Press brake capacity and the case for a fourth machine

00Summary of findings

Redbud is right that the press brake cell is the constraint. It is the only work center in the plant running above practical capacity, and it is holding back the rest of the shop.

It is not right that the cell is out of capacity. Forty-two percent of the brake cell's available hours are consumed by setup, not production. The cell runs 159 changeovers a week averaging 38 minutes each. The machine is busy; it is not producing.

That distinction changes the answer to the capital question. A fourth press brake would add 41.6 productive hours a week for $1.2 million. Reducing and resequencing setups would free 57.6 hours a week for $103,000 — more capacity, sooner, for about a twelfth of the money.

01The decision on the table

Redbud has a capital request in front of it for a fourth press brake at an installed cost of $1.2 million. The request is well founded in symptoms: brake lead times have grown from nine days to seventeen over the past four quarters, the cell is running scheduled Saturdays, and three quotes in the last quarter were declined because the promised date was unacceptable to the customer.

The question this analysis answers is narrow and specific:

Everything below is in service of that question. Where the analysis touches other parts of the business, it is because those parts bear on this decision, not because they need attention of their own.

02How this was done

Thirteen months of routing and labor-ticket data were extracted from Redbud's ERP at operation level — 41,900 completed operations across all six work centers. Setup and run time were separated using the labor ticket start and stop codes rather than the standards in the routing file, because standards describe what setup was supposed to take, not what it took.

Quote and order history came from the CRM, matched to the ERP records by job number. Cost figures are Redbud's own, taken from the standard cost roll effective January 2026, with overtime premium and premium freight pulled from the general ledger rather than estimated.

Two days were spent on the floor watching changeovers on all three brakes across both shifts. That observation is the reason this analysis reaches a different conclusion than the utilization report alone would support.

03Where the constraint is

Utilization across the six work centers is uneven, and only one cell is above the 85% threshold at which queues begin to grow faster than they clear.

This much confirms what Redbud already believed. The brake cell is the constraint, and no other work center is close enough to become the constraint if the brake is relieved — CNC machining at 76% is the next in line, and it has roughly 190 hours a week of headroom before it would matter.

The useful question is not whether the brake is the constraint. It is what the brake is doing with the hours it has.

04What the brake cell is actually doing

Three machines on two shifts give the cell 240 hours a week, or 12,000 hours a year. Separating the labor tickets into setup and run gives a picture that the 94% utilization figure conceals entirely.

Why setup is where it is

Two days of floor observation identified the causes, in order of the time they consume:

• Tooling is staged after the job arrives, not before. An operator averages nine minutes per changeover walking to and from the tooling crib. The machine is stopped for all nine.
• Jobs are sequenced by due date alone. Consecutive jobs frequently use unrelated tooling when a job two positions down the queue would have used the same setup. Roughly a quarter of changeovers are avoidable in principle.
• Programs are proved out at the machine. First-article correction runs six to fourteen minutes on the machine that is the constraint.
• No standard changeover sequence exists. Measured setup times for the same part number across operators vary from 24 to 61 minutes.

None of these are unusual, and none of them indicate a problem with the crew. They are the normal result of a shop that grew faster than its scheduling discipline — which is exactly the condition under which a capacity purchase looks obvious and is wrong.

05What the constraint costs today

The constraint's cost is not the capital request. It is what Redbud is already spending, every month, to work around it.

That is $1.62 million a year, or 4.2% of revenue. The figure is worth stating plainly because it reframes the capital request: Redbud is not deciding whether to spend $1.2 million. Redbud is already spending $1.62 million a year, and deciding how to stop.

One line deserves a caveat. The lost-contribution figure is the least certain of the three, because it assumes the declined work would have been won at normal margin had lead time not been the obstacle. Section 07 tests what happens to the recommendation if that assumption is wrong.

06Options

Four courses of action were evaluated. Each is stated in the same terms: what it costs, how many brake hours it frees per week, what share of the $135,300 monthly cost it recovers, and how long it takes to pay for itself.

Option A — buy the fourth press brake · $1,200,000

A fourth machine adds 80 hours a week of available time. At the cell's current 52% run ratio, that converts to 41.6 additional productive hours. It also adds four operators across two shifts at $185,000 a year loaded, which is netted out of the benefit above.

The machine works. It is simply the most expensive way to get the result, and it carries a sixteen-week lead time before it produces a single part. It also leaves the setup ratio untouched: the new machine would spend 42% of its life in changeover exactly as the other three do, which is why $1.2 million buys less capacity than $103,000 does.

Option B — setup reduction · $85,000

A structured SMED program across the cell: pre-staged tooling carts assigned to the schedule rather than the crib, quick-change clamping on the two highest-frequency die sets, offline program proveout, and a standard changeover sequence with the crew trained to it.

Target is a reduction in average changeover from 38 minutes to 22. That is a conservative target — the observed range already includes operators completing the same setups in 24 minutes, so 22 represents capturing existing best practice rather than inventing new practice. At the current changeover frequency this frees 42.4 hours a week.

Cost is $46,000 in tooling and carts, $22,000 in offline programming capability, and $17,000 in training and documentation.

Option C — sequence by tooling family · $18,000

Redbud's scheduler sequences the brake queue by due date. Grouping jobs that share tooling within an eight-day due-date window eliminates the changeover between them entirely. Modeled against thirteen months of actual order history, this removes 26% of changeovers with no due date pushed beyond its commitment.

Cost is configuration of the existing ERP scheduling module and two days of scheduler training. There is no new software. This option is cheap, fast, and the least disruptive of the four — it changes a sort order, not a process.

Option B + C — run together · $103,000

The two are complementary rather than overlapping: C reduces how often the cell changes over, B reduces how long each changeover takes. Together they take the cell from 159 changeovers a week at 38 minutes to 118 at 22 minutes — from 100.8 setup hours a week down to 43.2.

That frees 57.6 hours a week, or 2,881 hours a year — 38% more recovered capacity than the new machine, at 8.6% of the cost.

07How to challenge this

The recommendation rests on assumptions, and the ones that matter should be tested rather than trusted. The weakest is the lost-contribution line: it assumes Redbud can sell the freed capacity.

The bottom row is the one to hold onto. Even if Redbud wins no additional work whatsoever — if every freed hour goes unsold — eliminating the overtime and the premium freight alone repays the $103,000 in four months. The recommendation does not depend on the sales forecast being right.

08Recommendation

Sequence C first. It is cheaper, faster, and it reduces the number of changeovers that the setup reduction program then has to improve — running B first would mean optimizing changeovers that C is about to eliminate.

One caution. The measure of success is changeover minutes, not utilization. If the cell's utilization figure falls from 94% to 70% after this program, the program has worked — the machine is idle because it has caught up, not because it is starved. Managing to a utilization target after this change would recreate the batching behavior that caused the constraint.

09Scope and fee

This analysis was delivered at a fixed price agreed before the work began. No hourly meter, and the price does not change if the analysis takes longer than scoped.

Against a first-year benefit of roughly $864,000 at the planning assumption, the full engagement returns about twenty times its cost. Redbud is free to take the analysis and implement it with internal resources; the recommendation is written to be executable without us.