Register for the iXsystems Community to get an ad-free experience and exclusive discounts in our eBay Store.

Thermal and Accoustical Design Validation

Joined
Apr 4, 2016
Messages
92
Thanks
3
#41
Can you explain little further?
Sure*
This is the best picture of the splitter I've found. You have 7 connectors: 1 molex, 5 fan (facing downwards on the pic) and ONE for the mobo. But if you were making the mod as outlined you need 3 more fan connector (the left one) to connect the PWM wires.

*And I should say thank you making me actually think it over and realize my own mistake - you don't need two cables, you need 3 other mobo fan connectors :D
 

jgreco

Resident Grinch
Moderator
Joined
May 29, 2011
Messages
11,783
Thanks
3,021
#42
No. PWM is an output from the motherboard. You would need several connectors if you wanted tachy input from the fans.
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#43
No. PWM is an output from the motherboard. You would need several connectors if you wanted tachy input from the fans.
Just to make the confusion even bigger you were probably writing this while I was typing up my corrected version...
He (and me) needs 3 of these connectors.
 
Joined
Apr 28, 2014
Messages
44
Thanks
0
#44
Can you explain little further? I should be able to use the 1 splitter for the fan wall and plug the exhaust fans directly to the empty fan headers? This, of course, limits me to controlling all 3 fans on the fan wall to one PWM.
I guess this is where my confusion is coming from. To allow control of all 3fans I would need 3 connections to control the tachy via motherboard. To control all the fans with one PWM, I could use one of these cables as intended, correct? IE, plug connector to motherboard, connect all three fans and control the one PWM connection to what I want all three to be (all though I will not receive any RPM data)
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#45
I guess this is where my confusion is coming from. To allow control of all 3fans I would need 3 connections to control the tachy via motherboard. To control all the fans with one PWM, I could use one of these cables as intended, correct? IE, plug connector to motherboard, connect all three fans and control the one PWM connection to what I want all three to be (all though I will not receive any RPM data)
Yes, correct. That case just one RPM will be reported but - theoretically* - all your fans will spinning on same speed anyway.

*the manufacturing tolerance permitting
 

Bidule0hm

Server Electronics Sorcerer
Joined
Aug 5, 2013
Messages
3,715
Thanks
648
#46
NB: the connector is Wieson 2510C888-001, or Molex 47054-1000, or equivalent ;)
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#47
NB: the connector is Wieson 2510C888-001, or Molex 47054-1000, or equivalent ;)
:) I was just about to say. (No, I wasn't)

To control all the fans with one PWM, I could use one of these cables as intended, correct?
Still, I would use two mobo fan headers and separate the fans (3+2) and I would run the front 3 from one PWM and the 2 exhaust fans from a second PWM.
The beauty of it, that you can still use one splitter with all five.
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#48
'Failure to fix your wires properly or in the wrong order will result in a LOUD BANG, more than likely a bit of smoke and more than likely a dead PSU or worse PC!! '
This is an extract from the eBay seller's listing description :)
 

Ericloewe

Not-very-passive-but-aggressive
Moderator
Joined
Feb 15, 2014
Messages
16,139
Thanks
3,867
#49
NB: the connector is Wieson 2510C888-001, or Molex 47054-1000, or equivalent ;)
You know that by heart, or did you just have the BoM for your fan controller handy? ;)

The Molex 0470531000s are expensive, 40 cents a pop.
 
Last edited:
Joined
Apr 4, 2016
Messages
92
Thanks
3
#50
The Molex 0470531000s are expensive, 40 cents a pop.
You are lucky not being this side of the pond. Just bought 4 for a fiver :(
 

Ericloewe

Not-very-passive-but-aggressive
Moderator
Joined
Feb 15, 2014
Messages
16,139
Thanks
3,867
#51
You are lucky not being this side of the pond. Just bought 4 for a fiver :(
From Farnell or what? DigiKey and Mouser both had them around 40 cents last time I checked.
 

jgreco

Resident Grinch
Moderator
Joined
May 29, 2011
Messages
11,783
Thanks
3,021
#52
You know that by hard, or did you just have the BoM for your fan controller handy? ;)
Heh, if you were asking me, you'd want to be careful since the stock bin in the shop's pretty close by. :)

The Molex 0470531000s are expensive, 40 cents a pop.
You have a strange concept of expensive.
 

Ericloewe

Not-very-passive-but-aggressive
Moderator
Joined
Feb 15, 2014
Messages
16,139
Thanks
3,867
#53
Heh, if you were asking me, you'd want to be careful since the stock bin in the shop's pretty close by. :)



You have a strange concept of expensive.
They're a bit of injection-molded plastic with four metal pins. 40 cents on that is one heck of a profit margin.
 

jgreco

Resident Grinch
Moderator
Joined
May 29, 2011
Messages
11,783
Thanks
3,021
#54
They're a bit of injection-molded plastic with four metal pins. 40 cents on that is one heck of a profit margin.
No, fifteen cents (qty 10000) is a heck of a profit margin. 25 more cents is the chump price you pay for buying them one at a time, and it quite possibly barely covers the cost of having someone go get that and bag and tag it for you.
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#55
From Farnell or what? DigiKey and Mouser both had them around 40 cents last time I checked.
eBay but TBH this includes pp as well. Farnell?! Probably just a tenner or more for delivery :)
 

Bidule0hm

Server Electronics Sorcerer
Joined
Aug 5, 2013
Messages
3,715
Thanks
648
#56
You know that by heart, or did you just have the BoM for your fan controller handy? ;)
Actually the controller use the male version (and in fact you can use simple 0.1" header instead if you want to cut corners :)). But I have the PWM Fan spec, and as it's hard to find it (ok, not that hard but annoying enough) I've put it as an attachement file of this post if someone wants it ;)
 

Attachments

Ericloewe

Not-very-passive-but-aggressive
Moderator
Joined
Feb 15, 2014
Messages
16,139
Thanks
3,867
#57
Actually the controller use the male version (and in fact you can use simple 0.1" header instead if you want to cut corners :)). But I have the PWM Fan spec, and as it's hard to find it (ok, not that hard but annoying enough) I've put it as an attachement file of this post if someone wants it ;)
Yeah, it's hidden away in a dark corner of the internet.

It also has some odd statements that don't blend well with modern systems (tach pulled up to 12V on the motherboard? That's crazy, I don't think anybody does that.)
 

Bidule0hm

Server Electronics Sorcerer
Joined
Aug 5, 2013
Messages
3,715
Thanks
648
#58
Yeah, IIRC they say it can be 12 V but usually it's 5 V and 3.3 V is recommended for new designs.
 
Joined
Apr 4, 2016
Messages
92
Thanks
3
#59
OK, so the server's pretty much built I'm just waiting for some more (4) HDDs to come. While I was waiting I run a 4 passes memtest and now try to configure the fans. Obviously it's not a proper test yet as I am missing my drives but it's a good time to play around.

I have 3 Noctua NF-A14 PWMs as case fans and an NF-F12 PWM on the CPU with a X9SRE-F mobo. I set the threshold according to manufacturer specs and also modified and connected a fan splitter as described in OP.
I run into two small problems :
1) when setting the fan mode in IMPI's webgui it just won't stick and defaults itself to full speed. It's the same when do it from the console. I managed to set it once and kept changing things and I could not set it again. Even a reboot didn't help. Then again after countless tries it's on OPTIMAL so I won't touch it ;)
[UPDATE] after using the ipmi raw command from shell there is no problem setting it
Code:
ipmitool raw 0x30 0x45 0x1 0x0 (Set Mode to Standard)
ipmitool raw 0x30 0x45 0x1 0x1 (Set Mode to Full)
ipmitool raw 0x30 0x45 0x1 0x2 (Set Mode to Optimal)
ipmitool raw 0x30 0x45 0x1 0x4 (Set Mode to Heavy I/O)
     
ipmitool raw 0x30 0x45 0x0 (Get Fan Mode)


2) I don't know if this related but all fans (incl. CPU) spin around 300-450 when the server idle except one which spins almost full speed (1,350) and nothing changes it.
[UPDATE] this behavior disappeared after a few reboots and reinstalls and after a couple months switch off ;)

Any suggestion?

I'll will disconnect it and swap headers, although not on the mobo (as that obviously only used for reporting RPM) but on the splitter itself so I'll see whether the other one will spin up and this down.
 
Last edited:
Joined
Apr 4, 2016
Messages
92
Thanks
3
#60
Although the server itself was ready to go months ago but unfortunately I had a series of issues with the HDDs (long story)

Now that sorted I am running badblocks since yesterday and have some time to thinker with the fan settings. I've made the splitter when the server was installed and the only significant difference compared to OP's setup is that I only use 3 fans for the case.

1) As I was already playing around with fan thresholds in ipmi before I noticed that when you reinstall freenas and/or unplug the server from mains the set thresholds are gone. Shouldn't it be retained by the ipmi firmware? Should I setup a startup script to set the fan mode and thresholds at boot?

2) As this is quite a new experience in the unix world I need some help with the scripting as well
I have 6 HDDs but the problem is two of them are /dev/adaX and the other 4 /dev/daX

Code:
## Load the disks with reads
#echo "Reading from All Disks"
## Start 5 disk reads spaced 60s apart
#  for i in 1 2 3 4 5; do
#  sleep 60
##  for each disk
#  for disk in 0 1 2 3 4 5 6 7; do
#  dd if=/dev/da${disk} of=/dev/null bs=1048576 &
#  done
Should the above script be like this?
Code:
## Start 5 disk reads spaced 60s apart
#  for i in 1 2 3 4 5; do
#  sleep 60
##  for each da disk
#  for disk1 in 0 1 2 3; do
#  dd if=/dev/da${disk1} of=/dev/null bs=1048576 &
##  for each ada disk
#  for disk2 in 4 5; do
#  dd if=/dev/ada${disk2} of=/dev/null bs=1048576 &
#  done


Thanks a lot guys
 
Top