the time.) But this issue has not improved.
Wondering if moving the case fan to the front of the hard drive case would make a difference ?
Following are the results from Speedfan "in-depth online analysis" results from the SMART tab of the program.
This is for the HD that is running at 50C.
What do you all think?Your hard disk is a ST3250620A with firmware 3.AAE.
The average temperature for this hard disk is 33C (MIN=19C MAX=47C) and yours is 51C.
Your hard disk's S.M.A.R.T. attributes are now being analyzed and a full report about the reliability, health and status of your hard disk is generated:
Your hard disk is not below any attribute threshold. This is good.
Your hard disk was never below any attribute threshold. This is good.
Your hard disk is now being compared to real data used to define normal values for your specific hard disk model. This way, the analysis can automatically use proper operating ranges. The images give you an idea of how each attribute is within such range. Current and raw values are shown for easier reference for experienced users. There are 1471 hard disk models in the current archive.
Attribute Current Raw Overall
8 Raw Read Error Rate 113 57446099 Very good
7 Spin Up Time 95 0 Good
10 Start/Stop Count 100 23 Very good
10 Reallocated Sector Count 100 0 Very good
4 Seek Error Rate 70 10597527 Good
10 Power On Hours Count 100 523 Very good
10 Spin Retry Count 100 0 Very good
10 Power Cycle Count 100 37 Very good
10 Unknown attribute 187 100 0 Very good
10 Unknown attribute 189 100 0 Very good
1 Unknown attribute 190 53 807927855 Normal
4 Hardware ECC Recovered 69 8057636 Good
10 Current Pending Sector 100 0 Very good
10 Offline Uncorrectable Sector Count 100 0 Very good
10 Ultra DMA CRC Error Rate 200 0 Very good
10 Write Error Rate 100 0 Very good
10 TA Increase Count 100 0 Very good
All of the attributes of your hard disk have normal values. This is good.
The overall fitness for this drive is 96%.
The overall performance for this drive is 96%.
When the weather runs hotter than today, It may kill the Harddrive!! may be!!
"I added another case fan at the rear of the case (Some cheap case fans that runs all the time.) But this issue has not improved.
Wondering if moving the case fan to the front of the hard drive case would make a difference ?"
It depends where the front fan is located. Is it in front of HDD tray? Or below it? I.e are HDD directly on airflow path.
I favour under-pressurization as method of computer cooling. No intake fans, only exhaust. That way I can duct tape unnecessary intake openings to ensure all intake air passes by the components I consider most important to cool. If you don't have high wattage GPU, you don't need much ventilation in PCI area. You can pretty much duct tape the openings on lower back of computer case, and the openings on side panel (except CPU funnel opening, if present). Well, duct tape may be overkill... it's sticky and may leave glue on surfaces when removed. I actually use that tan coloured low-adhesive tape normally used for paintjobs... duct tape just sound cool. Tribute to MacGyver, hero of all ghetto modders!
There are some drawbacks in depressurized cases. First, since you have to duct tape openings to increase depressurization and airflow through wanted holes (instead of unwanted), you also reduce total airflow. While it's very easy to control where the air enters the case, you'll waste some fan power. Because resistance increases with airflow, adding more power wastes more power and so on. More suitable to run at low airflow because noise levels would be low (plus all fans at the back, where noise is less audible than on the front). Cooling to HDDs and selected (ducted) components maximized: general motherboard cooling minimal (which may or may not be bad, depending on hardware).
If you duct tape, the air resistance will go up and noise might as well. You could undervolt the exhaust fan... or remove it and duct tape the opening (so that PSU won't draw replament air from exhaust opening into the case).
Note, modern sidepanel funnels act as intakes and will reduce depressurization. If your CPU fan idles fast enough and won't increase rpm if CPU temps rise ~10 degrees, you may consider removing the funnel and duct taping side panel opening above CPU. If CPU fan doesn't ramp up rpm after you did this, it should increase HDD cooling with no increase in noise. Well, maybe turbulent noise at the intake openings at the front, but you can compensate by further undervolting of exhaust fans, removal extra exhaust fan or removing some of the duct tape and allow some "unnecessary" airflow that bypass HDD tray (while it won't help HDDs it will help keep the case temperature lower).
"When the weather runs hotter than today, It may kill the Harddrive!! may be!!"
It may, but it probably won't be instantaneous death. Also while absolute maximum give by manufacturer is a single value, exceeding it will not cause certain death, nor will staying slightly below it be safe. 30...40 would probably be optimal (reliability vs. noise) for most systems. I would measure required cooling performance as: HDD has to stay below absolute maximum temperature under extreme defragging at maximum imaginable room temperature. For 35 deg C ambient and 5 deg C diffence between seeking and idle, it's 55 -15 -5 =35 deg C at 20 deg C ambient and idle. So idling at above 40 deg C under cool ambient conditions is not something I consider safe... unless the room is air-conditioned and there's someone present to notice if AC equipment fails. Of course that assumes some extremes... while you should use worst values for calculations, I don't recommend actually defragging on hot summer afternoon. Also, keep computer case out of direct sunlight.
Wondering if moving the case fan to the front of the hard drive case would make a difference ?"
It depends where the front fan is located. Is it in front of HDD tray? Or below it? I.e are HDD directly on airflow path.
I favour under-pressurization as method of computer cooling. No intake fans, only exhaust. That way I can duct tape unnecessary intake openings to ensure all intake air passes by the components I consider most important to cool. If you don't have high wattage GPU, you don't need much ventilation in PCI area. You can pretty much duct tape the openings on lower back of computer case, and the openings on side panel (except CPU funnel opening, if present). Well, duct tape may be overkill... it's sticky and may leave glue on surfaces when removed. I actually use that tan coloured low-adhesive tape normally used for paintjobs... duct tape just sound cool. Tribute to MacGyver, hero of all ghetto modders!
There are some drawbacks in depressurized cases. First, since you have to duct tape openings to increase depressurization and airflow through wanted holes (instead of unwanted), you also reduce total airflow. While it's very easy to control where the air enters the case, you'll waste some fan power. Because resistance increases with airflow, adding more power wastes more power and so on. More suitable to run at low airflow because noise levels would be low (plus all fans at the back, where noise is less audible than on the front). Cooling to HDDs and selected (ducted) components maximized: general motherboard cooling minimal (which may or may not be bad, depending on hardware).
If you duct tape, the air resistance will go up and noise might as well. You could undervolt the exhaust fan... or remove it and duct tape the opening (so that PSU won't draw replament air from exhaust opening into the case).
Note, modern sidepanel funnels act as intakes and will reduce depressurization. If your CPU fan idles fast enough and won't increase rpm if CPU temps rise ~10 degrees, you may consider removing the funnel and duct taping side panel opening above CPU. If CPU fan doesn't ramp up rpm after you did this, it should increase HDD cooling with no increase in noise. Well, maybe turbulent noise at the intake openings at the front, but you can compensate by further undervolting of exhaust fans, removal extra exhaust fan or removing some of the duct tape and allow some "unnecessary" airflow that bypass HDD tray (while it won't help HDDs it will help keep the case temperature lower).
"When the weather runs hotter than today, It may kill the Harddrive!! may be!!"
It may, but it probably won't be instantaneous death. Also while absolute maximum give by manufacturer is a single value, exceeding it will not cause certain death, nor will staying slightly below it be safe. 30...40 would probably be optimal (reliability vs. noise) for most systems. I would measure required cooling performance as: HDD has to stay below absolute maximum temperature under extreme defragging at maximum imaginable room temperature. For 35 deg C ambient and 5 deg C diffence between seeking and idle, it's 55 -15 -5 =35 deg C at 20 deg C ambient and idle. So idling at above 40 deg C under cool ambient conditions is not something I consider safe... unless the room is air-conditioned and there's someone present to notice if AC equipment fails. Of course that assumes some extremes... while you should use worst values for calculations, I don't recommend actually defragging on hot summer afternoon. Also, keep computer case out of direct sunlight.
I will consider all the great points. I have in the mean time moved the rear case fan and stuck it right in front of 2 hard drives and it is running much cooler.whiic wrote:WOW! Good info! Thanks! whiic.
me wrote: Mid tower case.
Smart psu (2 fan inside, 1 is running and the other don't as far as I can recall).
CPU temp 30C. (now this has not changed before and after my tweaking.)
Case temp (unknown.)
2 HD Temps. 24C and 33C. (with case fan in front of the hard disks.)
Now what I'm thinking! is that since there is direct air flow on top of one of the hard disks, the temp sensor itself may be cooling and not reporting "actual temp" of the drive. But I could assume that is in fact running a bit cooler than before (47C and 52C).
I agree the need to be "a bit" under-pressurizing the case for effective cooling for the rest of the PC. So I'm thinking if this fan placement is helping cool the hard disk with direct air flow, I will need to add another case fan to help with overall cooling of the PC (since the fan in the front is pulling more air in from front of the case!).
Regards.
Noob.
Let me thank you all again for all the great info.
Back when 5400, and then 7200 rpm drives came into being, and 'airflow' was something server manufacturers and overclockers worried about, certainly not regular joe case designer, I lost a lot of drives relatively quickly in unventilated hellholes.
As a datapoint, my dad used to have an Inwin something-500 desktop case from the late 90s, which had no case fans and no airflow except what was generated by the PSU. I lost I think 3 hard drives in 4 or 5 years in that setup, all 20, 30 gig models with successively fewer platters.
Eventually, I cottoned on and I removed the front cover plate for the 3.5" spot between the floppy and the HD (of the 3 3.5 available, 2 were external and I was using the 2 at the two ends of the cage). HD temps dropped precipitately because it became the primary intake for the airflow generated by the PSU and that last replacement drive is still going.
Now, the temperature there was probably more than just a little warm, but the thing is.. losing an HD is a major pain in the ass, even if you back up religiously and we're not that religious. Much more so than losing a CPU or a motherboard, let alone anything else. That's why you will pry the intake fans blowing on my HDs out of my case over my dead body. It doesn't need that much, <50 is probably good enough, but <40 is better and >60 is lethal.
As a datapoint, my dad used to have an Inwin something-500 desktop case from the late 90s, which had no case fans and no airflow except what was generated by the PSU. I lost I think 3 hard drives in 4 or 5 years in that setup, all 20, 30 gig models with successively fewer platters.
Eventually, I cottoned on and I removed the front cover plate for the 3.5" spot between the floppy and the HD (of the 3 3.5 available, 2 were external and I was using the 2 at the two ends of the cage). HD temps dropped precipitately because it became the primary intake for the airflow generated by the PSU and that last replacement drive is still going.
Now, the temperature there was probably more than just a little warm, but the thing is.. losing an HD is a major pain in the ass, even if you back up religiously and we're not that religious. Much more so than losing a CPU or a motherboard, let alone anything else. That's why you will pry the intake fans blowing on my HDs out of my case over my dead body. It doesn't need that much, <50 is probably good enough, but <40 is better and >60 is lethal.
Hi all,
my experience with a customer's server room in the years from 2002 to 2005 with maxtor drivers and no room conditioning, with drives operating from 50° to 60° was of a 30% failing hard drives over an year, the drives were Maxtors and were mounted on four 16 bay fileservers.
The counterpart is another server room with ambient temperature never over 25° and drive temperature never over 40° (usually 36°) with Seagate drives in the years from 2006 to 2008, with three 12bay fileservers and the yearly percentage of failures =0%.
Both systems are broadcasting storage, so similar drive usage.
Thats all
my experience with a customer's server room in the years from 2002 to 2005 with maxtor drivers and no room conditioning, with drives operating from 50° to 60° was of a 30% failing hard drives over an year, the drives were Maxtors and were mounted on four 16 bay fileservers.
The counterpart is another server room with ambient temperature never over 25° and drive temperature never over 40° (usually 36°) with Seagate drives in the years from 2006 to 2008, with three 12bay fileservers and the yearly percentage of failures =0%.
Both systems are broadcasting storage, so similar drive usage.
Thats all
I've ran drives at over 50c for greater than 5 years straight, 24/7 with no issue's whatsoever.
Only time i ever had a drive fail on me due to heat was when the A/C went out on one hot summer afternoon and my HD temp got up to over 60c.
The manufacturers rated thermal limit while in operation was 55c.
So my conclusion? Stay under your manufacturers rated thermal limit and your fine.
One other thing i will add is temperature variance will hurt a drive way more than a stable high temp. Which is why you shouldnt power off your computer or let your HDD spin down after 15min of inactivity. Sure it may save a little power but at the cost of some lifespan of your drive. Data integrity for me is worth a few extra waisted watts each day.
Only time i ever had a drive fail on me due to heat was when the A/C went out on one hot summer afternoon and my HD temp got up to over 60c.
The manufacturers rated thermal limit while in operation was 55c.
So my conclusion? Stay under your manufacturers rated thermal limit and your fine.
One other thing i will add is temperature variance will hurt a drive way more than a stable high temp. Which is why you shouldnt power off your computer or let your HDD spin down after 15min of inactivity. Sure it may save a little power but at the cost of some lifespan of your drive. Data integrity for me is worth a few extra waisted watts each day.
No room conditioning means they get hot during the day, cool off at night, and have to deal with humidity and possibly condensation. All of which will have a huge impact on drive lifespan. So its not necissarily just the heat.deffie wrote:Hi all,
my experience with a customer's server room in the years from 2002 to 2005 with maxtor drivers and no room conditioning
I believe the 55 degree figure you're referring to is a non-operating environmental temp limit, not a drive temp one.Aris wrote:The manufacturers rated thermal limit while in operation was 55c.
So my conclusion? Stay under your manufacturers rated thermal limit and your fine.
It can be a bit difficult to find out the maximum drive temp, but you can sometimes figure it out by viewing current temp readings and SMART figures and see what the relationship is. From that I could infer my Samsung 2.5" drive hits its threshold at 80 degrees C.
*shrugs* Three random clicks from Google:sneaker wrote:I believe the 55 degree figure you're referring to is a non-operating environmental temp limit, not a drive temp one.Aris wrote:The manufacturers rated thermal limit while in operation was 55c.
So my conclusion? Stay under your manufacturers rated thermal limit and your fine.
It can be a bit difficult to find out the maximum drive temp, but you can sometimes figure it out by viewing current temp readings and SMART figures and see what the relationship is. From that I could infer my Samsung 2.5" drive hits its threshold at 80 degrees C.
http://www.wdc.com/en/products/products ... =en#jump66
Temperature (Metric)
Operating -0° C to 60° C
Non-operating -40° C to 70° C
Most of these specs are on the manufacturers web page.
No, i looked it up. Its the rated operating thermal limit. Non-operating thermal limit was much higher, somewhere around 70c.sneaker wrote: I believe the 55 degree figure you're referring to is a non-operating environmental temp limit, not a drive temp one.
This entire thread though really only matters to 3.5" drives. You almost never run into thermal issues with 2.5" drives. As the industry slowly continues to move towards the 2.5" form factor this entire thread will lose its meaningfulness. Especially when you consider that SSD's which aren't even made in the 3.5" form factor are projected to supercede HDD's within the next 5 years.
Where does the SMART temp sensor measure the temperature?
I'm using an external RAID case (SansDigital Towerraid TRM8) which for some reason prevents me from reading the SMART values. As I'm in the process of reducing the noise from the cooling fan I would like to double check the result.
My approach is far was to stick a temp sensor in between the drives. The temp there is about 35C while the ambient air is about 25C.
Any ideas how bad/good this is and how close this measurement comes to the SMART value?
I'm using an external RAID case (SansDigital Towerraid TRM8) which for some reason prevents me from reading the SMART values. As I'm in the process of reducing the noise from the cooling fan I would like to double check the result.
My approach is far was to stick a temp sensor in between the drives. The temp there is about 35C while the ambient air is about 25C.
Any ideas how bad/good this is and how close this measurement comes to the SMART value?