Potential reductions in plywood manufacturing costs resulting from improved technology

Potential reductions in plywood manufacturing costs resulting from improved technology Henry Spelter George Sleet Abstract Recent technology improvements in plywood manufacturing processes are described. The potential impact of these improvements on manufacturing costs and output capacity of a plywood mill was examined by a series of computer simulations. Results indicated that, compared to a mill equipped with technologies characteristic of the mid-1970s, a modernized mill could process similar wood input into the same product output with about 14 percent lower variable costs. By replacing medium-diameter (14-in.) bolts with small-diameter (9-in.) bolts, additional cost savings of 20 to 24 percent could be realized, depending on the technology applied. Also, annual output could increase by 13 to 28 percent without adding more lathes, dryers, or presses. Within the past decade, significant improvements in plywood manufacturing technology have emerged. These improvements are helping the industry to adapt to changing raw material conditions, where traditional large-diameter timber is becoming scarce (7) and even smaller sawlog-sized timber is too expensive to peel. By allowing the use of smaller, less costly timber and increasing throughput by reducing equipment bottlenecks, costs of manufacturing can be significantly lowered in a typical North American mill. The purpose of this study was to guide assessments of the plywood industry of the future by developing estimates through computer simulation of the potential cost and productivity impacts of these technologies. We begin by reviewing the technologies in various parts of the manufacturing process. Next, we describe the equipment configuration of a model mill that we use as a basis for comparison. Then, we present our estimates of potential cost reductions that may be expected when a mill is upgraded. (For additional discussion of plywood manufacturing practices, see Baldwin (1) and Sellers (14).) It should be noted that cost estimates were made by Spelter based on a simulation model called Plywood Mill Analysis program (PLYMAP) (15). PLYMAP is a command-driven FORTRAN program that prompts the user to supply a set of technical and economic specifications that define the manufacturing and economic conditions of a mill. The model then generates estimates of product recoveries, process flows, costs, and profits consistent with the data that were assumed. Machinery and process data used in the simulations for this report were obtained from individual mills and machinery suppliers. The cost estimates in this report are not official American Plywood Association estimates. Review of technological developments Plywood manufacturing involves placing heated blocks into a lathe, peeling the blocks, clipping the resulting ribbons to size, and stacking the clipped sheets. The sheets are dried to some target moisture content (MC) and those above target are returned for additional drying. Adhesive is applied and panels are assembled from the individual sheets. The panels are consolidated in a cold-press before being loaded into a hot-press where the adhesive is polymerized under heat and pressure. After pressing, the panels are set aside in hot stacks to allow the adhesive to continue to cure, then they are trimmed, some grades are sanded and patched, and then strapped in bundles for shipping. The review of technological developments follows this basic outline. Charging When the plywood industry was processing large old-growth timber, bolts were manually loaded into the lathe by projecting bullseye targets on the bolt ends, then positioning the bolt in the lathe with chain grapples or overhead cranes. When large bolts were used, The authors are, respectively, Economist, USDA Forest Serv., Forest Prod. Lab., One Gifford Pinchot Dr., Madison, WI 53705-2398, and Director, Quality Services Div., American Plywood Assoc., P.O. Box 11700, Tacoma, WA 98411. This paper was received for publication in April 1988. Forest Products Research Society 1989. Forest Prod. J. 39(1):8-15. 8 JANUARY 1989

TABLE 1. Estimated equipment technology adoption by North American plywood mills. a Units in service Manufacturing technology 1979 1980 1982 1984 1986 1987 losses because of misplacement were small. Although this procedure was slow, the downtime as a proportion of the peel cycle was also small. When mills began using smaller timber, automatic geometric lathe chargers were developed. These depend on mechanical fingers to sense bolt diameter and determine placement in the chucks. In 1979, the X=Y charger was developed. The X-Y charger scans bolt size and shape using obscure light (shadow) techniques, sound waves,or laser beams. The bolt is rotated a full turn as data are gathered on size and shape. Based on this data, the block is charged into the lathe to peel veneer from the largest right cylinder, maximizing veneer output. Table 1 illustrates the adoption rate of this technology over time. Peeling Powered backup roll. Veneer recovery is reduced by the tendency of the bolt to spin-out or split-out. Spinout occurs when the torque required to peel a bolt exceeds the amount that the bolt eds can withstand. The likelihood of spin-out increases when the forces to peel veneer are increased, or when the outer chuck in a dual chuck system retracts. Split-out occurs when the bolt breaks at a defect, such as rot or heart shake. Historically, backup rolls were used to keep blocks from flexing as the diameter got smaller. A logical modification was to power the backup roll so it would aid the spindles in providing torque to the block (5).By mid-1987, 110 mills were using powered backup rolls (Table 1). Powered nosebar. Slivers that stick in the gap between the knife and the nosebar cause another peeling problem known as plug-up. Plug-ups ruin good veneer and interrupt peeling. Traditional fixed nosebars are most prone to this problem, followed by small-diameter roller nosebars. By enlarging the roller nosebar from 5/8 to 2-1/2 inches or more andpowering it, the plug-up problem can be almost eliminated. The exit gap resistance is reduced, and slivers are simply driven through. An additional benefit is a further reduction in spin-out. Since its appearance in 1981, this technology has been adopted on 120 lathes (Table 1). Peripheral drive lathe. Another solution to the spin-out and plug-up problem is to power the bolt totally, or in part, by a gang ofrolls situated above the knife. The rolls can be spiked or smooth. Spiked rolls provide a greater degree ofturning power. The rolls are arranged at about 2-inch intervals within a sectionalized nosebar. Because the forces driving and resisting the bolt are nearly opposite, they neutralize each other, thereby eliminating forces that can lead to spin-out (6). An additional feature of this lathe is that the spikes introduce microchecks on the tight side of the veneer, relieving stresses ( tenderizing ) in the veneer. This helps the ribbon lie flat on the trays, which makes clipping more accurate. Also, less splitting occurs upon further handling, and the veneer dries faster, boosting dryer capacities by 10 to 15 percent (12). Tenderizing could be important to mills that peel to very small cores where weak, warp-prone, juvenild wood predominates. Peripheral drive lathes were originally developed for peeling 3-foot corestock(japanese plywood is 3 by 6). One 4-foot lathe is operational in the United States for this purpose and a second is on order. Now, 8-foot lathes are available. By reducing the tooth size on the disks, the marks placed on the surface are reduced and disappear upon sanding, making acceptable face veneer. A recent lathe variation is one that peels small bolts, 12 inches or less in diameter. To enable peeling, holes 12 inches deep are drilled at each end of the bolt during charging and the spindles are placed into these holes. This stabilizes the core from the inside rather than from the outside as with a backup roll. Peeling to 2-inch cores is possible. This lathe has not seen service in the United States. Hydraulic knife positioner. In any lathe, large forces are brought to bear on the knife and pressure bar during peeling. When changes in forces occur, they can affect the position of the knife relative to the block, thereby changing peel thickness (9). Depending on the amount of wear in the connections,significantvariations in thickness may occur. Changes in force take place when elements of the system are disturbed, such as when the pressure bar is opened to clear slivers or closed after roundup. In older lathes with screw drives (e.g., clutch, gear box, bevel gears, lead screws, and cross shaft),there are many wear points, which translate into large variations in veneer thickness if the connections are worn. Often, thicker peels are targeted to minimize downgrading of panels due to undersize. Recent digital linear knife positioners are much the same as gearbox drives, except the gearbox is replaced by precision direct current (DC) motors, and more accurate and slow-wearing ball screws are used in place of lead screws. The signal to these is digitally encoded. The peel can be changed infinitesimally, and thickness variation is improved, although problems of looseness remain. The hydraulic positioner is an improvement on either the gearbox or the DC digital knife positioner because it eliminates the lead and ball screws,cross shaft, and other wear points. It also limits play to only two FOREST PRODUCTS JOURNAL Vol 39, No 1 9

points: connection to the positioner and the trunnion mount to the cylinder itself. Field tests have demonstrated less veneer thickness variation. In practice, variation as little as ±0.002 inch is being achieved on well-maintained lathes (18). Another key benefit is increased productivity caused by faster charging and roundup. Table 1 shows the adoption rate for this equipment. Spindleless lathe. With the powered backup roll and powered nosebars, core diameters of as little as 3 inches have been achieved, making small-bolt processing more attractive. But another barrier to effective mall-bolt peeling is the charging time of the lathe. For small diameters, the time spent charging and rounding up the bolt can take up to one-half or more of the peel cycle. This reduces lathe productivity. An innovation of major consequence for small-bolt peeling is the spindleless lathe. Where traditional lathes hold and spin the blocks with chucks on each end, this lathe cradles the block between three full-length, individually driven rolls. Charging times of under 2 seconds have been reported (2). Since the lathe drives the surface of the block, the block must be somewhat round going in. This requires a second lathe to preround blocks. However, prerounding reduces roundup time during the main peel, bolstering lathe productivity. The maximum ingoing block diameter of the first-generation equipment is 8 inches. One such lathe is presently operational. A second-generation spindleless lathe, now available, can process bolts up to 14 inches in diameter, operate at faster speeds (900 feet per minute (fpm) vs. 650 fpm), and generate veneer more uniform in thickness (4). One second-generation lathe was installed in early 1988. Clipping Rotary clipper. Conventional guillotine-type clippers propel a knife downward into a cut, which momentarily stops the veneer flow. This stoppage can cause the leading edge of the veneer to fold under. At best, this causes a loss of good veneer. At worst, it causes a pileup and a loss of a large amount of veneer. Because of this limitation and the loss of clip accuracy at high speeds, guillotine clippers are usually run under 400 fpm, limiting green-end output. The rotary clipper involves a major change in clipper design because the blade, instead of being stationary, moves with the veneer ribbon flow. When a clip signal is received, the blade rotates downward, cleanly severing the ribbon at the lowest point. The movement of the blade is in the same direction as the ribbon flow, so there is no interruption of the ribbon. Speeds of over 500 fpm are obtainable with high accuracy caused by lower knife mass and inertia. Over 100 rotary clippers were installed or ordered by mid-1987 (Table 1). Drying Automatic dryer control. Accurate and consistent drying of veneer is made difficult by the inherent variability of veneer with regard to MC and thickness. Because of this nonuniformity, veneer exits the dryer at different MCs, and those sheets above a specified level are marked for redrying. The dryer control mechanism is the count of wet sheets. If the proportion of wet sheets from a batch of 100 differs from a set level, the dryer speed is adjusted. To keep redry rates low, many mills run dryers at speeds slower than optimum, resulting in much overdried veneer. Veneer that is too dry requires higher glue spreads to prevent dry-out of the glueline. A new dryer development monitors veneer moisture inside the dryer. The hot-air temperature is recorded before and after it passes over the veneer. The temperature difference indicates how much evaporation is occurring, hence it is a measure of moisture retained in the veneer. If there is no difference, the veneer is bone dry. A large temperature difference indicates that the veneer is still quite wet. A reading slightly above or below a target causes a microprocessor to automatically change the dryer speed. Redry rates continue to be collected and if they vary much from target, the temperature target is altered within the microprocessor. This allows veneer moisture variations to be handled. This system compensates for veneer variability by allowing the dryer to adjust to changes in MC. In mills where the system has been installed, throughput increased up to 10 percent (13). By mid-1987, seven mills had installed such controls (Table 1). Radiofrequency (RF) redryer. Maximum throughput for the primary dryer is realized at a fairly high (15%) redry rate. But the high redry volumes tie up significant dryer capacity. An efficient alternative is the RF redryer. When a separate redryer is available, then the optimum redry rate for the primary dryer is higher (as much as 25%). Redry loads are put onto an infeed chain in stacks about 26 inches high. The stack is automatically loaded and heated with microwaves for about 15 minutes. The RF energy drives off some of the moisture and redistributes the rest within the load. No operator is required other than the forklift operator to ready and remove loads, and there is virtually no redry loss caused by breakage or downgrading as in conventional dryers (16). In mid-1987, five mills had installed RF redryers (Table 1). RF vacuum dryer. A third development in drying is the RF vacuum dryer. Where the RF redryer is meant for redry only, the RF vacuum dryer is for primary drying. It consists of a large vacuum chamber 7 feet square and 25 to 30 feet long. Loads of unstickered veneer are set on carts and wheeled inside. A vacuum pump removes the air, and RF energy is used as a heat source. The water boils, but the temperature never exceeds 120 F because of the vacuum. There is no drying degrade because low temperatures are used and handling is minimized. But, there is a high electrical cost. The system was evaluated and rejected by a few plywood companies. However, now that industry target MCs are closer to 15 to 20 percent, the economics may be more favorable. Gluing The interaction of veneer temperature, MC, surface roughness, assembly time, and mill temperature complicates veneer gluing. With perfectly smooth, cool veneer, glue requirements are relatively small. As conditions deviate from the ideal, spread rates must be in 10 JANUARY 1989

TABLE 2 Assumed equipment operating limits in an older plywood mill Equipment Operating limits Ring debarker 150 fpm Lathe with roller nosebar Maximum spindle speed. 400 rpm Maximum sheet speed 1,200 fpm Charge time 3 seconds Guillotine clipper 375 fpm, full sheet 250 fpm, roundup Two 4-deck. 20-section jet dryers 375 F Automated spray layup 12.5 panels/min. Manual press loading 2 sec./opening creased to prevent dry-out or decreased to prevent overpenetration. Two significant gluing developments are foam extrusion gluing and high-moisture gluing. Foam extrusion. The essence of foam extrusion is to take a somewhat conventional adhesive and mix it with air until it is about six times its original volume. The foamed adhesive is pumped to an extruder head above the layup line where it is laid down in 1/8-inchdiameter ribbons, 3/8 inch apart, onto the veneer. Near the end of the layup line, a masher roll presses the panels, flattening the ribbons into a film covering all parts of the opposing veneer surfaces. This system allows thinner spreads, as little as 24 lb. per 1,000 ft. 2 of single glueline (MSGL) during cool weather. With traditional glue applicators, spreads are normally 33 to 46 lb./msgl. Glue waste is lower than with curtain coaters, spreaders, or spray systems. The adhesive is directed through nozzles and laid down in 49.5-inch widths, reducing trim and cleanup waste. The glue is isolated from the atmosphere, so there is no evaporation loss. The total loss, including trim, is about 8 percent with foams compared to 11 percent with curtain systems, 14 to 16 percent with spray systems, and up to 33 percent with roll coaters (11,14). Reports of 20 to 25 percent glue cost savings have been reported (3). By mid-1987, there were seven foam-based layup lines in service (Table 1). Problems with laps and gaps in the core have been reported for southern pine. Wavy, irregular veneers, often from the juvenile core of blocks, poorly align as they descend through the crowders. This prevents sheets from being laid down properly. the problem is a function of the crowders rather than foam gluing. Solutions could include unitizing random strips by means of a sheet composer or incising veneer with spiked rolls similar to those with peripheral drive lathes. High-moisture gluing. Veneer for plywood has traditionally been dried to average targets as low as 4 percent MC to avoid overpenetration of the glue into the veneer and minimize blows. Blows are caused by excess moisture that turns to steam inside the press. The trapped steam builds pressure until localized glueline failure or even panel rupture occurs. Various resin formulations are available that allow gluing at higher MCs. In mills where the program has been tried, 18 percent is now a typical redry level where previously there might have been an 8 percent redry level. With higher MC levels, compression loss and blows can become a problem. These problems, however, are offset by the process advantages of high-moisture gluing. including greater dryer throughput, less wood loss through shrinkage, less breakage of overdried veneer, lower glue spreads, and shorter press cycles. By 1987, over 30 mills were involved with a high-moisture gluing program. Pressing Compression controls. To obtain good contact between veneers, press pressures are set at levels as high as 200 psi. Such pressure causes densification of the wood with a resulting loss in volume. Densification can be lessened by reducing pressure during the press cycle. A recent study (17) showed that reducing press pressures incrementally can reduce compression loss without harming the gluebond. Many mills have installed controls that after a fixed period of full pressure reduce the initial pressure, but keep the positions of the platens constant for the remainder of the cycle. By 1986, 108 mills had installed some form of compression control equipment (Table 1). Panel watering. In the same study (17), it was shown that immediately wetting the panel after pressing restores some compression loss. About 1 percent of the original thickness of the panel recovers after this treatment. By mid-1987, 36 mills were watering their panels (Table 1). Analysis procedures TO estimate the economic effects of these innovations, the PLYMAP program was first simulated with process parameters reflecting older technologies. These parameters, shown in Table 2, represent the maximum throughput of each process center. For this and all other simulations, market prices and factor costs representative of mid-1987 West Coast conditions were used (trends developed in this analysis can be applied to southern mills, only the magnitudes will differ). The average bolt diameter processed was assumed to be 14 inches, the species was Douglas-fir, and production was assumed to be 4-ply, 1/2-inch-thick, CDX plywood. The effects of each technology were simulated individually by reassigning appropriate parameter values consistent with new equipment, then comparing the results with the reference simulation. Because of bottlenecks elsewhere, equilibrium operating rates of some machines may be below these limits. Wood input was also adjusted so that none of the unchanged process centers would exceed their capacities. Next, the combined effects of all the technologies were estimated with all model parameters allowed to range up to higher limits consistent with modern equipment. Because timber costs vary with diameter and these simulations were for 14-inch Douglas-fir bolts, separate simulations were made to determine effects of processing different diameter timber in tandem with new technology. One of these simulation sets involved spindle-driven lathes; the other, spindleless lathes with prerounded bolts. Cost estimates Older mill Within the limits set by the parameters shown in Table 2, the program was simulated over a range of veneer thicknesses centered on 0.125 inch. Because of FOREST PRODUCTS JOURNAL Vol. 39, No. 1 11

the variation in veneer thickness, there is a trade-off between higher recovery from thinner peels but more undersized panels, and lower recovery but more full-sized, higher priced panels. The target peel thickness that optimized these trade-offs was found to be 0.128 inch. For the range of equipment limits assumed, the mill processing 14-inch timber and 0.128-inch-thick veneer was limited by its clipper (Table 3). Excess capacities at the debarker, lathe, dryer, layup line, and press are indicated by below-par equilibrium operating rates. Annual output, based on a 250-day workyear and two shifts per day for the green end and three shifts per day for the dry end, was 121 million ft. 2 (3/8-in. basis). Gross wood costs, assuming $260/1,000 BF (log scale) for delivered No. 2 Douglas-fir sawlogs, were $75/1,000 ft. 2 (3/8 in.) ofplywood produced (Table 4). Net wood costs, after deducting revenues for chips and cores, were calculated at $61/1,000 ft 2 Wood is the most expensive cost item, accounting for over 55 percent of variable costs. This is based on over 50 percent losses of wood in processing and veneer recovery (Fig. 1). Most wood losses occur because of the 8 percent spin-out rate; the somewhat large (5.3-in.) target core; and the 18 percent trash, clipper, and fishtail loss. Labor costs were next highest, accounting for 30 percent of variable costs, based on total compensation of $13.5/hour. For this mill, approximately 2.4 hours of labor were required per 1,000 ft. 2 of plywood. Glue represented 10 percent of total variable costs. The spread was 39 lb./msgl. Glue costs were estimated at $10.64/100 lb., based on phenol resin costs of $33/100 lb. Electricity and fuel costs were the remaining 4 percent of the total. Because only the key power users in the mill were modeled, the energy and power cost of $5/1,000 ft. 2 may be understated, by about one-half. Powered nosebar and powered backup roll The powered nosebar and backup roll combination TABLE 3. Equipment operating rates at equilibrium for various simulations. Lathe Clipper full Press load Technology Debarker Charge time Ribbon speed sheet speed Dryer free time Layup panels time/opening (fpm) (sec.) ---------------(fpm)--------------- (panels/min.) (sec.) Older equipment 45 3.0 800 375 a (%) 10 9.7 2.7 Modified by Powered nosebar and backup roll 40.6 3.0 775 375 a 12 9.7 2.4 Peripheral drive 41.6 3.0 690 375 a 2 9.7 2.4 Hydraulic positioner 44.0 2.0 725 375 a 8 9.7 2.4 Rotary clipper 47.5 3.0 910 405 4 10.1 2.0 a Press controls 43.0 3.0 800 375 a 14 9.5 3.1 All modifications 14-inch bolts 47.8 2.0 880 490 4 12.1 1.0 a 13-inch bolts 50.1 2.0 950 490 2 a 12.1 1.0 a 12-inch bolts 51.0 2.0 980 465 2 a 11.6 1.4 11-inch bolts 52.5 2.0 1,040 445 2 a 11.2 1.8 9-inch bolts 53.0 2.0 1,200 a 365 9 9.5 3.8 All modifications plus spindleless lathe 14-inch bolts 47.2 2.0 510 425 9 12.0 1.0 a 13-inch bolts 44.2 2.0 517 425 7 12.0 1.0 a 12-inch bolts 51.5 2.0 525 420 3 11.9 1.0 a 11-inch bolts 53.5 2.0 540 410 2 a 11.6 1.3 9-inch bolts 57.3 2.0 565 375 2 a 10.7 2.1 a Bottleneck. TABLE 4 Operating costs for various simulations. Revenue from Technology Wood cost Cores Chips Net wood cost Labor cost Glue cost Energy cost Total costs ----------------------------------------------------------------------------- ($/1,000 ft 2 ) a -------------------------------------------------------------------------------- Older equipment 75.4 9.0 52 61.2 33.3 10.8 47 110.0 Modified by Powered nosebar and backup roll 66.2 1.6 48 59.8 32.8 10.8 48 108.1 Peripheral drive 67.5 2.1 49 60.5 31.4 10.8 50 107 7 Hydraulic positioner 71.7 8.6 43 58.8 32.5 10.8 49 107 0 Rotary clipper 75.4 9.0 52 61.2 31.9 10.8 4.5 108 4 Press controls 73.0 8.8 50 59.2 33.6 10.7 47 108.2 Foamed gluing 75.4 9.0 52 61.2 33.4 86 47 107.9 A11 modifications 14-inch bolts 63.4 1.5 42 57.7 24.3 81 41 94 2 13-inch bolts 58.8 1.7 42 52.9 24.4 81 4.1 89 5 12-inch bolts 54.1 2.1 42 47.8 25.2 81 42 85.3 11 inch bolts 49.4 2.5 42 42.7 26.2 81 43 81 3 9 inch bolts 39.8 3.8 41 31.9 30.6 81 51 75 7 All modifications plus spindleless lathe 14-inch bolts 63.1 0.6 56 56.9 24.8 86 43 94 6 13 inch bolts 58.3 0.6 56 52.1 24.9 86 43 89 9 12 inch bolts 53.3 0.7 57 46.9 25.0 86 43 84 8 11 inch bolts 48.3 0.9 58 41.6 25.6 86 43 80 1 9 inch bolts 38.0 1.3 59 30.7 27.7 86 45 71 5 a 3/8 inch basis 12 JANUARY 1989

has three effects: 1) it reduces the incidence and size of spin-outs; 2) it allows peeling to a smaller core; aned 3) it reduces veneer loss and lathe downtime caused by sliver plug-ups. Therefore, the spin-out rate was lowered from 8 percent to 3 percent; the average spin-out diameter was reduced from 9.5 to 6.8 inches; and the target coresize was reduced fromn 5.3 to 3.25 inches. Downtime caused by clearing slivers, assumed to occur in every 15th log in the base case,m was eliminated. These parameter assumptions were based on previous mill studies and reports (8,10). When these changes were made, the annual gross margin of the mill increased by $0.33 million or $1.7/1.000 ft. 2 (Table 5). Peeling to a smaller core increased the proportion of full-sheet veneer. Since fullsheet veneer is clipped faster, clipper capacity increased by 2 percent, but the clipper remained the bottleneck and prevented the full benefit of this addition from being realized. Wood input to the lathe was reduced by 10 percent in order not to exceed the clipper capacity. Peripheral drive lathe To model a peripheral drive lathe, core size was decreased to 4 inches, plug-ups and spin-outs were elimi Rotary clipper The major benefit of rotary clippers is the increase in throughput potential. We modeled the effect of this machine by relaxing the limit if 375 fpm maximum clipping speed. We note that rotary clippers are run in excess of 500 fpm. The result was an increase in annual gross margin of $0.44 million. The bottleneck shifted from the clip- per to the presses and dryers (Table 3). Because the Figure 1. Disposition of wood by waste and product categories. nated, and dryer capacity was increased by 15 percent. These changes increased annual output capacity by 2.6 million ft 2. Because of the faster dryer throughput, one of the two dryers could be run on a 2-shift per day basis. Resulting costs per 1,000 ft. 2 were $2.3 lower and gross margins were $2.2 higher. Annual gross margin increased by $0.36 million. Hydraulic knife positioner Quicker charging and roundup, increased yield, and more uniform veneer thicknesses were the primary benefits claimed for hydraulic digital control of the knife positioner. The rapid retract rate (12 in./sec. compared to 4 in./sec. for gear box drives) and split-peel capability allowing faster roundup saves up to 1 second per cycle. Increased yield results from more efficient sheet breaking, shorter runout at the end of the ribbon, and increased fishtail recovery. To reflect the first benefit, charge time was reduced from 3 to 2 seconds. Clipper losses were recalculated to reflect veneer savings. To reflect more uniform veneer thickness, veneer thickness variation was reduced from 0.007 to 0.004 inch. These changes increased annual gross margin by $0.59 million or $3.5/1,000 ft 2. Reducing thickness variability caused fewer panels to be downgraded because of thinness, and the optimal target peel thickness fell to 0.126 inch. Losses at the clipper and fishtail saw were reduced from 16.1 to 13.9 percent. But again, because of the clipper bottleneck, the plant was unable to realize the total benefit from the improvement. TABLE 5. Operating revenues for various simulations. FOREST PRODUCTS JOURNAL Vol. 39, No 1 13

TABLE 6. Typical adhesive mixes and costs. press-loading time per panel was assumed at 2 seconds, and each dryer was limited to four decks, both work centers reached their limits at a clipper speed of only 400 fpm. Compression controls After installing two-stage pressure controls on presses, compression loss declines by about 2.5 percent. Using three-stage pressure controls, about 3.5 percent reduction in compression can be achieved (17). With a reduced compression loss, the optimum veneer thickness drops to 0.124 inch. This results in a $0.23 million/year increase in gross operating margins or about $1.8/1,000 ft 2. Foam extrusion gluing To model this technology, glue spreads were decreased by 25 percent. Adhesive cost calculations were based on formulations shown in Table 6. Adhesive waste, net of trim loss, was reduced from 5 to 0 percent. A royalty fee of $.15/MSGL was included in the cost. As a result, annual gross margin increased by $0.24 million or $2.0/ 1,000 ft 2. Since the assumed change was from a spray system, the glue cost savings were approximately 20 percent. If the change had been from a spreader system, the savings would probably have been more. From a curtain system, they would probably have been less. Combined technologies By combining the capacities of the powered nosebar, hydraulic knife positioner, and rotary clipper, the bottleneck moves to the dryers and presses. Without dryend improvements, both the clipper and the lathe have to be operated below their capabilities. To reduce the press bottleneck, the mill could install automatic loaders to shorten loading time to 1 second per opening, and that was assumed here. By adopting a high-mc gluing program, the dryer bottleneck can be reduced. Veneer moisture targets were increased to 8.5 percent, and veneer with moisture higher than 16.5 percent was assumed to be redried. A redry rate of about 16 percent resulted. Note that high-moisture gluing is only one strategy to increase dryer capacities. Alternatives include installation of dryer controls to adjust dryer speed for veneer moisture fluctuations, thereby increasing dryer efficiency; or installation of RF redriers to free existing dryers for more primary drying. The result of these changes was that, while operating revenue per 1,000 ft. 2 stayed the same, costs fell by 14 percent and throughput increased by 27 percent. Subsequently, overall annual revenues and gross mar Figure 2. Annual mill gross profits by technology and logdiameter class. gins (revenues less variable costs) rose sharply, the latter from $4.26 to $7.97 million. The presses remained the ultimate bottleneck, while the dryers were near, but not at, capacity. Conventional lathes effect of changing timber diameter Sheathing mills are increasingly utilizing smaller sawtimber grade logs instead of larger, more expensive peelers because on a cubic foot basis, smaller logs costs less. With smaller size timber there is less clear (B or higher grade) veneer; as a percentage of total block, the core represents a bigger proportion of the wood. Because there is less wood to peel, there is increased time to charge the lathe and roundup the block, both of which waste time from the peeling viewpoint. Thus, these technologies are vital to the successful peeling of small bolts. The simulations indicated that gross margins per 1,000 ft. 2 increased with decreasing timber size (Table 5). For bolts below 11 inches in diameter, however, the rise in margin was not as great as the fall in throughput, hence, total annual gross margin was lower (Fig. 2). Thus, the optimum bolt size was 11 inches. Spindleless lathes effect of changing timber diameter To simulate the effect of the spindleless lathe, charge time was reduced from 3 to 2 seconds. Because bolts were prerounded by a lathe before being fed to the main lathe, the loss of peeling time due to roundup was eliminated. Target core was reduced from 3.25 to 2.0 inches, and glue spreads increased from 25 to 27 lb./msgl to reflect the rougher veneers obtained from smaller diameters. The maximum ribbon speed was also reduced from 1,200 to 900 fpm. Since there was less idle time at the lathe, the mill ran more efficiently. The cost savings from using small timber were retained without an offsetting sacrifice in throughput (through 12 in.). Below 12 inches, drying capacity began to constrain the green end because of the increased proportion of wetter sapwood from smaller blocks. But output dropped at a slower rate than costs, so annual gross margins continued to rise with decreas- 14 JANUARY 1989

Figure 3. Potential annual mill capacity bytechnology and log diameter class. D, P, L) dignify bottleneck. ing log diameter, peaking at 9 inches (Fig. 2). Compared to conventional lathes, the spindleless lathe yielded results that were about even in terms of margin for block diameter sizes of 14 to 12 inches. Below 12 inches, however, the spindleless lathe was superior because of the reduced idle time peels. The lowest cost per 1,000 for any combination of technologies was achieved by using the spindleless lathe with 9-inch-diameter bolts. Due to net wood costs of only $3111,000 overall operating costs were over one-third lower than the overall operating costs for older equipment (Table Summary and conclusions In reviewing these results, it should be noted that the cost levels shown are based on simulations to illustrate potential directions for improvement. We believe that the indicated changes from one technology to another are reasonably accurate and consistent with the assumptions made, but it is up to users to extend this exercise within the context of their own particular mills to determine the potential benefit to them. We suggest that by installing modern equipment, a mill can gain significant benefits in increased throughput, decreased wood costs, decreased labor costs, and decreased glue costs. For example, annual production can be increased by almost one-third with 14-inch bolts without adding ditional lines, presses, or dryers (Fig. 3). By reducing waste and peeling to a smaller core size, net wood costs can be reduced from $61 to ft? Likewise, high labor efficiency can reduce labor costs $33 to $24/1,000 ft. 2 Gluing innovations can lower glue costs from $11 to $8/1,000 ft. 2 Overall gross margins increase from $35 to $51/1,000 ft. 2, and combining this with the increased output boosts annua1 gross operating margins from $4.3 to $8.0 million. The most significant. impact of modern plywood technology is that it improves the economics of bolt peeling. The combination of rapid, accurate lathe charging and small-core peeling places within the reach of plywood manufacturers a significantly cheaper resource than traditionally used. Since wood is the larg- est component of cost, this ability enhances plywood economics more than any other individual change. In examining the impact of changing block size, the time needed to charge the lathe and roundup the bolt assumes critical importance. With conventional lathes, about 2 seconds seems to be the least that is required for charging. Adding in the 2 to 3 seconds it takes to roundup the bolt means that at least a 4- to 5-second gap exists between consecutive ribbons of veneer. This unproductive lathe time leads to a bottleneck that reduces operating efficiency in the remainder of the mill for diameters below certain levels (11 in. in these simulations for conventional lathes). Annual gross margins fall with each decline in bolt diameter beyond that size (although margins per 1,000 ft. 2 continue to rise). In the case of the second-generation spindleless lathe, with a reported charge time of about 2 seconds and no roundup time due to prerounding, annual gross margins continue to increase as timber size declines to 9 inches. Similar results could be obtained with conventional lathes if bolts were prerounded. As large-diameter timber becomes increasingly scarce and expensive,the emphasiswill shift to processing smaller bolts. Much research and development effort has gone into modifying the plywood process to economically utilize smaller bolts. The result of that effort is an array of technologies that can help mills to adapt to the changing economics of structural panel markets, and to produce commodities at a competitive cost. Literature cited FOREST PRODUCTS JOURNAL Vol 39, No 1 15