revisiting the 1000 faces of 0402
Werner Almesberger
werner at openmoko.org
Sun Jan 23 08:50:48 EST 2011
David Kuehling wrote:
> Last time I was confronted with footprint selection in geda-pcb [1], I
> found some interesting information in the footprint parameter list
> geda.inc [2]. Note how they modify footprints for resistors
> vs. inductors vs. capacitors. The generic 0402 footprint (at the
> bottom) is a compromise between all those (they use the metric naming
> for the first footprints. 1005 is actuaully a 0402):
Interesting, thanks ! But I must admit that I find the data somewhat
puzzling. If I understand the logic right, then, say.
> define(`PKG_INDC1005N', `PKG_SMT_2PAD_MM100( `$1', `$2', `$3', 70, 50, 80, 170, 100, 0, 0)');
would have pads 0.7 mm long (in the direction of the component's long
axis), 0.5 mm wide, and a spacing of 0.8 mm between pad centers, leaving
a gap of only 0.1 mm between pads. This seems awfully narrow. All the
vendor data sheets unanimously specify 0.5 mm for the gap.
I compiled the following "compromise" data:
- max of the mins/compromise/min of the maxes
- maximum across component types of the minimum value for each component
type
- the "compromise data"
- minimum across component types of the maximum value for each component
type
- what I have in stdpass.fpd
Pad length Gap Overall length Pad width
geda 0402 0.6/0.7/0.75 0.1/0.1/0.35 1.4/1.5/1.85 0.5/0.5/0.75
stdpass 0402 0.5 0.5 1.5 0.5
geda 0603 0.9/1.0/1.1 0.3/0.3/0.4 2.1/2.3/2.6 0.65/0.75/0.7
stdpass 0603 0.7 0.8 2.2 0.9
The low upper bound for the 0603 width is caused by PKG_RESC1608*.
Without this outlier, it would be 0.65/0.75/0.85.
Overall length and pad width are all within the expected range, with
0603 tending towards very narrow pads. What does differ a lot is the
gap size. A 0.1 mm gap just looks like asking for trouble.
On page 51 of "The Circuit Designer's Companion, 2nd ed.", Tim Williams
also gives a gap of 0.8 mm for 0603, noting that this leaves room for a
10 mil / 0.25 mm trace. Routing a trace across a 0603 component seems
to be a pretty common practice.
So either I misinterpreted that table, or the choice of gap size is
highly unusual.
- Werner
More information about the discussion
mailing list