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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Jacob Lacourse


IPC-TM-650 | 29 January, 2003

Hello Friends,

We currently inspect our incoming pallets according to IPC-TM-610 for bow and twist. Normally pallets will come in that are twisted and we send them back to the board house confident in our testing. Today we recieved a pallet that was bowed. According to IPC the discription of a bowed pallet is a specimen rectangular in shape that has a cylindrical curvature and four corners are touching one plane.

To test for maximum allowable bow percentage one must first determine on what axis the bow occurs one being length while the other being width. Based on this determination you will then be able to use the proper formula to test the specimen. For instance if you experienced a board with bow in the length direction you would use the formula RL=L(B)/100 where:RL=GO/NO-GO feeler guage, L= Length of the board and B=Maximum allowable twist pecentage determined by IPC-T-50). From the calculation you would get the maximum allowable bow that you would use to measure the outside perimter of the board in the length direction. This is where we have our problem the pallet is showing little or no deflection on the perimeter all the deflection occurs in the center of the pallet. My question to you is, what formula do I use to determine maximum allowable bow the Legth, Width, or a combination of the two?

Please advise if this is the right action to be taking for this particular circumstance? I beleive that the v-scoring could be causing our board to bow. The v-score spec allows for a remaining web thickness of .015". The v-score web does fall within spec but it does not remain in the center of the board thickness. for example if this | | represents the board thickness and this >< |. Please advise if the v-scoring might be our problem and if there are any IPC test methods to determine acceptability.

Much Thanks, Jacob Lacourse

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »

Jacob Lacourse


IPC-TM-650 | 29 January, 2003

Please ignore the v-score figure it did not come out the way I intended. The v-score is not uniform in that equal amounts of material are not removed from each side of the board.

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »



IPC-TM-650 | 29 January, 2003

Most bowing I see is caused by an asymetrical layer stackup of the PCB. (different copper thicknesses throughout the stackup) e.g. for a 4 layer board - 2 oz. copper on layer 1, 1 oz. copper on layer 4, etc... I would doubt that is the V-score since we use a lot of scored boards and they ARE NEVER RIGHT (see PCB quality thread) and we do not see bow/warp/twist on any of these that are layed up symetrically. I can't comment on the measurement because I continually fight suppliers Q.C. etc... on how to do it. I can say that bow will only affect one axis or dimension. If more than one dimension is affected it should be classified as warp or twist depending upon what you see.

Not much help but it is all I know


reply »



IPC-TM-650 | 31 January, 2003


I hope I am on point based on my reading.

The outer edges bow spec depends on a gradual deformation that finally builds up to the outer edges.

If you mathematically move this deformation back you will have allowable deformation at the point your are seeing.

If it is obvious you have a problem maybe.

If it solders good and works good in the app it is OK.

Now if you are buying pcbs, circuit boards, to the IPC spec and don't have a spec to cover this deformation you should probably accept the defect, and respecify.

YiEng, DDave

This message was posted Add this forum to your site! Click to learn more. the Electronics Forum @

reply »


IPC-TM-650 | 1 February, 2003

It sounds like your board is "dome shaped", some call it "cupped".

If so, this is more of a twist than a bow. Before moving on, if we receive a board that is so torqued that we can't figure-out how to measure it, we reject it.

Go back to your TM-650, Bow & Twist and look [going on memory here] for something like twist reference method. Basically, it uses dial indicators to measure the the height of the inside of the cup and then flips the board and measures the height of the outer edges of the board.

reply »

PCB Routers