At 09:52 PM 2/28/2002 -0500, Mike Reagan wrote:
>Y' all havent heard my response yet because e Im trying to visualize if
>using one paste screen for alternating top and bottom panels will work.  (
>it was the best expanation I heard yet why anyone would want to mirror a

Actually, it's not a particularly good reason, since it would be easiest 
and best to do the flipping with the photoplots. I'd think a fabricator who 
wants to do what printers call "work and turn" would be set up to do this 
with very little effort, less effort than it would be for us to make a 
combined PCB file. It's like step and repeat, even if we do it ourselves: 
it is better to do the repeating in the gerber.

(If you edit the PCB later, do you want to edit all 32 copies of the board? 
Do you want to deal with inner layer net assignment problems -- you would 
need to use *only* split planes, not the default plane assignment, Protel 
is not set up, in the photoplot routines, for multiple copies of a board, 
with some flipped. And I'd not want to see the added complexity; rather, 
giving us a tool to easily step and repeat and flip-combine photoplots is 
quite sufficient, I haven't investigated trying to do this with CAMtastic.)

I produce a price list for my wife's business on our copy machine. The list 
is printed on 8.5 x 14 inch paper (a standard size in the U.S.). The final 
list size is 7 x 8.5, being half of the paper stock. The front and back are 
placed on the same side of a sheet of 8.5 x 14 paper, side by side (I tell 
the printer driver that the page size is 7 x 8.5, and I make a simple 
pasteup with the two pages -- I could set up printing to print the whole 
sheet at once, both sides together, but it has not been sufficient trouble 
to make the pasteup to bother). I then place this original on the copier as 
if I were copying an 8.5 x 14 sheet, which I am.... Then I take the copies 
and place them back in the paper feed. This is a copier which turns the 
paper over when it copies, so the flip is automatically done. I then copy 
again, exactly the same, and then cut apart the price lists.

This is a very common technique in printing, where it saves plates and 
press set-up, one gets to print both sides of the sheet with only a single 
press set-up and plate.

Printed circuits should work exactly the same way; as long as the stackup 
is symmetrical and the two images are placed so that turning the panel over 
backs up the original in exactly the correct register, the two boards, once 
cut apart, will be identical in every way. Any process that works on one 
side at a time will save one tooling and setup cost by using work and turn. 
Since normally boards are panelized anyway, this process simply needs a 
different panelization, with half the boards flipped.

If the geometry of this is difficult for anyone, take a piece of paper. 
Fold the paper in half. Draw a shape (any closed curve or polygon) on 
one-half that paper. Most of us don't have light tables any more, so lay 
the paper on a window with light shining through and trace the outline that 
appears on the other half (at this point the other side of the folded 
paper) of the paper. You can then unfold the paper and write "TOP" within 
one shape and "BOTTOM" on the other. Put this drawing, unfolded, in a 
copier set to copy on the same size paper at 100%. Run a few copies. 
Leaving the original under the platen, put those copies into the paper feed 
tray, and copy them again. With some copiers you will have to turn the 
paper over for the second run. Normally, you would turn it around the axis 
of the fold line you made. If the two copies don't back each other up 
accurately, there is one of two problems:

(1) the paper needs to be turned differently when it is put back into the 
copier, or
(2) the copier is not accurate, which is normal. But in this case the two 
images should be close enough that you can see the idea.

You should end up with a single sheet of paper from each copy which has two 
polygons on it which could be cut out to make two identical copies of the 
shape, printed both sides.

Any part of the process which involves imaging or placing something on one 
side at a time will save a setup using this process, as long as both sides 
of the board can be treated identically. That's not necessarily true for 
all kinds of assembly, so at the point where the handling is not 
symmetrical, one would cut apart the panel along the axis of image 
symmetry, assuming that the "top" side has been placed all on one side of 
the panel and the "bottom" on the other.
Abdulrahman Lomax
Easthampton, Massachusetts USA

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
* Contact the list manager:
* Forum Guidelines Rules:
* Browse or Search previous postings:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to