[rescue] Sun Kit Needed for EE Student Here

Don Y dgy at DakotaCom.Net
Wed May 3 23:07:16 CDT 2006


Andrew Weiss wrote:
> On May 2, 2006, at 10:06 PM, Bryan Gurney wrote:
>> I can sort of agree with this; I took a look at a picture of my  
>> U2's guts (the U2 is under a stack of stuff sitting idle, and I'm  
>> too lazy to lift the cover off right now), and there aren't any big  
>> honking electrolytic caps visible (I can't see under the Creator3D  
>> FFB card).  There's guaranteed to be a few "beer cans" in the power  
>> supply, but for the most part the board is clean.  I think some of  
>> the smaller cylinders are electrolytics; perhaps they get a better  
>> lifetime than the "beer can" electrolytics?
> 
> There are tons of beer can electrolytics on a POWER3-II node board.   
> These aren't exactly bad machines.

The presence of caps in a machine means very little.
The *quality* of those caps can mean a LOT (witness the
PC mobo issue).

Unfortunately, caps are an essential evil.  And, they
represent a significant cost item in most electronic
products.  And, in addition to connectors, are one of the
biggest sources of failures (speaking in general terms).

I think one issue that separates the "consumer" products
from the more "professional" products is the design
philosophy used.  This also impacts component choices.

For example, a consumer device might opt to use 15V
caps on a 12V supply.  OTOH, when I design a (nonconsumer)
product, I would spec 25V parts for the same 12V supply!
I derate components more heavily when I am trying for
higher reliability (e.g., specifying 105 degree caps
instead of 85, etc.).

But, in those markets, I am looking NOT to see "returns"
(warranty repairs, etc.) and spending the extra money
up front pays off big time (save one repair and you can
buy LOTS of caps!)

I'll have to take notice of some of the failed motherboards
that I come across and see how true this really is...



More information about the rescue mailing list