Soo eS SC —( ss es ss oe ze i i Se eee eS SS rd ee SY a i ( rwOOCO~—C—OC Sy ie es 6 ent —— ee a thie oy, ee Pe rf = ee oS wT Ce. ss ee 2 | as .. .. ee ED nx we ee as tel a... E ee SSS ee SC SSS ee ae eee Pe CS r—“i—™—~—~——~—~—«<] ><‘The light-truck product family made the greatest revenne contribution to Apex and was under the heaviest price pressure. Apex appointed a Value Stream Manager for this product family, who drew a current state value stream map. This produet fa hipped to the State Street assembly plant in three variants: a short-hose assembly (S) for che short wheelbase truck, long-hose (L) for the long wheelbase model, and an alternative-fuel (ethanol) assembly (A) offered as an option on this vehicle, PRODUCTION CONTROL, e A fomns ciate aoe Lead Time = 'Odays 2days 2days 159 seconds ipa i Te 9 = 18sec. PART i: GETTING STARTEDApex started its fuel line value stream improvement at the right place: the ‘pacemaker’ process. The pacemaker involves production steps that are dedicated to @ particular family of products and responds to orders from external customers. A well run pacemaker sends smooth demand signals upstream to the pull loops of the remaining batch fabrication processes, which respond to requirements from internal customers. Apex managers and engineers made another good decision by minimizing their initial investment and keeping the cell simple. For example, they could have created @ more extended continuous flow by incorporating an end-forming press into the fuel line cell. Bur such a press would have requited substantial capital investment. (In the future Apex may decide to apply some of the cost savings from its improvements to purchase and add a press to the cell.) They decided to install a simple, inexpensive, flexible operator-based cell designed for State Street Assembly's needs. This is more likely to be highly reliable and well-suited to sending smooth signals up the value stream. ‘Apex managers chose a classic U-shaped layout for their new operator-based cell, as illustrated ‘on the next page. In only a few days they were able to move machines and configure the new cell to achieve striking reductions at this process in lead time and floor space required, awhile dramatically increasing the number of pieces produced per production associate. Original Current State ‘State Apex’s continu Progress with Continuous | No No Continuous Flow jonstable Production per Shift 62: (ectual/target) ‘690 Space 580 (sq. feet) Assembly) tidays | 37min (oad Timo (WIP x Taki) Number of Operators 6 7 Productivity | 13.5 20 (pieces/associate/n”) Functions Effectively as No No PacomakerA Closer Look — With Eyes for Flow Apex managers, engineers, and production associates were excited about their new fuel fine cell. After all, they quickly increased productivity by 50% while halving space requirements and dramatically slashing lead times. Yet when you look at the Apex cell with “eyes for flow” you should actually be disappointed. A walk through the fuel line cell will show why. Apex's fuel line cell — current state Saccembes cates, \\ ‘Tube Bender | tener (automatic) A | (automatic) Hourly Production Pian Asa ® A 1] 90 | 6 2| 2 | A ia 3| 2 | 20 = packout (romratars) 20 peter 4| 90 | o (oleh soa) 8 | 2 | ~ 6| 2 | 7| 1% | a | so | 2% 90 622 PART: GETTING STARTED 7‘The first step when we vi ta facility is typically to go see the current situation with ‘our own eyes and ask, “What is the problem?” At Apex, the first thing we notice is the production output chart at the entrance/exit of the cell showing planned and actual production. Looking at the output figures we wonder, “Why is there so much variation, and why does total production fall short of planned production?” More specifically, “Why is the cell achieving only two-thirds of planned ouput during many hours of the shift?” Is the problem incapable machines thar make had parts? Is it machines that won't run? Is a supplier shipping bad parts, or are parts missing? ‘And who reacts when these problems occur? Whatever the cause, the variation in output is clear evidence that cell performance can be greatly improved. We are even more certain of this when we note that in ¢wo hours out of eight the cell actually produced more than the plan, which is just as bad as being under the target. Four production associates were assigned to the cell the entire shift, so a change in staffing can’t explain the variations. Unless this was achieved by burrying, unacceptably cisking stress injuries and bed quality, there must be waste in the process. We begin to see a source of variation and waste when we closely examine the first two steps in the production sequence: the tube bender and the first assembly operation, "The first production associate has to leave her regular work area every 25 pieces, or about every 16 minutes if the cell is producing to takt time (as explained below in Question 2). This requires three minutes and means that either the material flow stops or the tube bender and the first two assembly steps are all decoupled from one another. This means no continuous flow. [As we continue 0 walk sround we notice that there are: various quantities of inventory between every operation and that the production associates are each anchored to their machine, which often means they have to wait while the machines © Variable inventory buffers between workstations are an inefficient way to balance uneven workloads. When a buffer gets too full, he supplying operation often takes an unofficial break — pethaps to get materials or do other out-of-cycle work — while the downstream station catches up. Operations are decoupled, allowing each to produce batches instead of one piece at a time.Decoupled operations, which we call ‘islands’, bake the waste of overproduction and the waste of waiting into @ cell, causing them to be repeated many times every shift, day, week, month, and year Tiny wastes often don’t seem significant to managers just visiting the process (and apparently are not visible to Apex managers), but think about them as they add up more than 600 times per shift! Decoupled operations also make it difficult to notice production problems as they happen. When a problem occurs the rest of the stations keep on working, By the end of a shift che unnoticed problems add up and the production volume falls shore of the target: Pacemaker processes, in particular, need to be manageable. Problems or abnormalities need to be spotted as they occur and support personnel must respond to them quickly. Production associates cannot react to and fix significant production problems, find and climinate the causes of those problems, and at the same time still achieve full production! inally, as we complete our tour we note that the Apex cell is aid out in a very wide “U*. This defeats one of the main objectives of a U-shaped cell layout: Permitting flexible deployment of operators by moving work arcas into elose proximity. Both the first and last production associates are moving back and forth over considerable distances to handle materials. Flow stops every time they leave a station to backtrack. Our conclusion, at the end of our walk through Apex’s cell, is that there is actually no continuous flow anywhere, Instead we see only cetratic and intermittent flow — as indicated by the small piles of inventory between each machine and the fluctuating output from hour to hour. Indeed, this cell is really just a ‘module’ of adjacent machines and operators producing at best ‘fake flow’ that the untrained eye. THREE FLOWS As you look at your own cells aceon Pore, Peau Pee) Pes Pred Cone ey eee) eed SC cee Reese abnormalities? 2. Does the material flow? Preece Pe eee Perce Pees Pere ey RU Cy ere Eenr) eer) rend TREE een performing one value- Pere Rcd eee10 Targets for Apex’s Fuel Line Cell Origi Current State State Continuous Flow No No Production cog per Shift = (actual/target) Space (sq. fest) Assembly Lead Time (WIP x Taki) Number of Operators Productivity (pieces/associate/hr) Funetions Effectively as Pacemaker While Apex’s new cell performance is much better than the original process village layout, « careful effort to achieve true continuous flow through proper process design and operation can double labor productivity, halve the needed space, reduce lead time by a further 90%, and dramatically improve both quality and responsiveness to customer requirements. Realistic targets for this cell, which we will show you how to achieve in che pages ahead, are shown in the right hand column of the table above. We'll get started by posing the first of eleven questions you should go through as you strive to develop true continuous flow in your own cells and lines. The questions require eareful work and attention by your entire team, but you will discover that the answers are invaluable once they are incorporated in your business.Question 1: Do You Have the Right End Items? imilies and assigned three end items to its fuel line cell. However, as you consider your own situation, you may have to think carefully about the right products to assign to your pacemaker process. Here are some guidelines we've found helpful. Apex has already determined their product 1) Flexibility. Sometimes demand is high enough to allow you to dedicate individual products to their own cells or lines like thi Product A Product B Ea | ‘~V “ 7 However, if demand gyrates between produets and you can keep changeover times short, you are often better off sharing products between mixed-model cells like this Products AaB Products A&B “ 7 “ 7 ‘The total capacity is che same in both cases but the ability of each process to accommodate shifts in demand between the two products is much greater in the second case. The demand for one product within family may vary, while the demand for a whole product family is often more stable. PARTE: GETTING STARTED 1112 2) Variation in Total Work Content. ‘The total work content — that is the operator time required co process one piece from start to finish — should not vary by more than about 30% between the different end items processed in the cell, especially when a moving conveyor is used. When the work content varies too much it becomes difficult to maintain flow and productivity. In such cases you may want to split the cell or assign some rare or low-volume end items to other cells. (Some facilities even create a separate line of cell 70 handle low-volume end items, until product engineers can reduce the content differences bevween the items via design changes.) 3) Similarity of Processing Steps and Equipment. When che steps required to build different products within the cell vary too much (ixe., when some products skip some processing steps) operators will have to “shift gears” every time they change to assembling 2 variant of the product. This reduces productivity and increases the chance of quality problems. Again, sometimes it is better to produce variants with markedly different processing steps in different cells. 4)Takt Time (Production Pace). units, Itis determined by dividing the cotal available production time per shift by the customer demand rate per shift (see the equation at right). As a general guideline, when takt time for a cell falls below ten seconds the operators’ jobs may become highly repetitive and stressful. When high demand calls for very short takt times you should consider using multiple footprints of the cell, possibly side-by-side, instead of a single high-speed cell. ‘This is particularly appropriate ifthe capital requirements of additional cells can be kept low through utilization of simple equipment. time is the rate at which customers require finished Conversely, when takt time slows to more than about 120 seconds, the number of work: elements sometimes gets so high that work motions can be difficult to standardize, In such cases consider adding additional but similar end items to the cell to bring down the take time. Of course, with some products it will simply be impossible to set cake times below 120 seconds because volume requirements are inherently low, even when several different end items ate run through the same cell or line, (With long take times it can get difficult ro have all pars at the line for the operators forthe different produce variations. Sometimes you have to increase the parts delivery frequency or deliver certain parts in the assembly sequence.) 5) Customer Location. When customers for a product are widely dispersed geographically, it may make sense to split up the work into multiple lines, each located near a different customer. This makes sense particularly when shipping costs and duties for finished units are high, when there are potential exchange-rate losses, when lead times for components are long, or when local infrastructure (supervision, buildings, etc.) is available at reasonable cost.Question 2: What is the Takt Time? Having decided what products to produce in the pacemaker, the next task for Apex managers was to determine che takt time. ("Takt’ is a German word for a pace or beat, often likened to a conductor's baton.) Tak time is a reference number that is used to help match the rate of production in a pacemaker process to the rate of sales. takt time Used to help synchronize pace of production with the pace of sales your available work time per shift customer demand per shift 27,600 seconds = 40 seconds 690 pieces The customer is buying this product at a rate of one every 40 seconds. Sales are usually calculated on a daily or weekly basis but most pacemaker processes are actually up and running only some fraction of each day or week. Since the point of take time is to pace actual production, the most sensible thing to do is to divide the number of products demanded daily or weekly into the number of shifts operated in that time period to determine demand per production shift. For example, the customer demand for Apex’s light eruck fuel lines is currently 6900 units per week and Apex operates its fuel line cell ten equal shifts per week. Thus the demand per shift is 690 units. ‘Once demand per shift is known the final step in the calculation of take time is to divide this number into the ‘effective working time’ per shift. This is start-to-stop shift time minus any scheduled operator breaks, meetings, cleanups, ete. Because takt time must represent the actual customer demand rate do riot subtract time for unplanned machine downtime, changeovers, or other internal problems. 1314 Apex operates two 8-hour shifts Monday through Friday, 6:00 AM to 2:30 PM, and 3:30 PM co Midnight. There are two 10-minute breaks each shift but no scheduled downtime for maintenance. This means Apex has 27,600 seconds of effective working time in each shift. 480 min, (8 hours) - 20 min. of breaks = 460 min. x 60 sec./min. = 27,600 seconds By dividing 690 units into 27,600 available seconds we determine the take time: 40 seconds. 27600 seconds i = 40 seconds per unit 690 units ‘This is the rate of customer demand, the all important ‘beat’ of the market. Notice that take time is expressed in ‘seconds-per-unit’ because it is easier for everyone to understand and use than decimals of minutes, Similarly, we use ‘scconds-per-unit’ rather than ‘pieces- perhour’ co describe actual production rates, or ‘cycle time’. Comparing take time and cycle time is the easiest way to answer the simple but critical questions: “How frequently does the customer need one piece?” and “How frequently do we actually make one piece at our pacemaker process?” There is one additional point that may be very important in your own takt time calculations, the amount of variation in eustomer orders. In Apex’s case the 6900 unit per week demand was relatively easy to determine because Apex is supplying @ massive automotive assembly plant whose own cakt time does not change frequently, Bue whet if long-term average demand and day-to-day actual demand are different? ‘We suggest that you check the range of daily customer demand variation by reviewing actual shipments (not orders) over the past twelve months. Your cell must be able to handie sustained demand, For occasional spikes in demand it is generally better to operate at a steady takt time (based on average long-term demand) and either hold a buffer stock of finished goods or run some daily overtime to ensure your ability to serve the customer, Changing takt time from day to day is inefficient, distupts the work pace, and increases the potential for quality problems. Lastly, regarding future demand for new products, it can be difficult co make accurate forecasts far in advance. When future demand is uncertain it may be wiser to add capacity in steps, as increased demand actually materializes, rather than designing your pacemaker now fora peak demand that may not appear.Cycle time is how frequently a finished unit actually comes off the end of your pacemaker cell. We often find processes that are operated at cycle times faster chan takt time, For ‘example, if you are running your facility three full shifts (perhaps to achieve high machine utilization) you will probably always need cycle times slightly below takt time because there is never any time available to catch up if your equipment or materials system fails. And to some degree these sorts of problems will always occur in manufacturing! However, keep in mind that when you chronically cycle much faster than take time you increase the chances of overproducing and may be using extra operators. (As the diagram ‘below shows.) Much worse, you conceal your production problems and reduce the incentive to find and eliminate their causes. It is important to maintain a certain tautness at the pacemaker to ensure that problems get noticed quickly and receive fest response by support staff. Cyeling much faster than takt may require more people take time eyele time 2 extra operator Note: The inevitability of problems in manufacturing is one of the reasons why many production facilities in the Toyota group of companies run their pacemaker processes for two shifts with a one to four hour gap between shifts. Then there is time to make up production losses with a little overtime at the end of each shift. PARTE GETTING STARTED 1816 Setting the Pace {As you go through the calculations to determine your own pacemeker takt time we need to explain one final point: Ie is seldom the ease that there is only one correct takt time! ‘The pace of production is one of the most critical considerations for the design of your processes. Here you will often have some choices to make. For example: A cell that has a takt time of 40 seconds over two shifts could also be run at 20-second takt in only one shift. In some cases itis easier and less costly to manage only one shift, particularly if running a second shift means extra support structure and paying night premiums. An added bonus is that the waste of waiting time is easier to sce and eliminate when cake time is shoreer. * The size, weight, and complexity of a product can influence what is a reasonable eycle time and the number of motions for each operator. Producing a light, low-complexity product with only a few work clements per operator to a ten-second takt time may be fine, But when operators are working on larger, heavier or more complex products it can be better to work to a longer takt time and assign more work elements to each operator. * When new products are introduced, substantial savings in capital investment can be achieved by adding them to existing cells rather than building additional cells. This will decrease the takt time for those cells, As you launch your new cell it is often much better to utilize a temporary and separately held ‘safety stock’ of specific finished goods to protect your cuscomer and to set your cycle time only slightly faster than take time. The tension this produces forces you and ‘your staff to address the causes of production interruptions. With experience you will gradually learn what's best for you. The key point for the moment is thar you must know what the takt time is and how it was determined.What is the Work? We've now examined and answered the first wo questions confronting your team as you strive to create continuous flow. We've specified the products appropriate to the pacemaker and determined the take time. Although you may be addressing some of these issues for the first time, Apex managers had in fact already answered them quite well. Yet the performance of the cell they created falls far short of what is possible. This therefore brings us to the more detailed aspects of cell design, wi Zeroing in on the actual work elements, timing each one of them, and immediately eliminating obvious waste in the work. At this step in continuous flow design you should forget about physical layouts or the number of operators and simply concentrate on the work content that needs to be done by people. fewer managers have mastered: Question 3: What Are the Work Elements for Making One Piece? When we look more closely at the fuel line assembly process we can see that cach operator performs a series of discrete steps, or ‘work elements’, that are required to complete one cycle at each station, By collecting all the work elements at cach workstation, we have the total operator work content of the cell, Understanding these work elements is « key ingredient for creating and maintaining continuous flow. A work element can be defined as “the smallest increment of work that could be moved to another person”. Thus “get one hose and place it in the fixeure” is a work clement while “get one hose” would not be. In the Process Study Form shown on the next ewo pages we have written down the work elements for the Assembly I and Assembly II steps in the creation ofa fuel line at Apex. Note: Always break work into elements. This helps you identify and eliminate waste that is otherwise buried within the total operator cycle. PART I:WHATISTHEWORK? 47Process: SLA Fuel Lines Process Study paces OPERATOR Steps Work Element Get bent tube & press into fixture Observed Get connector, place & clamp Get hose & place to fixture Start cycle Unclamp & remove Attach convolute Get tube & place to fixture Get hose & RH ferrule, assemble Place to fixture & clamp Get LH ferrule, assemble to hose Place to fixture & clamp Get valve & place to fixture Start cycle Unclamp & removeObserver: Mary Smith Lowest Repeatable Date/Time: Feb.10,2000 3PM MACHINE Cycle Time| Notes Operator has to hammer to fit Operator stacks own hoses (parts) Parts far away; Lots of walking Quality check needed? PART I: WHAT ISTHE WORK? 1920 ‘You can only complete a Process Study Form by carctully and repeatedly observing actual work. Do not rely on file data. As you observe the work you will probably find that operators perform it in a slightly different way from cycle to eycle. You will need to chink about the best way to do the job and try to record the elements as done this way. Use the “notes” area of the form to jot down problems, questions, and ideas. Ie is best to use a pencil for recording work clements. We find that we always have to correct the list several times before we are satisfied that we have understood and captured the process. Then when you think you've got it right, review the list with the operators who will show you what you've missed. Then re-write the “Shop Floor Courtesy 2S : Carefully observing @ process means that you will be spending time at the process you are trying to improve. You will be a guest in the “office” of the people who work thete. Please use the same courtesy you would when entering anyone’ office: ~ Introduce yourself _ . : ~ Explain what you are doing. » Say “thank you” It may be best to do this through the team leader or supervisor 80 you don’t irirript the operators and affect their safety, quality, and productivity. Always keep in. mind that you are not watching people or evaluating their performance. You are» watching the provess or how the product is made. You are figuring out What are the steps and then the times required to make the product : ‘Although we say these things when we gather observations on the shop floor, we know it takes time for this sort of thinking to sink into the culture of a company. | ‘When people are observed while they are working itis @ natural reaction to think Always remember ‘at she quenion erat Hew hee person doing?’ but tates “What is the work?” This isthe elective auenlon that people should focus on ~Paper Kaizen As you observe and record the work elements chere is something else you should begin to see: Waste! Operators walking considerable distances to get parts or waiting while machines cycle, Operators leaving their work.area to perform out-of-cycle tasks. None of these actions adds value from the standpoint of the customer. All disrupt the flow and all are waste. Here is an important step in completing your Process Study Form: Do not include any obvious wastes as work elements! They are not part of the work that needs to be done and an objective of the design process is to eliminate them. We call this approach of immediately leaving out wasteful steps ‘paper kaizen’ because you are climinating some waste on paper before the process is put in place, Of course, many improvements, like fine tuning parts presentation, will have to be done when the new cell is set-up and running. But there is usually a lot of waste you can eliminate in the paper kaizen stage. Here are some paper kaizen guidelin 1. Do not include any walking as a work element. Because the actual amount ‘of walking in your new process design is unknown and will be minimized, you should ignore it at this point. 2. Do not include out-of-cycle work for operators as work elements. This is a guideline you should never forget! Out-of-eycle work — like the first operator in the Apex cell stopping to feed the tube bender every 25 pieces — destroys continuo flow and makes it impossible to maintain efficient and consistent production to take time, Other examples include periodically getting bins of parts, making a quality check, or moving a container of finished parts ‘These tasks may need doing but should be given to support-saff, like team leaders and material handlers, who work outside of the takt-based continuous flow. If you are tkving out-of-eycle work to operators as a sort of break, then schedule an official break ‘or practice job rotation instead of burying such flow-killing waste in the work Some out-of-eyele tasks ean be converted to in-cycle tasks. For example, with only minor changes the tube bender in Apex’s cell can be loaded with a single work piece cach time the operator cycles through his or her work elements. This allows continuous flow and is easily achieved because the tube bender already has automation that can be adapted to process and eject finished pieces. PART I: WHAT ISTHEWORK? 2422 3. Do not include operators waiting for machines to cycle as a work element. Waiting for a machine to cycle is pure waste and must be eliminated. Always separate machine work from people’s work, It is OK to have machines finish cycling and sit idle until the operator returns on the next tak eyele, but an operator should never wait for a machine. You should develop your cells in a ‘way that allows operators to Ioad a machine, start it, and move on to the next workstation while the machine cycles. ‘4, Do not include time for removing finished parts from machines wherever you believe automatic eject could reasonably be introduced. When an ‘operator returns to a machine on the next takt eycle, he or she should ideally find an open and empty fixcure (an ‘empty nest’) into which the next part to be processed can be loaded. Stopping to remove parts before another part can be loaded adds extra handling and waste. Try to configure cells and ‘equipment to avoid this waste. Work Elements for the Apex Cell Using the guidelines for paper kaizcn, let's take another look at Apex’s fuel line cell. We've listed the fuel line work elements beneath the cell layout. Notice that the current operator activities that paper kaizen climinated are shown crossed out.Apex’s fuel line cell — work elements material flow ‘Assembly I fee com\t alee! el fine ebay [a [itn “rane [ond @ & LY waa com eo oom Gatbent abe, ‘Gat ae epace Gov acsonbly& pace pasoRtsdeto Get tion hose ine Got comscton, Bright fre, pace &clamp soenmnble ere Face dca torose Lede Gechose& place Pace to tur Start oye Starbeyele ‘amp Stott Govleft ferrules secontle to hose Acer Kate Arsach conolute Get vahe &place Get 85 tube & Pbce tobender o one Placa camp ea Start eyte mee Ge wat) Usclamp & mone ele PART I:WHAT ISTHE WORK? 2324 Question 4: What Is the Actual Time Required for Each Work Element? With all of the waiting on machines and manual removal of parts characterizing Apex’ current situation left out, we are now ready to time the work elements that are still necessary for making the fuel lines. “To collect accurate times for each work element you and your team need to go to the workplace and use stop watches. Avoid the temptation to use standard time data or time-and-motion tables because they do not capture the current reality of the shop floor. Likewise, avoid reliance on time-scudy data on file in the engineering office, Collecting the information yourself on site will help you understand the real situation and see waste thac will otherwise remain hidden, Note: Itis OK to begin with times from time-and-motion tables or data from past time studies when you are designing e completely new process and observing actual elements is not possible. However, as soon as the new process is operational you need to gather the real data as described here. Be sure to time each work element separately, not the total time required by an operator to perform a sequence of work elements, ‘This is because the total time fora sequence will include wasted time — in particular waiting — between work elements that should not be counted as work. (However, if several work elements in succession are only one second long you may find it impractical to measure cach one. It is OK to group such very short elements.) Once you have timed the individual elements then time the operator’s complete work cycle from start to finish. This time will almost invariably be higher than the sum of the work elements. The difference is the wasted waiting time between the elements. You will need to time every work element several times to get meaningful daca Indeed, if you are unfamiliar with this task we suggest that you time each element ten times. This is a lot of work, but practice is the only way to become proficient. As you do this be prepared for the reality that timing is labor intensive and time consuming, even if you are highly experienced.Sometimes as you record the time for one element, the operator will already be on to the next element before you can reset your stopwatch. So you will have to wait for another eycle to time that element. Similarly, if work element is interrupted while you are timing you must discard that time and wait for the next cycle. Just remember that there is no substitute for patience because the work element and timing data are the bedrock for everything else. You should time an experienced operator who is fully qualified to perform the job, but pick a typical operator if possible, not the fastest or slowest. ‘Then, after you have timed many eycles of each work element, select the lowest consistently repeatable time for each element. Do not use the average time since the lowest repeatable time is more indicative of ‘what is realistically possible. Finally, and critically important, don’t forget shop floor courtesy. Make clear to everyone that you are timing the work, not the operator, [As you record the times for each work element on the Process Study Form be careful co always separate operator work time from machine cycle time. For example, if an operator loads a machine for five seconds, takes one second to start the machine, and then the machine cycles for ten seconds, the operator work time will equal six seconds, not 16 seconds. While you are timing operator work elements go ahead and time the machine eycles as well and record them separately on the process study form as shown on the next page. Note: ‘Some experienced people we have worked with don’t use a stopwatch! Instead they count operator “motions,” valuing them at 0.6 second each, Perhaps you will fee! comfortable doing this yourself some day. However, we suggest you start with the method described above. STR rene ener Sunn Res ceed eee Secs Cee *Time several Cee Ree cea are eat) Pee) perform the job. Oey Cee) machine time. Select the lowest repeatable time for Ce eacasProcess Study Process Steps Assembly | Assembly II Process: SLA Fuel Lines OPERATOR Work Element Observed Get bent tube & press into fixture o|5|5 Get connector, place & clamp 434 Get hose & place to fixture Al4|4 Start cycle Unclamp & remove 3)2)2 Attach convolute 7|6\7 Get tube & place to fixture A515 Get hose & RH ferrule, assemble 5]4/4 Place to fixture & clamp Al4/4 Get LH ferrule, assemble to hose 4/3 /4 Place to fixture & clamp AIA Get valve & place to fixture 4l4 Start cycle Unclamp & remove AIS 14Observer: Date/Time: Page Mary Smith Feb. 10,2000 3PM Ww MACHINE 5 Lowest Times Ropeatable | Ce Time Notes 5 5 Operator has to hammer to fit 4 4 Operator stacks own hoses (parts) 4 Parts far away; Lots of walking 4 Quality check needed? | PART Ik WHAT ISTHE WORK? =—-27-Work Element Times for the Apex Cell ‘When we look at the list of all of Apex’s element times and machine cycle times an important finding jumps out. The operators ac Assembly T, Assembly Il, the Crimper, and the ‘Tester wait at their machines for a total of 21 seconds each time a fuel line is produced. 21 seconds may not seem like much wasted time on casual observation, but 21 seconds times 690 pieces per shift totals four hours of operator time spent watching machines eycle, No customer would think this creates value and it is completely avoidable because each machine only needs to cycle ance every 40 seconds (take time). There is nothing gained by the operator staying at the machine and reloading it as soon as it has finished ies cycle. Instead, the operator can perform other value-creating in-eycle work and return to the machine for reloading some time after it has finished cycling. Note: In reviewing the times in the list, we have estimated the time required to load the tube bender in a one-piece-at-a-time fashion because the work is not currently done this way. All “get & place” times at the other machines were measured at five seconds, so we will use that figure for the tube bender now. This estimate will be replaced with real data once the new cell is in operation.ine cell — work element times material flow ‘Ascombly 3pe8 TBpen esrorsetL etm \\ Tie somae Ly ee (automatic) [=| A |iauscmatic) sue fo @ BY Kas (oom) 20 pence ‘fae ota). (Get bert abe, Ger ibe pace pace Placoto hare Load auto feeder Get elon hase Got cxnecter, ‘aight ford, Remove hoot cap place &clamp ‘assole forte [attach hese to Oe tohose cure Gethose pace Place to Petur & Start ql Stareapie clamp pine eve ere & soso tohooe een Place & clamp Iropect crimpe ee-atesch cap Get vale & place Pllcomoute Inepect tofature overerings| Face shiping Stare eyele Alle contaner te vit) Unelamp &romve Mactine Cycle 5 sec PART I: WHAT 18 THE WORK?The Results of Paper Kaizen Now that times are known for all of the work clements, let’s see the benefits of our paper kaizen. We can show this by stacking up the times of all the work elements as the operators have been performing them. Then we can draw a second stack on the right showing the total work content after waste was eliminated by: ‘+ Introducing auto-eject at Assembly I, Assembly II, the Crimper, and the ‘Tester. (This eliminates operator time to unclamp, remove, and set aside parts.) ‘* Eliminating the waiting while machines cycle. * Converting the out-of-cycle work of loading the bender with a batch of 25 pieces into in-cycle work with the operator loading one picce every cycle. You can easily see that the paper kaizen has removed substantial operator effort from the Apex fuel line assembly process. Total work content co make one piece has fallen by 30 seconds, from 18 to 88. This improvement is actually more dramatic because the redesigned work elements have added the step of loading the bender every cycle instead of every 25 eycles. (This conversion of work from out-of-eyele to in-cycle will make the assembly process much more predictable and easier to manage against takt time and production requirements.) Because the amount of value creating work is the same in both the left and right stacks, it follows that in the future operators can devote a greater percentage of their time to actually creating value without working any harder. Apex’s first Operator Balance Chart With the work elements and times in hand, Apex’s managers can now create a critically useful tool, the Operator Balance Chart. The OBC is a picture of the distribution of work among operators in relation to takt time, based on real data that you personally “observe and record. It is quantitative, simple, visual, and takes the guesswork out of designing and operating efficient operator based cells, It enables engineers, managers, and production associates to work together using facts. This tool has been very helpful for developing our own “eyes for flow”. Whenever people start working with an OBC wwe always note how effective the tool is in helping them to understand, create, manage, and improve continuous flow.Paper kaizen for Apex fuel 125 120] current 115 Flace to container Tnopect 10) ane 105 Oyde 100/-| Remove cap & attach hose [= ine work content TOTAL TIME= 116 seconds, Eliminate, Use auto eject. Eliminate, Operator moves fonas machine cycles. Get accambly place to Thun TOTAL TIMI 86 second Pullconvolute over erimps: 85) — | Tropect crimps [| aeclanp reno 80 = cycle 75\- zs [=| Placa & clamp Liteide 70 Sez gesemby pace [=] “RH Ste to etre 65) A Eliminate, Use auto cect. ga = Eliminate, Operator moves = onasmachiecydes. 75| Remove cap & attach hose Place to tester fixture * Eliminate, 65 Unclamp & remove +} Eliminate, Use auto sect. FI Pull convelute over crimps Get risked pecs napoct orgs f= Place 6. clamp LH olde Cycle == Eliminate. Operator moves: Graeahike cycles BB} Getfinished piece place ; wines pace |= Get valve & place to fixture ) = 50 las he Get valve & place to fixture Place & clamp a : oe |Get LH ferrule & sosembte to hose | aor Place to fixture & clamp f= 40) 2 =] Place to fixture clamp = 35 [|r nose BRE fornlo, assem 35 = FE = coon tora soon = Got tubeé place = FE 30} 7 = Poce wo hacen tac F= ae = 25) 2s ee efuohes pee FE a fe attach convolute 20|=|—rreerpareRoe et —* Etninate Use autoejact. 20 te = Coe EE Elmnata Qpoarap moves gl | _ Oer ene rhee —S aaa 10/54 || Get connector, piace & clamp | — Get bent tube & place to ‘Assembly = #= Workpiece stays in machine, 10) = Standard Work-in-Progrese gs Get bent tube & place to ‘Assembiy! Sparttor ma [=] Gots tube place to bonder|—a2 10 In the OBC each horizontal line equals one second, with Apex’s fuel line take 1e drawn in at the 40 second level. The work elements for each operator are stacked one on top of another (which is why the OBG is sometimes called a ‘stack chart’). The height of each box in the stack represents the time for that element, Note that operator load and unload times are included but not machine eyele times, because the OBC is summarizing human work and not machine work. ‘Apex's managers thought they understood cell design, but looking at the fucl line OBC should give everyone at Apex a shock! Bach operator has considerably less work content then the 40-second takr. In two cases the operator work elements total less than half of take time, We have already seen that matetial is not flowing through the cell and that output is highly unstable, Now it is apparent that Apex is also using too many operators to produce fuel lines. This adds cost to the product and threatens everyone’s job. Bue please remember shat we are analyzing the process and not operator performance. In fact, the waste we've found has ies roots in the design and management of this process. Fortunately, the OBC can help Apex find a better way, as we will explain in the pages ahead. Apex’s Operator Balance Chart - current situation Got fnished ploce aeattach comoiice Get va place to facare Place 8 clamp OTe S} Get hose pace Pl convolte over crimp Face to container | Get cognac pace GaETiIaed parte (| Remove astatoe | z ‘comp Place tofecure clamp [=] “thopact ents a F=] Get bent tube & place : ani tonecerbi Cettectaitemsn sects] rceseampttate [=| RemmecapSattatone EY coves ube tiplace Pace tp Aaaemy i Settnared pace, ‘obender” ‘Rese naciiinearece Pace to taster ature @ © a ©Machines, Material and Layout for Flow. Production processes of any type require the coordination of Man, Machines, Materials, and Method (commonly called “The Four Ms’). In the previous part of Creating Continuous Flow we talked about Man (or more accurately, People) in terms of work elements. Now that the work elements have been identified and timed, we can put these aside for a moment and tuen our attention to arranging machines and workstations (Machines), and the materials management system and apparatus (Materials), in a way that permits continuous flow. Question 5: Can Your Equipment Meet Takt Time? Once Apex managers determined the products to make in their first cell (light tuuck fuel lines in three variants) and calculated the takt time (40 seconds), they needed to ask whether the machines they were moving together co create a cell could actually meet the takt time, To do this each machine must be able to complete its cycle on each part within take time. In fact, as we will see in a moment, the ‘effective cycle time’ of each machine should be considerably less than take time if continuous flow is to be achieved As Apex managers and engineers examined the five machines they had moved into the cell, they summarized their operations as follows: 1) Tube Bender. Stainless steel tubes from the previous process are bent to final shape in a tube bending machine, which cycles and ejects automatically as long as its feed hopper contains raw tubes. Because the hopper holds 25 pieces it would need to be refilled about every 16 minutes if the cell operates with a 40-second cycle time. 2) Assembly |, At this machine a connector and rubber hose are assembled to the tube. An operator loads and unloads this machine, which performs an insertion and crimping operation, 3) Assembly Il. At this machine a Teflon hose, ferrules (metal rings for connecting, tubes to hoses), and a valve are assembled to the tube. An operator loads and unloads the machine, which automatically performs the insertion and crimping ‘operations. This machine is changed over to run the two different lengths of PART Il MACHINES, MATERIAL, LAYOUT 33hose required by the product variants going through the cell. This involves the operator unpinning and sliding a clamp fixture to another position and repining it there. ‘The short- hose/long-hose changeover takes about 20 seconds, or half of take time. 4) Crimper. This machine crimps the left-hand and right-hand sides of the fuel line assembly, It is loaded and unloaded by an operator, 5) Pressure tester. The finished fuel line assembly is tested in an automated test fixture, which is loaded and unloaded by an operator. ‘Apex’s managers then prepared the following table to summarize the cycle time characteristics, or effective cycle times of cach machine. By ‘effective machine cycle time?” we mean: machine cycle time per piece + load and unload time (during which the machine cannot eycle) ‘+ changeover time divided by number of pieces between changeovers As the Apex team looked at these data they noted that the basic eycle time of each machine ‘was well below takt time, the longest case being only 16 seconds. However, when they Effective Cycle Times of machines in Apex’s cell Load, Start Effective Machine & Unload c/o Time/ Machine Machine Cyclo. Time Batch Size Cycle Assembly! | 46ec | 16sec o 20 sec Assembly t Time is actually 0.67 second with a minimum batch size of 30, based on packaging the customer requires as explained later. With @ 20-second changeover time, the changeovers on this machine consume just under one second per cycle when averaged over the smallest run,reviewed the effective machine cycle times they realized that the Assembly I machine raises a problem. If demand were to increase (meaning take time fell) or if the machines were not completely capable (that is, able to make a good part on every cycle), or completely available (that is, able to eyele every time they are instructed to) then Assembly II could easily become a bottleneck. Apex may need to consider improving the operator load and unload times at this step. At the other machines there are no bottlenecks that will inhibit continuous flow. In fact, we would expect these machines to be idle for several seconds during each 40-second take interval In our experience, in a world where machines are not completely capable or availible and demand does change, itis best to target effective cycle times for every machine in the cell at no more than 80% of the fastest (highest-volume) takt time, This ‘ensures that operators will not have to wait for machines to finish cycling the next 1 they come around to go through the work elements. It also provides a bit of ‘extra capacity co accommodate some demand increases without the need to add capital equipment or pay large amounts of overtime. ‘What if some of your machines can't meet the take time with a 20% margin? You have a number of options: * Kaizen the load, start, and unload process. * Eliminace waste in the machine cycle itself to reduce time. For example, shorten the trave} distance and time of moving machine parts. * Split apart some of the tasks a bottleneck machine performs and use more than one machine to perform them. Use simpler or single-task machines running simultaneously at shorter cycle times. * Install evo machines of the same type at the bottleneck point in the cell, and alternate between them on each cycle. * Create two cells instead of one. This may be particularly appropriate when there are advantages to locating separate pacemaker cells near different customers. Finally, if all else fails: * Remove the bottleneck equipment from the cell and operate it decoupled from the pacemaker in a batch mode. Normally this will require establishing supermarket pull system to regulate the production of the decoupled equipment. PART lk MACHINES, MATERIAL, LAYOUT 35.What About New Machinery? — Increments of Capacity {mn the Apex example, the company already has the machines it needs and wants to move ahead with its cell quickly to cut cost and improve responsiveness, However, you may be in a different situation, Some pieces of equipment are so anti-flow that replacing them ‘now is the best option. Or, in the extreme case, you may have a cellular layout in mind for an entirely new product and process requiring all new machines. What are useful guidelines for designing the needed machines in order to achieve continuous flow? ‘The first question is, “How should the materials be processed to achieve a quality part?” If there is a clear choice in technologies for making the part, then the design of the new Process and its equipment should begin there. ‘The next question is, “What is the desired process capacity?” The answer to this question is related to machine cycle times. The issue is whether you chose to design in the direction of one or a few complex ‘super machines’ that perform many activities with relatively long total cycle times, or you opt for a series of simple machines that each perform only one or a few of the processing steps with short cycle times. Because the cyclo time of a mult function machine will be longer, its capacity will be lower than a series of single-funi machines that each cycle more quickly. Demand rates and product portfolios can change frequently, even in stable industiies. For this reason itis often to your advantage to buy or build @ number of simple machines that cycle quickly. These types of machines provide each cell with more flexibility to respond to change. (This advantage holds if you keep operator load times per machine from getting too long.) For example, if the cycle times of two multi-function machines in a process are near takt time when installed (as shown in the machine balance chart below) there is not much potential to handle additional demand or additional products without buying another machine. Capacity increases in this example must be made by additional machines in increments of 30 seconds of cycle time, because the longest machine cycle time process determines the overall increment in which capacity can be added Limited flexibility of multifunction machines take time 40 seconds Ability to handle “~<——— increased demand and more products machine cycle time 30 seconds machine balance chartIf this same process instead uses a series of simple machines that each cycle in ten seconds {as shown below) the call or fine has the ability to handle demand increases and additional products without buying more equipment. And capacity can now be added in increments of only ten seconds. Single-function machines also tend to be more reliable and less expensive than larger multi-function machines. Also, procurement lead time is shorter. In fact. you might be able to design and build such simple machines yourself! Greater flexibility of simple machines takt time 40 seconds “~ Greater ability to handle increased demand and more products <— machine cycle time 10 seconds Maximize the Utilization of Machines or People? Many of us have been taught that efficiency is achieved by maximizing machine utilization. ‘As we analyze equipment capacity we need to reconsider this flawed maxim. The physical elements of production are people, machines, and material. There are tradeoffs between these elements when you design a process. If you try to maximize the utilization of one element, the utilization of the other two tends to decline. For example, if you try to maximize the utilization of machines — running them constantly and as fast as possible — you will need extra people to run machines all the time. You will also need extra in-process material between processes to cover problems and keep the machines running. Likewise, if you seek to maximize your utilization of material — having virtually no inventory on hand — you'll need extra people and equipment to handle demand fluctuations and breakdowns. Interestingly, trying to maximize the utilization of people is unique because humans are flexible. Ifthe item that a machine makes is not yet needed by the next processing step then it is OK for that machine to sit idle and not yet cycle, (This is particularly true with simple equipment.) But an operator can move to a differant machine to make something that is needed now. The machine is not very flexible, the operator is. This means that in pacemaker cells and lines you should design operator work content not to simply maximize equipment utilization, but for the best operator utilization. At times this ‘may appear to underutilize the equipment, which sits idle for some time during each takt interval, but producing faster than takt is overproduction — the worst waste of all PART Ili MACHINES, MATERIAL, LAYOUT 37Question 6: How Much Automation? Ler’s take a closer look at how automation can be used to make an efficient continuous flow. You can utilize automation to help you achieve an efficient and flexible continuous flow of material. But designed or used the wrong way, automation can also inhibit flow. To avoid this you must consider how the material flow and the operator flow will interact with one another. As 2 guide to thought, we often use a simple chart highlighting what we call ‘the great divide’ in automation (see below). ‘To allow operators to move on and add value while a machine works you will need at least Level 2 Automation in your cell. This permits machines to automatically complete their cycle once they're started without the need for further human attention. With Level 2 Automation operators ean work in a sequence like thi Remove a finished workpiece from Machine 1. \ Place a new workpiece in Machine 1. | Start Machine 1 (which then cycles unattended). { Carry the finished workpiece to Machine 2 (che next processing step). | Repeat the sequence at Machine 2, Levels of automation fee) Da Ve anc Ws overs Machine eee 4 le Auto Ce ae Auto Auto eC 3/9 The Great Divide LEVEL 4 Auto Auto Auto e 5 Auto: Auto Auto Auto‘This work progression allows operators to handle multiple processes within the takt time, as shown in the diagram below. While the machine cycles, the operator moves on to the next processing step. The operator never waits for the machine. We think of it this way: When the operator waits at a machine while it cycles the operator is working for the machine. When the operator moves on to perform subsequent work elements while a machine eycles the machine is working for the operator. Sometimes we still see operators staying at machines to monitor them. Again, the operator is working for the machine. This is pure waste, In these cases engineers need to develop sensors that detect problems, alert someone, and if necessary even automatically stop the machine. Then people come to the machine only when they are really needed. The cost of simple sensors to detect problems is almost always fur below the cost of keeping an operator ata machine, And human inspection is never 100% effective anyway. Multi-process handling instead of waiting a ' St ee eer ate i ‘ a i i ot ; ites i 3 Standard WIP ' / x : (one piece) ’ Unload Machine & Load Cycle PART Ik MACHINES, MATERIAL, LAYOUT 39Corrs Cie eet PO ee Dre aeLts Ror Urs Cae) SRL) ee) SCR during the rest of the Cac any Cm ny Pee DR Seg STRUCT COC ei Ly automation) th operator is or) CC eGR nor built into the process. Having a machine cycle unattended while the operator moves the finished piece to the next machine means that one piece of ‘standard ‘work-in-process’ inventory is left behind in the machine. Parts are being processed one piece at a time, but the one piece of ‘standard WIP" frees the operators from the machines. (During daily operation of the cell the supervisor or team leader will need to regularly ensure that these pieces of standard WIP arc in place or the efficient flow will break down.) Note: Proper machine guarding needs to be in place when machines are able to cycle without the operator's hands remaining on the switch. Look for ways to configure safety devices so they do not interfere with the operator's smooth mation. Level 2 Automation May Not Be Enough Level 2 Automation works well when parts can be unloaded and loaded with one hand. However, there is a problem when the parts being processed require both hands to unload and load. ‘Then the operator has to work in this sequence: Sct the new workpiece aside. ' Remove the finished workpiece from the machine. { Set the | Pick up the new workpiece. | Place the new workpicce in the machine. ished workpiece aside. ‘Start the machine, which then cycles unattended. | Pick up the finished workpiece, Bring the finished piece to the next machine, and repeat these steps. \‘As you can see, waste occurs because the operator must double-handle both the finished workpiece and the new piece. This is the situation in Apex’s cell, where both hands are needed to handle the long fuel lines, and it is one of the reasons that Apex's operators each stay at one machine. as they walk through Apex’s cell, it appears more efficient to have them stay in one position and simply remove a finished part, hang it onto a rack for the next operator, and then get and Iead a new part, In fact, the parts are still double-handled, but the waste is obscured because it is spread over two operators. Apex managers also felt chat by keeping the operators at the machines the operators could reload the machines as soon as they are open again, which results in high machine utilization and feels efficient, But of course, this completely ignores takt time, causes waste, and destroys flow. Double-handling means that many pacemaker processes need Level 3 Automation to achieve an efficient continuous flow. Level 3 Automation means that the finished workpiece is automatically ejected from the machine at the end of the cycle. As a result, the machine presents an ‘empty nest’ each time the operator returns with a znew part. A new part can be loaded without having to double-handle both parts. Because great precision is generally not required to eject a part, Level 3 Automation is usually quite inexpensive, A workpiece can often be ejected by utilizing energy from the final cyclical movements of the machine, for example via a part eject atm (ot ‘kicker’) connected to a moving ram. (We urge you to avoid steadily blowing ‘compressed air to eject parts, This creates a noisy workplace and high energy bills.) More Than Level 3 Automation May Be Too Much ‘An interesting point in the table of automation levels is that when you go beyond Level 3 Automation to Level 4 or Level 5 Automation capital costs and technical complexity increase dramatically. Properly loading parts is easy for humans but can require delicate technologies to automate. Similarly, transferring parts automatically may require robots or transfer lines. This may scem attractive on paper, but the complexity often reduces process reliability to below 70%. In addition, what can you do with an expensive, highly automated line when customer demand changes? ‘When demand falls off you are stuck with an underutilized piece of high-depreciation equipment. When demand rises above capacity you need to buy another expensive automated line, PART Il: MACHINES, NATERIAL, LAYOUT 41a2 In our experience, Level 3 Automation achieves many of the benefits of full auto- mation but without the expense, procurement lead time, maintenance requirements, downtime, and volume inflexibility. We call the decision to pursue Level 4.and 5 Automation ‘crossing the great divide’. In some cases you may need to do it, but you should look very hard at the alternatives first. Question 7: How can the physical process be laid out so one person can make one piece as efficiently as possible? Here is an elegant tactie for cell design. Arrange the machines, workstations, and material presentation devices as if only one operator makes the product from beginning to end, even if you will never run the cell this way. When you design a process so one person can move through it and efficiently perform all of che work elements, you automatically design a process that avoids isolated islands of activity, minimizes inventory accumulation between processes, eliminates excessive walking, removes obstacles in walking paths, and brings the people-driven, value-creating steps as close to one another as possible, ‘This is certain to be the most efficient layout, even though the correct number of operators has yet to be detcrmined (and will vary anyway when cakt time changes significantly). On the following pages are two simple checklists for organizing machines and workstations so that operators can perform their work elements as efficiently as possible, Among these guidelines is that machines and workstations shouid be close together and that the inside width of a cell should be kept to about five feet if possible, With this in mind, many cells naturally end up in a narrow U-shape. The more workstations or machines in the cell, the longer the U. OF course, there are often product, machine, or part presentation issues that affect cell design and thus many different cell shapes are po: Note: There is a school of thought that material should flow through cells in a right-to-left direction relative to the operator, because more people are right handed and it is more efficient and natural to work from right to lett. However, many efficient processes flow to the left and many flow to the right. Simply evaluate on a case-by-case basis whether a particular direction makes more sense.Guidelines for Cell Layout O Place machines and workstations close together to minimize walking distance. Remove obstacles from the efficient operator walking path. Try to keep the inside width of a cell at around five feet to allow flexibility in reallocating work elements among team members. With a width of no more than five feet, team members can easily walk across the inside of the cell during their work cycle. iminate spaces and surfaces where work-in-process inventory can accumulate. Maintain consistent heights for work surfaces and points of use. Locate the leadoff and final processes near one another. This minimizes return walking for the next cycle and allows one ‘operator to easily handle both the leadoff and final process. When this is possible it greatly aids line pacing. Avoid up-and-down and front-to-back transfers of the workpiece. possible, keep the sides of the machines open to allow horizontal transfers on the shortest path between them, Use gravity to assist operators in placing parts and moving materials whenever possible. Install flexible utility drops from the celling to make layout adjustments easior. Keep hand tools as close as possible to the point of use and orient them in the direction that they are used by operators. Use dedicated hand tools instead of tools that require bit changes, and combine two or more tools wherever po: PART Ill MACHINES, MATERIAL, LAYOUT 43,Guidelines for Cell Layout Q1 Absolutely ensure safety and good ergoriomics. A lean process is designed to support the operator and value-added working. Poor ergonomics is undesirable from a human standpoint and contributes to waste, Keep manual, operator-based work steps close together to allow flexible work element distribution and value-added operator work, Segregate Level 5 automation and continuous-cyele operations (like ovens) from manual operators or operator based work flow, ‘as shown in the diagram below. Note that an automated segment incorporated into a cell in this, manner must be highly reliable. If reliability is poor and hard to improve, begin by placing the automated segment in @ separate area and regulate its production with a pull loop. Incorporating automated segments into cells <— Automated segment Manual area {operators work here) material flowGuidelines for Machines 2 Use smail equipment dedicated to a single task rather than large, multi-task equipment, C2 Introduce auto-eject (Level 3 Automation) whenever operators must use both hands to handle the part. Install one-touch automation where possible, One-touch automation ‘means that an operator can place a part in a machine, initiate the machine cycle, and move on. Avoid batching. Ideally, machines should be able to process one-piece-at-a-time in less than takt time. Incorporate sensors to signal abnormal conditions and even automatically stop machines if necessary, so operators don’t need to watch machines during their cycle. Design in maintainability. This means machine designs that are easily accessible for maintenance and repairs, and can be fixed quickly. At the pacemaker process, strive to devise machine changeovers between different end items that take less than one takt time cycle. Guidelines for Mater ‘The essenti Is Management companion to these guidelines for layout and machines are guidelines for material handling. As Apex’s team works on the physical design of the fuel line cell ic also needs to take a look at its system for getting the required parts and materials to their ‘point of use’ in the cell. When you use the following materials management guidelines it will help your cell and line operators perform their work elements as efficiently as possible. PART Il: MACHINES, MATERIAL, LAYOUT 45Guidelines for Materials Management 1 Present parts as close as possible to the point of use, but not in the walk- ing path of the operator. Present parts so operators can use both hands simultaneously. ‘Try to keep all part variations at the operators’ fingortips at all times to eliminate changeover time. * Use fail-safe storage mechanisms when different parts look almost the same to prevent the wrong parts being assembled. *When you cannot keep all part variations near the point of use because they are too bulky or numerous, increase the delivery frequency for those parts or sequence their delivery to match the end-item assembly mix running through the pacemaker. For example, doliver sequenced parts every ‘pitch’ of work. (Piteh is discussed under Question 10.) Do not have operators get or restock their own parts. With the Possible exception of refilling screw pouches inside the workstation, use a material handler on @ regularly scheduled, standardized route (typically every hour) to deliver parts and take away finished goods. Keep no more than two hours of materials at the point of use. |Fthe material handler fails to deliver on schedule, the cell or line will soon stop, forcing managers to address erratic material flow. Do not put additional parts storage in or near the process because this makes the operation of the cell or line harder to understand and encourages operators to get their own parts. This can reintroduce the evil of out-of-cycle work. Utilize kanban to regulate parts replenishment. The material handler ‘comes regularly but will only bring those parts that have actually been used as indicated by kenban. No kanban, no parts, ‘Size parts containers for the convenience of the operators or as a ‘multiple of finished-goods packout quantity, not for the convenience of the material handler or the supplying process. The operators are the ones creating the value, Everyone else is at best incidental work!Guidelines for Materials Management Do not interrupt operator work cycles to replenish parts. Parts should be replenished in small containers from outside the cell and wherever possible should slide to the point of use by gravity-feed racks or chutes. Design these so an empty container taken off the flow rack causes the next full container to slide into position. The operator then slides the empty container down a return chute to exit the cell or line. boxes opened for Flow rack with return operator (one kanban / on each box) exay t inside cell t a al side x empty containers material handler operator Overhead view of flow rack delivery 7) Oo Flow racks can also flow over benches and through machines, not just between them. PART Il MACHINES, MATERIAL, LAYOUT 47Applying the guidelines to the Apex cell When we examine the current Apex cell through the lens of our guidelines, we see the following: * Too much walking distance from beginning to end. © The inside of the cell is too wide, * The leadoff and final operations are far apart, with the consequence that a single ‘operator will have to walk a long way to get to the beginning of the next cycle, ‘+ There is ample space for work-in-process accumulation between every machine. * The cube bender needs modification to load one piece every cycle. * The “out” tray for the bender obstructs the operator walking path. Current Apex fuel line module layout 580 square feet, 56 pieces in process AL™ JA Tes, Spee feet A K uy diets nae a} a 5 ‘teatro | 44] re (automatic) So pesleetsine ‘cain goals) Before applying guidelines aaReorganizing the cell in accordance with the guidelines produces a very different configuration as shown in the diagram below. his layout will be fine-tuned when deployed on the shop floor. However, we have now improved che physical layout to «& point chat we nced to ask how to integrate the work elements and the operators in the newly configured cell flow Example layout for efficient one-operator, one-pit 252 square feet, 5 pieces in process Assy Il angled to keep cell inside oe ade width about 5 ft “a Parts & materials presented on flow racks from outside cell oe Accontiy! single piece of in-process stock left in machine as operator moves on feet e a : {~___— no obstructions in walking path no space for WIP accumulation —— << tester angled to bring cell start “uve Bendr | Gatton) [oe esis fo and end together << A So pcicoetne ‘ileed goo) After applying guidelines PART I: MACHINES, MATERIAL, LAYOUT 4950a OODISTRIBUTING THE WORK ‘With the physical process now rearranged so that work can be performed in the ‘most efficient manner, it is time to bring the operators and the customer back into the cell design process. Question 8: How Many Operators Are Needed to Meet Takt Time? ‘The total operator work time to make one fuel line afver Apex’s paper kaizen is ‘88 seconds. If this were equal to or less than takt time, the cell could mect the customer's need with one operator. However, the customer currently requires ‘one finished fuel line every 40 seconds, so more than one operator will be needed. ‘The appropriate number of operators should not be determined by estimate or by negotiation between the different members of the team. Instead you should begin the process of determining the operator need with this equation: Total Work Content (after paper kaizen) Number of Operators TaktTime In the Apex example this works out as follows: 88-seconds work content = 22 Operators 40-second tak time ‘There being no such thing as two-tenths (0.2) of an operator, this means that three operators will be nceded if Apex is to meet customer demand with no further improvement in operating methods. While this is already one fewer operator than Apex is currently using, two-tenths is not far above two, which ‘means using three operators would result in low productivity. This cell can be run with two operators if the entire team is willing to sct this as a goal. PART V: DISTRIBUTINGTHE WORK 51‘The table below provides a guideline to evaluate the initial ‘number of operators’ calculation. This table assumes that operator loading will be in the 90 to 95% range. That is, 90 10 95% of each takt interval will be filled with work and operators will not need to watch machines or wait for parts. ‘The table also assumes that your ccll or line will be introduced and managed as described in this book. Note: If you intend to use a cycle time more than 10% below takt time be sure to make your first number-of-operators calculation based on takt time. We think you should have the potential productivity of your cell or tine clearly in mind. ‘Then you can redo the calculation using your planned cycle time as the denominator, instead of the takt time, keeping in mind the serious problems associated with cycling too much faster than takt time. ‘We have found that in each facility or company the commitment to create continuous flow should be accompanied by a consensus on how fully jobs will be loaded against takt time. This discussion shouldn't reoccur every time a new cell is designed and it certainly should never become a game where management is always trying to increase loadings and production teams are trying to reduce them, Fortunately, the Operator Balance Chart makes the actual loading process quite precise so the whole team can discuss the issue with facts. Guidelines for determining the number of operators In a cell Remainder in catcuaton otor Guideline / Target aper taza Do not add an extra operator. Further reduce waste & incidental work. Do not add an extra operator yet. After two weeks of cell operation & kaizen, carefully evaluate if enough waste & incidental work can be taken out. ‘Add an extra operator if necessary and keep reducing waste & incidental work to eventually eliminate the need for that operator in the call,How can Apex reach the 2-operator goal? By committing their whole team of managers, engineers, and production associates to a 2 —4 week troubleshooting and debugging process at the cell When you start to mun the cell with two operators you may only reach 65% of target production and be tempted to add another operator. This will be @ critical moment of decision because if you do add a third operator now the pressure to kaizen out more waste is likely to disappear, Instead, we urge you to run the cell with two operators for extra time at the end of the shift or on the weekend to make the necessary output carget while you continue to Kaizen, In our experience, a team willing to commit to intense kaizen in pursuit of a target number of operators based on our guidelines can usually reach 80% of the required output within a week of startup and 90% within two weeks. The final 10% is the hard part and will requite all elements of production and support to come together. However, as you reach this goal — and we absolutely know that you can — you will have created an operation that can run for months at low cost with true continuous flow. Even more important, this will set a new standard for your entire organization. Operator Loading Options When you are in the upper range of our operator guideline (with an initially calculated need over 0.5 additional operator per cell) you will face a choice of, how to distribute a Jess-than-full work content among the operators as you begin to operate your cell, ‘The traditional approach is called ‘balancing the line’ (see Option A on the following page). The work content is divided evenly between the operators, pethaps from a desire to be fait. But traditional line balancing bakes the waste of waiting into the process by spreading it eventy across all operators. The cycle time of each operator is indeed balanced, but each operator is only partially loaded. This practice not only makes it harder to eliminate waste later but also creates che potential for overproduction. For lean producers this is just a bad as failing to meet demand. PART: DISTRIBUTINGTHEWORK 6&3.54 ‘After a few days of running, the operators in a traditionally balanced cell usually begin to function as isolated islands. Small batches of inventory start to pile up between eperators and continuous flow falls apart. Managers may try to overcome this problem by placing ‘kanban squares’ between cach operator. These only hold one piece of work and operators are not allowed to make another piece until the downstream Kanban square is emptied by the next downstream operator This techaique can reestablish a single-piece-flow of material, but at the price of permanently incorporating the waste of waiting time in the process, It is a band-aid that doesn’t address the root cause. A bewter solution is to redistribute the work elements and fill every operator but one with work elements consuming almost the entire takt interval (Option B below). This is the lean option, By loading most of the waste of waiting on one ‘operator, the opportunity for kaizen is brought to the surface. Thea, once the cell is able to operate with one less production associate, the organization realizes a tue cost savings. balancing the line vs. the lean option OO OF OF A B Option A: ‘Balancing the line' by dividing work evenly between the workers in the call tands to bake the waste of waiting into the cell by spreading it evenly across all the operators. Operators are “balanced” but not fully loaded. Option B:This approach redistributes work to fully load every operator but one. ‘This exposes the waste and makes it easier to improve.Question 9: How Will You Distribute the Work Among the Operators? Apex’s team has decided to operate the cell with wo operators, each loaded with work content that equals about 95% of takt time. Apex now needs to de ide which work elements to give which operator. Work element allocation in cell can be done in many ways, Managers and engineers need to be aware of the range of possibilities, both for our Apex cell example and for different types of cells chey may construct in the future. Here are some approaches to consider: 1. Split the Work among the operators so each performs one takt time worth, of the total work content, often moving between several machines, 2. The Circuit, where one operator performs all the work elements to make a complete circuit of the cell in the direction of material flow. A second operator follows a few stations behind. 3. Reverse Flow, in which the operators make a circuit in the reverse direction of the material flow. 4. Combinations of splitting the work and a circuit or reverse flow. 6. One-Operatorper-Station, in which cach operator stays at one workstation. 6 ‘Tho Ratchet, in which each operator works two machines and “ratchets” the work piece ahead each time the operator moves to a downstream machine, Let’s look at each of these approaches in more detail PART IV: DISTRIBUTINGTHEWORK 5B.1. Split the Work Because tie machines in Apex’s cell are now arranged in a narrow U-shape, operators can easily cross from one side of the cell to the other during their work cycle. This means chat a number of split-the-work combinations are possible. It also means that the sequence of work elements the operator performs ean be different from the processing sequence required to make the product. ‘The ability to create across-the-cell work combinations is particularly helpful in finding new combinations of work elements that stack up to take time when work elements need to be redistributed after a kaizen or due to a change in customer demand. The variety of combinations made possible by keeping manual workstations close together is another reason ¢o isolate fully automated segments of production from the operator-driven portion of a cell. In splitting up the work it is a good idea to assign the same operator the first and last work elements in the material flow because this creates an automatic pacing effect for the whole cell Splitting the work Splitting the work meens that each operator is given one portion of the total work content. material Om 0) O00) O00, ae & Y Oo 7 Senge tet Need eet aeeeuiee 3 . 2 1 ©. i “o—-0 +0 0O—+0 0-—+0 finishea product2. The Circuit Assigning all work elements to every operator and having the next operator follow che first wich a gap of a few stations provides a number of advantages. It provides a natural pacing effect. It is easy to implement. Ie can reduce walking distances since operators have a short retum path to start the nexe eycle at the completion of the circuit. And it automatically rotates jobs and makes the work more interesting, However, circuits also have limications: « They are generally limited to two operators because it is quite difficult to coordinate more. It's better to try another approach if the total work content requires more than two operators. * Circuits will generally not work if mare than 40% of the total operator work content occurs at a single workstation, because the operators will jam up at the high-content station. If simple, quick-loading, single-purpose machines ‘can be substituted for a complex, high-cycle time machine you may be able to spread the work over more stations and make a circuit feasible. * They require skilled operators because each operator must be qualified to perform every element, while operating all the equipment and making all necessary quality checks. Circuit work distribution Each operator performs all the work elements. Oe A —— return walking distance PART I: DISTRIBUTINGTHE WORK 87.3. Reverse flow Reversing the direction of operators in a circuit where parts require both hands is a ‘good choice when the machines do not yet have auto-eject (Level 3 Automation). ‘The long-term solution is to modify the machines with automatic part ejection, but this can take time and may even have to wait until the launch of a new product. Because we want to maintain an action orientation and introduce continuous flow right away, reverse flow can be a good interim step. Reverse flow, once you try it, creates a more natural working/walking sequence in this situation, With reverse flow the operators move in the opposite direction from the workpieces progressing through the cell, Operators start at the finished goods container and work upstream to the leadoff point in the cell. Reverse flow does require that one additional piece of in-process product be kept in a holding position bevween exch operation, but the advantages outweigh this drawback. ‘The diagram illustrates how reverse flow works in a three-machine cell (which we have used for ease of presentation): * Standing at Machine 3, the operator unloads a finished part and puts it into the packout container. The operator then reaches to the holding position between Machines 3 and 2, takes the waiting part, places it in Machine 3, and starts the machine cycle. ‘* The operator walks to Machine 2 empty-handed. ‘* At Machine 2 the operator unloads the finished part and places it in the holding position between Machines 2 and 3, the position just emptied to load Machine 3. ‘Then the operator reaches between Machine 1 and 2, takes the part from the holding position, loads it into Machine 2, and starts the machine. * The operator walks to Machine 1 empty-handed, © At Machine 1 the operator repeats the sequence by removing the completed part and piacing it in the empty holding station between Machines 1 and 2. He then cakes a piece ont of the raw-material container, puts it in Machine 1, and starts the eycle. * The operator walls back to Machine 3 empty-handed to complete the reverse circuit.Reverse flow Operators move in the opposite direction of the material flow. finished product Machine 3 EA | 4 raw material Machine 1 +s, part holding ‘e = positions 4 ‘ ' i i / material flow ——> operator flow You have probably noted that each workpiece is still double-handled in reverse flow. This is undesirable and can only be eliminated by introducing auto-cject. However, while the machines are being modified we suggest using reverse flow. PART IV: DISTRIBUTING THE WOR4. A Combination In many cells a combination of splitting the work and a circuit or reverse flow will make the most sense. For example, some operators will work at specific stations or combination of stations, while one or more pairs of operators work in circuits or reverse flows in other parts, of the cell. This can achieve many of the benefits of circuits in cells requiring more than two operators, In addition, when a pair of operators include both the leadoff and the final work. elements in their cixcuits, it provides an excellent pacing mechanism for the whole cell. Combination work distribution material oo A _— product 5. One-Operator-per-Station Some processes involve only manual work with no automated equipment. In this situation the number of workstations can be the same as the number of operators, or there can be ‘one operator on each side of a workstation, Each operator performs all of his or het work. elements at the single workstation and chen passes the work on to the next station, ‘The assignment of work elements is easy, but it may be harder to evenly balance work and fally load the operators because of the limited ability to devise flexible work combinations, ‘A moving conveyor to transfer the workpiece is often a good idea for maintaining continuous flow with this type of work distribution, because without it there can be a tendeney for regression co betching, One-Operator-per-Station distribution Empty station i for volume material flow increase6. The Ratchet In this arrangement the number of workstations is one greater than the number of operators. Each operator works at two workstations and moves back and forth between stations every takt increment. Except for the leadoff and final workstations, ‘ovo operators will work in each station, one after the other. When operators move to the downstream workstation they carry the workpiece along. Moving upstream is done empty-handed. Thus the work ‘ratchets’ forward with each eycle of the operators, ‘The work content for each operator must be divided becween the two workstations such that each operator divides the takt time in the same ratio between the two stations. For example, if after 50% of the takt time one operator moves downstream to complete his or her cycle at the second workstation, all operators must move after 50% of the take time. Because all operators move at the same time, the ratchet provides a strong pacing mechanism and is excellent for day-to-day kaizen because line imbalances are immediately visible. However, ratchets are only practical when work elements can be evenly divided within takt time at every work station. This generally implies mostly load-and-start machines, short machine cycle times, and light machines that can be moved easily to even up the work elements. A good example is sowing, complex articles in the garment industry. The Ratchet Example: 4 operators & 5 workstations Workstation responsibility in the Ratchet Operator 1: workstations A +B Operator Operator 2: workstations B + C Operator workstations C + D workstations D +E PART IV: DISTRIBUTINGTHE WORK 6162 Selecting a Work Distribution for Apex With these basie approaches in mind, we may now ask what the best work: distribution will be for Apex. In looking at che situation ic is apparent that a ratchet approach is not appropriate because it would be very hard to divide up the cycle time on each machine into, say, half of take time, The one-operator per-station distribution is not suited to this application because all of the machines are to some extent automated, permitting operators to move on and. perform other work elements while the machines cycle. In addition, we note that this is a small cell with a target of two operators, so there is no need for a combination of splitting up the work and circuits or reverse flow. ‘This leaves the first three choices on our list — split operations, a circuit, and reverse flow. We prefer the idea of the two operators moving all the way through the cell exch takt interval, which makes it easier to maintain continuous flow, provides a pacing mechanism, and adds variety to the work. Because it will take a little time for Apex to design and install che auto-eject devices essential for an efficient circuit distribution in this case, we will begin with a reverse flow and move as quickly as possible to a circuit. Reverse flow does involve a little more ‘work content duc to the double-handling of parts, so there will be a need for a licele overtime each shift. This should create pressure to eliminate more waste and roduce auto-ejeet as soon as possibleConnecting to the Customer and Regulating the Flow ‘We have now integrated Man (People), Materials, Machines, and Methods to design a true, efficient continuous flow, But we still need to determine how to schedule the cell and how to react when there are changes in customer demand. Question 10: How Will You Schedule the Pacemaker? To this point we have been working on the flow of products through the cell and the flow of materials and operators to support the product flow. We now need to design a suitable information flow from the customer. Ieis not realistic to expect customer demand to be completely smooth, However, if we tty to adjust the output of the cell to respond from hour to hour to every twitch in demand it will be very hard to sustain any type of flow, Costs will go up and quality will fall, Similarly it is not realistic to expect the mix of product types demanded to be ‘constant. Yet if we try to change over from one product type to another, as exch item goes through the cell we may encounter material handling and productivity problems. ‘The alternative of producing large batches of one product type between changeovers reduces these problems, but at the price of slowing response to customer requirements for different types, and sending large waves of component orders upstream. This will require holding subscantial inventories of finished goods downstream from the cell — in the hope that we will have on hand what the customer wants — and of the necessary parts and components upstream of the cell. Both of these inventories increase the lead time through your value stream. ‘To achieve and maintain continuous flow and a lean value stream we need to schedule and operate a cell, particularly when it is a pacemaker process as in the Apex case, with as litcle volume fluctuation as possible. We call this ‘leveling the volume’ of work, Similarly, we need to decide on the most appropriate batch sizes to run before changing over to another product type. We call this “leveling the product mix’. Both volume leveling and mix leveling must be part of your cell design process. PARTV: CONNECTINGTOTHE CUSTOMER 63Leveling the Volume If the customer's schedule to a pacemaker process fluctuates widely over extended periods, you will nced capacity (people, machines, material) well above average long-run demand in order to always meet customer needs. But, for the more typical peaks and valleys in demand, when long-term average demand is fairly stable, establishing 2 supermarket of finished goods between the pacemaker process and the customer can allow you to level production requirements in the cell while still satisfying the customer. The cost of the supermarker — linked to the cell by pull signals — is often much less than maintaining extra capa Life would be simpler if customers were the only cause of demand fluctuations. But in most processes internal variations in performance also create waves of demand variation. For example, the pacemaker process may at times experience machine problems, quality defects, or missing materials that cause its output and material requirements to vary. You have the choice of either responding very quickly co these problems to maintain level output, or adding more inventory of finished goods downstream and parts upstream from the cell. g the volume of work more level production day-to-day Pa 4M Finished Goods Customer ‘Supermarket volume fluctuations 1Rapid response to problems is the better choice. ‘To respond rapidly you need to become aware of problems as soon as they occur. Speed of awareness is relaved to how much production you release to the pacemaker at one time, For example, if production control releases a daily schedule to the pacemalser process, managers and production associates will tend to find out how they are doing only toward the end of the day. At that point it may be hard to catch up and the whole discipline of working to takt time will have been lost. What you need is a much smaller ‘management time frame’, by which we mean the amount of work you release to your pacemaker and the amount of finished goods you take away. By reducing your management time frame and responding quickly to problems, it should become possible over time to produce a consistent volume without needing to hold extra finished goods inventory “just in case”, Ideally, Apex’s production control would release one takt worth of production (one fuel line) to the pacemaker cell every 40 seconds. At the same time, material handling would tale away one finished fuel line, This would be true continuous flow. Apex would be able to discover any production problems “instantly” within a single takt time. What is your management time frame? ‘¢What increment of work are you releasing? ‘+ How often do you know your performance to customer demand? PART: CONNECTINGTOTHE CUSTOMER‘The problem, of course, is that releasing and taking away one take ata time at each of your pacemaker processes would require an army of material handlers. Apex will need co employ a more practical approach based on the concept of ‘pitch’, Pitch is the takt time multiplied by some number of pieces that gives « practical time increment for releasing and taking away work at the cell. An often-used multiplier is the finished goods packout quantity. At Apex the packout quantity for the cell is 30 pieces per container, We get from Apex’s packout quantity to a pitch with the following formula: 30 pieces per container Pitch = x 40 seconds per piece (takt time) 1200 seconds or 20 minutes 20 minutes would be the basic scheduling increment, or “heartbeat”, for Apex’s pacemaker cell, Based on this pitch, Apex can establish a material handler route through the plant, Every 20 minutes he or she will bring the next 20 minutes of schedule for the cell and simultancously take away 20 minutes of production — ‘one container of 30 picces. ‘This technique of ‘paced withdrawal’ will allow Apex’s managers to always know within 20 minutes if they have a problem. Ifa fast problem-response mechanism is in place it should be possible to correct problems and get production back on track without the need for large amounts of overtime, Apex managers can truly manage their cell by pitch if they: 1) Know the target (30 pieces cvery 20 minutes). 2) Check progress regularly to spot abnormalities (every 20 minutes). 3) Quickly respond to abnormalities. Remember that the operators will be working very nearly to the full tak time. ‘They cannot solve the problem and stay on pitch, Instead the team leader will usually need to step into the cell to address the problem, or switch places wich aan operator who can rework defective parts.Managing a cell by pitch is most feasible when changeover times are very short, Ideally the changeover time will be less than take time, This goal should be addressed by engineers who design the machines for the cell. Fortunately, this is the case for the Apex cell where the longest changeover time is 20 seconds. Whenever you encounter changeover times that arc significantly greater than takt time you will need to assign changeovers to specific pitch increments and skip one ot more pitches each time there is a changeover, ‘There are many possible pitch intervals, and you need to find pitch times that make sense for your own situation, We sometimes begin with longer pitches and ‘move to shorter ones as the capability of the pacemaker process is improved, Note: Pitch and paced withdrawal are used to detect problems in a pacemaker process, not to make operators work harder or faster. Leveling the Mix Producing large batches of one product family variant in the cell makes it hard to supply the customer with a variety of product types in a short lead time with- out holding extra finished goods. Batching also tends to send surges of demand for particular components up your value stream. This forces upstream processes to hold more inventory co handle the surges. To minimize batching, surges, and upstream inventory, you need to level the product mix of your cell. You can measure the degree of batching in any process by looking at its ‘EPE’, which means “every-part-every-interval”. EPE indicates how often a process can produce cach of the high-running product types it makes. For example, if a process can produce all ofits high-running end items every day, then EPE becomes “every-part-every-day”, An appropriate EPE for pacemaker processes is “every-part-every-ship-window” or “every-part-every-day”, whichever is smaller. Apex makes two shipments of fuel lines per eight hour shift (with all three product family variants potentially included in each shipment). The goal for Apex’s cell should be an EPE of cvery-parcevery-tourhours. Apex’s managers will need to calculate changeover frequencies accordingly (with six changeovers per shift). PARTV: CONNECTINGTOTHE CUSTOMER 67The Load Leveling Box One good tool for leveling production volume and production mix at the pacemaker is load leveling box. This simple device has a column of slots to hold the kanban for cach pitch interval and a row of slots for the kanban for each type of end item running through the cell. In Apex’s load leveling box the columns would represent 20-minute increments {the pitch). It would contain three rows of stots for the S, L, and A fuel line variants. On each box of finished fuel lines in the cell’s finished goods supermarket there would be one kanban card indicating the following: # Type of fuel line (S, L, or A) * Quantity in the box (30 pieces) * Supermarket address * Cell address {As production contro! pulls a customer order from the finished goods supermarket to stage it for shipping, the kanban cards are removed from the boxes and placed into the load leveling box for the pacemaker cell in a level mix for each ship window (four hours). The material handler serving the cell then withdraws kanban from the box at the pitch increment (20 minutes) and delivers them to the cell to initiate production, Apex's load leveling boxKaizen the Ship Frequency Many manufacturers are required to make daily (or even more frequent) shipments to their customers. This is a blessing because it establishes a rapid ‘clock speed” for your pacemaker process, within which it must be able to produce all of its high- running part numbers. If you get to the point where the EPE for all of your processes (both pacemakers and upstream fabrication) equals the ship frequency, then your entire value stream will flow: Your facility will be a ‘money pump’ that converts raw materials into final products that can garner cash before the materials bills come due. However, if your customers require only infrequent shipments or fal to steadily reduce the ship interval there will be little pressure on your pacemaker and upstream processes to improve. Steadily increasing ship frequencies is therefore a highly worthy kaizen project for you and your customers. PART: CONNECTING TOTHE CUSTOMERQuestion 11: How Will the Pacemaker React to Changes in Customer Demand? We have now given lengthy attention to the Apex fuel line cell in order to integrate ‘Man, Machines, Materials, and Methods. After all this work it would be great to finish kaizening the process during implementation (as explained in Part VI) and then min it as designed forever! Unfortunately, this will be impossible for two reasons. First, Kaizen is never finished and second, customer demand always changes. In particular, changing volume requirements are something that anyone who is designing a pacemaker process needs to consider from the beginning of operations so that they can be ready to respond. Responding to Change in Customer Demand Many facilities are only subject to small day-to-day fluctuations in demand. A year-long history of shipments will often show that demand is stable for periods of time. The customer takes alittle more today, a little less tomorrow, and over time it balances out. ‘Smaller day-to-day demand fluctuations are best handled by use of a finished goods supermarket. The supermarket allows managers to set the take time and the number of operators and makes it possible to manage continuous flow for extended periods. But eventually the average customer volume requirement will change and the pace- maker process will need to respond. If volume drops and you keep the same number of operators, productivity will fall and the potential for overproduction will increase, If volume increases then the takt time for the cell will decrease (the take time bar moves down the OBC). You will need to add people and perhaps machine capacity to continue producing to take, Responding to an increase in demand take 1 take 2‘The periodic need to increase or reduce the number of operators is one of the main reasons for utilizing cells and arranging them in a narrow U-shape. The great variety of walking pattems offered by the U creates many options for redistributing the work elements over a different number of operators, When designing a cell, engineers should prepare ‘one-up’ and ‘one-down’ scenarios for responding to changes in volume requirements. They do this by drawing OBCs for the current number of operators and for work distributions with onc less and one more operator. We call these different positions ‘toggles’ because they are like the different positions on a toggle switch, You ean alter the output of a cell by ‘toggling’ the number of operators up or down in response to changing customer demand. Over the next 18 months of operating the fucl line cell at Apex, the customer demand, number of operators, and output per aperator (labor productivity) may look as follows. Apex’s customer demand/shift “unr Aprstene ulpSept OstOxe Jantar prone PART: CONNECTINGTOTHE CUSTOMER 7772 Apex’s number of operators amount of over per shift Janae AprJune dup Sept OstDee anlar AprJune Apex’s hourly output per operator © increases due to “eg _ongoing kaizen andar Apelune—dulpSept OctDee denM prune Note: The hourly output per operator (labor productivity) remains constant (with a bit of improvement toward the end from continuing kaizen} despite changes i customer demand and the number of operators. This is quite an achievement and, in combination with true continuous flow, is a hallmark of a lean pacemaker process, We call this achievement ‘labor linearity’, meaning that at each toggle position — with different numbers of operators in the cell — the work content for every operator stays very close to takt time.‘Apex has three options for reacting to changes in customer demand, listed in increasing order of volume change: © Absorb day-to-day customer fluctuations with a finished goods supermarket. * Run a little overtime each shift (which can only be done if there is a time gap between shifts, making this difficult in three shift operations). It is better to run a little overtime than to stop production a little early because operator productivity stays high, © Toggle the number of operators. ‘This list of options also corresponds to the frequency with which you should resort to them, Ies better to try to deal with demand variation through overtime and the use of & finished goods store. These two options can be used daily. But in our experience the minimum time frame for maincaining a toggle (ata particular cake time and staffing level) is a week, and we strongly recommend two weeks to ¢ month, And when you are starting up a new or redesigned cell, 2s Apex will shortly, you should cry to maintain takt time even longer, say three months, to give yourself time to kaizen the process and develop your process management routines. Note: Trying to change the takt time as you toggle is difficult because you must also change work distribution and redeploy people. For this reason, lean facilities try to maintain their takt times for some period of time, Toyota, for example, currently recalculates takt times in its vehicle assembly plants monthly. Increments of Capacity Increasing output above current process capacity will require not only more operators but more machines. Fortunately, lean cells typically use small, simple, inexpensive equipment that permits flexibility in responding to changing customer requirements. In many cascs you can increase capacity incrementally by adding a machine to the cell. We call this ‘inerements of machine capacity’. If even more capacity is needed you can choose to make some of the product in an existing adjacent cell, adding some machinery there if necessary. Finally, if lange inereases in output are needed it may be appropriate to add another cell. We call this, adding ‘increments of cell capacity’. ‘These options require redeployment of people, plu capital spending and significant lead times. You want to make sure that you have achieved the full benefits of kaizen in your cells and have high machine reliability before you consider spending major sums on more eapacit PART\: CONNECTINGTOTHE CUSTOMER 73