Issue link: https://iconnect007.uberflip.com/i/1293772
OCTOBER 2020 I SMT007 MAGAZINE 53 keep it on the line for you. They may have to take it off, reprogram it, and pay a new setup fee. Now, you have these parts that have been loaded and unloaded, and you've lost whatever parts were on the lead tape that was loaded, so maybe you don't have enough spare parts anymore. We also have customers who want to do their own kitting. That's when we either end up with mounds of individual lit- tle chunks of cut tape or an exact number. We beg them and beg them to let us do kitting because that will screw them in assembly. Johnson: For those who see themselves in what you just described—when they do finally let you do the kitting—you then do what? Kolar: A good rule of thumb we use is if something is going through a pick-and-place machine—unless it's a really expensive part— we want to make sure there's a six-inch lead on the tape; how many components that is will vary by part. If it's an 0201, that might be about a hundred parts; if it's an 0402, maybe 70 is enough. You need to plan for a mechani- cal six-inch lead to go through and know that you're going to lose those parts. If they're really expensive parts, you can get away with fewer. We like to ideally do at least 3% over. If parts get really expensive or are hard to acquire, they may only get one over, or they may get none over. We make sure the assembler is aware in advance of those rare parts. People like to think that reusing kits saves a lot of money, but you spend more time digging back through those old kits. It's worth it for expensive parts. It's not worth it for passives. The other thing I've seen people not think about when they send kits to assemblers is when you're building multiple boards at once and they share the same components, it's an issue if you send them on a single reel. That reel is trapped to one build at a time. If they have multiple assembly lines, unless they cut off from it as cut tape, you're trapping them. We do a lot of projects where we build subtle vari- ants of boards where some things are populated, and some aren't, based on the availability of parts or based on experiments. It's tempting when you're ordering cut tape to say this board needs 20, and this needs 20, so I need 40, and let me maybe order 60 on one tape. Again, that's tempting to do if you expect the assembler to deal with it, and you send them one kit for two boards. Unless they're going to run them one after the other, you have an issue. If you want to run them at the same time, they have to go through and split all those up, and you need to realize they're not going to have enough excess for all of the parts. That's typically where we're paying atten- tion strategically to how they will run. Even if there's a lot of shared parts, we're kitting them separately. We're ordering them independently and kitting them as completely separate kits. It makes our assembler's lives a lot easier, and it gets things through a lot more quickly. We have to balance the cost benefit vs. the poten- tial time delay of sharing parts. Johnson: We just keep bumping up against the theme that designers have much more deci- sion-making power than they realize. Kolar: Yes, and they need to be willing to push back on the customer. It is also important to take the time to learn what you don't know. When we start new designers, if they haven't been in the industry or they have an EE, they don't know anything about manufacturing. We don't expect them to. They have to learn that somewhere. Warren: Case in point: We had a designer who had never done board layout before, and he placed 0402 resistors within 0.4 mm of an RJ45 Dan Warren