Fan Direction (was: Re: PC floppy cable twists...)

Allison ajp166 at
Fri Oct 21 16:16:24 CDT 2005

>Subject: Fan Direction (was: Re: PC floppy cable twists...)
>   From: Mark Tapley <mtapley at>
>   Date: Fri, 21 Oct 2005 13:24:57 -0500
>     To: cctalk at
>At 6:09 -0500 10/21/05, Allison wrote:
>>....  All of my systems
>>cooling and air flow get attention even if it requires some card stock
>>and tape duct work.  Fans blow in and filters keep cat hair and
>>other deutrius from getting in those places where it's going to do bad
>>things. I find that tends to keep the inside of my PCs clean too. I
>>turn the fan around and add a filter. Beats having a hairball clogging
>>the cpu fan and crashing the system.
><see my previous posting, somewhere back in the archives...>
>	Although this in general is a good idea, it doesn't always 
>work. NeXT cubes were  designed with the fan exhausting air, so the 
>crapola sucked in through the optical drive slot wiped out the 
>optical drive before long. NeXT eventually approved the procedure to 
>mechanically turn the fan around (do *NOT* simply reverse the 

That would be a bad thing.

>	Problem is the designers actually did do _some_ thinking. On 
>its way out the fan in the original configuration, the air went over 
>the logic board(s) first, then the power supply and drives last. 
>Reverse the fan, it's now hot air hitting the logic board. On my 
>system, this caused the SCSI controller to fail occasionally.
>	Solution in my case was to put the fan back to its original 
>configuration, then tape over the floppy and optical cut-outs on the 
>front of the case. (N.B. my system may be unique in that the floppy 
>is mounted with a custom mount to an optical-sized cutout, with no 
>attempt to seal well. It may be that either taping or reversing the 
>fan would have solved the problem.)

Before I did it there was some thought applied and a few changes like 
an added fan in a differnt location.  In the end I was both pressurizing
the case and insuring adaquate airflow over the hardware.

In the case of most PCs the airflow pattern is at best bad and its
difficult to make even small improvements.  But some of the S100 
cases the airflow was poor to non existant. In those trying to get 
a laminar airflow or even air to flow between the boards in a crowded 
case was hard to achive.  The worst is any of the altairs, one look 
and you see why.


More information about the cctalk mailing list