At 04:41 PM 6/2/2005, Steve Allen wrote:
Hello All,
I've struggled over the years to determine the required area for my board
designs. Typically, I add up the area required for all the land patterns
and multiply the area by a scaling factor. I adjust the scaling factor
based on the details of the design. The scaling is adjusted by my gut feeling.
Yes, that's how I've done it for years. However, I used different scaling
factors.
Currently, I have a mixed package (through-hole and SMT) with 549
parts. I need to keep both assembly and materials costs down. That means
4 layers or less and all components on a single side. There's a number of
traces that will need to be fairly wide (.05in) and an area where I'll
have to pour large sections of copper. There are a number of
connectors. Land Pattern Area = 44.5 sq. in.
First of all, my original formulas were developed for single-sided design,
through-hole. But SMT single-sided is generally a little more
space-efficient (because many holes may be avoided and thus there is more
routing space on the reverse, while a little less on the component side.
Here are my guesses:
2-layer
Difficult = 115 sq. in.
Medium = 125 sq. in.
Easy= 135 sq. in.
Unstated was the meaning of "Land Pattern Area." I defined the area of a
footprint as the space dedicated to the part if parts were packed at 100%
efficiency, i.e., 100% density, which meant, for example, with through-hole
parts on 0.100 inch grid, the parts were placed as close as would be
allowed physically and the holes were on-grid. SMT parts were similarly
considered except that space must be allowed for fan-out vias. With *some*
SMT designs, one might allow SMT parts closer than that, if most vias could
be avoided, as in a memory array.
Once the 100% density figure was obtained, design density was then
calculated. Essentially, 50% density was considered non-penalty, no extra
time was allowed for design. So, *if* the 44.5 sq in figure was a true 100%
density figure, non-penalty, i.e., easy would be anything over 89 sq in for
the board. So all the board sizes given would have been considered "easy."
The maximum density obtainable would vary, of course, with the nature of
the design. With a memory array, one might be able to get above 90%. But
then there is the packing efficiency on the actual board to consider.
Mr. Jenkins' comments are worthwhile. When board size is not fixed, it is
best to start out with a generous board outline. However, my habit was to
always design fairly densely, in sections with some space between them.
Thus, I could tighten the board fairly easily if it was desired to make a
smaller board, or add a few parts in changes without making a big mess.
Tight sectional density, with the sections properly defined, should also be
better for noise control.
If there is board space left over after doing this, which would be normal,
I'd give the customer two basic options: the smallest possible board or a
larger board with some prototype pattern filling up the remainder. Which
one is better will depend on many factors, but the first option will result
in lower board cost, as well as possibly smaller external hardware (like
case), resulting in cost savings. And the second option may be better in
early design stages. If section design has been kept tight, it will be
relatively easy to shrink the board later, providing that space remains.
But if the case is being designed simultaneously, and the client is willing
to risk some redesign costs, and is reasonably confident in the existing
design, the tighter board may be best.
4-Layer
Difficult = 105 sq. in.
Medium = 115 sq. in.
Easy= 125 sq. in.
With 4-layer, I used 60% density as the limit of non-penalty design. Again,
all the figures given are quite generous, given the assumption I made about
the definition of footprint area.
Is this reasonable? Any other ideas?
It is probably far too generous, if the goal is to come up with a design
time estimate.
If the goal is to make the design itself as easy as possible, starting with
a larger board size is advisable. In no way will this *add* to design time.
But even the "difficult" design estimate I'd consider easy.
There may be standard case sizes to consider....
____________________________________________________________
You are subscribed to the PEDA discussion forum
To Post messages:
mailto:[email protected]
Unsubscribe and Other Options:
http://techservinc.com/mailman/listinfo/peda_techservinc.com
Browse or Search Old Archives (2001-2004):
http://www.mail-archive.com/[email protected]
Browse or Search Current Archives (2004-Current):
http://www.mail-archive.com/[email protected]
