Antonio,
a) You shouldn't care how many machines they have on their line at all. Let's assume the number of parts to be placed on a pick&place is A, and the number of parts to be placed on a chip-shooter is B. They need to find the limiting machine, So, A/(p&p cph) will give them a time and the pick&place will be kept busy for that time. If this time is shorter than the screen printing time, (20-30 seconds at my times) then your screen printer is the limiting factor, you can only run 120 boards per hour, with yields and part changes, let's say 90 boards. I don't know what yield ratio they use... So, everything starts from the 90 boards per hour parameter. They know what 1 hour of line time costs to them.
If there was a case where B/cph >> A/cph, then you need a second chip shooter. Then each chip shooter will place B/2 parts and possibly catch up with the pick&place machine.
If the line has 2 chipshooters and a pick&place, it will be pretty hard to use the one unused chipshooter for another job because for a high speed production that machine has to be used as a conveyor, and if the production is too slow where the operators can walk the boards by that machine, then it shouldn't be called a "3 machine line" anyway.
If your CM is focused on the number of machines in their quote, I would look for another Cm, NOT another CM with a 2 machine line.
b) each machine has a "tact-time" for different types of components. In an ideal world, for example, a chipshooter will give a tact time of 0.18 seconds for a 1206 chip capacitor. Then the formula is adding up (tact-time for type z component)x(number of type z components) for all different types of components. Then you come up with a total time spent on a machine. Then you add feeder changes, i.e. each reel has 2500 parts, each board uses 25 parts, every 100 boards I will have a feeder replaced. if the CM puts 5 minute down time for a feeder change, then you'll question why they don't have a full feeder ready, or maybe that machine can be loaded in a way that once there's an empty feeder, it switches to the second feeder bank.
So, there are endless ways to bring the "actual run time" to the "ideal run time". CMs have their own priorities, some are high-volume/low-mix, they won't want to change a machine setup every 4 hours, once they start running a board, they run that board for days, that way they can lower the costs. Some others are low-volume/hi-mix, their offline feeder setup group will have more people than any other group.
You have to find the right CM focused on your type of work. Otherwise, anybody can run your boards but some CMs will simply say NO since it's out of their focus. (yes, even nowadays)
c) I believe they add up each machine's capacity. 2 chipshooters each 25kcph, one pick&place 5kcph, there's your 55kcph. But if all the components on your board are large QFPs or BGAs, you'd better get prepared for a 15kcph capacity, or in your case, just add the chipshooter and pick&place.
The cph values may be calculated just like the gas mileage on a car. With a skinny driver when the wind blows fromm behind on a smooth surface road, 60 miles per gallon. Not when you drive the car with a heavy foot, canoe on top, full trunk... Same idea, if the parts are placed as a matrix on a board where the board hardly moves during placement, since it's the same fastest placed component, the feeder carriage does not move either. I don't want to say car manufacturers or SMT machine makers are not telling the truth, they may be given best case scenario numbers, that's all...
I can help you further if you want off-shore manufacturing.
Regards Erhan finepitch-services@lycos.com
reply »