Hi Poh,
I have done the evaluation pretty much as Lainie Loveless describes in
an earlier response and agree with Lainie as to methodology. This method
is as good as you make it. I will stress that there are multiple,
interdependant, variables to be considered and you should design the
evaluation around isolating those variables that you are really trying
to control. The tools commonly refered to as FMEA (Failure Modes and
Effect Analysis), MSE (Measurement System Evaluation), and DOE (Design
Of Experiments) are very helpful in this regard.
More practically, I found using clear double-sided tape a plus. This
allowed me to use the pads on my board as my measurement reference for
individual components. I used the same board for all runs to eliminate
board to board variation, since the pads were my measurement reference.
You might consider purchasing "dummy components" for use in the
experiment they are cheaper and usually more consistant in dimension. I
bought mine from Topline but there are several vendors for these. Also I
would suggest using a few of your common style components 0805, 1206,
MELF's, PLCC's, ect... to check the machines ability to cope with
different body styles. Finally, make sure the system of measuring
repeatability and accuracy is up to the task.
You've got a really good project to use the above analysis tools and
don't forget to throw the terms around a lot, it really impresses
management. Hope this helps. Have fun and best of luck.
Steve Tucker
Mfg. Eng. AlliedSignal Aerospace
>----------
>From: Lainie Loveless[SMTP:[log in to unmask]]
>Sent: Tuesday, September 30, 1997 10:12 AM
>To: [log in to unmask]
>Subject: Re: [TECHNET] Assy: Placement Machine Measurement
>
>> Could anyone advise how to measure the machine placement accuracy and
>> repeatibility for the Pick and Place machine.
>
>Hi Poh --
>On a CS 3AV, this is how we do this:
>
>1. Get a real of 1210's (decent size for this test and cheap) and
> cover a board with double stick tape in the area where you'll be
> placing your test parts.
>
>2. Create a program with known coordinates (something along the lines of:
> R1 x=1 y=1 Theta=0 Tool=1
> R2 x=1.5 y=1.5 Theta=90 Tool=2
> R3 x=2 y=2 Theta=180 Tool=1
>
> ...and so on). We place maybe 8-10 parts for each spindle. Choose
> your fiducial (maybe lower left) as both you "Zero" and your "Skew."
> This will give a 0 skew angle for calibrating.
>
>3. If you have multiple heads, do one at a time. If you have
> multiple spindles on a head (we have 2 on each), then do the
> following to ensure you are calibrating the correct spindle:
> For a real of 1210's, say you're using a .100 dia tip, and its name
> is T1. Assign T1 to R1 and T2 to R2 (and so on). T2 can be defined
> as a .XXX dia tool, but put another .100 tool in its holder (if
> you're using 1210's, for example). This way, you can control
> which tip places which part and use the same (one) real of parts.
>
>4. If you have a transfer option (to a conveyor or reflow, for
> example), disable this, so the finished board remains loaded in the
> machine (to help reduce the chance of disturbing the placed parts).
>
>5. After you run the program, go back and visually digitize the parts
> (find their centroid coordinates of the part, where you usually
> do the pads).
>
>6. Find the differences for each placed part between what you
> programmed and where it actually placed.
>
>7. Separate the deltas into groups corresponding to each spindle
> tested. We have a spec of .004 as the widest margin of error
> (repeatability) and an average of +/- .002 (accuracy). If our
> deltas were above these values, we would either re-calibrate or
> adjust the offsets, as necessary, and then re-perform this test.
>
>Hope this helps. Tell me if any of this is unclear or does not
>apply.
>
>Regards,
>
>Lainie Loveless
>General Atronics Corp
>Wyndmoor, PA
>(215) 242-7336
>
>##############################################################
>TechNet Mail List 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 web site (http://www.ipc.org/html/forum.htm) for additional
>information.
>For the technical support contact Dmitriy Sklyar at [log in to unmask] or
>847-509-9700 ext.311
>##############################################################
>
##############################################################
TechNet Mail List 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 web site (http://www.ipc.org/html/forum.htm) for additional information.
For the technical support contact Dmitriy Sklyar at [log in to unmask] or 847-509-9700 ext.311
##############################################################
|
