Lugs per Minute and Pieces per Minute: Measuring Sawmill Throughput

Lugs per minute says how fast a lugged chain can present places for boards, and pieces per minute says how many boards actually arrive; the gap between the two, and the machine that causes it, is what a sawmill throughput study measures. UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for aerospace and heavy industry from its Spokane Valley, WA facility, integrating Allen-Bradley PLC and motion control with in-house CNC machining, heat treating, and stress relief. This article gives a method, not a table of rates: it defines the measures, shows how lug fill and running time turn a lug rate into a piece count, explains how a bottleneck is located, and works through the one published, neutral lug-rate study it could find, a single British Columbia mill, whose figures apply to that mill only. It closes with the counting and control logic that produces the data, and with where the build chain, from design and machining through drives, controls, tuning, and monitoring, shows up in the numbers.

What do lugs per minute and pieces per minute measure?

The two terms measure different things:

• Lugs per minute is the rate at which lugs, the attachments that each hold one board, pass a fixed point on a lugged chain.
• Pieces per minute is the rate at which boards actually pass that point.

For a chain running at a steady speed, lugs per minute is the chain speed divided by the lug spacing (this article's definition and arithmetic). Thoews, Maness, and Ristea, in their simulation of one British Columbia sawmill, set the trimmer's process time from its associated conveyor speed, as it is an inline process. They note that the process times of a machine center are often equal to the rate of the conveyor speed, and that variability occurs where there are manual set-up or alignment times within the process.

Their paper states the same trimmer settings in m/s, ft/min, and lugs per minute: the base case is 0.20 m/s (40 ft/min), later described as 40 lugs per minute. It never states the lug spacing, and this article does not derive one from that equivalence. A lug rate quoted without a lug spacing, or a speed quoted without the rate, is not yet a complete specification (Thoews, Maness, Ristea 2008, pp. 229–242).

How is lug fill defined and measured?

Thoews, Maness, and Ristea define lug loader utilization as the percentage of full lugs, containing a board, from the total number of lugs of the lug loader conveyor, throughout one shift. It is the fill fraction that converts lugs into pieces. In their base case:

• Trimmer lug loader conveyor: utilized 92.59% on average (reported as 93% in their Table 2).
• Trimmer operator: busy 97.56% of the time on average (98% in Table 2), with busy defined as the time the operator was physically taking a board and placing it into the trimmer lug loader.
• Large buffer: the average volume of material in the large buffer leading into the trimmer area was 5.72 m³ (2,422 fbm).
• Edger: utilized 96% on average, defined as the percentage of time the edger had a board or cant in process.

Each utilization is a ratio with a stated numerator and denominator, and the buffer figure is an average volume. As engineering reasoning, each can be measured on a running line: lug fill is counted by comparing a board-presence signal with a lug count over the same shift, and an operator or machine utilization needs a defined busy state, as the paper gives, before it can be logged (Thoews, Maness, Ristea 2008, pp. 229–242, Table 2).

How do lugs per minute, fill, and running time combine into pieces per shift?

Three numbers set the piece count at a lugged machine center:

pieces per shift = lugs per minute × lug fill × running minutes per shift

(this article's formula). Using the Thoews paper's trimmer figures, this article's arithmetic gives:

• Base case: 40 lugs/min × 0.93 = 37.2 pieces per running minute.
• Scenario 2, lug loader at 50 lugs/min: 50 lugs/min × 0.89 = 44.5 pieces per running minute, about 20% more.

The paper's simulated boards out rose from 13,148 to 14,019 per shift between those two cases, an increase of 6.6%. The arithmetic does not close on the rate and fill alone. The paper does not report running minutes per scenario, and the boards-out total and the trimmer's lug fill are separate outputs, and this article cannot allocate the gap between the 20% and the 6.6%. As engineering reasoning, a lug rate is a capacity, and a mill that plans on it without measuring fill and running time can overstate what a faster chain delivers. In the Thoews model, the faster trimmer lug loader also drew down the large buffer that fed it, from 5.72 to 3.93 m³, as the next answers show (Thoews, Maness, Ristea 2008, pp. 229–242, Table 2).

What is a bottleneck, and how did one mill's simulation find it?

Thoews, Maness, and Ristea define bottlenecks as processes whose limited capacities reduce the capacity of the whole system. Their abstract states that sawmills often invest in a new machine center and then find that the processing bottleneck just moves somewhere else, and their model was built to test changes on the whole system. Their method, in the order they report it:

• Data. Time and motion studies at machine centers with variable set-up or process times, with at least 30 times measured for each time element; downtimes from operator log books, plus short downtimes collected by observation, since operators did not record them, which was particularly important at the trimmer with its frequent short failures; conveyor lengths from mill drawings and conveyor speeds measured; buffer capacities estimated by mill personnel.
• Model. A discrete event simulation of the mill's lumber production process, including both breakdown lines, the resaw, the edger, and the trimmer, built in Arena.
• Validation. Among other tests, logs processed per minute in the model were compared with 25 random headrig shift reports and 30 random canter shift reports, and the authors found no statistically significant difference.

The diagnosis came from buffer levels and utilizations. With both lines running, the trimmer was the bottleneck; with only the canter or only the headrig running, the trimmer was no longer a bottleneck, because the buffer was no longer at its maximum capacity, and the mill was constrained by the primary breakdown. In engineering terms, a full buffer ahead of a machine with high utilization marks the constraint, and an empty buffer with falling utilization marks a machine that is waiting on something upstream (Thoews, Maness, Ristea 2008, pp. 229–242).

What happened when that mill's model raised the trimmer from 40 to 50 lugs per minute?

The five trimmer scenarios in the Thoews study form a worked case of the method. The figures below are simulated results for one mill:

Scenario	Trimmer change	Boards out per shift	Trimmer conveyor utilization	Large buffer volume
1 (base)	0.20 m/s (40 ft/min)	13,148	93%	5.72 m³
2	0.25 m/s (50 ft/min)	14,019	89%	3.93 m³
3	0.25 m/s and 10% less trimmer downtime	14,280	81%	0.97 m³
4	0.41 m/s (80 ft/min) and 10% less downtime	14,693	66%	0.66 m³
5	Scenario 4 and edger feed times halved	17,294	85%	3.19 m³

The authors found a significant difference between scenarios 1 and 2: when the feed speed was increased from 40 to 50 lugs per minute, average total board output rose 6.6%, a figure the paper later rounds to about 7%. The authors describe 50 ft/min as the maximum speed at which the trimmer operator could physically operate the machine, and the authors conclude that a second employee would be needed at that rate to keep boards flowing to the operator. In their words, experienced trimmer operators can manage speeds of about 0.20 m/s (40 ft/min), while 0.25 m/s (50 ft/min) will result in additional operator errors, missed boards, and increased machine-center downtime from boards crossing up and blocking the conveyor. The mill later raised the trimmer lug loader to 0.25 m/s and reduced some trimmer downtime, and its actual output rose 10%, slightly more than the 8.6% the model predicted for scenario 3 (Thoews, Maness, Ristea 2008, pp. 229–242, Tables 2–3).

Why did faster lugs stop paying off in that mill?

Between scenarios 2, 3, and 4, the Welch confidence intervals for the change in boards out included zero, and the authors found no significant difference. On subsequent analysis they attributed this to the bottleneck of the whole system shifting from the trimmer back to the edger:

• In scenario 4, with the trimmer at 0.41 m/s, the buffer ahead of the trimmer was 9% full on average and trimmer utilization was 66%, while edger utilization was still 92%, indicating that the bottleneck had moved to the edger.
• In scenario 5, with edger process time halved, the edger fed the main buffer faster, the buffer rose to 3.19 m³ (1,353 fbm), the trimmer was busy 88% of the time, and the bottleneck moved back to the trimmer.

The authors conclude that improvements to both the edger and the trimmer must be made at the same time to balance the flow between machine centers.

The paper contains three internal slips worth knowing before reusing it. Its discussion credits scenario 4 with an increase in edger speed to 0.41 m/s and about 30% more output, but scenario 4 is defined as a trimmer speed of 0.41 m/s, and its Table 2 gain is 11.8% (14,693 ÷ 13,148, this article's arithmetic); the 31.5% in its conclusions belongs to scenario 5. It gives 6.6% and "about 7%" for the same scenario 2 result. It also uses ft/min and lugs per minute as if they were the same unit, defining scenarios 1 and 2 at 40 and 50 ft/min and reporting the same change as 40 to 50 lugs per minute, without ever stating a lug spacing, so neither figure can be converted to the other from the paper (Thoews, Maness, Ristea 2008, pp. 229–242, Tables 2–3).

What did the optimizer find, and under what constraint?

The authors then used OptQuest for Arena to search for the best combination of trimmer and edger settings, with trimmer downtime reduced by 10% as before. The optimizer had four control variables:

• edger board process speed, ranged from 2.44 to 4.88 m/s;
• edger cant process speed, ranged from 0.76 to 1.52 m/s;
• waterfall conveyor speed, ranged from 40 to 80 lugs/min;
• trimmer lug loader speed, ranged from 40 to 80 lugs/min.

The minimums were the mill's current settings, the maximums were twice those, and a constraint held trimmer utilization at 80% or more, chosen because faster trimmer lug speeds had lowered trimmer utilization in the scenario analysis. The optimum found was a trimmer conveyor speed of 70 lugs/min with the waterfall at 40 (16,708 boards per shift, 27% over the base case) when the edger's current feed and align times were kept, and 78 lugs/min with the waterfall at 67 (17,996 boards, 37% over the base case) when those times were assumed to be part of the edger's process time. These are model optima for one mill's raw material and product profile. The authors caution that a change in actual raw-material inputs, such as a shift to larger logs of one species, would invalidate the use of the empirical data in the model. As this article reads them, they are not recommended rates for another mill (Thoews, Maness, Ristea 2008, pp. 229–242, Tables 4–5).

How does a design rate in logs per minute size the rest of a mill?

A throughput target can also be set at the front of the mill and carried downstream. In a 1979 USDA Forest Products Laboratory design and investment model for two small-log sawmills, Harpole, Williston, and Hallock assumed that logs are sorted by length to reach an average feed rate of six logs per minute at the quad-band headrig. They state that this implies 8 per minute for 8 ft logs and 4 per minute for 18 and 20 ft logs, and a 50 percent increase in headrig productivity compared with unsorted log processing. The model then scales dry kiln capacities and all items of processing equipment to accommodate those log-flow requirements, on a two-shift, 250-day-per-year operating basis.

Two of its handling choices address flow limits directly:

• Trimmer bypass. A trimmer bypass system limits the number of pieces handled by the trimmerman and minimizes the wasteful need for "slashing" to clear a flooded trimmer.
• Restacking. In both model mills, lumber is unstacked after kiln drying to remove stickers and restacked as a solid package; in the conventional mill, the paper says, this facilitates the feed rate required by high planermill speeds.

These rates are 1979 design assumptions for the two model mills of one study, not measurements, and not current or typical rates. As engineering reasoning, the method is what carries over: set a rate at the constraint, then size every conveyor, buffer, and machine to that flow (Harpole, Williston, Hallock 1979, Research Paper FPL 310, pp. 2–3).

Should throughput be counted in pieces, volume, or value?

Piece counts are one of three measures, and a mill that counts only pieces can miss the other two:

• Pieces. The Thoews model reported boards out per shift, and the mill's actual result was reported as a 10% increase in lumber volume output.
• Volume and recovery. The same authors validated their headrig logic with the lumber recovery factor, defined as the volume of lumber cut from a log in board feet divided by the log's volume in cubic meters.
• Value. The unscrambler and lug loader article covers the hardwood research on scanner-based edging and trimming. It includes the earlier study, summarized by Regalado, Kline, and Araman, which found manually operated edging and trimming in three hardwood mills reaching 62 to 78 percent of optimum value.

As engineering reasoning, a throughput study reports all three: a trimmer that runs more pieces per minute can still lose value on each piece, and a change that raises pieces per minute can lower recovery or value if operators or scanners have less time per board. The Thoews authors report that the trimmer operator has only a few seconds to flip each board, evaluate the trim, and set up the machine (Thoews, Maness, Ristea 2008, pp. 229–242; Regalado, Kline, Araman 1992, pp. 29–34).

What sensing, counting, and PLC logic measure lugs and pieces on the line?

Every figure in a throughput study has to come from a counter, and the counters can be built into the line's controls. At a functional level:

• Lug count. An encoder or proximity target on the lug-chain drive counts lugs and gives chain speed; lugs per minute is the count over a clock interval.
• Piece count and fill. A photoelectric or laser board-presence sensor at the lug loader discharge, read against the lug count, gives full lugs and empty lugs for each shift, the ratio the Thoews paper calls lug loader utilization.
• Downtime logging. The Thoews authors collected short downtimes by observation; operators did not record them. As engineering reasoning, a PLC that time-stamps every stop and restart of each machine center records those short stops automatically.
• Buffer level. As engineering reasoning, presence or level sensing in the buffers gives a running line the buffer-level data that, in the Thoews model, located the bottleneck.
• Task timing. Logix 5000 controller tasks can be configured as continuous, periodic, or event, and a periodic task performs a function at a specific time interval; as a design choice, rate and fill counters can be accumulated in a periodic task.

UTEC Industrial integrates Allen-Bradley ControlLogix and CompactLogix controllers with PanelView and FactoryTalk operator interfaces and EtherNet/IP networks, the platform on which these counts can be displayed and logged by shift (Thoews, Maness, Ristea 2008, pp. 229–242; Rockwell Automation 1756-RM094N-EN-P-2025, Ch. 5 pp. 39 and 41).

Where does the build chain show up in lugs per minute?

A line's measured throughput is the product of decisions made along the build chain:

• Design and engineering fix lug spacing, chain speed range, buffer sizes, and the rate each machine center is built for. The Thoews optimizer ranged the trimmer lug loader and the waterfall conveyor up to twice their current speeds, and, as engineering reasoning, a design basis that leaves no speed margin rules such a change out.
• Parts machining, fabrication, weld fatigue, and stress relief decide whether a multi-strand lug chain keeps its lugs square across the strands. Crossed and blocked boards were one of the failure modes the Thoews authors expected at 50 ft/min; as engineering reasoning, lugs that are out of square across the strands add a mechanical cause of the same fault.
• Drives and tuning set how smoothly the chain holds speed. On a Kinetix 5700 servo axis, Rockwell Automation notes that actual bandwidth values depend on the application and can require adjustment once motor and load are connected.
• Controls and monitoring produce the lug, piece, and downtime counts above, and trending them by shift shows whether a change moved the bottleneck.

The sawmill material-flow article places the trimmer and its feed conveyors in the mill's full sequence from log yard to planer (Thoews, Maness, Ristea 2008, pp. 229–242, Table 4; Rockwell Automation 2198-UM002E-EN-P, Kinetix 5700, p. 211).

What should a mill measure before specifying a faster line?

Before asking for a faster machine center, a mill can measure by shift, on the line as it runs today, the quantities the Thoews study used as model inputs or reported as outputs:

• Rates: lugs per minute at each lugged conveyor, and the lug spacing that goes with it, which the Thoews paper does not state.
• Fill: full lugs over total lugs at each lug loader, over at least one shift.
• Utilization: busy time at each machine center and operator station, with the busy state defined first.
• Buffers: the volume or count held ahead of each machine center, and how often each buffer is full or empty.
• Downtime: every stop, including the short ones operators do not log.
• Logs per minute at each breakdown line, the measure the Thoews authors used to validate their model against shift reports.

With those numbers, the constraint can be located before money is spent on a machine that, as the Thoews authors warn, may only move the bottleneck somewhere else. UTEC Industrial performs factory acceptance testing and on-site commissioning, where a new conveyor's lug rate, fill counting, and stop logging can be demonstrated against the purchase order (Thoews, Maness, Ristea 2008, pp. 229–242).

Related Articles

• Unscramblers, Even-Enders, and Lug Loaders Explained — the lug loaders whose fill rate sets throughput
• Sawmill Material Flow from Log Yard to Planer: Handling Machines — the full mill flow in which a trimmer bottleneck sits
• Material Handling in Lumber and Wood-Products Mills — the lumber-mill view of handling equipment and controls
• Machine Vision in Material Handling: What It Does and How It Works — how vision identifies and measures boards and parts on a line
• Sawmill Chain Transfers: Strand Count, Pitch, and Speed — chain pitch, speed and strand matching behind lug rates

References

• Thoews SE, Maness TC, Ristea C (2008). "Using flow simulation as a decision tool for improvements in sawmill productivity." Maderas. Ciencia y Tecnología, 10(3), 229-242.
• Harpole, G.B., Williston, E., Hallock, H.H. Investment Opportunity: The FPL EGAR Lumber Manufacturing System, Research Paper FPL 310. USDA Forest Service, Forest Products Laboratory, 1979.
• Regalado C, Kline DE, Araman PA (1992). "Value of defect information in automated hardwood edger and trimmer systems." Forest Products Journal, 42(3), 29-34.
• Rockwell Automation 1756-RM094N-EN-P-2025: Logix 5000 Controllers Design Considerations. Rockwell Automation, 2025.
• Rockwell Automation 2198-UM002E-EN-P (2018): Kinetix 5700 Servo Drives User Manual. Rockwell Automation, 2018.

---

Ready to Discuss a Material Handling System?

UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for heavy industry, from the stress-relieved structure and drives to the Allen-Bradley PLC controls, tuning, and monitoring that run them, at its Spokane Valley, WA facility. Send UTEC the application, loads, and duty cycle to start a system review.

Request a Quote →

Questions? Call (509) 922-1832 or email sales@utec.co