Intel Communities: Message List - Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Thu, 28 Mar 2013 15:14:41 GMT

I am waiting here. Please help me. Stack with noise.

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Fri, 22 Mar 2013 13:21:38 GMT

Hi,

I have updated the file now how to apply the changes to system. Please someone guide me. My system is creating enormous noise by fan.

Thanks

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Mon, 28 Jan 2013 16:25:17 GMT

I know this post is ages old, but ....

My Intel S5000PSL motherboard (S5000.86B.04.00.0066.101220061333) is very loud. lm-sensors/fancontrol is not able to control the fan speed. IPMI is able to read out the speed, but I do not know how to set rpm speed accordingly to temperature/load.

17:49:07^root@datastorm:~/Files/iasl > ipmiutil sensor -i 000d -v

ipmiutil ver 2.85

isensor: version 2.85

idx = 0xd

-- BMC version 0.55, IPMI version 2.0

_ID_ SDR_Type_xx ET Own Typ S_Num Sens_Description Hex & Interp Reading

000d SDR Full 01 01 20 m 04 snum 51 CPU 2 FAN = 3d OK 2013.00 RPM

Entity ID 29.3 (Fan), Capab: arm=man thr=write evts=state

Volatile lo-noncr 396.00 lo-crit 330.00

SdrThres lo-noncr 396.00 lo-crit 330.00

nom 4455.00 nmax 8415.00 nmin 495.00 smax 8415.00 smin 0.00

ipmiutil sensor, completed successfully

17:51:42^root@datastorm:~/Files/iasl > ipmiutil sensor -i 000c -v

ipmiutil ver 2.85

isensor: version 2.85

idx = 0xc

-- BMC version 0.55, IPMI version 2.0

_ID_ SDR_Type_xx ET Own Typ S_Num Sens_Description Hex & Interp Reading

000c SDR Full 01 01 20 m 04 snum 50 CPU 1 FAN = 3b OK 1947.00 RPM

Entity ID 29.1 (Fan), Capab: arm=man thr=write evts=state

Volatile lo-noncr 396.00 lo-crit 330.00

SdrThres lo-noncr 396.00 lo-crit 330.00

nom 4455.00 nmax 8415.00 nmin 495.00 smax 8415.00 smin 0.00

ipmiutil sensor, completed successfully

17:51:44^root@datastorm:~/Files/iasl >

Via IPMI I am able to read out the current RPM values. It seems like it is also possible to set thresholds via ipmi (see ipmi docs), however, none of the returning values make any sense. There is no point of reference 396? 330? 4455? 8415? What do those numbers mean? This shouldn't be that hard to adjust.

Re: loud system

webadmin@intel.com — Thu, 27 May 2010 14:05:52 GMT

Tom,

I have looked in the SDR file for instances of your case and there are several values for fans that are running at 100%. Editing this file would definitely help your noise level. We use Rexflow fans which have good airflow and a quiet noise level. I have tested the processors at max load with my reduced values and experienced no heating issues, in fact the temps are quite below the threshold even with all cores pegged at 100%. Good luck.

loud system

webadmin@intel.com — Wed, 26 May 2010 18:06:00 GMT

hoi aryan

1. Wrong system fans are connected on the board.

i connected the original SC5650WS fans to the headers according to the description in the manual

2. Incompatible chassis and wrongly configured SDR.

the updatepackage detects my chassis SC5650WS correctly

3. Compatible chassis with a wrongly configured SDR

the right chassis is chosen rigth during the sdr update.

4. Bios Settings wrongly configured.

the settings are configured as mentioned: acoustig, 300m or less, OLT

i purchased the S5520SC+SC5650WS combo because it was advertised as a quiet workstation.

the first time i plugged the powercord into the PSU i was quite surprised about the 'standby noise' of the PSU.

but i was told that this is standard, no way to change that.

the second surprise was the noise.level of that system running idle. wow! what a noise. that two 120mm chassis fans

can not be missed even on low rpm. they are purely heavy-duty-server-fans blowing all beside.

i admit it is save to have a cool system but that box sounds like a 10yr old overclocked AMD box!

actually i am testig some third party fans and i dont have any headache so far

the system runs on the same cool temperatures as with the 8000rpm fans

this is what i call low-noise - cool system

@parkman

as i wrote above i mounted other fans but maybe i will try to modify the sdr if i have time.

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Wed, 26 May 2010 15:05:12 GMT

Hello TomMix,

@ParkMan

Firstly, thank you very much for the tips on how to modify the sdr.

@ TomMix.

When the fan are spinning too loud, it can be due to:
1. Wrong system fans are connected on the board.
2. Incompatible chassis and wrongly configured SDR.
3. Compatible chassis with a wrongly configured SDR
4. Bios Settings wrongly configured.

I have checked, the chassis that you are using is on the Tested list.

http://www.intel.com/support/motherboards/server/s5520sc/sb/CS-030490.htm

From the link:
http://download.intel.com/support/motherboards/server/s5520sc/sb/e39546005s5520scqsugback.pdf http://download.intel.com/support/motherboards/server/s5520sc/sb/e81423002_sc5650scws_qsg_back_ol.pdf

For Chassis SC5650WS:

System Fan 1
Chassis Fan 3
Chassis Fan 5
Make sure to refer to the guide for correct connections.

EFI Update Procedures:
=======================

Video on how to do the update:

http://www.intel.com/support/motherboards/server/sb/CS-030611.htm

Download this onto your desktop:

Website: http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=19007&ProdId=3073&lang=eng

Extract the zip files.
Copy all the extracted contents in the folder to a FAT-FORMATTED USB (ROOT DIRECTORY of the usb)
Insert the USB to the target machine.
Boot the system into the BIOS (F2)
Go to Boot Manager
Select the EFI Shell and press Enter
It will automatically boot onto the EFI shell.
Once on the EFI shell, the shell will look for the file startup.nsh automatically in the USB.
If it does not find the startup file, try another USB.
It will load the firmware package automatically and you just need to follow the instructions.

If you have the right 4-pin power connector is connected onto the board, then it wont ask you for the chassis.

Anyway if it asks you, make sure to choose the correct chassis when doing the SDR firmware update. Follow the instructions.

After the update of these firmware, i will go onto the bios and Advanced, System Acoustic and Performance Configuration, Throttling Mode set to Open Loop, Set fan profile as Acoustic and Altitude set 300 or less.

http://www.intel.com/support/motherboards/server/sb/CS-022809.htm

Other than this, next steps will be to verify the sel log and verify the release notes of the firmware if there is anything mentionned.

Hope these are useful,

Regards,
Aryan

Reference:

http://communities.intel.com/message/28833#28833

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Wed, 26 May 2010 14:38:43 GMT

@parkman

holy moly! that HC_BMC.sdr is one nice config file!

so far i can see it is highly customizable, and a really silent WS is possible.

it will take some time though!

btw. the captcha is somewhat insulting

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Tue, 25 May 2010 21:12:03 GMT

Ok here is the fix. This is an example of how to change the Fan 5 profile setting. There are multiple settings for each fan profile, and settings for the Processor fans which need to be changed as well, but as long as you follow the same basic process it's fairly simple, just repetative.

Here is an example of code taken from the SDR file:

TAKEN FROM HC_BMC.sdr

//=========================================================================//
// Chassis Fan Control Domain 1 with profile 1,2,3,4
// For OTHER Chassis
// Fan Control Domain Number #1 (PWM1: System Fan5 Rear Fan)
// OLTT&CLTT
//====================================================================//
_SDR_TYPE      C0
_SDR_TAG       'OTHER'
_REC_LEN       001D

   // Sensor Record Header
    00C3               // Record ID
    51                 // SDR Version
    C0                 // Record Type
    18                 // Record Length

   // Record Body Bytes
    570100             // Manufacturer ID
    0C                 // Record Subtype Value (Fan Speed Control SDR Record, rev 2)

   // Temperature Fan Control Record
    01                 // Record Version
    01                 // Fan Control Domain Number #1 (PWM1: System Fan5 Rear Fan)
    14                 // Normal Control Value (14h = PWM 20%)
    FF                 // Boost Control Value (FFh = PWM 255%)
    01                 // Ramp Step (1=1%)
    02                 // Scan Rate (2 Cycles)
    0F                 // Fan Profile Support

   // Stepwise Linear type Temperature Sensor Sub-record
    20                 // Sensor Owner ID
    20                 // Temperature Sensor Number -- (T1)
    64                 // Temp Sensor failure control Value (64h = PWM 100%)
    00                 // Sleep Control Value
    41                 // [7:6]-Supported in S1 (1=yes); [5:3]-Reserved; [2:0]-Control Type (001b=Stepwise linear)
    10                 // Temperature Info [7]-Reserved; [6]-Domain Maximum=0; [5:3]-Negative Hysteresis=2; [2:0]-Positive Hysteresis=0
    03                 // Count of Temperature/Control Value Pairs (03h = 3)

0F // Temperature 1 15c
64 // Control Value 1 (Percentage) (64h = PWM 100%)

19 // Temperature 2 25c
64 // Control Value 2 (Percentage) (64h = PWM 100%)

23 // Temperature 3 35c
64 // Control Value 3 (Percentage) (64h = PWM 100%)

-----------------------------------------------------------------------------------------------------------

It clearly shows that it intends to run the fans at 100%. The value of 64 in the left column is what controls this. The intel cases have much better values for these.

-----------------------------------------------------------------------------------------------------------

10 // Temperature 1 16c
19 // Control Value 1 (Percentage) (19h = PWM 25%)

1A // Temperature 2 26c
1C // Control Value 2 (Percentage) (1Ch = PWM 28%)

24 // Temperature 3 36c
32 // Control Value 3 (Percentage) (32h = PWM 50%)

-------------------------------------------------------------------------------------------------------------

So basically I took these values and replaced the 64 which is 100% for each instance of "other chassis" , to corresponding values like 19, 1C, and 32 which were equal to Intel's values. Obviously you don't want to change to critical temp and fan fail 100% values, just the ones that apply to normal fan operation.

Hope that helps

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Tue, 25 May 2010 19:15:58 GMT

please please post the information. im also a victim of the useless fanspeed controll of an S5520SC in a S5650WS chassis.

please post the information how to alter the sdr data.

Re: Found the solution to loud fans when using "other chassis" option when updating Firmware on a server board

webadmin@intel.com — Tue, 18 May 2010 13:45:07 GMT

I'm having problems with loud fans. They are running at full speed after bios update. Would greatly appreciate it if you could post your solution. Thanks

Kangsi

Server board : S5520SC 2 x CPU : W5590 ; Chassis : CoolMaster