Airflow direction in rack-mount cases
Allan Herriman
mount cases (e.g. ETSI 300 119 and NEBS), and it's usually front to back,
bottom to top, or (least preferred) left to right.
But recently I noticed some equipment from Brocade in a 1U case that drew
air in at the back, and blew it out the front. This is actually mentioned
in the datasheet, so it was not an assembly error (with the fans in the
wrong way around).
Why would designers go against the standards and do this? Could there be
some benefit to the equipment?
Curious,
Allan
John Larkin
<allanh...@hotmail.com> wrote:
>Hi, I'm aware of some standards that dictate airflow direction in rack
>mount cases (e.g. ETSI 300 119 and NEBS), and it's usually front to back,
>bottom to top, or (least preferred) left to right.
>
>But recently I noticed some equipment from Brocade in a 1U case that drew
>air in at the back, and blew it out the front. This is actually mentioned
>in the datasheet, so it was not an assembly error (with the fans in the
>wrong way around).
It's called a "hand warmer."
>
>Why would designers go against the standards and do this? Could there be
>some benefit to the equipment?
Standards are for sheep. If a rack is internally plenum-fed with cool
(and sometimes super-temperature-controlled) air, it makes sense to
intake from inside the rack and dump the warm air into the room. We do
a bunch of electro-optical stuff that way, with the most temp-critical
stuff in the back of the box. The fan's in front, so its own heat
heads straight out.
Left-right will just spin hot air inside a rack. And probably restrict
flow badly. Ditto bottom-top.
Small benchtop instruments usually blow out the back, so as not to
annoy users with the air and noise. Whatever works best.
John
Allan Herriman
news:er826459ghl18nabn...@4ax.com:
> On Tue, 24 Jun 2008 11:02:25 -0500, Allan Herriman
> <allanh...@hotmail.com> wrote:
>
>>Hi, I'm aware of some standards that dictate airflow direction in rack
>>mount cases (e.g. ETSI 300 119 and NEBS), and it's usually front to
>>back, bottom to top, or (least preferred) left to right.
>>
>>But recently I noticed some equipment from Brocade in a 1U case that
>>drew air in at the back, and blew it out the front. This is actually
>>mentioned in the datasheet, so it was not an assembly error (with the
>>fans in the wrong way around).
>
> It's called a "hand warmer."
>
>>
>>Why would designers go against the standards and do this? Could there
>>be some benefit to the equipment?
>
> Standards are for sheep.
The standards exist so that designers and users can have a better chance
of getting their equipment to play well with others.
I've seen a lot of rack mount equipment, and the back to front airflow is
quite rare. Brocade are doing something that is different.
> If a rack is internally plenum-fed with cool
> (and sometimes super-temperature-controlled) air, it makes sense to
> intake from inside the rack and dump the warm air into the room. We do
> a bunch of electro-optical stuff that way, with the most temp-critical
> stuff in the back of the box. The fan's in front, so its own heat
> heads straight out.
>
> Left-right will just spin hot air inside a rack. And probably restrict
> flow badly. Ditto bottom-top.
Bottom to top may be installed with 1 or 2U high angled panels between
the equipment to make the airflow in at the front/bottom and out and the
top/back. Bottom to top is quite commonly used with vertical "line
cards" in 4-6 U chassis.
The north American CO style with open backs will have farms of racks
organised so that "cold aisles" and "hot aisles" alternate, with the
equipment drawing air from the cold aisle, and blowing hot air into a hot
aisle.
The ETSI ones are designed to have closed backs (which allows two racks
to be placed back to back or against a wall). Typically, air will be
exit out the top, and enter at the bottom front of the rack.
> Small benchtop instruments usually blow out the back, so as not to
> annoy users with the air and noise. Whatever works best.
Thanks for your input.
Regards,
Allan
mpm
I've seen many systems that blow from bottom to top - especially for
racks that sit on a raised "computer-type" floor where the space
underneath the floor acts as an air-conditioned plenum.
I've also seen (and unfortunately, had to work on) systems where the
boards would overheat and shutdown when placed on extender cards. The
use of extender cards made it possible to access the electronics, but
simultaneously removed the board from the forced air - causing the
regulators to shut down. I can still recall the occasional burn of
the inside of my forearms when not paying really close attention to
where the heat sinks were..... Ouch.!!
(Don't ever design something this way if you can help it.) -mpm
Actually, now that I think about it, the Larcan 40KW solid state VHF
TV Transmitter blows from the top down. (But really, it's just
blowing on the finals, which happen to be in the upper part of the
cabinet. I don't think its intended to cool anything else in the
cabinet - so that might not be a fair comparison??.)
JosephKK
Thinking a bit, bottom to top is helped by convection flows. Front
panel versus back panel requires consideration of who and why some is
at the front panel or back panel. The front panel is usually for
meter readers and such who are very intolerant of warm air in their
faces, back panels are usually full of connectors and such and
represent a restricted flow. Sideways flows in racks are useful only
in carefully managed situations. In all cases consider various users
and maintainers and total heat transport.
John Larkin
<Zarat...@thusspoke.org> wrote:
>On Sat, 28 Jun 2008 11:01:33 -0700, JosephKK <quiett...@yahoo.com>
> The idea is to create a positive pressure inside the rack module being
>designed. One draws air in directly through the fans, and creates said
>positive pressure. The fans should never be driving the exiting air, they
>should always drive the incoming air.
The big advantage of negative pressure is that it lets you locate and
size various inlet holes to blow just the right amount of air on
specific hot spots. You can't get that sort of tunability with a fan
blowing in.
Plus, the heat from the fan motor heads right out, not in.
John
Eeyore
Allan Herriman wrote:
> Hi, I'm aware of some standards that dictate airflow direction in rack
> mount cases (e.g. ETSI 300 119 and NEBS), and it's usually front to back,
> bottom to top, or (least preferred) left to right.
That's how I like to do it.
> But recently I noticed some equipment from Brocade in a 1U case that drew
> air in at the back, and blew it out the front. This is actually mentioned
> in the datasheet, so it was not an assembly error (with the fans in the
> wrong way around).
>
> Why would designers go against the standards and do this? Could there be
> some benefit to the equipment?
They're nuts.
Aside from anything else, why anyone would think operators would appreciate
having hot air blown at them is beyond me.
Graham
Eeyore
StickThatInYourPipeAndSmokeIt wrote:
> The idea is to create a positive pressure inside the rack module being
> designed. One draws air in directly through the fans, and creates said
> positive pressure. The fans should never be driving the exiting air, they
> should always drive the incoming air.
What a crappy idea. How about a negative pressure in the rack ?
Graham
Phil Allison
"Allan Herriman"
** For the most effective cooling, fans need to blow into an enclosure and
directly onto the hottest parts - ie all the heatsinks fitted to semis.
In the world of professional audio power amplifiers (where dissipations can
exceed 1kW per amp), the front panels are mostly used for user controls and
displays which leaves little room for any fans - so they wind up at fitted
back along with the in /out and AC supply connectors.
The result is that " back to front cooling" is the pretty much the norm -
though examples of the reverse exist and also back to dual side outlet
cooling too.
The worst idea out is " bottom to top cooling" - cos although this the
way air likes to flow and seems intuitively correct what you inevitably get
with a rack full of such amplifiers is the internal temperatures rising
alarming from bottom to top as well - since each amp is supplied with the
heated air from the ones below !!!!
..... Phil
Phil Allison
"Eeysore autistic mental defective"
> Allan Herriman wrote:
>
>> But recently I noticed some equipment from Brocade in a 1U case that drew
>> air in at the back, and blew it out the front. This is actually
>> mentioned
>> in the datasheet, so it was not an assembly error (with the fans in the
>> wrong way around).
>>
>> Why would designers go against the standards and do this? Could there be
>> some benefit to the equipment?
>
> They're nuts.
>
> Aside from anything else, why anyone would think operators would
> appreciate
> having hot air blown at them is beyond me.
>
** Yeah
- ya gotta feel sorry for all those poor cable monkeys slaving over hot
internet routers, day after day.
.... Phil
John Larkin
Here's a power amp with front intake, rear exhaust, and fans in the
middle. Something to offend everybody.
ftp://66.117.156.8/L500_Front.jpg
ftp://66.117.156.8/L500_Side.jpg
John
John Larkin
<Zarat...@thusspoke.org> wrote:
> With the difference being that there are no heat producing elements in
>the front section of the module.
>
The blue heatsink in the front has 32 300-watt power fets on it, clamp
mounted to copper heat spreaders. This is an MRI gradient amplifier,
about 17 KW peak power output.
>>
>>ftp://66.117.156.8/L500_Front.jpg
>>
>>ftp://66.117.156.8/L500_Side.jpg
>>
>
>
> PVC wiring, And it actually has side exhausts.
More on the back.
John
Phil Allison
"Phil Allison"
>
> ** For the most effective cooling, fans need to blow into an enclosure and
> directly onto the hottest parts - ie all the heatsinks fitted to semis.
>
> In the world of professional audio power amplifiers (where dissipations
> can exceed 1kW per amp), the front panels are mostly used for user
> controls and displays which leaves little room for any fans - so they
> wind up at fitted back along with the in /out and AC supply connectors.
>
> The result is that " back to front cooling" is the pretty much the
> outlet cooling too.
>
> The worst idea out is " bottom to top cooling" - cos although this the
> way air likes to flow and seems intuitively correct what you inevitably
> get with a rack full of such amplifiers is the internal temperatures
** There is ONE even dumber way to arrange the fan cooling in a rack mount
amplifier and that is to have the air leave by a identical patterns of slots
in the top and bottom of the case.
When two such amps are stacked - the air leaves from the top of one and the
bottom of the other, providing there is some space above and below them.
But when three such amps are stacked ???
I know two Australian manufacturers who built their amps like this -
Linear Transfer and PTM.
..... Phil
JosephKK
<jjla...@highNOTlandTHIStechnologyPART.com> wrote:
I see your point.
JosephKK
<myceli...@underyourshrooms.org> wrote:
>On Sat, 28 Jun 2008 12:24:27 -0700, John Larkin
> The problem with negative pressure systems is dust. Every little crack
>gets filled with it, and that proves that it does indeed get in, so a
>filtered system would lose cleanliness in a negative pressure system that
>doesn't sport a fully welded chassis and gasket sealed top and bottom
>lids. Which are nice anyway. Fan temps are negligible in systems where
>fans are needed. Otherwise straight convection would work. Since fans
>work the heat they add to systems is negligible in either flow case or
>they wouldn't be in business.
And i see your point as well.
JosephKK
That presumes operators walking in the equipment stacks, an iffy
supposition at best nowadays. And the tech's will be in the back
where the connectors are.
Eeyore
JosephKK wrote:
> Eeyore wrote:
> >Allan Herriman wrote:
> >
> >> Hi, I'm aware of some standards that dictate airflow direction in rack
> >> mount cases (e.g. ETSI 300 119 and NEBS), and it's usually front to back,
> >> bottom to top, or (least preferred) left to right.
> >
> >That's how I like to do it.
> >
> >> But recently I noticed some equipment from Brocade in a 1U case that drew
> >> air in at the back, and blew it out the front. This is actually mentioned
> >> in the datasheet, so it was not an assembly error (with the fans in the
> >> wrong way around).
> >>
> >> Why would designers go against the standards and do this? Could there be
> >> some benefit to the equipment?
> >
> >They're nuts.
> >
> >Aside from anything else, why anyone would think operators would appreciate
> >having hot air blown at them is beyond me.
>
> That presumes operators walking in the equipment stacks, an iffy
> supposition at best nowadays. And the tech's will be in the back
> where the connectors are.
Not my own expereince by a LONG way.
Patching and set-up / configuration is done at the FRONT of the racks. The rear
is for permanent connections.
Graham
JosephKK
No, it is best done from a remote graphical terminal addressing a
server that is collecting data on a regular basis from all the network
equipment.