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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Screen printing the board from........

Ron Costa

#16614

Screen printing the board from........ | 4 March, 1998

Hello everyone! Does any know anything about bare board size variations? Is there a spec. or tolerance? I'm running small lots of boards and during the screen printing process I find that I cannot paste each board perfectly.I've tried other screen printers as well as a new stencil generated from the bare board itself. This board is about 13" x 15" and has a variety of SMT components such as 50mil down to 20mil QFP's and TSOP's. I'm using a chem etched stencil with trapezoidal aps. Could it be my stencil? Someone has recommended a laser cut stencil. What is the difference, accuracy wise between Chem. etched and laser cut? Would appreciate any feedback. Thanks, Ron

reply »

Steve Gregory

#16618

Re: Screen printing the board from........ | 5 March, 1998

Hi Ron, Your question really opens up a "Pandora's box", but when you say you're not able to print perfectly, what is it exactly that you mean? Is the print lacking volume? Is it mis-registered? Are you bridging? Also, what printers have you used? Do they have vision? Or is the board being located with tooling pins? Let us know some more specific details about the problems you're experiencing, and the types of printers you have used, and I think I can be more helpful with some suggestions that may point you to a fix. In my experience, it is pretty rare nowdays to have a board so out of spec from the cad data that you've been provided to generate the stencil, that you can't get a decent print...that's provided the stencil has been made correctly within certain guidelines that are necessary to get good paste release from the apertures. I'm at NEPCON now, participating in the TAC line. We're building a double-sided PCMCIA board that has a 15.7mil 256-pin TQFP, 20-mil TSOP's, 256-ball BGA, Micro BGA's, a Flip-chip, 0402's, 0603's, 0805's, SOT-23's, and some 50mil SOIC's. The board is turning out pretty good (considering we only had 3-days to set the line up and run product for the show!) This is a flip-line too, and it's been quite a challenge to get all our machines to talk to each other and "tweek" things in so we don't look like idiots....(GRIN). This board is only only .020" thick, and is pretty flexible, so that in itself can introduce some variables as far as having a nice stable surface that is repeatable to print on, but we're doing it. One of my specialties is printing, so if you can provide a few more details about your situation, I'm sure I can help you... -Steve Gregory-

reply »

Justin Medernach

#16615

Re: Screen printing the board from........ | 5 March, 1998

| Hello everyone! | Does any know anything about bare board size variations? | Is there a spec. or tolerance? | I'm running small lots of boards and during the screen printing process | I find that I cannot paste each board perfectly.I've tried other screen printers as well as a new stencil generated | from the bare board itself. This board is about 13" x 15" and has a variety of SMT components | such as 50mil down to 20mil QFP's and TSOP's. I'm using a chem etched stencil with trapezoidal aps. | Could it be my stencil? Someone has recommended a laser cut stencil. | What is the difference, accuracy wise between Chem. etched and laser cut? | Would appreciate any feedback. | Thanks, | Ron Ron, Steve hit it all right on the head. I too am a self-proclaimed printing wiz. There are several aspects to be concerned about. Steve pretty much nailed them all but I'm going to get into some aspects of stencil design and manufacture. There are basically two types of ways to manufacture a stencil. There are subtractive and additive processes. Subtractive processes are laser cut and chem etch. EFAB is an additive process. The subtractive processes are pretty much standard. Both will give you roughly the same positional accuracy although the laser cut stencil may be a hair more accurate depending on the fabricator's chemical etching and photo imaging processes. The big difference between the two will be in the release characteristics. Laser cut stencils do tend to release better than chemically etched stencils. It is difficult to control the trapezoidal etch with chemicals. It is a natural part of the laser cutting process. As the laser makes its' way through the steel it loses intensity and thus will form a trapezoidal aperture. This design does facilitate a slightly better release but good reductions and pad dimensions can eliminate the need for laser cut stencils entirely. We build product with 16 mil pitch using chem etch stencils with a very high yield. Your problem may not be entirely related to the printing process but may be hidden in your substrate. What types of defects are you encountering? This is important. What is your underside wiping frequency? Do you qualify this frequency or is it some number determined by the ever popular "engineering experience." If you can get your hands on proceedings from NEPCON East or West from '97 do so. Read my paper in each on "Improvising the Fine Pitch Screen Printing Process." There is a ton of good information in them and they should prove helpful. Remember, whenever possible, benchmark success. If it's not possible make your own success through TQM, then SPC, then DoE. It's a methodology that works. I've done it three different companies with the same result. Try it out. Post more info. Good Luck and Best Regards, Justin Medernach Mfg. Eng. Flextronics Int'l

reply »

Ron Costa

#16616

Re: Screen printing the board from........ | 5 March, 1998

| | Hello everyone! | | Does any know anything about bare board size variations? | | Is there a spec. or tolerance? | | I'm running small lots of boards and during the screen printing process | | I find that I cannot paste each board perfectly.I've tried other screen printers as well as a new stencil generated | | from the bare board itself. This board is about 13" x 15" and has a variety of SMT components | | such as 50mil down to 20mil QFP's and TSOP's. I'm using a chem etched stencil with trapezoidal aps. | | Could it be my stencil? Someone has recommended a laser cut stencil. | | What is the difference, accuracy wise between Chem. etched and laser cut? | | Would appreciate any feedback. | | Thanks, | | Ron | Ron, | Steve hit it all right on the head. I too am a self-proclaimed printing wiz. There are several aspects to be concerned about. Steve pretty much nailed them all but I'm going to get into some aspects of stencil design and manufacture. There are basically two types of ways to manufacture a stencil. There are subtractive and additive processes. Subtractive processes are laser cut and chem etch. EFAB is an additive process. The subtractive processes are pretty much standard. Both will give you roughly the same positional accuracy although the laser cut stencil may be a hair more accurate depending on the fabricator's chemical etching and photo imaging processes. The big difference between the two will be in the release characteristics. Laser cut stencils do tend to release better than chemically etched stencils. It is difficult to control the trapezoidal etch with chemicals. It is a natural part of the laser cutting process. As the laser makes its' way through the steel it loses intensity and thus | will form a trapezoidal aperture. This design does facilitate a slightly better release but good reductions and pad dimensions can eliminate the need for laser cut stencils entirely. We build product with 16 mil pitch using chem etch stencils with a very high yield. Your problem may not be entirely related to the printing process but may be hidden in your substrate. What types of defects are you encountering? This is important. What is your underside wiping frequency? Do you qualify this frequency or is it some number determined by the ever popular "engineering experience." If you can get your hands on proceedings from NEPCON East or West from '97 do so. Read my paper in each on "Improvising the Fine Pitch Screen Printing Process." There is a ton of good information in them and they should prove helpful. Remember, whenever possible, benchmark success. If it's not possible make your own success through TQM, then SPC, then DoE. It's a methodology that works. I've done it three different companies with the same result. Try it out. Post more info. | Good Luck and Best Regards, | Justin Medernach | Mfg. Eng. | Flextronics Int'l Justin, I've been screen printing boards for a long time myself and have never come across this problem before. I use TX-309 wipes after each print. The problem is: I can get one board to paste very close but never perfect. The print will be dead on,on one side of the board and slightly off, on the other. No matter what I try I cannot print this board and another similar board perfectly across the whole board.I get various shorts through out the board.Is it possible that the boards vary slightly in size from one to the other? Justin how can I get a copy of the proceedings you spoke of? What exactly do you mean by the "engineering experience?"

reply »

Ron Costa

#16619

Re: Screen printing the board from........ | 5 March, 1998

| Hi Ron, | Your question really opens up a "Pandora's box", but when you say you're not able to print perfectly, what is it exactly that you mean? Is the print lacking volume? Is it mis-registered? Are you bridging? Also, what printers have you used? Do they have vision? Or is the board being located with tooling pins? Let us know some more specific details about the problems you're experiencing, and the types of printers you have used, and I think I can be more helpful with some suggestions that may point you to a fix. In my experience, it is pretty rare nowdays to have a board so out of spec from the cad data that you've been provided to generate the stencil, that you can't get a decent print...that's provided the stencil has been made correctly within certain guidelines that are necessary to get good paste release from the apertures. I'm at NEPCON now, participating in the TAC line. We're building a double-sided PCMCIA board that has a 15.7mil 256-pin TQFP, 20-mil TSOP's, 256-ball BGA, Micro BGA's, a Flip-chip, 0402's, 0603's, 0805's, SOT-23's, and some 50mil SOIC's. The board is turning out pretty good (considering we only had 3-days to set the line up and run product for the show!) This is a flip-line too, and it's been quite a challenge to get all our machines to talk to each other and "tweek" things in so we don't look like idiots....(GRIN). This board is only only .020" thick, and is pretty flexible, so that in itself can introduce some variables as far as having a nice stable surface that is repeatable to print on, but we're doing it. One of my specialties is printing, so if you can provide a few more details about your situation, I'm sure I can help you... | -Steve Gregory- Steve, How's the show? I hope to make it next year. I have tried only HTI-9,and HTI-3's. These are the only screen printers we have The HTI-9 has vision that aids in setting up but,does not read fiducial marks automatically.I have not tried any other vendors equipment. I do build other boards from a different customer that are similar in complexity and size and don't have any trouble with them. I build double sided 20mil pitch boards on a regular basis with minimal defects.Today,I built a new board with 15mil pitch.I used a laser cut stencil that worked great. The first board out of the machine was perfect not one short.I was very happy with the outcome of the stencil,for we always used chem etched and chem traps on the aps. I wipe under after every print.It's just these two particular boards. Thanks for replying and have some fun while your there! Ron

reply »

Ron Costa

#16617

Re: Screen printing the board from........ | 6 March, 1998

| | | Hello everyone! | | | Does any know anything about bare board size variations? | | | Is there a spec. or tolerance? | | | I'm running small lots of boards and during the screen printing process | | | I find that I cannot paste each board perfectly.I've tried other screen printers as well as a new stencil generated | | | from the bare board itself. This board is about 13" x 15" and has a variety of SMT components | | | such as 50mil down to 20mil QFP's and TSOP's. I'm using a chem etched stencil with trapezoidal aps. | | | Could it be my stencil? Someone has recommended a laser cut stencil. | | | What is the difference, accuracy wise between Chem. etched and laser cut? | | | Would appreciate any feedback. | | | Thanks, | | | Ron | | Ron, | | Steve hit it all right on the head. I too am a self-proclaimed printing wiz. There are several aspects to be concerned about. Steve pretty much nailed them all but I'm going to get into some aspects of stencil design and manufacture. There are basically two types of ways to manufacture a stencil. There are subtractive and additive processes. Subtractive processes are laser cut and chem etch. EFAB is an additive process. The subtractive processes are pretty much standard. Both will give you roughly the same positional accuracy although the laser cut stencil may be a hair more accurate depending on the fabricator's chemical etching and photo imaging processes. The big difference between the two will be in the release characteristics. Laser cut stencils do tend to release better than chemically etched stencils. It is difficult to control the trapezoidal etch with chemicals. It is a natural part of the laser cutting process. As the laser makes its' way through the steel it loses intensity and thus | | will form a trapezoidal aperture. This design does facilitate a slightly better release but good reductions and pad dimensions can eliminate the need for laser cut stencils entirely. We build product with 16 mil pitch using chem etch stencils with a very high yield. Your problem may not be entirely related to the printing process but may be hidden in your substrate. What types of defects are you encountering? This is important. What is your underside wiping frequency? Do you qualify this frequency or is it some number determined by the ever popular "engineering experience." If you can get your hands on proceedings from NEPCON East or West from '97 do so. Read my paper in each on "Improvising the Fine Pitch Screen Printing Process." There is a ton of good information in them and they should prove helpful. Remember, whenever possible, benchmark success. If it's not possible make your own success through TQM, then SPC, then DoE. It's a methodology that works. I've done it three different companies with the same result. Try it out. Post more info. | | Good Luck and Best Regards, | | Justin Medernach | | Mfg. Eng. | | Flextronics Int'l | | Justin, | I've been screen printing boards for a long time myself and have never come across this problem before. | I use TX-309 wipes after each print. The problem is: I can get one board to paste very close but never | perfect. The print will be dead on,on one side of the board and slightly off, on the other. No matter what I try | I cannot print this board and another similar board perfectly across the whole board.I get various shorts through out the | board.Is it possible that the boards vary slightly in size from one to the other? Justin how can I get a copy of the proceedings you spoke of? | What exactly do you mean by the "engineering experience?" Ron, By "engineering experience" I'm referring to something we engineers tend to do too frequently. We accept something that we really don't understand yet. This is a problem I have specifically. I tend to focus on a solution before I have thoroughly qualified it. I was poking fun at myself more than anything. There is a certain amount of stretch associated with the printed circuit fab process. This is called "true Position". there is a tolerance for true position but I think it's somewhere around plus or minus 4 mils. One thing you can try is to send the boards out to a lab, if you don't have an accurate coordinate measuring machine and have the true position of component locations on the fab compared to their cad coordinates. Then you will be able to ascertain any component migration. The board size you referred to is rather large but not out of the ordinary by any means. I had experience with a board like this when I worked at General DataComm. It was a processing card 10 x 14. It had 2 rows of 8 QFP208s going across the 14 inch axis. Some would always print dead on and some would be off. We compensated for this by going to a more expensive higher Tg FR4 material. I really don't like the solution now as I've been in the contract industry for a year and a half now and have processed several similar fabs. I believe the problem lies in your PWB suppliers process. You may be able to get the position data of the boards through your stencil manufacturer. Lean on him and let him know your problem. I suspect you have as you had the stencil made from a fab. Have him measure the device coordinates on the board. He will verify for you. Your fab supplier probably doesn't have the right combination of prepreg, Cu, and epoxy resin for the process. Hope I could be of help. If you want to discuss, give me a call and give your fax number. I'll send you my paper. It's not the greatest but it's got some good concepts to build from. My number here is 978 392 3200. Best Regards, Justin Medernach Mfg. Eng. Flextronics Int'l PIC East

reply »

SMT in-printer dispensing

Software for SMT