Hi everyone,

I have some questions about local fiducial recognition from Quad4c mark
2.

First a bit of background information.
The local fiducials we are using are 1mm in diameter, and we are using a
single fiducial per fine pitch part.
The spindle runout is less than 4 mils.
The upward and downward looking cameras have been calibrated and
the downward looking camera is parallel to the y axis as is the upward
looking camera.

Now some information on the trials we have done.
We have two products we have been trying to accurately place QFP100
parts (among other parts) on. These products are A and B.
The A board has a HASL finish and we initially calibrated the Quad so
that the placement of the QFP100 was satisfactory. While doing this
calibration the image on the monitor was as follows: Draws two blue
circles with a cross in the centre, when the fiducial is found the image
shows the circles over the fiducial in a binary mode (i.e. the fiducial is
white and the background is black).

The B board has a gold finish. When parts were placed on this board
(after we calibrated for the A board) the parts were off the pads
significantly. The image of the local fiducials on the monitor was the
same as described above.

We then went and adjusted the lighting levels (changing the levels up
and down but eventually leaving them as they were) at the CC and reran
the B board. The result was a consistent failing of the local fiducials. The
image whilst trying to find the fiducials can be described as follows:
Draws a blue box with a cross in the centre. The system will search
changing lighting levels and changing back and forth from binary mode to
normal mode.

Then we switched the machine off and reloaded the file. After we
restarted the machine the results (while not perfect) were approaching
an acceptable level.
At this stage the fiducial image on the monitor was again different. It can
be described as followed: Draws the two blue circles searching for
the fiducial with a small blue cross at the centre. When the fiducial is
found the image shows circles over the fiducial in a normal mode (i.e. the
image did not turn binary at all).

What we are having difficultly understanding is how the fiducial search
algorithm appeared to change giving us the different results. Maybe
someone can shed some light on this for us?

We wish to use the correct algorithm, so can anyone describe what the
algorithms are actually doing to help us better understand which one
would be best to apply to our local fiducials?

Also does anyone else use local fiducials on these machines?

If so, are they being used successfully (i.e. is there still misplacement)?

I look forward to any response.

David
David Simmons
Production Engineer
Tait Electronics Ltd
558 Wairakei Rd, PO Box 1645
Christchurch
New Zealand
Direct Phone: (64) (3) 358 6663 Fax: (64) (3) 358 3603
Cell (025) 228 7066
Email: [log in to unmask]
http://www.tait.co.nz

################################################################
TechNet E-Mail Forum provided as a free service by IPC using LISTSERV 1.8c
################################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in the body:
To subscribe:   SUBSCRIBE TechNet <your full name>
To unsubscribe:   SIGNOFF TechNet 
################################################################
Please visit IPC's web site (http://www.ipc.org) "On-Line Services" section for additional information.
For technical support contact Hugo Scaramuzza at [log in to unmask] or 847-509-9700 ext.312
################################################################