CM centurion 590 vs. Lian Li PC-A17B

[fitbrit]

Did you fill up the exhaust slots with 120MM fans?

Granted 60c is pretty dam hot. but having a 120MM fan blowing across them should keep them cool.

 

In my configuration, I block up any vents that do not have a fan, and use only 120MM exhaust fans to suck air through the front.

I do not have ANY fans in the front of the drive and my drives hover around 38-42c.

 

If taking the side off helped, then you need to exhaust the air faster.

 

Thanks for the quick reply. No not all the slots are used up with fans yet, but I really thought this case would be even cooler than my last one, which had some of the drives in a plastic enclosure within the case, without air flowing over the drives. Also, ntoe that it's only the drives in the middle 4-in-3 'cage' that are getting hot. I'm wondering whether it might be due to the rubber grommets used to hold the drives; they may be insulating them too much from the case. Still, I guess it's time to buy some more fans soon. Right now I have a single 350 mm fan blowing air into the case and it's doing its job okay. If more fans don't work, I'll be tempted to get another ThermalTake iCage; I'll give up a drive slot, but that thing keeps my drives at or below 40 C in summer and 33 in winter.

[WeeboTech]

Look at my picture below. I have no fans pushing air into the machine over the drives.

I have 3 120MM fans pulling (exhausting) air from inside the case outside.

I taped (with clear packing tape) the bottom vent (power supply is upside down exhausting air from inside).

I also taped the right side vent and the back vent.

Forcing the three 120MM fans to pull air.

I also taped up the 120MM vents that were not filled with fans.

This creates a suction pulling air from the front over the drives.

 

If the 4 in 3 drive cage is overheating, then there is blockage and or not the proper airflow over the drive.

Try the exhaust fans.

[fitbrit]

I did put my PSU in 'upside down' for the same reason. My next step will be to get three additional 120 mm fans. Right now I have the one at the back, which I forgot to connect last night, and one at the top, closest to the drives.

 

I'll put in the second one at the top and add the door fans too. I'll investigate further when I return home; I left a drive clearing this morning at 46 C with the room fan blowing onto the drive. The 1 TB parity sync seemed to have completed without issue overnight.

 

Thanks for the tips - much appreciated.

[WeeboTech]

For the record, I do not have door fans installed.

I used clear packing tape to cover them, (Left Side), Clear Packing tape on right side vent, Clear Packing Tape on Back PCI Vent.

It's important to be sure where you are "pulling air from".

 

I've had cases with substandard design work very well by careful blocking and exhausting.

 

Overall I would say I have 60-90 CFM in exhaust fans alone (No intake fans).

[lionelhutz]

 

I used the 590 and installed a black faced Istar 4in3 raid cage. It worked well and even looks good. I'll eventually add 2 more and end up with 12 drives in the case.

 

I like the case. It's a good looking case and pretty easy to use. I just slid the Istar in from the front and clicked 2 of the quick holders into it from the side. I have not removed the right case side yet. I'd guess that the case sides are ligher than the Lian Li because the premium cases always have heavy sides. The sides are also a little trickier to get onto the case, but really not much. I have another case and it has a "hook" at the bottom that drops over the case then I just close it up and slide it forward. I have to hold the side over the opening and wiggle it around a little to get it on the 590. The build of the main chassis though seems good. I've seen cheap cases that are held square only with the sides on and that isn't the case with this one.

 

I also put the power supply in upside down. I think you can get 3 of the 4 screws in doing that. Seems like a common mod, if you can call it that.

 

The only mod I did was change the fan on the Istar. It was too noisy for my liking so I put a 80mm onto the outside of it instead of the 80mm that was mounted inside the back cover. I've seen the drives get up to about 38C but the fan I installed really isn't that powerful. It was a low speed one I had lying around. I might change it again when I get the next cage.

 

Honestly, I'd have a hard time justifying the $180 for the Lian Li when the 590 is $70.

 

Peter

 

[WeeboTech]

Just for the record, It's 90f(32c) degrees where I live (outside temperature).

 

Inside is probably a bit higher as I have no ac and the fans are not running in the windows.

 

My two drives are 41c (Segate 7200RPM parity) and 35c (WD Green 5400 RPM).

 

Right now the drives are lightly used with my rtorrent client running, but still spinning and used nonetheless.

I suppose the 8G of ram helps.

 

I guess this means those exhaust fans are working pretty well.

 

[fitbrit]

 

I guess this means those exhaust fans are working pretty well.

 

 

That, or unRAID's temperature monitors, or your estimates are wrong. To have a spun up drive at ambient temperature (if the inside of your home is over 32 C and the drives are at 35 C) almost defies thermodynamics!

 

I've got the side door off and the fan off now, and I'm monitoring hte  temp increases. If the temp doesn't increase past 50 C like this, it's definitely an exhaust problem. If it does, I'll see what I cna do about cleaning up the cables, getting more fans etc. Thanks for the data points and providing a target to reach for.

[WeeboTech]

It's 93f(33c) outside out in the sun. I walked into my computer room and I'm pouring sweat.

I would imagine it's between 85-95 in here now. (phew must turn on the window exhaust fan).

 

Here are my unraid drive temps.

ST31000340AS / 3QJ08RZW - 41°C 

WDC WD10EACS-00Z / WD-WCASJ0437718 - 37°C 

 

To be sure, I compared this against my unix wokrstation which has 2 300GB 10,000RPM drives.

Of course the fans on the Supermicro workstation sound like a jet engine.

However the drives are pretty dam cool. I'm quite surprised.

 

 

[root@hercules ~]# smartctl -a /dev/sda 
smartctl version 5.36 [x86_64-redhat-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

Device: MAXTOR   ATLAS10K5_300SAS Version: BK00
Serial number: J80CZE4K
Device type: disk
Transport protocol: SAS
Local Time is: Thu Jul 17 17:17:48 2008 EDT
Device supports SMART and is Enabled
Temperature Warning Enabled
SMART Health Status: OK

Current Drive Temperature:     35 C
Manufactured in week 06 of year 
Current start stop count:      1074003968 times
Recommended maximum start stop count:  1124401151 times
Elements in grown defect list: 0

Error counter log:
           Errors Corrected by           Total   Correction     Gigabytes    Total
               ECC          rereads/    errors   algorithm      processed    uncorrected
           fast | delayed   rewrites  corrected  invocations   [10^9 bytes]  errors
read:     103324      777         0         0          0       1178.422           0
write:         0        0         0         0          0       1627.653           0

Non-medium error count:      366

Last n error events log page

SMART Self-test log
Num  Test              Status                 segment  LifeTime  LBA_first_err [sK ASC ASQ]
     Description                              number   (hours)
# 1  Background short  Completed                   -       9                 - [-   -    -]

Long (extended) Self Test duration: 5760 seconds [96.0 minutes]
[root@hercules ~]# smartctl -a /dev/sdb
smartctl version 5.36 [x86_64-redhat-linux-gnu] Copyright (C) 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

Device: MAXTOR   ATLAS10K5_300SAS Version: BK00
Serial number: J80BY3XK
Device type: disk
Transport protocol: SAS
Local Time is: Thu Jul 17 17:17:55 2008 EDT
Device supports SMART and is Enabled
Temperature Warning Enabled
SMART Health Status: OK

Current Drive Temperature:     35 C
Manufactured in week 06 of year 
Current start stop count:      1074003968 times
Recommended maximum start stop count:  1124401151 times
Elements in grown defect list: 0

Error counter log:
           Errors Corrected by           Total   Correction     Gigabytes    Total
               ECC          rereads/    errors   algorithm      processed    uncorrected
           fast | delayed   rewrites  corrected  invocations   [10^9 bytes]  errors
read:      80927     1560         0         0          0        689.636           0
write:         0        0         0         0          0       2197.002           0

Non-medium error count:      134

 

on my other linux machine I have 2 SATA and 2 IDE drives I have a mixture of fan arrangements over the drives there.

See below, the numbers are not so out of wack.

In fact all numbers on three different machines are pretty close to one another.

 

=== START OF INFORMATION SECTION ===
Device Model:     WDC WD3200JB-00KFA0
Serial Number:    WD-WCAMR1415458
Firmware Version: 08.05J08
User Capacity:    320,072,933,376 bytes
root@gatekeeper ~> /usr/sbin/smartctl -a /dev/hda  | grep emp
194 Temperature_Celsius     0x0022   114   095   000    Old_age   Always       -       36


=== START OF INFORMATION SECTION ===
Device Model:     WDC WD3200JB-00KFA0
Serial Number:    WD-WCAMR1621102
Firmware Version: 08.05J08
User Capacity:    320,072,933,376 bytes
root@gatekeeper ~> /usr/sbin/smartctl -a /dev/hdc  | grep emp
194 Temperature_Celsius     0x0022   114   099   000    Old_age   Always       -       36

 

 

I design my machines with low voltage processors with as many and as large of fans that I can use.

This unix machine has two low voltage xeons with huge swiftech heatsinks with 80MM fans

So you get an idea here is the output of lm_sensors for that machine.

This particular machine has a 120MM fan in power supply

and 2 silverstone fans one as input over the drives and one as exhaust under the power suppliy near the CPU

 

http://www.newegg.com/Product/Product.aspx?Item=N82E16835220023

 

What I like about these fans is the mount is 80mm but the blades are 92MM.

This particular machine runs my radiostations irc bot and music playing daemon.

it also runs vmware and I use it for dvd rips and re-encodes.

 

root@gatekeeper ~> sensors 
w83792d-i2c-0-2f
Adapter: SMBus I801 adapter at 0500
fan1:     2636 RPM  (min =    0 RPM, div = 64)              
fan2:     2109 RPM  (min =    0 RPM, div = 64)              
fan3:     1757 RPM  (min =    0 RPM, div = 64)              
fan4:        0 RPM  (min =    0 RPM, div = 128)              
fan5:      878 RPM  (min =    0 RPM, div = 64)              
fan6:        0 RPM  (min =    0 RPM, div = 2)              
fan7:        0 RPM  (min =    0 RPM, div = 128)              
temp1:     +46.0°C  (high = +127.0°C, hyst =  +0.0°C)   
temp2:     +47.5°C  (high = +80.0°C, hyst = +75.0°C)   
temp3:     +46.5°C  (high = +80.0°C, hyst = +75.0°C)   

 

 

In comparison.. my readyNAS NV+ was just accessed, the drives were spun down.

 

Disk 1	Seagate ST3750640AS 698 GB , 36 C / 96 F , Write-cache ON		OK
Disk 2	Seagate ST3750640AS 698 GB , 38 C / 100 F , Write-cache ON		OK
Disk 3	Seagate ST3750640AS 698 GB , 38 C / 100 F , Write-cache ON		OK
Disk 4	Seagate ST3750640AS 698 GB , 38 C / 100 F , Write-cache ON		OK
Fan 1	2142 RPM		OK
Temp 1	38.0 C / 100 F [Normal 0-60 C / 32-140 F]		OK
UPS 1	Not present		OK

 

These drives will get up to 42-46 in the next hour or so.

 

 

My point is,

if you design your case with meticulous airflow and enough CFM it can stay cool and very close to ambient temps.

 

This is one of the reasons I'm an advocate of the CM-590.

You can design it with allot of exhaust and at a low price on noise (with the right fans).

I do believe you have to block certain vents to force air intake from where you want it.

 

[Equilibrium]

Ditto, I cannot emphasize that enough; good airflow is the key low temps and low noise. Just because you've got 25 fans inside your PC - pointing every which way - doesn't mean it will run any cooler than a well planed out 2 or 3 fan system.

[fitbrit]

Thanks for the data and opinions, guys. I'll be getting a couple more fans and will let you know how it's going. I've got a top fan because I figured that the hot air from the drives would rise anyway. In theory, if there are a number of fans, all blowing out except for the front ones over the drives, it should be more cooling than a lower number of fans, right?

 

I'm thinking that instead of blocking the slots, I'll add exhaust fans where possible, and block the remainder. I'm using a Celeron 420, which is very low power, and have onboard video too. Overall, there shouldn't be that much heat being produced, other than the 9 drives. So once the unused slots are occupied by fans/blocked, I'm hoping to see a big difference.

I can attest to the importance of airflow, as my old Macintosh was very well designed. The CPU would actually overheat if the cover was off; I should have remembered that little anecdote.

[fitbrit]

So with the rear exhaust fan connect, things improved somewhat. The drives didi heat up again, but much more slowly. Still not perfect yet. I'm awaiting the arrival of some fans and blocking the vents.

 

Something that may be important, my WD green drive never gets too hot. My other WD drives run cool too, but the Seagate and Hitachi drives are the ones that get hottest, the Hitachi ones moreso. Now I can get more 1 TB WD Green Power drives for $150 each, so I think I'll replace my Hitachis with these. In some cases (no pun intended), the WD drives are 10 C cooler!

[SSD]

You are right that the WD green drives are cooler. They spin at 5400 RPM as opposed to the Seagates and Hitachis (and most other normal desktop drives) that spin at 7200 RPM. Slower rotation equals lower heat.

[fitbrit]

You are right that the WD green drives are cooler. They spin at 5400 RPM as opposed to the Seagates and Hitachis (and most other normal desktop drives) that spin at 7200 RPM. Slower rotation equals lower heat.

 

I knew that about the WD green drives, but what I said was that my other (normal 7200 rpm) WD drives also run cooler than the Seagates or Hitachis.

[Billped]

What you know is that they report lower temps, not that they run cooler.

 

I have put temp probes against drives reporting 5-10C deltas and the probes showed the same, so there is variance in where manufacturers position their sensors, the quality of the sensors, or both.

 

 

Bill

[fitbrit]

What you know is that they report lower temps, not that they run cooler.

 

I have put temp probes against drives reporting 5-10C deltas and the probes showed the same, so there is variance in where manufacturers position their sensors, the quality of the sensors, or both.

 

 

Bill

 

Ah! I see. Thanks for shifting the box so I could think outside of it.

[Billped]

BTW, I am not stating that they don't run cooler, but that the SMART info isn't enough to allow any such definitive statements.

 

In other words, my post was more annoying than useful, but the logician in me can't help it.  I'm the same guy that corrects people who say they saw "200%" improvement when going from 10MB/sec to 20MB/sec.  For me, it's an autonomic function. 

 

To be a bit more helpful ... the SMART info is fantastic for relative measurements even when absolute results are suspect.  So if no fan gives you 90C (an unlikely value) and with a fan drops it to 70C, then you can feel comfortable that the improved airflow has helped substantially.  I also use the "touch test", where I simply put my finger on the top of the drive when it is writing.  If it is uncomfortably warm, you have a problem, otherwise it is probably OK.

 

 

Bill

[fitbrit]

I get what you're saying. I said almost the same as this to weebotech a few posts up, but in a more annoying fashion.

[SSD]

I knew that about the WD green drives, but what I said was that my other (normal 7200 rpm) WD drives also run cooler than the Seagates or Hitachis.

 

I have noticed the exact opposite.  My WD 7200 RPM drives report slightly higher temps than the Seagate equivalents.  Can't say for sure if this is real or based on sensor placement - but I looked carefully at them side by side and noticed that the Seagates are cut slightly differently and allow a bit better airflow.  I have a lot of fans in my case (see My Rig link in my sig) so a little better aerodynamics on the drive's part would have a positive impact on temps for me.

 

I have noticed that positioning of the drives in the case can have a significant impact on heat.  Drives that are either at the top or bottom of the drivestack, with airflow restricted on one side, tend to be significantly hotter than those in the middle where my fans blow air over both sides.  I've found having the bottom of the drive blocked to be worse than having the top blocked, but both are bad.

 

I've seen a number of people post screenshots where the temps are shown.  There can be a 5-8C range in temps for the drives.  Usually the hottest one is above 45C - a real danger zone ESPECIALLY, as Bllped notes, if the sensor is not 100% accurate - so it could be even hotter.  If a user researches and finds the hottest drive is in prime airflow conditions, and is not so hot to the touch, he could blame it on the sensor.  But if it is at the bottom of the case, the fan is not blowing on it, and cables are squeezed between it and its neighbor, it would be foolish to not believe the sensor!

 

My (anecdotal) experience with reported drive temps is that they are pretty consistant across all my drives (Seagate, WD, and 1 Maxtor).

 

Bottom line:  the sensors are all we have to go on.  If we want to be safe, we should believe them to be the minimum temperature of the drive.  If, on a case by case basis, we try to improve the temperature of a hot running drive and find that it is reporting high temps no matter what we do, only then should we fault the sensor.

[fitbrit]

Update:

 

I took the build apart yesterday. I completely rerouted all the wiring,using the space between the motherboard mounting 'floor' and the rear case panel extensively. This freed up a big tangle of cables right behind the over-hot drives. I also installed more 120 mm fans, so now there are 3 fans blowing air over the drives (intake) and 5 are exhaust. However, 2 of the fans stop spinning once the hard drives spin up during boot time, so I think there are some power issues happening.

Anyway, the changes have resulted in a maximum reported temp of 49 C, with some drives at 37. This is an improvement, and I think once I have the other fans working and unoccupied holes blocked off, it'll improve further. Thanks for all your inputs.

[WeeboTech]

49c is still pretty hot. I would like to read an update once all fans are spinning and the other holes blocked.

 

Remember there is an open grill on the bottom under the power supply.

So if you have the power supply mounted upside down (Fan facing the PCI slots), tape off the grill under the power supply.

[fitbrit]

49c is still pretty hot. I would like to read an update once all fans are spinning and the other holes blocked.

 

Remember there is an open grill on the bottom under the power supply.

So if you have the power supply mounted upside down (Fan facing the PCI slots), tape off the grill under the power supply.

 

For sure 49 C is still too hot, but it's a big step in the right direction. I do have the PSU upsidedown but I figured it'd block the holes pretty well that way. There are sooo many holes in this thing that need covering up. Even the fan slots are made for 140 mm fans, so I have 1 cm of holes to block around my 120mm fans. I think that once my power supply problems are addressed, I'll see an improvement too; the 4-in-3 cage with the hottest hard drives has a fan, but it sure spins slowly. I wonder whether it's slowing down because not enough power is reaching it. As a stop-gap I'll move my cooler WD drives into that cage and see if they go up in temp too.

[WeeboTech]

There are sooo many holes in this thing that need covering up. Even the fan slots are made for 140 mm fans, so I have 1 cm of holes to block around my 120mm fans.

 

There aren't that many places to block.

Under the power supply, Small spot in back, 80mm square on right side.

I also noticed the fan slots can take 140MM. I was going to go for those, but went for others instead.

I choose fans that moved 30cfm of air.

 

So as far as slowly spinning fans. Either it does not spin fast, It's too blocked, or there is not enough exhaust air being moved

 

I think moving the cooler drives to the hot spot may help.

[fitbrit]

So, things are much better now. Highest temperature is about 42 C, and that's just one of my drives. I cleaned up the wiring a lot more, but not too much more as I have to change PSU shortly anyway. All the fan slots are filled with 120 mm fans. I have 8 altogether, including three on the three drive cages. 2 of the fans have variable speed controllers, and these do a great job of cooling even when barely audible. I can speed them up if I need more airflow out and they really work well. I haven't blocked holes yet; I'll wait until I have my new PSU.

I've also rewired stuff so that each molex vcable from the PSU has a more even mix of hard drives and fans to power. I've also limited my fan connector from the PSU to powering only 2-3 fans. Incidentally my kernel panic problem has also disappeared as a result of the rerouting of power cables and connectors. I also replaced a 750 GB Seagate with a WD green power drive. It's clearing at 35 C reported right now.

 

All in all, I'm really happy with my 590 now and have learned a few valuable lessons. Thanks for all your help, guys.

[WeeboTech]

Report back after you block the holes (underneath and the right side). I bet that will help too.

[fitbrit]

Report back after you block the holes (underneath and the right side). I bet that will help too.

 

I've now replaced one Seagate and two Hitachi 750 GB drives with 1 TB WD green power drives in my system. I have eight 120mm fans running, two of them (a door and a top slot) are Silverstone Scythe fans with controllers (rated at different max. cfm) I've blocked off the side, bottom and even rear holes.

I can gladly report that with the Scythe fans running near silent, my hottest disk reaches 44 C during parity check. With the fans audible, but not annoyingly so, the temps drop to 40-41 C for the hottest. I can drop the temps more by turning up the Scythes to louder levels, I'm sure, but I'll only do so when I'm not in the same room. I plan on replacing most of my other drives with Green Power ones eventually. Now if only there were not so much downtime when clearing disks! Thanks again for all your suggestions and helping me work through my woes.