wathey at salk.edu
Mon Oct 4 16:37:15 PDT 2004
On Mon, 4 Oct 2004, Jim Lux wrote:
> At 03:19 PM 10/4/2004 -0700, Jack Wathey wrote:
> Very nice.. particularly the efforts you put into the cooling air. The
> comment about flow against a pressure drop is very sound.
> I realize it's a time thing, but had you considered removing the fans from
> the power supplies?
The linear flow rate through the whole rack is typically about 120 to 200
fpm, which is sigificantly less than the linear flow rate through a PS.
If the PS fan was there *only* to cool the mid-tower box it was meant to
go into, then removing it would do no harm. But a PS generates a fair
amount of heat of its own, so I left the fans. And as you say, it would
have taken time.
> What's a rough order of magnitude cost on the blower/VFD? Does this VFD have
> a "servo" input that could be used to automatically change blower speed in
> response to temperature? (some VFDs have an analog input that can be used to
> set up a speed = linear function of input voltage, and by cleverly setting
> the gain and offset, you can do quite nicely)
The blower was about $1300, the Teco inverter about $1000. The inverter
has an rs232 interface for computer control, which I don't use at present,
but hope to some day. It also has the analog control feature you
More information about the Beowulf