PC Services

"From Payroll to Body Scanners"

Established 1994

PC Services (Electronics)

A Resolution Too Far

Tel: 0118 946 3634
Email

PCB CAD, Grids and A Resolution Too Far

PCB CAD and grids

I was laying out a PCB and had a really silly ratsnest problem where no matter what I did I could not get a track to go on a 45 degree angle and join two segments of a track together, so I checked its details which gave the following co-ordinates in mm shown in the picture

So the first From co-ordinate is 47.849,965,625 mm or

0.047,849,965,625 m

Now let us bear in mind some figures for the size of a copper atom, from various sources

1. https://www.copper.org/resources/properties/atomic_properties.html

   General and Atomic Properties of Copper
   Atomic Number	29
   Atomic Diameter	2.551 x 10-10 m (255.1 pm)

2. https://www.webelements.com/copper/atom_sizes.html
   gives diameter between 230 and 290 pm (10^-12 m)
3. Some go as low as 228 pm (10^-12 m)

Assuming diameter of a copper atom is 250 pm, that can be rewritten as

0.000,000,000,250 m

So comparing the two numbers against each other

0.047,849,965,625 m

0.000,000,000,250 m

The last digit of that position is a resolution (if only in steps of 5) of

5 / 250 = 0.02 (or 2%) of the diameter of a single copper atom !!

So even if we made a PCB or panel with a dimension 1 m or even 10 metre, how many sub-atomic particles would be out by over the whole distance and how many could measure it, or actually get a problem with it being out by more than 0.005 mm (5 micron).

This is about the smallest feature size you could reliably get using light or even single colour lasers, before going to UV or x-rays. Most PCB manufacturers only go down to 0.15 or 0.12 mm feature size on standard production facilities.

Note going down to multiples of 0.005 mm means removing the last SIX digits of that co-ordinate giving us -

0.047,850 m     ( 47.85mm )

NO standard PCB manufacturer could make anything to anywhere near the accuracy of the CAD package stipulates.

Is it not time for PCB CAD packages to be written by people who understand the physical limits of the real world, and how the work flow operates instead of graduates with a mathematics and graphics background.

W have not yet got 3D printers that can deposit sub-atomic particles or even single copper atoms to make PCBs, but obviously the CAD packages think so.