revisiting the 1000 faces of 0402

Werner Almesberger werner at openmoko.org
Sat Jan 22 08:08:23 EST 2011


A while ago, I tried to determine the proper footprint sizes for
passive standard components (0402, 0603, etc.). What I found was
that major component manufacturers wouldn't even quite agree on
the exact size of their packages, let alone on footprints. [2]

I then picked the parameters published by Vishay as a reference,
since they yield relatively compact footprints. They also agreed
roughly with what we had used in GTA02, and I had not heard of any
major difficulties with 0402 or 0603 there.

When making boards with these footprints, I found that they were a
little short for comfortable manual soldering. Therefore, I
introduced a set of -M variants (0402-M, 0603-M, etc.) that were
elongated by a total of 0.1 mm each.

Recently, Adam showed me the footprint sizes he prefers, based on
what VIA use. They're HUGE [1]. So I went back to checking what
the rest of the industry recommends, this time also including
the pad length and the gap between pads.

Here's the data (if reading this with a mail reader that uses a
proportional font, you'll either have to switch to a monospaced
font or copy & paste what follows into an xterm or equivalent):

Package

    |<----Px---->|      Px = Package total length = 2*Tx+Bx
    +--+------+--+      Py = Package width
 Py |  |      |  |      Tx = Terminal length
    +--+------+--+      Bx = Body length
    |Tx|<-Bx->|

Footprint

   |<-----Fx----->|
   +----+    +----+     Fx = Footprint total length = 2*Cx+Gx
Fy |    |    |    |     Fy = Footprint width
   |    |    |    |     Cx = Copper length
   +----+    +----+     Gx = Gap length
   |<Cx>|<Gx>


0402, reflow

        Source          Px   Py   Tx   Bx   Fx   Fy   Cx   Gx

        stdpass [11]    1.0  0.5  -    -    1.3  0.6  0.4  0.5
          *-M           1.0  0.5  -    -    1.4  0.6  0.45 0.5
        Adam [1]        -    -    -    -    1.7  0.66 0.69 0.33
        GTA02 [2]       -    -    -    -    1.4  0.4  ?    ?

        AVX [3]         1.0  0.55 0.2  0.6  1.7  0.55 0.6  0.5
        Panasonic [4]   1.0  0.5  0.2  0.6  -    -    -    -
        Vishay [5]      1.0  0.5  0.25 0.5  1.3  0.6  0.4  0.5
        Yageo [6]       -    -    -    -    1.5  0.6  0.5  0.5
        Yageo [7]       1.0  0.5  -    -    1.5  0.5  0.5  0.5
        T. Yuden [8]    1.0  0.5  0.25 0.5  1.4  0.5  0.45 0.5

0603, reflow

        Source          Px   Py   Tx   Bx   Fx   Fy   Cx   Gx

        stdpass [11]    1.55 0.85 -    -    2.0  0.9  0.5  1.0
          *-M           1.55 0.95 -    -    2.1  0.9  0.55 1.0
        Adam [1]        -    -    -    -    2.6  0.99 1.04 0.53
        KiCad SMT0603   -    -    -    -    2.2  1.15 ?    ?    (process ?)

        AVX [3]         1.6  0.81 0.35 0.9  2.3  0.8  0.85 0.6
        Panasonic [4]   1.6  0.8  0.3  1.0  -    -    -    -
        Vishay [5]      1.55 0.85 0.3  0.95 2.0  0.9  0.5  1.0
        Yageo [6]       -    -    -    -    2.6  0.8  0.9  0.8
        Yageo [7]       1.6  0.8  -    -    2.3  0.9  0.8  0.7
        T. Yuden [9]    1.6  0.8  0.35 0.9  2.3  0.7  0.7  0.9
        Lite-On [10]    1.6  0.8  0.4  0.8  2.3  0.8  0.8  0.7

Package Dimen   Average Ranking
0402    Fx      1.48    1.5 (2)         1.3, 1.4, 1.7 (1)
        Fy      0.55    0.5, 0.6 (2)    0.55 (1)
        Cx      0.49    0.5 (2)         0.4, 0.45, 0.6 (1)
        Gx      0.5     0.5 (5)
0603    Fx      2.3     2.3 (4)         2.0, 2.6 (1)
        Fy      0.82    0.8 (3)         0.9 (2)         0.7 (1)
        Cx      0.76    0.8 (2)         0.5, 0.7, 0.85, 0.9 (1)
        Gx      0.78    0.7 (2)         0.6, 0.8, 0.9, 1.0 (1)

Conclusions:

0603 shows considerably more variation than 0402. This is probably
because equipment was less accurate in the heydays of 0603, so
larger tolerances were needed. I see little reason for designing
for such obsolete processes, also because many of the more complex
components, e.g., QFN, require the same or higher precision than
0402.

The footprints in stdpass.fpd are at the short end of the range of
what the industry recommends, so I'll make them a bit longer. The
width is already quite conservative and yields pleasant results with
manual soldering.

Anyone already using stdpass.fpd will not notice a change immediately,
because KiCad keeps a copy of the original footprint definition in the
board file. However, if explicitly updating the footprints (with Edit
footprint > Change Module(s)), KiCad will use the new definitions.

Since the new footprints will change the point where traces connect
to the pads, some light re-routing is necessary after an update.

Actions:

- elongate 0402 from 1.3/1.4 to 1.5 mm
- eliminate 0402-M, since the new 0402 will be even longer than 0402-M
- elongate 0603 from 2.0/2.1 to 2.2 mm
- eliminate 0603-M, since the new 0603 will be even longer than 0603-M
- defer updating 0201 and the larger sizes until later

Details:

        0402                    0603
        Fx   Fy   Cx   Gx       Fx   Fy   Cx   Gx
old     1.3  0.6  0.4  0.5      2.0  0.9  0.5  1.0
old *-M 1.4  0.6  0.45 0.5      2.1  0,9  0.55 1.0
new     1.5  0.6  0.5  0.5      2.2  0.9  0.7  0.8

[1] http://en.qi-hardware.com/wiki/File:0603_0402_stencil_aperture_dimension.png
[2] http://lists.openmoko.org/pipermail/gta02-core/2009-September/000563.html
[3] http://www.avx.com/docs/Catalogs/accuf-p.pdf
[4] http://industrial.panasonic.com/www-data/pdf/AOA0000/AOA0000CE1.pdf
[5] http://www.vishay.com/doc?20035
[6] http://www.yageo.com/exep/pages/download/literatures/PYu-R_Mount_5.pdf
[7] http://www.yageo.com/exep/pages/download/literatures/UPY-C_GEN_13.pdf
[8] http://www.yuden.co.jp/ut/product/pdf/dlmk_e.pdf
[9] http://www.yuden.co.jp/ut/product/pdf/djmk_e.pdf
[10] http://optoelectronics.liteon.com/en-us/api/DwonloadFileHandler.ashx?txtSpecNo=DS-22-99-0151&txtPartNo=LTST-C190KRKT
[11] http://downloads.qi-hardware.com/people/werner/tmp/stdpass.pdf

- Werner




More information about the discussion mailing list


interactive