Table of Contents:
During the 1980’s, I worked in the machine shop of a factory that made showers, valves and industrial metering equipment. I had several jobs there over a period of 14 years, including Machine Setter, Production Engineer and Quality Control Inspector. I mostly worked in a machine shop, where metal components were made.
Following recent cyber hacking activity at a well known manufacturing company, suggestions are being mooted to return to the old style of manufacturing and stock control. I ask: is that a good idea?
There’s been murmurings in the news lately regarding the impact of cyber hacking on manufacturing companies, such as Jaguar Land Rover. There have been various comments bandied around slating the choices for cellular manufacturing, or as we used to call it “Just In Time” or JIT for short.
But what is it? And moreover, why am I writing drivel about it?
I was there
Well, not at Jaguar Land Rover (I wish!). But I did work in a factory that made the transition from a stock-orientated manufacturing process to a cellular manufacturing process. As previously mentioned, this was referred to as a “Just In Time” system. Let me explain that transition and what it meant to us in the factory.
We’ve worked like this for years!
Picture this: it’s early 1980’s and I’m working on the shop floor of a factory that made showers and other bathroom-related equipment. These were the big heavy chrome plated mixer valve affairs and we made all of them on site, complete from the raw metal stock to the finished product shipped out to retailers.
The factory floor dealt with making the shower bodies, the components that went inside them, the shower heads (the type that screwed to the wall in institutions) and various other parts for products that we no longer made. One of the company mantras was that we would manufacture spares for showers and equipment that were discontinued, for at least ten years from the date of discontinuation. Therefore there were a lot of different components, thousands of different types, shapes and sizes, ranging from little brass washers to large chrome-plated flow measurement devices.
Generally speaking, we did high volume runs. It did depend on the component, but we would manufacture 100’s if not 1000’s of the same component over a number of days. Those components would get processed through the factory floor until they were complete and passed inspection (some components had multiple operations performed on them) and then they would be whisked off to the warehouse to be stored until needed. And we wouldn’t have to make any more for a while. Sometimes, quite a while!
Stock. Lots of stock.
The stock warehouse was huge. There were thousands of parts in there, all of them indexed. There were old people in there that claimed they hadn’t seen the light of day for many years (that’s a joke, by the way), but they would find you a single component in an instant, mostly from memory.
How was the stock accrued?
The planning department
What we made and when we made it was controlled by the planning department. A group of people had access to the stock control system, which was based on an IBM AS/400 mainframe computer. They accessed the mainframe using “dumb terminals”, green (or orange, sometimes) terminals that were just a screen and a keyboard connected to the mainframe using coaxial wiring (VT420’s using a BNC token-ring network). The early 1980’s was pre-personal computer era, at least for businesses (and definitely for a shop floor) and so ordering and stock control was done using the dumb terminals and the bespoke software on the mainframe.
The planning department had an overview of the complete stock, demand and sales areas of the company. They would be responsible for scheduling foundry runs (we had our own foundry to cast and stamp components), manufacturing runs (all of the metal parts – and some plastic parts – were made in-house), plating runs (we chrome plated our own parts), and assembly runs throughout the whole manufacturing processes.
And we made them by the bucket-load (sometimes quite literally!). There were very few times that we had a batch run under 500 parts, and if we did, that was sent to the “Short Order Shop”.
The Short Order Shop
A brief mention of the Short Order Shop, as I worked in in that department for a few years as a machine setter/operator. We used to make parts for mainly obsolete products (in order to maintain that company philosophy of holding spares for ten years after the discontinuation of a product). The batch numbers weren’t huge, the most we ever did was around 200 parts. The machine operators for the Short Order Shop were considered to be slightly more skilled than the “standard” machine operator, and that was reflected in their pay.
Working in that department was one of the best times I’d ever had in manufacturing workplace.
The shop floor supervisors
Once the planning department had planned the manufacturing run schedules (i.e. you make a number of these components on this date), they sent the schedules to the machine shop floor supervisors. There were three supervisors working three shifts (day, twilight and night) who supervised the whole shop floor. They were responsible for allocating what work went where in the shop, in terms of which machines were available to make a component.
All of this information was hand written onto a tab index card, which was placed in a slotted vertical rack. The rack was organised by machine type (drills, lathes, mills etc.) and the tab cards had the component (and part number) on one, the setter name on another and the operator name on yet another one. There was also a status card to indicate that the machines were being set, or set ready to run, or broken. Sounds a bit complicated, but it worked well.
The shop floor supervisors were manged by the shop floor manager (of which there was only one). He was responsible for overseeing the whole shop floor operation, including quality, safety and ensuring components got to the next stage of their manufacturing journey on time. He was also responsible for all of the people that worked on the shop floor, the setters, operators, swarf cleaners etc.
I wrote very briefly about him here. Bless ‘im.
The machines
At the time of pre-cellular manufacturing, there were no CNC machines in the machine shop. Yes, they existed, but the company chose not to invest in any at the time.
So there were sections of (mostly manual) machines. Pillar drills, milling machines, lathes (both manual capstan and air operated peg-board), the beloved (by me, anyway) AMII and Didomats, bespoke machines (such as one solely built to drill and tap 1/2″ BSP holes in one particular shower body). There were a few “automatic” machines, in which the operator would clamp a component in a chuck, close a big steel door and a single cutting tool – driven by a cam and follower – would shape a profile. These machines were set up by a person (hence the term “setter”) to perform an operation on a component. Once the component had been passed by the Inspector, an operator ran each and every component through the machine until the batch was finished. The manufacturing runs would sometimes take days, over several shifts.
There were a number of purely mechanical bar-fed machines, that made components from round (or square) bar stock that was fed from the side of the machine. These machines used rotating cams to move cutting tools in and out of the workspace to perform a cut on a component. The cams were times so that once a tool had finished its cut, another tool would move in and make the final finishing cut. These machines made mostly spindles and shafts for shower bodies. They were very noisy (due to the rattling of the cams), they used the same cutting oil as the AMII (as in oil, not water-based cutting fluid) and they required constant supervision to ensure that the bar stock was being fed. There were no automatic bar feeders in those days, so a person was employed to replenish the bars, clear the piles of swarf (metal chips generated from cutting) and look after the machines in general. One person would look after 12 machines at a time.
The time and motion study people
Speaking of time, how did the planner know how long it would take to make a component? Because when you’re planning to make a finished product, you’ll need to know how long that’ll take.
There existed a team of people that were employed to time the casting, the manufacture, the polishing, the assembly of every single element of making a finished product. They would time how long it would take to perform an operation, do some calculations and then come up with a time for each component or element of manufacturing.
Therefore the planning department could calculate the amount of time it would take to make a batch run of components and therefore plan the machine activity for the whole year. They had a huge year planner on their office wall with that information on it, so you could see far in advance what the machine shop machine would be doing months down the line. Post-it notes on a chart. 👍
The setters
The machine setters set the machines up to make a particular component run. Most of the components had to have multiple operations on them, each one leading to another one – most likely on a different set of machines. The people that set the machines tended to stay setting a particular type. The chap that set the mills and drills did only that, the chap that set the peg-board lathes did only that – for the entire length of time they worked in that machine shop. There were very few people that could set multiple types of machines, although they did exist. I was one of them, but managed to get stuck on the big Diedesheims (mainly because no-one else wanted to set them!).
Each day, the setters would consult the tab card system in the machine shop office, which would tell them what component to set on which of their machines. If they had a set to do, that is. Due to the large amount of components made, once a machine was set, it could be days or even weeks before it had to be set for something else. If all of your machines were set, you did maintenance or adjustments. Or just looked busy 😉
The supervisor’s office also contained filing cabinets. Quite a few of them. These contained the setup packs for each operation on each component. Once allocated a set to setup, the setter would extract the setup pack from the filing cabinets (indexed by part and operation number). The setup pack had details of the tools, fixtures and gauges required for that components machining operation. The setter would consult the documentation, book the various tools, fixtures and gauges out from the machine shop stores and set the machine.
Once set, the setter submitted a sample to the local Quality Control Inspector, before the start of the manufacturing run. If the inspector passed the component, the setter would inform the machine shop supervisor that the machine was ready to go. A tab index card was then placed in the slot, indicating it was available to be operated.
Quality Control
I touched briefly on Quality Control in this post. The role of the Inspector at that time was to ensure that the components being manufactured met the standards required in a QC Form that related to that particular components operation.
In the days before cellular manufacturing, there was an inspector allocated to each machining area. There would be an inspector for the drills/mills section, one for the cam machines, one for the Diedesheims etc. Each inspector had an “Inspection Station” – a plywood desk mostly – that held specialist measuring equipment and a granite surface block.
Before any production started, the components had to have what was known as a “first-off”. Called so, as a setter would set a machine, machine a component and submit that component – the first one from the batch – to the inspector for checking, to make sure it was correct before production started. The setter provided the component, the drawing and any gauges that were required to check the components.
The QC Form dictated the quality checks that were to be made on the component for that operation. Gauges were used (for convenience, mainly) to check critical dimensions such as O Ring diameters, shaft diameters or thread sizes. The inspector had the use of what was called “The Standards Room” if needed: an air conditioned office that contained a more specialist array of measurement equipment, such as a projector (and later on a co-ordinate measurement machine).
Once the first-off was satisfactorily completed, the machine would be ready for production.
The operators
Again, I touched on the operator element briefly in this post. The operators were allocated by the machine shop supervisor – using that tab card system – to a machine. The operators would report to the supervisor’s office at the start of their shift, get allocated a job and then report to the setter responsible for that machine to start their work.
Operators worked via piecework: i.e. they were paid a basic wage, but worked to produce a certain number of components, dictated by the operation. If they achieved the amount of components set out by the paperwork (e.g. 25 per hour), they would be paid the full operators wage. If they didn’t, then they forfeited some of their wage. It was therefore in the operator’s interest to achieve – and maintain that number.
The number was calculated by the time & motion people. The operator could work out how many components they had to make over the course of their shift. The focus was on quantity, rather than quality, shall we say.
Quality was where the inspectors came in. They would regularly check the components that an operator was producing – usually visiting each machine at least once an hour. There was a checklist on the QC Form which had to be checked and initialled by the inspector. Completing the checklist usually meant that the components were good to go.
And sometimes, they weren’t good to go
The operators at that time were focussed on getting parts in and out of those machines in order to get paid, and that machine could be any one over the whole machine shop. An operator would have an easy-ish job one day drilling one hole in a component once a minute, to having to suit up in oil-proof clothing and stand in front of one of my beloved AMII machines (which would have been very warm and very fast-paced), producing 680 elbows per hour – all individually placed by that operator.
The operators weren’t obliged to check any components for quality. The setters might offer advice about things to look out for, but on a fast-paced high production machine, the operator didn’t have time for that!
It was mainly down to the inspectors on their patrols to pick up any faults or issues. Having said that, if a tool broke inside a machine, it would normally go with a bang and the operator would stop (usually. Some didn’t!) and go and get the setter. Nine time out of ten, there would be tool wear and a dimension would creep out of tolerance. Not noticeable to the naked eye, the error would be picked up by the inspector on their patrol and the operation stopped. The setter would be called for and an assessment made of the amount of time it would take to correct it. If it was a short time, the operator would hang around and then resume a few minutes later. If it was to be a longer adjustment, then the operator would be sent back to the supervisor’s office to get another job.
It was the inspector’s task to ascertain whether the out-of-tolerance components would be OK to use. If not, then the inspector had to dig out the defective ones (usually with the help of the setter) and scrap them. It was a matter of pride to the setter whether they had any scrap. Scrap meant a crap set. And the shame that came with it! 😁
Then everything changed.
Turning Japanese
The transition to cellular manufacturing as a whole was not an overnight task. However, there had to come a point in a manufacturing process when you stopped making many thousands of the same components to making hundreds (or tens, sometimes).
What is Cellular Manufacturing?
Wikipedia tells all about cellular manufacturing, but this is what it meant to us based on the machine shop floor.
The machine shop as a whole would be split into five “cells”. Each cell would have a Cell Leader and a dedicated group of setters and operators, who would set, manufacture and QC check a range of components. The Cell Leader would plan the cell’s manufacturing timetable just for their range of components, based on a pre-planned demand.
The setters would set the machines required for the jobs and the operators would run the machinery. The batch sizes were considerably smaller: no longer would you have a batch run that would last days or weeks – these would run hours, before being broken down and reset for something else.
We called it “The Japanese Idea”, but in reality, it was American, made popular by the Japanese in the late 1970’s.
Impacts
Several things had to happen before the machine shop could even engage in cellular manufacturing.
Training!
There were a lot of new things, for a lot of people!
- Cell Leaders had to be employed. A series of interviews and whatnot went on for those posts, resulting in five new appointments. The Cell Leaders were to eventually replace the machine shop supervisors and the machine shop planners.
- There was a new corporate operating system for planning, ordering and stock. The old dumb terminals would eventually be replaced by PC’s that would run Windows 3.11, be networked using ethernet cables and would run the new corporate software (SAP software) along with the newly released Microsoft Office.
- New PC’s needed to be installed in the machine shop. Rather than put them out in the dirt, swarf and coolant-soaked desks on the shop floor, new offices were built to house desks, tables and computers for the Cell Leaders.
- New training programmes had to be devised and delivered to Cell Leaders, setters and some operators. Accessing any information they required on the new SAP system, involved not just knowing how to use SAP, but how to use those new PC’s.
- Some of the setters that got put into a new cell didn’t have the necessary experience required to set all of the machines. That meant a bit of cross training between setters, cross being the operative word in some cases.
- There was some investment in new machines for the machine shop. CNC machines were becoming popular at that time, so several were bought and installed. The earlier purchases were based around the bar stock machines, as these now featured auto bar loading, along with the CNC driven machine tooling. There eventually appeared CNC lathes and milling machines, replacing some of the aging equipment that had been in that shop for many years.
Generally speaking, the setters that set certain types of machines (e.g. the Diedesheims, the peg-board capstan lathes) were allocated to cells that encompassed those machines. With the advent of the CNC machinery, new people were recruited to set them and a program (ha!) of retraining for some existing setters to be able to set the CNC machines ensued.
The operators were allocated to the cells by the machine shop manager. Some cells got more operators than others, due to the amount of operations that had to be performed on a component, but they were allocated and near enough stayed at those numbers – at least for the duration I was there.
Full steam ahead
Once all of the above elements were in place, cells began to establish themselves in the machine shop. It was a different way of working, as setters and operators alike focussed only on a certain range of components. Operations that would have been split into two, or even three separate operations could now be merged into one: the operator would machine a component on three different machines placed in close proximity.
The operators were still on piecework. The older times that were allocated to some components no longer applied, so the new way of working had to be reassessed by the time & motion department. It has to be said that we had to “train” some operators when they were being timed. If they went too fast (or we had set the machine to operate quickly) then the timing could be reduced, meaning they would have to produce more to achieve their piecework payment. If they went to slowly however, the time & motion guy cottoned on quickly and cried foul!! We had to “advise” the operators beforehand to not go too fast, but not too slowly either. It was a bit uphill with some 😒
It took some time for some Cell Leaders – especially those who had not had experience on the shop floor – to get to grips with batch sizes and planning. The CNC machines (and some of the peg-board machines) were perfectly fine with short batch runs, as they could fairly quickly be “turned around” – meaning once a batch had finished, stripping the tools and fixtures out and resetting it with different tools and fixtures for a new component. Some machines – like my beloved Diedesheims for example – took quite a long time to set. They were high volume machines, designed to do batch runs of thousands, or tens of thousands. When you had to make 100 elbows, it took eight hours to set the machine and twenty minutes to run the batch. We had to make some adjustments there, but it sorted itself out over time.
Gone was the (sacred) tab card index! Mind you, it went along with the machine shop supervisors: both index and supervisors surplus to requirements now that the Cell Leaders ran the cells. All three of the supervisor retired, none wanted to become a Cell Leader.
Some setters became very busy! Those that worked on the mills and drills (for example) became less so, as the machines were physically moved to the locations of the cells that needed them. Most setters were able to set up a drill, or a mill in those days.
It took a little while (and a not inconsiderable amount of grumpiness on most people’s part), but things settled down. The operators got to know the components, the setters got to know the operators.
Quality Control
One of the major things that cellular manufacturing changed was the responsibility for quality. Once the realm of the inspector, who was responsible for the components being made on his patrol – this was now the responsibility of the Cell Leader and ultimately: the operator. Extra time was factored in by the time & motion people to include periodic checking of the components (by the operator) with the supplied gauges. In some cases, extra gauges were made for the component, to include dimensions that were previously measured with measuring equipment (by the inspector).
It was not the demise of the inspector, however. First-offs still had to be produced – and approved – by an inspector and there were less frequent patrols by the inspectors. The QC Forms – once the sole domain of the inspector – now became part of the operator’s function. The operator would fill in the check boxes and initial it to signify compliance. There were more than a few “non-compliances” caught in the first few months, resulting in disciplinary actions, thus reducing the incentive for lying about the quality of your bits.
And the “shame” of producing scrap still existed.
Some pros, some cons
So what did the switch over to cellular manufacturing actually achieve?
The mass manufacture of components had its advantages: there was a lot of stock kept, so you had a bigger buffer of available bits. You didn’t have to remanufacture the same components quite so often, so the turnaround times for the machine shop machines was less. It was less wear of the machine tools as well, as they weren’t being put in and out of machines frequently. They tended to last longer if kept cool, lubricated and a reasonable cutting feed. Swapping them in and out of tool holders caused wear on the shanks, leading to vibrations and tool movement (i.e. breakage).
But mass manufacturing meant keeping a lot of stock = warehouse space = (effectively) dead real estate. There’s a lot of money tied up in stock, so reducing the amount of stock reduces the amount of money you’ve had to invest and the amount of space that it inhabits.
Cellular manufacturing meant you only made the components – and more crucially – the number of components that were required in the short term. Whereas in mass manufacturing, you might make a few thousand components that would never sell: the warehouse might end up with crates of components that were just gathering dust. Cellular manufacturing’s Just-In-Time (JIT) philosophy reduced that risk to next to nothing, as you made what was required. The cost saving was huge, working on JIT alone.
Cellular manufacturing provided a bit of a break (from the operator’s point of view) from the monotony of doing the same thing every day, day-in, day-out. Gone were the weeks long batch runs, in were the shorter runs that meant an operator might do several different jobs in one shift.
On the flip side of that (for both the operators and the setters), it meant that you made whatever your cell was planned to make: which was a finite range of components. Each cell had their own range of components that they made, so you could end up doing the same job week after week, but for small batch runs.
The plus side of that was that you could train all of your operators to manufacture that finite range of components. Each operator would get to know each component: it’s machining quirks, the quality “gotchas” to look out for. The operator was given an element of responsibiity and empowerment, which paid off in the fact that some operators helped the setters clean the machines and set up for the next batch run.
What did I think?
I’m glad you asked. I worked as a setter, an inspector and as a production engineer for a few years, seeing the transition from mass manufacture to cellular manufacture.
As a setter, it was harder work. Given that I was obliged to work on the bigger Diedesheims, the amount of time between setting it (and getting the first-off), to having to break it down to set the next job reduced dramatically. In the mass manufacturing world, the time between sets would be days, weeks. Cellular manufacturing sometimes meant hours between sets. I found that as long as we knew what was on the horizon, we could plan in advance to get the tooling and paperwork ready in good time. That meant relying on the Cell Leader to plan accordingly. Personally, I didn’t mind it. It kept me “gainfully employed”.
As an inspector, it was less work. Apart from doing the first-off, the responsibility was on the operator to do “in-flight” checks. The inspector was there as a backup and to do the occasional inspection round. There were less inspectors (we lost some due to retirement etc. that didn’t get replaced), and we had bigger rounds to do, but it was less work and less responsibility.
As a production engineer, I thought it great fun (some people did not!). The production engineer’s task was to design and make (or have made) tooling for machines to produce components for new products. It was our job to do the first sets and troubleshoot any problems. We’d then have to do test runs with operators and finally sign the job over to production. I found that to be the busiest job of all, but that’s mainly down to the nature of the task: introducing new tooling, paperwork and quality methodologies. It would be exactly the same whether it was cellular or mass manufacturing.
In the news
As I mentioned at the beginning of this never-ending drivel: following recent cyber hacking activity at a well known manufacturing company, suggestions are being mooted to return to the old style of manufacturing and stock control. Is that a good idea?
Would it have “saved” Jaguar Land-Rover?
Saved: meaning instead of the huge financial loss caused by the lack of availability to make or get parts to build cars (and therefore sell them for money).
In the short-term, carried stock would have mean that JLR could have continued to make (and therefore sell) cars.
There are of course caveats: all of JLR suppliers would have to not be affected by any cyber hacking and be able to supply JLR with parts.
A point in time would arrive – be it sooner or later – when JLR would need to manufacture more parts in order to sustain production, meaning their computer and manufacturing systems would need to be up and running in order to make parts. However long that time period is would solely depend upon how good their disaster recovery plans are and how quickly they could recover their manufacturing systems.
Given that JLR’s amassed parts stocks would be dwindling, the method by which production would start would probably depend on what they were running low on, and replacing that stock first.
Conclusion
In my uninformed opinion: would a return to mass manufacturing have softened the impact of JLR’s cyber hack? Yes, I think it would have, to an extent, anyway.
The real question would be whether JLR would want to invest a considerable amount of money and real estate into making and holding a lot of stock on the off chance that their production processes were disrupted in some way (like a cyber hack).
I’ve no idea of the amount of money they may have lost as a result of lost production. But would it be more than the cost of holding stock for years? 🤷
TL;DR
Jaguar Land Rover’s computer systems were cyber attacked stopping production. I (rhetorically) ask: if they’d held some more stock, would things have been better for them?
I also launched into a long, long ramble about my experiences with mass and cellular manufacturing. You can skip that bit, if you like.
