Loading leadless parts(passives)by hand

We are a small division. We are starting up a leadless line with an old screen printer and a convection oven. At this time we don't intend to get any sort of pick and place assistance and wil be loading by hand. We find the biggest problem with this(besides speed) is misplaced parts. Does anyone know of a set up where as the operator loads parts a read out can be obtained to determine values of passive components. Has anyone ever taken this approach. I am trying to convince management this isn't the way to go and we should invest in pick and place semi automatic loading cabapiblity. There are a few nice ones out there but even they don't guarentee the parts going down are the correct value. Any Ideas?

reply »

| We are a small division. We are starting up a leadless line with an old screen printer and a convection oven. At this time we don't intend to get any sort of pick and place assistance and wil be loading by hand. We find the biggest problem with this(besides speed) is misplaced parts. Does anyone know of a set up where as the operator loads parts a read out can be obtained to determine values of passive components. Has anyone ever taken this approach. I am trying to convince management this isn't the way to go and we should invest in pick and place semi automatic loading cabapiblity. There are a few nice ones out there but even they don't guarentee the parts going down are the correct value. Any Ideas? | Clarissa: It sounds like your division is taking a very thoughtful approach to developing SMT capabilities. There�s a school of thought that espouses "getting the basics right, before moving on to more complex things." In other words, if you can�t figure-out how to get one component in the correct spot using manual methods, how in the world are you going to get it correct using automated methods? Sounds boring!!!

So anyhow, you�re slow at misplacing parts. Let�s accept the "slow part" and focus on getting better, because often when you get good at doing something, you get faster also. When you say misplaced parts, do you mean that you:

1 Can�t find the parts? 2 Put the correct part in the correct location, but aren�t accurate? 3 Put the wrong part in the wrong location, but put it there accurately? 4 Something else?

Probably "all of the above." So after you split your problem into these four groups, you can figure-out approaches to solving each element of your problem. Sounds boring. OK, let�s push on to problem #3. We need to make money at building lots of short-run and prototype boards. Gotta stop this cursed analyzing stuff. Assuming you purchase the correct parts and get them into properly labeled bins, here are some hand work tips for you:

1 Provide proper documentation, that makes it painfully easy for any assembler to know what�s required. 2 Train your assemblers before they go to the line. 3 If assemblers are not suited for this work, find something else for them to do or only assign them a limited number of tasks. 4 Keep bins close. 5 Use 5AR or 7SA tweezers. 6 Work only one quadrant of the board at time, build a bunch of boards, and then move onto the next quadrant or have a bunch of people focusing on assembling their particular quadrant of the board.

You say: "Folks that sell pick and place machines don't guarantee the parts going down are the correct value." Say they would guarantee the parts going down are the correct value, would you hold it against them if placed parts were the incorrect value, because:

1 You put the wrong part on a feeder or pick-up location? 2 Some monkey mixed-up the orientation of some of the SOICs in a tube? 3 Your supplier mislabeled the parts you received? 4 And on and on ...

"Any ideas?" I think you�re asking your supplier to solve a complex task that you are best suited (and paid, I assume) to solve.

TTYL Dave F

reply »

Hi,

For hand placement you should see about 150-200 ppm of errors (solder, placement, values etc) on a board populated of types such as 1206, 0805, SOT, Tantalum etc. . As for solely wrong components (component at wrong location) you can usualy reduce this by having a line process with as many stations as there are different types of components. Hand placement versus wrong component should only ave. 20-30 ppm . If you do go with a machine and get set-ups inspected before start and during reloading the overall ppm should drop to 10 to 20 ppm for the same board. As for wrong values they should drop to zero!

I hope this helps!

Mario Lapointe Eng. Tech. Dir. SMT-ASSY

reply »

| Hi, | | For hand placement you should see about 150-200 ppm of errors (solder, placement, values etc) on a board populated of types such as 1206, 0805, SOT, Tantalum etc. . As for solely wrong components (component at wrong location) you can usualy reduce this by having a line process with as many stations as there are different types of components. Hand placement versus wrong component should only ave. 20-30 ppm . If you do go with a machine and get set-ups inspected before start and during reloading the overall ppm should drop to 10 to 20 ppm for the same board. As for wrong values they should drop to zero! | | I hope this helps! | | Mario Lapointe Eng. | Tech. Dir. | SMT-ASSY | Mario: Interesting analysis. What does "Hand placement versus wrong component should only ave. 20-30 ppm." mean? Is that a subset of errors? Thanks. Dave F

reply »

| Mario: Interesting analysis. What does "Hand placement versus wrong component should only ave. 20-30 ppm." mean? Is that a subset of errors? Thanks. Dave F |

What I mean Dave is that if you have a good (well actualy a great set-up) with all the good stuff that you mentioned in your reply it is fair to shoot for a 20-30ppm of errors due to wrong components eventhough it is a hand placement process. If presently he/she sees more than this then perhaps it is an indication of needed set-up changes, such as less parts type per employee, color code, alternate component size, color etc. I know it sounds low (20PPM) but I have seen it done with a board with the mix that I earlier descripbed.

Mario

reply »

Clarissa,

Dave had good suggestions. We load several boards manual at this time. We set our components up in clearly marked bins. The documentation includes a chart with the part number, bin number, reference designator and color code for the part. We then provide a drawing showing the position of the part, color coded to the chart, and the reference designator. This has work well for us for accurate placement of the part. I have did all the board drawings on Paintbrush. It is rather tedious, but you work with what you have. Besides it started out to only be one or two boards.

We place the parts using two different methods. One is using a Fritsch semi automated pick and place. The other is using a vacuum pen. One of the assemblers said if she just had to choose one, she would pick the vacuum pen. I know it beats tweezers and the cost is not that high. Ours is on an EFD dispenser which also dispenses paste when needed for touch ups. Nancy | | We are a small division. We are starting up a leadless line with an old screen printer and a convection oven. At this time we don't intend to get any sort of pick and place assistance and wil be loading by hand. We find the biggest problem with this(besides speed) is misplaced parts. Does anyone know of a set up where as the operator loads parts a read out can be obtained to determine values of passive components. Has anyone ever taken this approach. I am trying to convince management this isn't the way to go and we should invest in pick and place semi automatic loading cabapiblity. There are a few nice ones out there but even they don't guarentee the parts going down are the correct value. Any Ideas? | | | Clarissa: It sounds like your division is taking a very thoughtful approach to developing SMT capabilities. There�s a school of thought that espouses "getting the basics right, before moving on to more complex things." In other words, if you can�t figure-out how to get one component in the correct spot using manual methods, how in the world are you going to get it correct using automated methods? Sounds boring!!! | | So anyhow, you�re slow at misplacing parts. Let�s accept the "slow part" and focus on getting better, because often when you get good at doing something, you get faster also. When you say misplaced parts, do you mean that you: | | 1 Can�t find the parts? | 2 Put the correct part in the correct location, but aren�t accurate? | 3 Put the wrong part in the wrong location, but put it there accurately? | 4 Something else? | | Probably "all of the above." So after you split your problem into these four groups, you can figure-out approaches to solving each element of your problem. Sounds boring. OK, let�s push on to problem #3. We need to make money at building lots of short-run and prototype boards. Gotta stop this cursed analyzing stuff. Assuming you purchase the correct parts and get them into properly labeled bins, here are some hand work tips for you: | | 1 Provide proper documentation, that makes it painfully easy for any assembler to know what�s required. | 2 Train your assemblers before they go to the line. | 3 If assemblers are not suited for this work, find something else for them to do or only assign them a limited number of tasks. | 4 Keep bins close. | 5 Use 5AR or 7SA tweezers. | 6 Work only one quadrant of the board at time, build a bunch of boards, and then move onto the next quadrant or have a bunch of people focusing on assembling their particular quadrant of the board. | | You say: "Folks that sell pick and place machines don't guarantee the parts going down are the correct value." Say they would guarantee the parts going down are the correct value, would you hold it against them if placed parts were the incorrect value, because: | | 1 You put the wrong part on a feeder or pick-up location? | 2 Some monkey mixed-up the orientation of some of the SOICs in a tube? | 3 Your supplier mislabeled the parts you received? | 4 And on and on ... | | "Any ideas?" I think you�re asking your supplier to solve a complex task that you are best suited (and paid, I assume) to solve. | | TTYL Dave F |

reply »

| Clarissa, | | Dave had good suggestions. We load several boards manual at this time. We set our components up in clearly marked bins. The documentation includes a chart with the part number, bin number, reference designator and color code for the part. We then provide a drawing showing the position of the part, color coded to the chart, and the reference designator. This has work well for us for accurate placement of the part. I have did all the board drawings on Paintbrush. It is rather tedious, but you work with what you have. Besides it started out to only be one or two boards. | | We place the parts using two different methods. One is using a Fritsch semi automated pick and place. The other is using a vacuum pen. One of the assemblers said if she just had to choose one, she would pick the vacuum pen. I know it beats tweezers and the cost is not that high. Ours is on an EFD dispenser which also dispenses paste when needed for touch ups. | Nancy

Nancy: Our ol' bud Steve Zeva says this:

"I use Lavenir Viewmate and Paintshop Pro...both you can download offa' the NET for free. I'll pull up the PCB silkscreen in Lavenir, change the colors so I have a white background and black lines, then hit the "Print Screen" key on my keyboard. That copies your screen to your clipboard. Then open up Paintshop Pro and do a "Paste-New Image" and you got your gerber silkscreen into something that's easy to edit and add things like jumper wires or any other rework instructions to...piece o'cake!"

Graphicode also makes a free gerber viewer.

TTYL

Dave F

reply »

| Clarissa, | | We place the parts using two different methods. One is using a Fritsch semi automated pick and place. The other is using a vacuum pen. One of the assemblers said if she just had to choose one, she would pick the vacuum pen. I know it beats tweezers and the cost is not that high. Ours is on an EFD dispenser which also dispenses paste when needed for touch ups. | Nancy

Nancy: We use vacuum pens also, but given a choise our operators would rather use tweezers all day long. As with any tool, tweezers need to be maintained. Check this:

"Caring for your Tweezers

By W. R. Smith, FBHI, FNAWCC, CMC, CMW

When you start loosing parts from your tweezers, it is time to check the condition of your tweezers! Having spent at least the first two years of my watchmaking life on the floor searching for lost parts, I feel that I qualify to comment on the problem. There are three major reason why parts snap from tweezers--the tweezers are too stiff, which limits the worker's ability to feel the part; the inside of the jaws are too slick from long use, and the jaws have been bent enough to form a V with the apex toward the worker when a part is picked up. Any one of these can cause a problem.

There are several tricks that will greatly lessen the number of parts lost by being snapped from tweezers. First, one must address culprit of tweezers stiffness, which prevents the user from feeling the part as well as is desired. In the 10-man watch repair shop where I worked as a young man, every watchmaker used Dumont 3-C tweezers and ground their legs in the region of the joint to weaken them so the parts could be felt better.

To do this, lay the flat portion of the tweezers leg near the joint on the face (the curved periphery) of a bench grinder stone and grind all the way past the joint to the very end. Do this carefully to keep the thickness of the blade equal on each side. When properly ground, the user will be able to feel the part being picked up much better and apply far less pressure.

Another trick of great help is to open the tweezers and slip a fine India stone inside them. While holding one jaw flat against the stone, stroke it to and fro to flatten the inner surface of the jaw and to remove the polish there that occurs from use.

Lastly, when tweezers become open at the end, their points can usually be bent back together without breaking them by pressing each against a piece of wood while held in a position that will close the point.

Hope this will help some users of tweezers.

W. R. Smith, FBHI, FNAWCC, CMC, CMW. Gateway Clocks 7936 Camberley Drive Powell, Tennessee 37849, U.S.A. Phone: 423-947-9671 WRSmith2@AOL.COM

Page address http://www.bhi.co.uk/hints/automsp.htm Last Updated 10 January 1999

� 1999 British Horological Institute. We welcome your comments and suggestions. Please contact us at: British Horological Institute. Upton Hall, Upton, Newark, Notts. UK. NG23 5TE Telephone (01636) 813795. Fax (01636) 812258. E-Mail info@bhi.co.uk

Reproduction of part or all of the contents of any of these pages is prohibited except: These pages may be downloaded onto a hard disk or printed for your personal use without alterations. This copyright notice must appear on each copy. These pages may not be included in any other work or publication, or be distributed or copied for any commercial purpose except as stated above.

reply »

| Clarissa, | | Dave had good suggestions. We load several boards manual at this time. We set our components up in clearly marked bins. The documentation includes a chart with the part number, bin number, reference designator and color code for the part. We then provide a drawing showing the position of the part, color coded to the chart, and the reference designator. This has work well for us for accurate placement of the part. I have did all the board drawings on Paintbrush. It is rather tedious, but you work with what you have. Besides it started out to only be one or two boards. | | Nancy

Nancy: OK So you have the bins with the labeling and the color code that matches-up with the marking and color code on the drawing for parts to be installed at each station. How do you deal with the inevitable and on-going changes to your set-up?

Thanks

Dave F

reply »

Dave,

We have had very few changes in the board. I just edit my drawing and chart.

Nancy | | Nancy: OK So you have the bins with the labeling and the color code that matches-up with the marking and color code on the drawing for parts to be installed at each station. How do you deal with the inevitable and on-going changes to your set-up? | | Thanks | | Dave F |

reply »

Dave,

Thanks for the information. If I ever have to do that again I will be prepared. We are looking to set up a SMT automated line in house so hopefully I won't need it.

Nancy | | Clarissa, | | | | Dave had good suggestions. We load several boards manual at this time. We set our components up in clearly marked bins. The documentation includes a chart with the part number, bin number, reference designator and color code for the part. We then provide a drawing showing the position of the part, color coded to the chart, and the reference designator. This has work well for us for accurate placement of the part. I have did all the board drawings on Paintbrush. It is rather tedious, but you work with what you have. Besides it started out to only be one or two boards. | | | | We place the parts using two different methods. One is using a Fritsch semi automated pick and place. The other is using a vacuum pen. One of the assemblers said if she just had to choose one, she would pick the vacuum pen. I know it beats tweezers and the cost is not that high. Ours is on an EFD dispenser which also dispenses paste when needed for touch ups. | | Nancy | | | Nancy: Our ol' bud Steve Zeva says this: | | "I use Lavenir Viewmate and Paintshop Pro...both you can download offa' the NET for free. I'll pull up the PCB silkscreen in Lavenir, change the colors so I have a white background and black lines, then hit the "Print Screen" key on my keyboard. That copies your screen to your clipboard. Then open up Paintshop Pro and do a "Paste-New Image" and you got your gerber silkscreen into something that's easy to edit and add things like jumper wires or any other rework instructions to...piece o'cake!" | | Graphicode also makes a free gerber viewer. | | TTYL | | Dave F |

reply »

Regarding a Fristch semi (using tube feeders on the left side and 200 bins on a rotary belt. From working with one about 3 years ago, we found that we had an average of 240 components per hour placed versus 320 without. The difference being that you are constantly waiting for the machine to follow the operator ex. the locating ring and open and close the bins. For proto and even small runs the set-up and programing was too much time versus just jumping into it and doing it manualy.

We use mostly Sipel SMD 10-sa for tweezers we find them very versatile. For tantalum we just spread them open a lillte more. But teewers are the best way to go when going manualy in my opinion.

Mario

reply »

Maybe it is all what you get use to. Our assemblers like the vacuum pen. Since we do more than passive components, some parts would be difficult to hold and line up. As for as our Fristch, we use it as a manual and just have our bins on the side. The most usuful part is the visual system.

How are tweezers ergonomically? Any complaints?

Nancy | Regarding a Fristch semi (using tube feeders on the left side and 200 bins on a rotary belt. From working with one about 3 years ago, we found that we had an average of 240 components per hour placed versus 320 without. The difference being that you are constantly waiting for the machine to follow the operator ex. the locating ring and open and close the bins. For proto and even small runs the set-up and programing was too much time versus just jumping into it and doing it manualy. | | We use mostly Sipel SMD 10-sa for tweezers we find them very versatile. For tantalum we just spread them open a lillte more. But teewers are the best way to go when going manualy in my opinion. | | Mario |

reply »

Yes Nancy when it comes to PLCC or anything bigger then Som, so16 we too use vacumm pens. The initial comment was regarding passive compoents.

As for the fristch semi : the vision system is the only thing that is still being used for: 20 mil QFPs 208 and 304

reply »