[Beowulf] A Cluster of Motherboard.

[Vincent Diepeveen]

At 07:10 AM 11/10/2005 -0500, Robert G. Brown wrote:
>On Sat, 5 Nov 2005, oxylabtech at gmail.com wrote:
>
>> Hello
>
>> It would like to know if is possible to make a cluster of
>> motherboards? I don't speaking in a cluster of PC's, but of
>> motherboards. Somebody already saw? If it is possible, Could someone
>> help me?
>
>Yes, it is possible and it has been done a number of times.  Several of
>them are documented, more or less, in the list archives -- GIYF here.
>
>If I recall, most of them just attach the mobos directly to shelving and
>use a special wiring harness to drive them off of oversized power
>supplies.  All pretty much OTC parts, although the counter in question
>is that of Home Depot for a lot of it.  Fans are often mounted to blow
>over the shelves.  Somebody posted a whole series of pictures on at
>least one such effort.
>
>With micro ATX boards with integrated network and maybe video, you could
>probably work out a vertical mounting that gave you near-blade
>densitites.
>
>Two minor pieces of advice:  Try to stay safe -- electrical devices
>start fires, so watch the connections to e.g. wooden shelving.  Try to
>stay within codes (include FCC codes for EM emissions) both to stay safe
>and because if you pack things too close together without shielding you
>could conceivably get some electrical oddities that destabilize things.

Isn't shielding 50Hz fields pretty expensive, or do we speak of other type
of shielding here, so not getting rid of the electro magnetic field?

I tend to remember mu-metal is one of the metals that gets used, as it
completely eliminates 50Hz fields. One of problems is it that not much of
it gets made, so a square meter thin mu-metal is already a couple of 
hundreds of euro.

What other ways are there to get rid of the EMF?

I mean *solutions where with a tesla meter i can measure there is not much
of the radiation left*, i do not mean methods where one must believe on the
blue eyes of the salesman that it shields; i was threatened once by a
courtcase when i concluded in the commission (high voltage power) publicly
that a certain device wouldn't shield against EMF fields when a tesla meter
still would measure the same amount of radiation after bringing the device
to work under high voltage power cables (which have the tendency that more
cancer underneath it happens, which doesn't necessarily mean it is the
power cables themselves causing it, could be some other effect of course,
just the increase in cancer is there).

In our search to shield off against such fields we only found mu-metal, but
perhaps we missed some obvious solution that's getting used in electronics,
or the prices quote we got were simply far too high (which happens a lot of
course). I would really appreciate comments there, as i'm sure in mass
market electronics such as clusters with huge fields and a lot of electric
power, 
you really need to shield of equipment, so there must be many experts here.

Please note that general agreement seems to be that it is wise to avoid new
situations where persons must nonstop live in a field above 0.4 microtesla.

How is this with electronic components?

Also note governments accept that peoples working conditions may be a lot
worse than 0.4 microtesla.

Those building regulations are far newer than the outdated regulations from
public healthcare. Some governments still use norms like 100 microtesla, a
field you have basically when you put your head against a 400000 volt wire.

In general many electric devices have problems, or show disturbance
regurarly with fields above 1-2 microtesla. That was however observed with
amateuristic methods.

Vincent

>Jim Lux, IIRC, had a bunch to say on these latter points in previous
>discussions to this point.  He's the list's EE guru as much as anybody
>and builds all sorts of "interesting" microclusters.
>
>    rgb
>
>> Thankx³
>>
>> --
>>
>> Fernando.
>
>-- 
>Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
>Duke University Dept. of Physics, Box 90305
>Durham, N.C. 27708-0305
>Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu
>
>_______________________________________________
>Beowulf mailing list, Beowulf at beowulf.org
>To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
>