On Thu, Feb 07, 2008 at 10:57:49AM -0600, David Carr wrote:
> I have the distinct (dis)pleasure of making some PCBs with QFN/MLF packages
> that have exposed copper paddles. 

I'm designing a board that will use one of these beasts. Actually I've already 
used components with copper paddles, but for SOIC (Hittite GaAs switches) and 
TSSOP (Linear Technology's LT3510 IIRC).

> I think I'll need to use stencils to properly apply paste for these packages. 
>  

It's better, but maybe not necessary for a prototype: I actually soldered the 
LT3510 by hand (I only needed to build 2 boards), putting some solder paste 
well past its limit date on the pad and heating the board from top (hot air 
gun) 
and bottom (heater plate). It was slow and I was afraid of overheating the 
chips, 
but it worked. However I may have been lucky...

For the Hittite part I had a stencil made (30 boards). The stencil manufacturer
told me that the solderpaste pad had to be split (and actually modified it).
The circuit has had other problems (many shorts from the PCB manufacturer) but 
the reflow of the paste deposited through the stencil has worked flawlessly.

> It appears that PCB simply uses SMT pads (slightly shrunk) as the apertures 
> for the paste layer.  

I had the impression that they were not even shrunk. I shrunk them by
hand by editing the Gerber file.

> I think this is a simple effective design for most packages --- except exposed
> ground paddles.  These require that the paste be broken into smaller
> "blocks" with say 75% coverage of the paddle.  (See Analog AN-772)

Even 75% seems a bit high from what I remember.

> 
> Has anyone on the list dealt with this issue in the past?  If not, I wonder
> if there is any way to distinguish paddle pads from regular ones so that I
> could modify the paste layer code to do the right thing.  I guess we could
> do something like look for square pads with a size > Xmm on a side.  That
> seems a big fragile though.

In my case, the ground paddles were not square. I think that the right
criterion is the largest of the two dimensions is larger than, say,
a couple of millimeters (the paddle for the Hittite part pad is 
1.85x2.49 mm). There are also weird cases, like SOT89 paddle+pin.

        Regards,
        Gabriel


_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to