Altair Fan

Randy McLaughlin cctalk at randy482.com
Fri Jun 17 13:27:55 CDT 2005


From: "Dwight K. Elvey" <dwight.elvey at amd.com>
Sent: Friday, June 17, 2005 12:56 PM


> >From: "Pete Turnbull" <pete at dunnington.u-net.com>
>>
>>On Jun 17 2005,  7:47, Cini, Richard wrote:
>>
>>> Rules-of-thumb like this are great to know. The fan, I'm sure,
>>was
>>> installed (like the IMSAI) blowing IN because the inside was caked
>>with
>>> dust. So, I made it an exhaust fan (no filter).
>>>
>>> As I work on the IMSAI, which has no filter either, I will swap
>>it
>>> around, too.
>>
>>Don't do that.  If it's designed to blow in, changing the direction to
>>blowing out will alter and probably reduce the cooling.  Why?  Because
>>turbulent air is lots better at cooling than a laminar flow, and will
>>reach more parts of the case.  You get turbulent air from the "blow"
>>side of a fan, but laminar flow towards the inlet side.  You may also
>>be directing the airflow away fom some component that previously was
>>cooled, eg a PSU.  Moreover, if you alter the flow direction so that
>>air is being sucked in through all the other orifices, it will be
>>entering via the fronts of disk drives etc (well, probably not on this
>>machine but I'm thinking of the general case).  That's the last place
>>you want the dust.  Far better it drops on the motherboard where you
>>can vacuum it out.  Better still, fit a filter.
>>
>
> Hi Pete
> You have to remember, these machines were designed
> by electrical engineers, not an air flow expert. Yes,
> the original used in blowing air.
> I did a quick check of the temperatures of the boards
> in my IMSAI with the in blowing air and some boards
> seemed to get no flow at all. Reversing the flow made
> a big difference. The powersupply in the IMSAI, at least
> is large enough to need little more than some air flow.
> The only parts that might get uncomfortable are the
> filter caps. Any additional heat here is harder on them.
> Still, after many hours of use, with almost a full
> boat of 8K boards, I find the caps to be mildly warm.
> While not specifically schooled in air flow, I did
> work for a company that made burnin ovens. I learned
> a lot about air flow there. We would often get request
> from customers to fix the airflow in in-house designed
> chambers that had really poor temperature distribution.
> Mil spec states how uniform the chamber should be.
> Like I said, many of these were designed by electrical
> engineers.
> Dwight

I recently had a few emails back and forth with Todd Fischer on the airflow 
characteristics of the Imsai.

The fan was an option for the Imsai and is not required for normal operation 
at all, personally I strongly recommend installing a good fan.

In the past I know of many people that would block some of the openings 
around the chassis and have are flow one direction or the other to cool 
specific cards.

It is a good idea to try and test the temperatures that are built up and not 
just use the old here is the right answer don't worry approach.

After it has run for a while open the case and just use your hand to find 
hot spots, it is common for some voltage regulators to get extremely hot. 
If so try rearranging the cards to increase airflow where needed.

In general most S100 chassis prefer lots of cards to run cooler, more cards 
alter airflow as well as pulling down the power-supply voltages which make 
the regulators run cooler.

Todd Fischer's advice is to leave all the ventilation holes open to keep the 
power-supply cooler.

Many other manufacturers believed it best to not have ventilation right 
beside the power-supply forcing air over the S100 cards before reaching the 
power-supply.

As I said my recommendation it to set it up then test it, change things and 
test again.


Randy
www.s100-manuals.com 




More information about the cctalk mailing list