Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Placement Accuracy

ggspook

#23632

Placement Accuracy | 4 March, 2003

Hi Mike Interested on your thoughts about (Pam) Placement Accuracy Measurement I am looking into establishing Accurate repeatability of my equipment we check our calibrations Regular and the calibrations are in spec but this does not guarantee me repeatability

Scenario I have programmed a line for a class 3 medical product checked calibrations Machine all in spec I have run 15 cards and the 16th card shuts the line down for 2x 240 QFPS, s misplaced. My machines are ok so where do I go from there What do I measure to predict this occurring? Interested in your thoughts I hope I am not too vague

John

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

#23633

Placement Accuracy | 4 March, 2003

Hello John, Your second statement is a little vague about the 16th card shutting down the line with the 2x240 QFPs misplaced... i'll try to interpret, assuming the cards are actual product, which I think they are, your 16th card failed because 2 out of 240 QFPs were misplaced. Is this a one time occurance or can you repeat this? You are assuming your machines are in spec based on the calibrations done by the manufacturer. Based on experience, these calibrations are not always correct! There are a number of possibilities that could cause errors like this. First and most likely, a bent nozzle could easily create a deviation that incorrectly places the QFP. Second, a placement head could be sporadically failing in a specific motion sequence. Thirdly, axis table motion could possibly be out of spec, or something that's not tested with the calibration. Honestly John, there are too many possibilities without knowing more about the equipment and what the calibrations have fixed or verified. To answer your question, what to measure to predict this occurring: You can only determine complete machine placement accuracy (and capability) with a test, even multiple tests, designed to exercise all axes, heads and nozzles during the placement routine. Tests like this need to use glass plates and glass components so that variability in materials is removed from the placement process. Only then can the machine tell us what's wrong. And, only with multiple components and multiple plates exercising the machine, can you begin to statistically determine its performance and placement capability. Let me know if you require additional assistance. Thanks!

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

ggspook

#23634

Placement Accuracy | 4 March, 2003

Thanks Mike Honestly the whole subject of repeatability is Is confusing i.e. I do perform placement tests using glass plate�s Fuji IP3 and Pam on the Fuji chip shooters which calibrate to 3 sigma but I cannot guarantee placement repeatability

As you say the variables are enormous we are in the process of a six sigma and lean management transition so you know were I am coming from

I have been collating figures but what do I base them on The machine manufactures don�t really know either They only guarantee placement spec of .025 IP3, .1 FCP6 .04 FQP2. So what do I measure am I benchmarking what the manufacture is saying

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

#23643

Placement Accuracy | 4 March, 2003

You're welcome John, it can get confusing but what's important is that machine's exhibit accuracy AND repeatability. Generally, high Cp values represent high repeatability (for instance in darts: hitting a triple 20 over and over means you're very repeatable but not accurate because generally you aim for the bulls-eye) and high Cpk values represent high accuracy (again in darts, hitting the bulls-eye over and over means you are accurate and repeatable), so we try to translate this explanation to machine capability testing. I'm not exactly sure how the Fuji PAM works but if you're using the Accuspec for the IP3's, I've seen and used that before. Whether the tools can test for repeatability or not, it is very important to know your equipment capability. You need to find what the machine's operation and quality performance specification limits are, for instance a Fuji CP6 for instance should be able to perform to +/- 100 micron @ 3 sigma. They may be able to guarantee that they can place to .025 on an IP3 but can they guarantee they can do it within a quality performance specification? What you need to measure is the accuracy and repeatability performance on a machine and see if it meets what the manufacturer says it can do. If it cannot, then the manufacturer needs to get involved to fix the machine. I don't think you should be benchmarking the manufacturer, you should be verifying that the machine does what the manufacturer says it can do. That's what is important to you! Your goal should be to calibrate, measure and verify it's performance on a regular basis so you can count on less defects, higher utilization and more capacity. Let me know if you require additional assistance. Thanks!

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

bill yates

#23731

Placement Accuracy | 10 March, 2003

PAM uses precision cut black plastic parts against a white contrasting background why is this not acceptable??

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

#23736

Placement Accuracy | 10 March, 2003

Hello Bill, As I mentioned in an earlier thread, I was not familiar with the PAM system but have since been given a quick run down on how it performs its measurement routine. It is an acceptable way to measure and optimze Fuji chip shooter performance and probably works very well, but the conflicting issue is that the same camera that is used for normal production fiducial recognition is used for the accuracy measurement of component placement. If the camera motion axes are worn or do not immediately show signs of a mechanical deficiency, the accuracy measurement and optimization process could easily be jeopardized. It's not usually recommended to use a system to test itself when there are other low risk solutions that can independently and objectively measure capability. Let me know if you require additional assistance. Thanks!

This message was posted Add this forum to your site! Click to learn more. the OnBoard Forums

reply »

#24057

Placement Accuracy | 8 April, 2003

Running the PAM board is a great tool. Using another machine to get you CPK is very inportant. We use the DT651 which does everything the PAM rom card does plus it gives a CP and CPK value. You can also use it to verify feeders in the X and Y. SMT research is developing a machine that goes even further and calibrates the XCYC and XOYO.

John you never said how this 16th board shut down the line for being misplaced? Was it just a visual observation? Running the Glass plate for manual calibration isn't always accurate. Once again the DT651 can do these calibrations and give CPK results. If you want to check you placement more accurately run the PAM board through the IP. Then run it throught the CP using and look at the X and Y placement average. This won't give good 3 sigma, but usable X,Y and theta.

reply »

ii-feed SMD Intelligent Feeder

PCB contract manufacturer