MANUFACTURING SOFTWARE

Single user is free.

Buggy, check the forums. Problems with Inventory fields and importing data. Particulaly Manufacturer data.

Visual Jobshop is a little better, but no indented BOMs.

Cheers

As a side comment. We once had an MRP dumped on us by the new MD. Consumed 12 months and many man hours in installation and subsequent maintenance. It never-ever gave us anything in return. (could never even be made to understand what a part reel of wire was) These things tend to be liked by bean counters as they can offer total accountability for parts, materials, labour and costs. Of good benefit to companies making (say) mechanical products that use a set number of nuts, bolts and washers, are carved in Granite and sold without change for years. MRP can fall over when faced with a typical 'chaotic' electronics company and it's continual design changes due to obsolete and unavailable parts, design modifications, improvements etc. Following 2 MD's then brought with them ideas of 'cell' manufacturing and then Kanban. I'd had enough and left.

Yep. Keep it simple. I use plain databases here (Access) and in a larger company we used Agile which worked nicely. But that's more for huge enterprises although John has a ton of different products they make so something that large could fit.

Kanban may not be all that bad. Introduced something similar together with a production manager. I have never understood the concept of "kitting". It just doesn't make sense to me.

It does, when there is a base model with several, or even dozens of variations. The base model is considered a component as far as production planning is concerned. It is very useful for small runs and quick turn job shops. More than once we had an emergency order from an important customer, and could pull a complete radio together in under two weeks, instead of the six week, or longer order cycle. Boards that were already in production were diverted to the emergency order, and the ones made for that radio would replace them. We generally shipped everything at least two weeks early, so the delay was never seen by the customers.

At one point we had almost 100 base chassis in the production stockroom, so all we needed was the IF and tuner modules, plus any options, and that customer's EPROM for the front panel. that way, 90% of the system assembly work was done before it hit final assembly. A month or so later, most of them were gone, and we were back to a few in stock. that extra stock let us claim an early delivery bonus from a large contract. If the design wasn't kitted, we would have never made it, and the bonus was $500,000.

In your type of design it wouldn't be very useful.

Many of my designs are similar. Big ultrasound machines with lots of options to pick from. With or without Doppler, ColorFlow, various disk drive sizes, filter modules for different transducers and so on. All our large systems are custom-configured, there is no such thing as standard models like you have with cars on dealer lots. Still the Kanban-style pull system was a lot better than kitting. Less inventory, less shop space need, plus lots of saving in labor since nobody was kitting anymore. So we left it to the floor manager to determine the number of starts. They were also allowed to pull whatever they thought they needed from stock, whenever they needed it. Of course this require a decent MRP system to stay on top of the stock forecasting.

There was initially some resistance because kitting seems to be the usual scenario in the US. But when leadtimes did not get worse and cost went down big time people began to like it. Special order leadtimes actually shrunk because system production never had to wait anymore for a kitting session to complete. They could start building right when the order bell rung. Yes, we did have a big old ship's bell that was rung whenever Sales called in a firm order. Could be heard clear across the parking lot ;-)

Some contracts were for delivery over a number of years, to replace older equipment. The kitting allowed the required parts to be reserved, so that all the units were identical. We had several hundred variations, plus some customer specific modifications. The MRP package tied all of the departments together seamlessly. We ahd over 200 people at that plant, with over 100 different jobs running at once, and it worked for us. It replaced a cumbersome and mistake prone card file system.

A board or module level was considered as 'Kitted'. The only time we actually kitted a board order for an outside assembly of VME SMD boards. We sent out $80,000 worth of parts and got back ten very poorly made boards that needed $10,000 worth of rework to salvage. The board house claimed they used PickNPlace, but different boards had wrong parts in different places, some of the ICs were installed with the wrong orientation, and the paste solder they used looked like it was ten years old. Lots of scaly joints with slag and balls all over the place. I had to reflow thousands of bad joints on each board, but I managed to salvage all but one, because of damage they did to the PC board. It was framed and hung in the ME office to show to anyone who suggested they try it in the future. I wanted the name of the company that screwed it up to be put on a brass plate, but they were too chicken to do it. All I know is that the crappy work was done by a board house was in Orlando, Florida.

The bell was over the head of production's desk, but they only rang it for million dollar orders, or when they had bad news. You KNEW that when they offered a 'free lunch', something had hit the fan.

For older system support we had service inventory. It was physically in the same stockroom as the rest but Service could enter hold qties. Of course they had to watch that inventory was at a reasonable level just like all the other departments had to.

My clients pretty much all do it the same way. No kitting. They send the whole set of parts over, boards get stuffed and the balance of parts comes back with the stuffed boards. Last round was at WD Burch here in CA and they did a great job again. Plugged the board (new design) in and it worked.

No free lunches out here when something hit the fan ;-)

Were the folks up here in Grass Valley (Norcomm) able to help you with your SMD assembly?

Jim

Not this time. I had sent the files to Norcomm as well but they had to turn the job down so we went with the company in southern Cal. It was a fairly large job, 10 proto boards with over 500 components each. I've tried to keep parts variety to a minimum by using lots of same value jelly-bean components but it's still a lot of parts.

However, I keep Norcomm on file for the next design. It is always good to rely on personal testimony like yours when selecting service providers, so thanks again for the hint.

I thought kitting meant just that -- sending in all parts and getting the assembled boards back? Or did I miss a pun?

robert

Like Lloyds of London:

Best regards, Spehro Pefhany

In the US it usually means kitting with the exact number of parts needed. I don't think that is efficient. So, in a production environment I prefer handing the folks there some authority to pull what they think they'll need. In the case of the ultrasound systems that allowed them to keep spares in case something broke off, or pull less than required for the whole month so more shop space remained available. One requirement in medical is training in stock keeping because they could also return stuff that wasn't needed back to the stock room.

With the circuit boards we just sent the whole stock of parts. Labor cost for kitting: Zilch. Time delay due to kitting: Zilch.

That is OK if you don't have several hundered, to a thousand different jobs in process at one time.

...

Well, they got the bell, but it doesn't say if they got the gold.

Hmmmm....

Cheers! Rich

In those cases we still did it but the stock remained at the contract assembler. Later the contract assembler gradually took over purchasing, further reducing our overhead costs.

Almost all of our work was in house, and there was a lot of custom aluminum chassis and covers. Every module was shielded, and all wiring fed through feedthru capacitors to reduce noise and RF leakage as much as possible. It just didn't make sense to farm out numerous small runs that were similar, and when it was tried, it was a disaster. A lot of mistakes, and late deliveries meant a lot of in house overtime to correct the problems, not to mention replacing the wrong or damaged parts. As an engineer to order job shop, things were very different than a normal electronics manufacturing situation.

Yes, that's a different scenario. Yet even there a pull system can work well. If the guy in production thinks he'll need 50 feedthroughs during the day plus some other stuff he could just go and get them. Provided there is an ironclad thoroughness in checking the stuff out in the computer. Otherwise Kanban can land you in really hot water. I had the impression that companies sometimes clung to the old concept of kitting because only very few people would be trusted with materials management. That's one of the things I changed when I was "da boss".

We had three people in the stockrooms pulling all parts for each job. The parts and paperwork were put into an antistatic tray and moved in the computer from planning to production. If each assembler pulled their own parts there would have been 75 people in each other's way. This system prevented a job being started that was missing parts. It also allowed the head of production to walk up to someone and hand them a critical order that had a higher priority than what they were already working on. This usually only happened when a work order had parts on backorder, and the parts had just cleared incoming inspection. Common parts were in the main stockroom, while bulk parts and sheet metal were stored across the hall in the secondary stockroom.

The assemblers had to be certified for each item they could build. When they finished a job it was moved in the computer to the next level, and they picked up their next job from a list for their department, based on required lead times for shipping and what they were certified to build. Some items were a single level, while others were multiple levels. Single level would be the wire room or cable line, but PC boards were mounted into their cases and the harness attached. Then it went to the module line for testing and calibration or alignment. It was a complex MRP system, but management, the head of production, and the workers could track the location of everything in the company.

The test department was allowed to keep a small stock of parts for select in test and production repairs. These were exchanged when your stock was depleted. Customer support and out of warranty repairs were considered a separate division, and had their own inventory. It was moved from California to Florida, which turned out to be a wise choice. The old man who ran it died about six months later, and it would have been a nightmare to shut it down and move it if he had still been out west.

No problem. We had a chat about that and the individual managers agreed to keep foot traffic lean by urging their folks not to go all at once. Typically only 2-3 people from each department would actually show up in the stock room. I had a talk with the stock room clerks to see how they liked the new pulling scheme. "A lot", they said, citing mostly the absence of huge carts during a major kitting session.

That's exactly one thing we wanted to become able to do. If a set of circuit breakers was missing because it was on back order they could now begin to assemble the system anyhow. No more waiting because a kit was incomplete, meaning we had less of that typical quarter-end ship crunch. Which also meant less overtime. Which meant less cost and a happy CFO.

We could even do that after someone had pulled such a critical item because the MRP knew where it went. A quick decision among the managers and a part was pulled back from area 1 and brought to area 2, and the MRP was updated accordingly.

That is sad. Happened to the manager of our tool and die shop as well, he suffered a major stroke and did not recover. I still miss him.