Which cases offer the best air circulation for 9 or more internal 3.5" hard drives?

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
I received a Nanoxia Deep Silence 6, but I'm frankly disappointed in it. Even though it's a very large case, and even though there are 140mm case fans positioned to blow air through the stack of hard drives, the fact is this: the hard drives are packed so closely together that very little of the fan air appears to penetrate to the other side of the stack.

P.S. Oops. I should have posted this in the chassis forum. However, I see no way to delete this post or move it there. Moderator, please feel free to move it. Sorry.
 

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,625
2,043
113
SuperMicro has some nice towers, but they're in the $400 range even on ebay. I don't have any, but I would assume like other SM chassis the airflow is directed, redundant, and more than enough for the drives it was made for.

What you got appears to be a standard desktop/modders ATX desktop case, in which case to run 24/7 with 9+ spinners you're going to have to modify it to get the air flow you want. I have some cheap ATX cases I did this with too, and even cheap-brand 4U cases are like that... to get the optimal air flow you need to modify, or spend $$.

Now, have you actually ran them and utilized them how you plan to? (IE: STreaming movies) and checked their temps? Judging by the pics on amazon I think your case is fine, and you may be over thinking it if you haven't checked temps. Fan placement directly in front of those drives looks really good.


My 4U Rosewill actually blocks 20-30% of the 120mm fans, but I can still feel the air going through.

Another option is to go to 3000rpm PWM Noctua Industrial fans, which is what I do for certain setups. MUCH MORE AIR, and it pushes it quietly too. They have 140mm too!! The benefit here is you get the air when you need it, and not when you don't so you get silence and then you get a jet if things ramp up high. This is how I believe ALL systems should be made, and in recent years they have
 

rubylaser

Active Member
Jan 4, 2013
846
236
43
Michigan, USA
I would definitely check the temps as stated above. And, do you plan to have the drives ever spin down when not in use or will the all always be spinning?
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
Another option is to go to 3000rpm PWM Noctua Industrial fans, which is what I do for certain setups. MUCH MORE AIR, and it pushes it quietly too. They have 140mm too!!
It's good to hear that the 3000rpm PWM Noctua are quiet too. The 140mm Nanoxia fans are 1100RPM and don't seem to be pushing much air, so I ordered a 1500RPM 140mm Noctua (NF-A14 PWM) and a 2000RPM 140mm Noctua fan (NF-A14 iPPC 2000) to try out and hopefully push more air. I did notice the 3000rpm 140mm Noctua on Amazon, but I thought it would likely be too loud. If I'd known it would be quiet, I would have ordered one of those to test out also.
 
  • Like
Reactions: T_Minus

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,625
2,043
113
@ 3000rpm or 2000rpm it isn't quiet like watching TV quiet if you have a # of them, at-least not for me. BUT these are PWM fans, so 3000rpm is RARELY if ever seen, that's the beauty of them :)
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
What's a good program for putting load on the drives so that I can measure how hot they get when fully burdened? As a strawman, I was thinking of perhaps using badblocks, because badblocks can be run on all the drives in parallel.

Also, what's the best tool in FreeBSD for monitoring drive temperature? So far the ones I'm considering are smartctl, smartd, and/or MRTG.
 
Last edited:

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
By the way, I think I may have figured out at least one answer to the OP: the Antec Twelve Hundred case. It consists of twelve front-facing 5.25" drive bays. So, I could mount a 3.5 drive on adapter brackets in each 5.25" drive slot and be guaranteed of getting at least some decent airflow over each drive from the fans that are blowing from front to back.

What I don't know is how quiet it might be overall. It doesn't have the sound dampening material that the Nanoxia has. I may have to just buy one and try it.
 
Last edited:

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,625
2,043
113
I have made PCs with ANTEC P182/3/# over the last 6 years for friends, family, myself (I think I still have 2 or 3 of those in use). Everyone with the ANTEC oversized PSU, and everyone is still going strong to this day. It's def. a quiet case, but it's not the 1200, so I'm not sure if it will be the same, but the quality on the ones I have is good especially for being 5+ year old designs.

I think it's worth a shot, especially if it has more drives and/or better cooling from the get-go... that is IF they're too hot right now, I don't think they will be. Personally, I'd be replacing fans if they were and case last resort.
 

chinesestunna

Active Member
Jan 23, 2015
621
191
43
56
By the way, I think I may have figured out at least one answer to the OP: the Antec Twelve Hundred case. It consists of twelve front-facing 5.25" drive bays. So, I could mount a 3.5 drive on adapter brackets in each 5.25" drive slot and be guaranteed of getting at least some decent airflow over each drive from the fans that are blowing from front to back.

What I don't know is how quiet it might be overall. It doesn't have the sound dampening material that the Nanoxia has. I may have to just buy one and try it.
I've ran a 1200 for past 4 years and it could be quiet if you turn all the fans down. Of course with all the air holes and vents if something inside the case is loud such as graphics card then it will seep though.
Personally I find drives don't need that much air to stay cool, esp 7200 or below sata ones with less than 5 platters. I've ran Hitachi 7K2000 with 5 platters in SM 5in3 bays and that 92mm fan pulling in the back and they all kept below 45C even during active read and write
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
As near as I can tell, very little of the air pushed by the Noxania fans against the stack of 3.5" drives actually penetrates into them. When I put my fingers on the side of the drives that's opposite the fans, I'm feeling very little if any airflow. The drives are packed pretty tight, so it's notionally similar to aiming a fan at a brick wall. Most of the air moved by the drive fans seems to be blasted backward toward the fans or travel sideways. So, I can believe the side of the 3.5" drives that is facing the fans might be getting some cooling, but the rest of a drive isn't getting as much help.

At what location on the hard drive does Western Digital put the temperature sensor? If it isn't on the hottest part of the drive, then there's going to be a temperature gradient between where the sensor is measuring the temperature and what the hottest temperature is. Which temperature is the one that matters? I'm guessing it's the hottest temperature. When you pack a bunch of drives closely together, and even more so when the cooling is asymmetric, that's going to change the temperature gradient. Most likely the hottest parts will get even hotter, especially relative to where the temperature sensor is. If the temperature sensor is closer to the periphery of the drive, it becomes a less good proxy.

So, after thinking about it, I'm realizing I don't know enough to confidently evaluate the HDD temperature data, if I were to collect it, unless the temps reported were obviously low on the face of it, as they appear to be in chinesestunna's rig. There may be a lot of implicit assumtpions baked into a seemingly simple "operating temperature" number.

Given all that, I'm inclined to err on the side of more cooling ventilation than what's needed rather than risk not enough cooling ventilation. Maybe that's just me, though, and someone else, with more knowledge about the background assumptions, might well reach a different conclusion.
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
I'm not advocating for no cooling, but here's an interesting study on drive temperature vs. failure rates: Hard Drive Temperature - Does It Matter?
Also I would look up the operating temperature ranges for the drives, if they are 5C-60C, then honestly running at 45C probably won't hurt them
Thanks for the link. Unfortunately, it's clear the author took a narrow dataset and then reached an overgeneralized conclusion. Looking at their data, all their drives were being run at 38C or less. Also, it starts off by saying "The Backblaze Storage Pod is designed to provide good airflow over the disk drives," so you'd expect much less temperature gradient than if the drives were packed together to form an almost monolithic brick. It could easily be that there's a temperature threshhold that none of their drives reached and so sure, (1) at or below 38C, and (2) in their physical configuration and with "good airflow over the disk drives," maybe longevity is completely dominated by other factors. The data doesn't really speak to the longevity of drives operated outside those two conditions. Absence of evidence isn't evidence of absence.

Anyhow, I went and ordered the twelve hundred. If noise turns out to be an issue, I'll see what I can do to mod the Nanoxia to spread the drives out better and possibly alter the fan configuration and/or type of fans. Based on what I'm seeing with the Nanoxia, I happen to think the re-do is justified, even without hard data, but I'd be the first to admit that because of what I don't know it's purely risk mitigation, and a judgment call at that.
 
Last edited:

chinesestunna

Active Member
Jan 23, 2015
621
191
43
56
Sure, I'm not saying it's scientifically rigorous or conclusive but just a reference point. I think the infamous IBM GXP75 "DeathStars" were partially victims of high temp, to achieve the speed and density IBM packed 5 platters into that drive and unlike enterprise SCSI devices at the time, most were installed in desktop builds where HDD cooling were an afterthought.
Have you taken temperature readings of the drives in that Nanoxia case? One "con" I can say for the Antec 1200 drive cages is that it's not dense enough IMO, you only get 1 3.5" space per 1 5.25" bay, so each pod of 3x5.25" will only hold 3 drives. Air flow definitely good though, no doubt about that
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
Sure, I'm not saying it's scientifically rigorous or conclusive but just a reference point. I think the infamous IBM GXP75 "DeathStars" were partially victims of high temp, to achieve the speed and density IBM packed 5 platters into that drive and unlike enterprise SCSI devices at the time, most were installed in desktop builds where HDD cooling were an afterthought.
Have you taken temperature readings of the drives in that Nanoxia case? One "con" I can say for the Antec 1200 drive cages is that it's not dense enough IMO, you only get 1 3.5" space per 1 5.25" bay, so each pod of 3x5.25" will only hold 3 drives. Air flow definitely good though, no doubt about that
No, I don't have the temperature readings, in part because I'm still waiting for the last three 3TB drives to arrive, but mostly because I realized that without knowing more about sensor placement, etc. (see earlier post), it wouldn't help me decide anyway.

Believe me, I'm not happy about all this. If the two Lian Li A-75's I purchased last Friday turn out to be like the Nanoxia, I'll probably eat the postage and return them. I wish I could return the Nanoxia, but I'm outside the return window. For sure, now that I'm better aware of the issue, I'll choose better the next time.
 
Last edited:

chinesestunna

Active Member
Jan 23, 2015
621
191
43
56
No, I don't have the temperature readings, in part because I'm still waiting for the last three 3TB drives to arrive, but mostly because I realized that without knowing more about sensor placement, etc. (see earlier post), it wouldn't help me decide anyway.

Believe me, I'm not happy about all this. If the two Lian Li A-75's I purchased last Friday turn out to be like the Nanoxia, I'll probably eat the postage and return them. I wish I could return the Nanoxia, but I'm outside the return window. For sure, now that I'm better aware of the issue, I'll choose better the next time.
Honestly I think you're being a bit too careful and hung up on the drive temp situation. The temperature sensor position does make a difference but let's not forget drive enclosures are aluminum so they will will transfer heat well through the drive and temperature variation shouldn't be too great, and assuming engineering best practices they probably put it near the source or sensitive point like the drive motor.
Anyways, good luck whichever direction you go
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
Honestly I think you're being a bit too careful and hung up on the drive temp situation. The temperature sensor position does make a difference but let's not forget drive enclosures are aluminum so they will will transfer heat well through the drive and temperature variation shouldn't be too great, and assuming engineering best practices they probably put it near the source or sensitive point like the drive motor.
Anyways, good luck whichever direction you go
Interesting how we see it so differently. Compared to the cost of the entire setup, or even the cost of a drive, to me switching cases to eliminate a plausible (though unproven) source of early drive failure seems like a no brainer. Just look at the very poor reviews of the WD 3TB MyBook, and all the reported failures there, and ask yourself whether it's the drive or the enclosure or something else that's causing the problem: WD My Book 3TB Desktop USB 3.0 External Hard Drive Storage WDBACW0030HBK-NESN - Newegg.com And that's just one drive by itself, so even with a plastic enclosure wrapped around it, it has far better heat dissipation than if it was trapped between hot drives on either side of it. I can 100% confirm that those are WD 3TB Green drives inside those myBooks, because I harvested 3 of my green drives from myBooks I had purchased on sale (for less than the cost of OEM green drives by the way).
 

chinesestunna

Active Member
Jan 23, 2015
621
191
43
56
Interesting how we see it so differently. Compared to the cost of the entire setup, or even the cost of a drive, to me switching cases to eliminate a plausible (though unproven) source of early drive failure seems like a no brainer. Just look at the very poor reviews of the WD 3TB MyBook, and all the reported failures there, and ask yourself whether it's the drive or the enclosure or something else that's causing the problem: WD My Book 3TB Desktop USB 3.0 External Hard Drive Storage WDBACW0030HBK-NESN - Newegg.com And that's just one drive by itself, so even with a plastic enclosure wrapped around it, it has far better heat dissipation than if it was trapped between hot drives on either side of it. I can 100% confirm that those are WD 3TB Green drives inside those myBooks, because I harvested 3 of my green drives from myBooks I had purchased on sale (for less than the cost of OEM green drives by the way).
I'm not disagreeing with you, but I just think you might be chasing an issue that doesn't exist - Let's assume same drives in Antec 1200 with really good spacing and airflow would run at 35C fully heat soaked from constant read/write with 25C ambient, so +10C over ambient. Those same drives running in Nanoxia or another less well ventilated enclosure might run at 40-45C, or +15-20C over ambient. Does the Antec 1200 cool better? Yes. Are those values all within comfortable operating ranges of the drives? Yes.
Currently you don't have any data on Nanoxia as far as what temperatures the drives will reach and just anecdotal observations that it doesn't feel windy and drives are packed close together. Like you mentioned in a previous post, you're unhappy that you're spending more $ on Lian Li A75 to fix this issue and might have to send them back + postage due to same issue, why not do some real testing with drives before jumping to conclusions?
 

EffrafaxOfWug

Radioactive Member
Feb 12, 2015
1,394
511
113
Just to throw some more fuel on the fire - we had a stack of about 300 hard drives that were "cooked" at ~85° operational and ~60° ambient when pone of our datacentre operators forgot to turn on the circuit for the HVAC after a total power outage. Since the DC guys were liable for any hardware failures post-incident we tracked all of the hardware; failure rate of the drives that were cooked was no different from that of drives that spent their entire lifetime at 18-20° ambient.

I've also got a variety of drives at home that have had their service temps routinely exceeded. Again I don't see any increase in failure rates but my own dataset is much smaller.

As such I don't lose any sleep over drive temperatures. My £0.02.

Back on topic - I'm able to fit all my data, both operational and backup, onto enclosures with 8x3.5" bays. Until the enclosure from the Brontastor box appears on the market (which can fit 10x3.5" drives) I'd want a rackmount hot-swap for 9 bays or more. As much as I'd love to recommend some Lian Li cases, the lack of hot-swap backplanes kills their utility as a NAS IMHO.
 
  • Like
Reactions: Chuckleb

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
I'm not disagreeing with you, but I just think you might be chasing an issue that doesn't exist - Let's assume same drives in Antec 1200 with really good spacing and airflow would run at 35C fully heat soaked from constant read/write with 25C ambient, so +10C over ambient. Those same drives running in Nanoxia or another less well ventilated enclosure might run at 40-45C, or +15-20C over ambient. Does the Antec 1200 cool better? Yes. Are those values all within comfortable operating ranges of the drives? Yes.
Currently you don't have any data on Nanoxia as far as what temperatures the drives will reach and just anecdotal observations that it doesn't feel windy and drives are packed close together. Like you mentioned in a previous post, you're unhappy that you're spending more $ on Lian Li A75 to fix this issue and might have to send them back + postage due to same issue, why not do some real testing with drives before jumping to conclusions?
What you're saying is perfectly reasonable if the temp sensor is physically located in the hottest part of the HDD. If it is, I would happily do that. However, I don't know that it is, and I'm not willing to assume that it is. I did a brief google, but I couldn't find where the sensor is located. If anyone happens to know for sure, please post. I could try adding an external sensor to what is likely the hot spot, but it's not worth the effort or expense or time if changing cases simply eliminates the issue.
 

NeverDie

Active Member
Jan 28, 2015
307
27
28
USA
Just to throw some more fuel on the fire - we had a stack of about 300 hard drives that were "cooked" at ~85° operational and ~60° ambient when pone of our datacentre operators forgot to turn on the circuit for the HVAC after a total power outage. Since the DC guys were liable for any hardware failures post-incident we tracked all of the hardware; failure rate of the drives that were cooked was no different from that of drives that spent their entire lifetime at 18-20° ambient.

I've also got a variety of drives at home that have had their service temps routinely exceeded. Again I don't see any increase in failure rates but my own dataset is much smaller.

As such I don't lose any sleep over drive temperatures. My £0.02.

Back on topic - I'm able to fit all my data, both operational and backup, onto enclosures with 8x3.5" bays. Until the enclosure from the Brontastor box appears on the market (which can fit 10x3.5" drives) I'd want a rackmount hot-swap for 9 bays or more. As much as I'd love to recommend some Lian Li cases, the lack of hot-swap backplanes kills their utility as a NAS IMHO.
Just to throw some more fuel on the fire - we had a stack of about 300 hard drives that were "cooked" at ~85° operational and ~60° ambient when pone of our datacentre operators forgot to turn on the circuit for the HVAC after a total power outage. Since the DC guys were liable for any hardware failures post-incident we tracked all of the hardware; failure rate of the drives that were cooked was no different from that of drives that spent their entire lifetime at 18-20° ambient.

I've also got a variety of drives at home that have had their service temps routinely exceeded. Again I don't see any increase in failure rates but my own dataset is much smaller.

As such I don't lose any sleep over drive temperatures. My £0.02.

Back on topic - I'm able to fit all my data, both operational and backup, onto enclosures with 8x3.5" bays. Until the enclosure from the Brontastor box appears on the market (which can fit 10x3.5" drives) I'd want a rackmount hot-swap for 9 bays or more. As much as I'd love to recommend some Lian Li cases, the lack of hot-swap backplanes kills their utility as a NAS IMHO.
Well, OK, what make and model was the hard drive, and how long were they cooked for? It's not as though all different drive models are equivalent. Some are designed to handle higher temperatures and/or use higher quality components less prone to failure. Time at temperature is known to be a factor, especially wrt to cap failures. For that matter, different capacitors have different heat ratings. It's in the part specification. For example, it's common for cheap LED bulbs to die prematurely within a year or two due to cap failure, even though the temperature was essentially the same the whole time as it was on day 1. It can take time for the electrolyte to boil completely off. Once it does, it's toast.