MANUFACTURING SOFTWARE

Joerg wrote:

That never really happened. Parts were pulled as the work orders came out of the stockroom printer, with the worst lag being one working day. Top level work orders had the days required to complete each job, and the MRP system generated each days run of lower level work orders. The work orders were printed by stock number, so one person could fill all the similar orders at once. On the rare occasion they got behind, one of the people in incoming inspection would help for a couple hours. All the carts were lined up in the order they were needed, and only a couple hours work at a time. Then someone from planning would take a couple empty carts and trade them for ones loaded with work orders that were ready for Assembly. We had about 200 carts and 25 "Mules" which would clamp to a 19" rackmount chassis and could be rotated 180 degrees to work anywhere on the four sides. the only time I remember seeing a logjam of carts was the end of one quarter, and I turned in 14 carts full of tested PC boards in a very long evening of working overtime by myself. I cherry picked the boards, and completed 75 boards that evening. It let us claim a half million dollar early delivery bonus that was offered at the last minute. It took QC two days to process all my work, and they didn't want to talk to me for months. They were VERY upset that they couldn't even get to their desks because the isles were blocked with a dozen carts that were there when they left, plus the 14 I added that night. i was told I could take the next day off if I wanted to, but I couldn't resist seeing their faces when they came in the next morning. :)

Most of our work involved building PC boards, and assembling them into working modules. Our lead time was six months, so an out of stock part was rare. When we had problems, we would qualify a ne part, only on the condition that the vendor could supply what we needed from existing stock. The circuit breaker you mentioned would have been no problem, because it would have been a plug in component, but it would still have to be in the chassis before the unit could go to final test and calibration. You would have to see one of these telemetry systems. Modular tuners, for the customers requirement, up to 12 IF bandwidths internal, and 12 more were an option in an external chassis. 12 different video bandwidths, and seven different time constants for the system AGC. The video combiner was able to track to .01 dB. Unfortunately, cameras were not allowed inside the plant. I have some older product I picked up surplus, so I'll try to get some pictures. I'm not sure if I'm well enough to drag it off the shelf right now.

... It

The only advantage I saw was that anybody who lasted through the 30 day probation had already learned how everything worked. I'm not arguing with you, just point out that both systems CAN work well, when used properly.

Chuck didn't wake up one morning. The other tech went to his house to find his van in the driveway. When he didn't answer the door for 10 minutes he called for help. The landlord unlocked the place for the rescue squad, but he'd been dead for hours. He was a cantankerous old fart, but I know what you mean.

One day can be way too long in medical electronics. With the new pulling system they'd take a few chassis weldments, transformers and miscellanea and a few of the guys would start assembling ten minutes after the order was there.

Well, that's how it looked at our place before switching from kitting to pulling. Almost every quarter. The problem was that OT had to be paid time-and-a-half so labor costs could really rocket up. I can't remember even one serious quarter-end crunch after the switch.

Our leadtimes were between a few hours and a couple weeks, roughly. Before it was one month fixed because that's when kitting time was.

They can. But with kitting I've only found it to work well if you knew the stock room and kitting folks well enough so they'd let you have some of those transformers before finishing the kit. The occasional box of chocolates did wonders. Pretty much like the relationship between engineering and marketing which in my case resulted in marriage ;-)

It's no comfort now but at least he did not have to die a slow death in a nursing home.

I've done that with the common parts. Often they'll have them in stock waiting. I usually had to design around availability though. These I bought before the design was complete to insure I had a buildable design. Sometimes I had a better supply (could apply a firmer twist to the arm) than the contract manufacturer. I never had more than a couple of designs and variations in the mill and quantities were small.

On small qties that works ok. Somehow my consulting projects split into the extremes. Not much in the middle. Some are parts of big machines where only a few hundred a made per year. The other stuff is real mass products where every penny needs to be turned around which actually is a lot of fun during the design for me.

Well... Mine have been at two extremes too. Either a prototype or test fixture, where only a few are made, or production, where I had nothing to do with the manufacturing end (until something broke).

Except for making sure that the design is optimized for production or in MBA speak "DFM" my clients handle production. But also only until something happens. Like that late night request from China after a part went out of datasheet spec range and they needed another source right now because they had a line stop situation.