fiducial mark placement

Subscribe to Thread

I have always been taught to read fiducial marks that are d... - Dec 09, 1999 by

IF YOU READ FIDUCIAL MARKS THAT ARE PARALLEL TO EACH OTHE... - Dec 09, 1999 by JAX

If possible, three or more fiducials is optimal. That way... - Dec 09, 1999 by

In answer to your question no. Fiducials should always be a... - Dec 11, 1999 by Mark Wiegold

alan

#8252

fiducial mark placement | 9 December, 1999

I have always been taught to read fiducial marks that are diagonally opposed to each other. I am sure this is to get a better correction for skew. My question is if you use two fiducial marks that are parallel to each other even though placement might still take place, is it as accurate as the diagonal read ??

reply »

JAX

#8253

Re: fiducial mark placement | 9 December, 1999

IF YOU READ FIDUCIAL MARKS THAT ARE PARALLEL TO EACH OTHER THE PLACEMENT WILL NOT BE AS GOOD. ONE OF THE PROBLEMS WITH THIS METHOD IS IT COMPLETELY THROWS OUT THE CONCEPT OF BOARD STRETCH IN ONE AXIS. THE OTHERS MAY VARY DEPENDING ON WHAT MACHINES YOU USE.. ( ALGORITHMS, BOARD MAPPING, ECT�.)

reply »

Andy Yates

#8254

Re: fiducial mark placement | 9 December, 1999

If possible, three or more fiducials is optimal. That way scale, translation and projection can be estimated.

Andy

reply »

Mark Wiegold

#8255

Re: fiducial mark placement | 11 December, 1999

In answer to your question no. Fiducials should always be as far apart from each other in both directions (X,Y) as possible. This will allow ANY placement machine to calculate the EXACT location and rotation of the PCB, and therefore correct the mount data accordingly.

If you only use one direction (ie parallel fids) the machine can only ESTIMATE the mount positions. If you did try this method you will find the larger the PCB the larger the ERROR in estimation.

P.S. I've never seen a board with parallel fids!.

reply »