| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 233.1 | ...off the top of my head. | GONE::PIERSON | | Fri May 23 1997 17:22 | 32 |
| 1) New heat sink (larger) means redo shock & vibe.
2) Also EMI
3) also Thermals
4) redo sw quals
[And i may be forgetting summat...)
(Each of those is a task, with hours/dollars associated.)
Also need to know HOW slow is acceptable.
Also, JUST slowing the CPU still leaves the rest of the SBC dissipating
power. May not be able to get to where they need to get to.
Either needs a huge opportunity, or a noticeable amount of NRE.
Need to know what enclosure this would be in.
(ie: what airflow?)
If its NOT in a DMCC box, they still need airflow to cool the rest of
the configuration.
IF this is going into a DMCC Enclosure, the SBC fan does not add much,
if any, perceptible noise over and above the enclosure fan.
[I suspect the _real_ issue is the fan, per se, (ie: mtbf) rather
than noise, right?]
Is the customer aware that the SBC fan on k2 is wired into server
management and will alarm if it fails? (Does that address their
concerns? It may not....)
regards
dwp
|
| 233.2 | Don't worry about MTBF, worry about mechanical noise | BBPBV1::WALLACE | john wallace @ bbp. +44 860 675093 | Fri May 23 1997 19:15 | 11 |
| Real issue is actually noise not MTBF. This is a studio sound desk
(music, broadcasting, etc) application. Extraneous mechanical noise is
unacceptable. This is going into a custom enclosure which with Noname
is currently passively cooled (no fans). A standard DMCC enclosure is
out of the question in this application because of the fans.
ta for info to date, will pass to Account Manager. Who should he get
back to with info re volumes etc ?
regards
john
|
| 233.3 | Quiet Option Cards too? | NNTPD::"[email protected]" | Gordon McNab | Mon May 26 1997 14:37 | 13 |
|
John,
We're sitting in the lab right now listening to several DE500s
switching between AUI and Thinwire. It's a relay and is audible
above the noise of the box fans!
There's another thing to look out for in this config.
Gordon
[Posted by WWW Notes gateway]
|
| 233.4 | DE500 noise | HELIX::SONTAKKE | | Tue May 27 1997 17:15 | 3 |
| RE: .3
But that should only happen once when you have powered up the system
|
| 233.5 | Yes, I hope so. | caly70.ayo.dec.com::Gordon | Gordon McNab | Wed May 28 1997 09:11 | 9 |
|
When we run it with VMS we get quite a lot of clicks as the machine is
coming up, but when it's started then it's OK. If the carrier gets lost
from the network that is chosen would it try the other two connectors
for carrier signals while VMS is running?
Gordon
|
| 233.6 | | GONE::PIERSON | | Wed May 28 1997 23:39 | 14 |
| >de500 switching at power on
support engineers get the boards to do strange things.... 8)>>
>...acoustic noise...
ahmmmmm. Is the AXPpci 33 (noname) in one of our boxes?
If so, its' got a fan.
(If not, i congratulate their thermal designer... It CAN be
done, but...)
regards
dwp
(In any case, the control room is isolated from the recording
studio. yeah. I know. The Customer is Right...)
|
| 233.7 | If they'd waited for us to sell boxes... | BBPBV1::WALLACE | john wallace @ bbp. +44 860 675093 | Thu May 29 1997 11:52 | 14 |
| No, it's not in one of our boxes. This design goes back a while; the
customer had finished their design long before Digital were able to do
such exciting things as sell bought-in boxes. And anyway their box has
a lot more knobs and displays than ours (it's a mixing desk :-).
Last time I looked, there was a large piece of metal bolted on to the
21066 which then connected to the (metal) desk chassis, thus the whole
thing is used as a passive heat sink. There's plenty of thermal
capacity...
Keep the ideas coming.
regards
john
|
| 233.8 | | GONE::PIERSON | | Mon Jun 02 1997 17:02 | 36 |
| '...plenty of thermal capacity..'
There is IF someone did the numbers..
AND did the mechanical attach right.
Same trick will work with 21064 (the numbers needed to do desigh are
in the EBM4x UM) and with the 21164A (the numbers are not in the EBM2x UM.
but are avilable from Hudson).
_BUT_
(and this applies to the AXPpci 33 as well...)
What About the REST of the stuffe in there?
AXPpci 33 is a good case in point: the 3.3v regulator is the hottest
thing on there. did 'the customer' sink THAT to the box?
Ditto the memory, ditto the disks.... All Those are designed for
airflow coolling.
IMO (unless this is hugish opportunity, volume wise), the 'best'
thing to do might be to point the customer to the info on the Digital
Semi Publick site and let their metal benders have at it, so long as
they think about diska and memoryes and cache and such as well. If
they don't feel competent to tackle it, there are 'hired gun'
consultants who would assist them.
(As it happens i was at a presentation from some of the Sony folk, whose
names i do not recall, when they were stateside, 18 mos(?) ago...
And, presumably, this thing needs meet only 'office' (25-30C) in
contrast to some '_emmbedded_' which may go plant floor and need 40-50C
capability....)
regards
dwp
|