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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


SMT EQUIPMENT FOR LOW VOLUME HIGH MIX

HERMAN

#24302

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 25 April, 2003

Hello everyone, could you advise me how to select smt EQUIPMENT for low volume high mix??. Before we run high volume production but now we are changing to low to medium volume, we�re using fuji CP6 and universal GSM, and righ our best time for a change over is 30 minutes average, depending on the product we can last up to 1 or 2 hours. I�ll appreciate your help

reply »

Stefan

#24309

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 26 April, 2003

HERMAN, I think, you are doing very well with 30 minutes change over time. Your stuff got used to the machines, setting up feeders, program preparation and conveyor changes. The Universal GSM is as flexible as Mydata or Siemens for that matter, but you�ll have to get used to their different feeder styles and programming methods. Some of your people may not be very receptive to new machines, in particular, when they got very good in handling the existing machines. You would have to tell me more reasons than flexibility to convince me, that you require different machines. P.S. Writing in caps locked letters will get at least Dave F.'s attention. I bet, he'll comment on that.

reply »

#24322

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

It always amazes me when a high volume manufacturer has to change to LVHM environment. They seem so lost as to how it should be done. It really is two different worlds and there is more to it than equipment selection. But to answer your question... Changeover time is the most critical. Thirty minutes is unacceptable when jobs are only running 1 hour or so. That equates to 66% machine up time. Your goal should be SMED (single minute exchange die). This can only be done with feeder carts that can be prekitted off line completely. This also means an investment in extra feeders. From a philosophy point of view you may need to make some changes as well. Kits must be complete and delivered to the floor well in advance. Someone needs to be responsible for the kitting of the next job. It is crucial that this is done completely before the previous job ends. Common components between jobs need to be easily identified if duplicate reels are not in the next kit. Lights on the feeders can help identify these parts. There is too much to write about on this subject in this forum as the details go on and on for this application. Get a demo from some of the companies that specialize in machinery for this environment and see which solution best matches your current philosophy. Check out the C7 series from Contact Systems (www.contactsystems.com) as we developed this platform for this need.

reply »

Larry

#24323

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

Herman, I'm betting you will get a lot of response on this one. Frankly, everybody has their favorite machine. The trick is to consider the types and quantitys of boards you will run. Sometimes, it is better to set up a load for several jobs that have a high quantity of common parts. You should look for software that allows you to combine boards for a single load, interchangable feeder carts or magazines, Quick programming, and an intellegent feeder system. Take a look at the Mydata, Asemblion, Juki, and Samsung machines. There are others in this class as well. I would advise you give the board data for several of your top running boards to the manufacturers and ask them to give you the running speeds and changover times. Then verify this data when you go to look at the machines. The critical thing is that you go and run the machines at the manufacturers sites- you woldn't buy a car without driving it and the same is true of a Pick and Place machine.

reply »


JAX

#24326

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

Herman, Before I get into this I would like to know something. What machine takes the longest to change over? That should be equal to your change over time. If it is not one of your pick and place machines this is all beside the point. I am not sure you need to look at new equipment specifically designed around change over speed. What good does it do to go out and buy Mydata or Contact Systems or "pick your favorite" when the change over time you save is only lost in run time. If you cannot build boards just as fast as you can now with these new machines, the changeover time is irrelevant. You will probably do better working on how to reduce the change over time on your current equipment. This is achieved in many ways, most listed by previous postings. Common Set-Up's, Pre-Loaded feeders, etc... Remember, you do not have to have the entire line changed over before you start running. You can have someone changing over while the machines are running. Does the thirty minutes involve certain tasks assigned to insure quality?.... Set-Up verification, etc.. Maybe you can do these tasks ahead of time, offline, to speed up the process.

reply »

HERMAN

#24327

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

Thanks for your responses, as gregp says, one hour in changeover is not good when you run only 10 panels, we alredy do from 10 panels to 2500 boards per run, depends on the product, we already have a pre-kitted area mounted in their respective feeders, but, the material it�s been counted, then they cut the reels with the exact amount of material needed to run this order plus 2% of attrition, but if the feeder was bad or the cp6 fail, this 2% is not enough and we need to get more material, in this time we lost at least 15 minutes per material, sometimes we last up to 3 hours to close that order + 1 hour in the change over is a total of 4 hours. I just have worked with fuji and panasonic, and only in the shows I�ve saw others like siemmes, and my guess is that those equipments that work at lower speed, should need less attrition, so I can close faster the order, thats what I think, Am I right??

Regards Herman

reply »

HERMAN

#24328

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

Hi Jax, I�ve took times, and average, we last from 25 to 40 minutes to load a CP6 depending on how big is the load, but quality guys last other 25 min. per equipment to audit, we usually have 2 CP6 and 1 GSM or 1 CP6 and 2 GSM, at any case, we need to stop the line to do either to load or audit, and I know some equipment you just insert the feeder cart and you can audit the loaded material without stopping the equipment. We audit every shift and every change over. Other point is when you need to change the feeder because is empty; in those equipment you don�t need to stop to add the new reel into the feeder, and if you read the reply I did to gregp, we have problems to close the orders because of the attrition, so all of you are right, we need to get better numbers regarding time, more efficiency in prekitting area, and calculate better the attrition, also we are implementing a cascade changeover team, that will be responsible to accelerate the changeover process, but still with this, in some analisys I have done, I believe we can improve up to 35% productivity if we change to other kind of eqiupment specialized for this specific enviroment. thanks for your advise Regards Herman

reply »

#24329

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

Possibly. It sounds like attrition is a contributor to your problem. Are you recieving the material with the tapes cut from the customer (consignment) or is this something you do in the kitting process? If you are doing it, how do you maintain proper leaders and trailers on the cut sections of tape?

In general, a steady production flow is better. While 40,000 cph with 50% uptime (CP6/GSM) is equivalent to two high speed flex machines (say 25,000 cph) at 80% uptime, the latter is preferable for many reasons. 1. There is less likelihood of machine starvation. Because the first line runs at a higher cph, there is a greater chance that a single operator will not keep up with printed boards or component replenishment. 2. If a machine is starved (or down for any reason) you are losing production at a faster rate. 3. An even production flow allows the operators to be better prpared for the next job because of reasons 1 and 2. These are just a few off the top of my head. Remember the old fable the Tortoise and the Hare. Not only does the tortoise beat the hare in the race but he's not even breathing heavy at the finish line.

reply »

HERMAN

#24331

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 28 April, 2003

HI STEFAN, THIS IS WHAT YOU MEAN WHEN YOU SAID WRITTING WITH CAPS LOCKED??, OK THANKS FOR YOR ADVISE, YOU ARE RIGHT, THE PEOPLE ALWAYS WILL REJECT TO DO CHANGES, SPECIALLY AS YOU SAID WHEN THEY ARE USED TO WORK WELL WITH THIS MACHINES, RIGHT NOW I HAVE LOTS OF PROBLEMS BECAUSE OF THIS, SPECIALLY WITH DIRECT LABOR, BUT IN THIS TIME, IF WE DON�T IMPROVE OUR PROUCTIVITY, THIS INDUSTRY WILL MIGRATE TO ASIA, WHERE THE COST LABOR IS CHEAPER, SO THERE ARE MORE REASON DOWN THIS PAGE, TAKE A LOOK IN THE OTHER REPLYS AND TELL ME YOUR COMMENTS. THANKS AGAIN

reply »

#24349

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 29 April, 2003

My opinion to hold on to your machines was based on the current market situation. First of all, I don�t believe that a customer with ten boards will go to China for prototype production. Truth is, that the remaining market here is the prototyping market, which got highly competitive. I don�t know of any machine, which is particular made for prototype runs. Almost all feeders require a leader tape, the Mydata Aegilis feeder doesn�t, but still you loose a few components passing the pick up window and you have to select the right peeler blade for the right tape, which may be a bit challenging in the inventory. A possible solution for small runs could be, that the short tapes are treated like a matrix tray. You stick the tapes onto a plate, a peeler removes the cover tape and the components are picked one by one. The next obstacle is the vision system. If the board fiducials require teaching, you would require an off-line fiducial camera-vision system. The same applies to the components. The machine has to �learn� the feeder pick position, preferable the camera should check each component before pick up (newer Sanyo�s). However, the pick position identification could only use a backlight vision. If components, like BGA�s require front light, these components should be test picked before you run the first board and placed back into the feeder. Feeder change over tables are common to most of the pick & place machines. Split feeder tables on chip shooters are on the up rise. Although time is saved with the change over, the components were not test picked from the new feeder table. Feeders and tables should have some sort of intelligence ( barcode ) to know, if the right component sits in the right feeder location ( Mydata, Siemens and other ). Most of the machine manufacturers concentrated their market on the high volume production. When even contract houses rejected jobs under $ 100 k volume, there did not seem to be the demand for a do-it �all prototype machine. Improvements were made to reduce the change over time and demonstration of change over times within seconds were made at Sony for example, but these programs were tested and fine tuned in a prototype lab first. If you can�t wait that long for a new machine to be developed, I would buy a used machine, from a manufacturer you are already used to. You may want to designate the machine just for the purpose to do the tricky stuff, like test odd shaped components. Another machine may be set up with bulk feeders (if your customer permits upside down resistors) and you could dump your short strip of taped components into it. You may also want to look into the possibility to purchase an old machine with mechanical board and component alignment. These machines do at least down to 0805 without vision. If you designate this machine only for a certain purpose and don�t try anything else with it, it will save you the time on your high volume machines. If after all you can convince QA that you are faster in correcting set up errors upon the first board coming off the machine, than auditing the set up, you may gain more productive time.

reply »

#24355

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 30 April, 2003

Herman, LVHM is exactly what we have always done with our SMT machine. We use the Siemens 80F4 auto pick and place and have a HUGE set-up time for each run. And attrition problems (though i'm finding they are packaging related rather than machine related). Either way, if you want quick set up times, and quick change-overs, don't go with the Siemens 80F4! Chrissie

reply »

Ozzy

#24452

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 8 May, 2003

Chrissieneale,

Have you invested in additional feeders and feeder tables to prepare for the next job offline while running the current job?

If your changeovers take as long as you suggest then this should be a quick ROI.

reply »

kenBliss

#24456

SMT EQUIPMENT FOR LOW VOLUME HIGH MIX | 9 May, 2003

Hi HERMAN

I did not see any financial opinions so I thought I would offer one. The goal of the company is to make money, how you achieve that has many variables. Buying new equipment when the existing will do the job costs big money, provided the existing equipment is not a bottleneck. The bottleneck is simply what ever is the constraint that controls the overall output of the line or the plant. Based on what you stated it appears that you believe you have the wrong machine because of the long changeover times. but I think you have more that 50%of your capacity is done time. You Pick and Place machine should be up placing parts at least 80% of the time

I would do several things. 1. determine if you could toss all the current rules at your factory about setting up and changing over a line.

2. Prestage all the feeders needed for the next job with full reels, stop trying to send only the exact amount of parts onto the floor. You are not in the �control inventory business� you are in the manufacturing business. You will use what every you need to use to run the job and return the rest to stock. Stock telling you the maximum you will use is futile and is killing your bottom line. You can continue to work towards minimizing attrition but first get the job done efficiently.

3. When a job is almost done, have a programmer standing by the machine, a person to change the stencil screen, a feeder audit person if needed and a person to load raw boards.

4. The current job running stops.

5. Everyone does there job at the same time to complete all the changeover in a few minutes, feeders should take no more that 20 seconds each. The feeder auditor should be reviewing the work as it is happening, meaning the feeders that have been installed should be checked as soon as they are on the machine do not wait until all of them are on the machine.

6. about 15-20 minutes you should be up and running

7. Any problems everyone should be standing there to tweak anything.

8. Prestage feeders that will be needed during a reload, so reloads only take about 20 seconds each.

As my company specializes in this specific area I also suggest you goto our website. Check out Bliss University at http://www.blissindustries.com it explains in more detail how to use feeder, reel, and stencil carts to make this entire process easy.

I would be pleased to walk you through the entire process over the phone if you like, give me call

Ken Bliss Bliss Industries 510-490-8401

This message was posted via the Electronics Forum @

reply »

ICT Total SMT line Provider

Reflow Oven