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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Local fiducial spacing...

jdumont

#34881

Local fiducial spacing... | 10 June, 2005

Hi all, we are just getting into using local fiducials for BGAs. My question is, in addition to having to be on the same diagonal (opposing corners) do the fiducials have to be the exact same distance from the center of the part? I always thought yes but some boards that just came from engineering have them at different center distances...

Thanks in advance JD

reply »


RDR

#34883

Local fiducial spacing... | 10 June, 2005

Depends on your machine.

reply »

jdumont

#34887

Local fiducial spacing... | 10 June, 2005

Can you elaborate? If the point of local fiducials is to find the center of the part how could it matter on some machines and not others....?

reply »

pr

#34888

Local fiducial spacing... | 10 June, 2005

Are the local fids. necessary? Anyway, I can't think of any reason the local fids. would have to be spaced exactly the same distance from the part. We place BGA's and QFP's using board fids, and have yet to find a problem that local fids helped fix. Usually bad XY data, bad part recognition, bad support,not placing it at the end of the program or not pre camera rotating, were more likely culprits. I just never found them to be much help. good luck, pr

reply »

#34889

Local fiducial spacing... | 10 June, 2005

I agree with PR. We run down to 12 mil pitch with out local fids. We run with 2 global fids for a panel of up to 40 boards. This has worked very well for us.

Jerry

reply »


FD

#34892

Local fiducial spacing... | 10 June, 2005

This depends on how the software of your particular machine was written (designed) to handle local fids. Most machines do not have a requirement about the position of the local fids.

The purpose of local fids are the same as global fids. All the placements are to be in relation with the position of the fids.

If the board isn't clamped in the same position on the conveyor then the reading of the fids will detect this shift and all the placements are offset so that the placements are aligned to their pads and the placements maintain the positional relationship to the actual position of the fids.

The same thing for local fids. If the reading of the local fids detect that the board isn't in the correct position, then this one placement will shift accordingly.

The local fids should not need to be equal distance from the center of the pads. Your global fids are not equal distance from each placement.

And I really hate it when board designers only put one local fid in the center of the placement.

reply »

KEN

#34896

Local fiducial spacing... | 10 June, 2005

JD, fiducials are merely land marks. The vision processor will measure the actual location and compare it to the CAD location. It does not care if the design is symetrical, asymetrical or what. In the case of a single center design fiducial (which is useless), you may have to teach the same point to represent two fiducials (software dependent).

When are local fids helpful? Really long boards with dimensional stability issues (stretch and shrink).

reply »

Sven

#34900

Local fiducial spacing... | 12 June, 2005

Hi all, I agree with all the postings above, but I have a few ammendmets: (How is my english so far?) The purpose of the so called "local fiducials" was in the old days, when "board stretch" was a factor. At least people think that was a big matter... In those days the fine pitch component had a lead span of 50 mil. There are absolutlely no reason for use of a "local fid.s". Unless Youre are dealing with a low acrisy (how is my spelling or grammatics?) placement machine, that shifts in X-Y axis during travelleing. On the other hand, if You have such a poor placement machine; concider other alternatives.

We place 0.4 mm pitch TQFP:s without any problem and just using 3 global fid:s (our pcb:s are up to 400 x 300 mm). Sincerly,

reply »


RDR

#34912

Local fiducial spacing... | 13 June, 2005

Some machines may require that the two locals be used to calculate a center point where some (probably most) just use them as a second positional identifier. The fiducials are programmed as just XY and just prior to placement the machine will read them and adjust the specific placement as required. This can be necessary if the PCB has moved during the placemnt operation or if the initail recognition may be inadequate to precisely place the part from say poorly defined fids or a PCB that does not have optimal locations for the globals. I along with the other posters so far have never used them since the globals have always been adequate.

Russ

reply »

Stefan

#34923

Local fiducial spacing... | 13 June, 2005

I guess, since nobody is using local fiducials anymore, we forgot how to program the fiducials. Was it in the component-package file or in the placement file? Nevertheless, once you find the place, where you enter the fiducial position, you'll find that it doesn't matter, if they have the same distance from the center. Regardless of the machine you are using, two local fiducials are not doing any better than one. In order to determine the angular offset, the local fiducials are too close together to improve the angular correction. Local fiducials may improve the X,Y accuracy, which requires only one fiducial. The angular correction of all components are most likely calculated from the global fiducials.

reply »

jdumont

#34939

Local fiducial spacing... | 14 June, 2005

So I set up a board yesterday and did a test run. It appears that the spacing does matter on my machine. It placed the part off-center the same amount as one of the fiducials. Im going to look more into fixturing and placement pressure adjustments I think...

reply »

bobpan

#34940

Local fiducial spacing... | 14 June, 2005

If the placement is off the same amount everytime...(which I suspect it will be)...just adjust your placement co-ordinate for the part and be done with it.

reply »

#34942

Local fiducial spacing... | 14 June, 2005

What machine are you running. Someone here can probably help you more if you give us more detail.

Jerry

reply »

jdumont

#34943

Local fiducial spacing... | 14 June, 2005

Its a Contact Systems 3Z...

reply »

Stefan

#34947

Local fiducial spacing... | 14 June, 2005

I would rather take Bobpan's advise. If the placement is calculated from the center of two fiducials, than the placement should be off by 1/2 of the distance of the two fiducials, but not by the offset of that one fiducial. However, your approach in checking the placement pressure and other parameters, have not much to do with the fiducial positions.

reply »


FD

#34970

Local fiducial spacing... | 15 June, 2005

You didn't state if you are using all CAD data or not.

If you are not using any CAD data then all bets are off.

If you global fid XY data and all your placement XY data are all from CAD data, but you are manually teaching your local fids, then they will always place off. Whenever you manually teach you always have human error, machine error, board stretch and such be part of the taught data.

So always use CAD data for your local fids, or don't use them at all.

reply »

Reflow Ovens thermal process improvement

Plasma Prior to Conformal Coating