Illustrative sample. The plant, the weeks and the footage are synthetic.

Set up a mold-maintenance program
Maintenance

Set up a mold-maintenance program

Approved fixes get an update every week. Paused, Declined and Done do not. In this sample the weekly reports are already written, so changing a status here does not change them.

We watch how often the eight worst molds stop a press mid-run, from the press cameras and the ERP downtime log. Target: about 4 stops a week.

$54,500USD/ yr EBITDA

Extra profit a year once this is running, after what it costs to run. Profit, not revenue.

~2fewer tool stops / week

Times a week a mold quits mid-run for a broken part or a cleaning it never got. Those recovered press hours are capacity, not cash, until sales fill them.

high confidence

Capex

$18,000

Payback

4 months

Difficulty

2 / 5

Time to implement

~10 weeks

The finding

  1. This plant molds plastic parts on about 26 injection presses, and the tool that shapes each part is the mold, a precision steel block the melted plastic is forced into under high pressure.

    • Every cycle, the pressure, heat, and friction of the shot wear the steel a little, so a mold needs regular cleaning and service to keep making good parts.
    • Here those tools get cleaned only at changeover and pulled for service only after they start throwing flash (plastic squeezed out where the mold halves meet) or drag marks, which means bad parts are already being made before anyone reacts, and half the tools have no working shot counter to say when service is due.
    • Operator feedbackMolds get cleaned only at changeover, and pulled for service only after they start making bad parts.
    • Shop recordsHalf the tools have no working shot counter, so nobody knows how many cycles a mold has run since its last service.
    • Shop recordsFlash and drag marks that "come and go" on the same handful of tools, which everyone knows are bad but keep running.
    • Walkthrough videoNo spare ejector pins or springs on the shelf, so a broken pin means an emergency order and a press down for days.
    • Read togetherThe recurring tools in the scrap report and the missing shot counts are the same problem seen twice; without a count, the only trigger for service is the scrap that is already made.

  3. Run a written mold-maintenance program off shot count, not the calendar, so worn tools get serviced before they make scrap.

    • Put a working shot counter on every tool, give each mold a digital history (shots run, what was cleaned or replaced, what defects it throws), and set three tiers of service: operators clean and inspect daily, technicians service at shot milestones (say every 50K, 250K, and 500K shots, set by the steel and the resin), and an engineer reviews the defect trend each month.
    • Stock the wear parts (ejector pins, springs, gate inserts) so a break is a shelf part, and clean tools without tearing them down using dry-ice blasting (frozen CO2 pellets that knock off residue without scratching the steel).
    • Roughly $18K to stand up.
    • The mechanismA shot-count history finally tells the tool room which mold is about to make scrap, so the service happens in a planned window instead of after a week of flash; the scrap half is cash the moment a tool stops flashing.
      • The program takes back roughly a fifth of a point of tool-driven scrap, about $27K, and recovers around 500 press hours a year of unplanned mold downtime at about $55 an hour, another $27K.
      • Roughly $55K a year, for about $18K once, paid back in about four months.
    • Industry evidencePreventive-versus-reactive mold-maintenance benchmarks put the unplanned downtime recovered in the hundreds of press hours a year for a fleet this size; the downtime half is recovered press hours, capacity rather than cash until sales fill the freed time.
      • If the flash comes from an unstable process, that is the scientific-molding card's job, so pull the scrap codes first and sort them into tool faults (flash, drag) and process faults (shorts, sink).
      • Both the scientific-molding and cycle-time cards wait on this one, because you cannot tune a process or shorten a cycle on a worn, flashing tool.

The math

Scrap from worn tools removed+$27,000

0.225 points × $120,000

Press hours recovered+$27,500

500 hrs × $55/hr

Scrap plus recovered press time$54,500/ yr

$27,000 + $27,500

Scrap points from worn tools removed
points
Press hours of mold downtime recovered
hrs
Program cost (one-time)
$
Annual resin volume (lb)
Resin cost / lb
$
Value of a freed press hour
$/hr

Deep dive

    • A maintenance program removes the scrap and stoppages that come from a worn or dirty tool: flash off the parting line, drag, sink, dropped parts, a mold that lets go mid-run.
    • If those same defects are really coming from a process running so close to its limit that a small drift makes a bad part, cleaning and servicing the tool changes little.
    • The whole $55K rests on this split.
    • What it looks like on the floor if this is wrong: you clean and service a tool, flash returns within the run, and the scrap codes were shorts and sink (process faults) all along.
    • The cheap check is to pull the scrap codes first and sort tool faults from process faults.
    • Half the value is recovered mold downtime, which is capacity, not cash.
    • If the plant has backlog, those hours run sold work and become margin; if the presses are already waiting on orders, the downtime half of the saving is theoretical.
    • What it looks like on the floor if this is wrong: molds stop breaking mid-run and uptime climbs, but the extra hours sit idle because there is nothing to load onto them.
    • The whole program runs off shot counts.
    • If half the tools have no counter or a dead one, the shot-milestone triggers have nothing to fire on and the program slides back to a calendar, which services the tool that sat and misses the one that is worn out.
    • What it looks like on the floor if this is wrong: nobody can tell you how many cycles a mold has run since its last service, so the intervals get guessed.
    • Preventive service is hours the tool room is not spending today.
    • If it is already swamped reacting to breakdowns, the PM will keep losing to the emergency in front of it unless the work is scheduled, protected, or partly outsourced.
    • What it looks like on the floor if this is wrong: the maintenance schedule exists on paper but the day fills with breakdowns and the PM slots never get run.
Load-bearing
The card breaks if this is wrong.
Supportive
The card weakens but survives if this is wrong.
  • Risk

    A maintenance program removes the scrap and downtime that come from worn or dirty tools. If the defects are really coming from an unstable process, cleaning and servicing the tool does not remove them.

    Early warning

    Flash or shorts return within a run of a freshly serviced tool, and the scrap codes are dominated by process faults (shorts, sink) rather than tool faults (flash, drag).

    Mitigation

    Sort the scrap codes before you start, and hand the process-driven scrap to the scientific-molding card.

    At stake

    Most of the scrap half of the value, about $27K.

  • Risk

    Half the value is recovered mold downtime. If the plant has no backlog to load onto the freed presses, those hours stay idle and never become margin.

    Early warning

    Molds stop breaking mid-run and uptime climbs, but the schedule has gaps and the extra hours run nothing.

    Mitigation

    Treat the downtime half as capacity, not cash, and pair it with a sales plan to fill the hours; bank the scrap half, which does not depend on demand.

    At stake

    The downtime half of the value, about $27K.

  • Risk

    Shot-count logic only works if service fires off the counter. Left to memory, PM slides to changeover-only cleaning and the worn-tool scrap creeps back.

    Early warning

    Overdue PMs pile up, tools get cleaned only at changeover, and nobody can name the current shot count on a tool.

    Mitigation

    Automate the triggers off the shot counter and tie a monthly supervisor review to overdue-PM count.

    At stake

    The whole $55K erodes back toward the starting point over a year or two.

  • Risk

    Without wear-part spares stocked, a broken ejector pin or spring means an emergency order and a high-runner press sitting idle for days while the part ships.

    Early warning

    Emergency spare-part orders, jobs bumped off the schedule, and a press down waiting on a $40 pin.

    Mitigation

    Min/max stock the common wear parts on the top tools so a break is a shelf part, not an order.

    At stake

    Days of lost press time per event, plus late shipments to customers.

A. Why "when it breaks" is the expensive way to run a tool room

A tool that only gets attention when it fails has already made a run of scrap and idled a press before anyone touches it. An emergency repair costs several times the same repair done on schedule (you pay for the rush, the overtime, and the bumped jobs), and the run that was in the press when the mold let go is usually scrap. A shot-count program moves the work to before the failure: you service the tool while it is still making good parts, on a schedule tied to how hard it has actually been run.

B. Shot count, not the calendar

Two tools both "serviced in March" can be in completely different shape by June if one ran three shifts and the other sat. Calendar maintenance services the tool that sat and misses the one that is worn out. A shot counter measures the thing that actually wears a tool, cycles, and triggers service off that. Set the milestones by the steel and the resin: a hardened tool running an unfilled commodity resin goes far between services; a softer tool running a glass-filled resin (the glass fiber is abrasive and eats steel) needs attention much sooner.

C. The three tiers, and who owns each

  • Operator, daily: clean the faces, check the vents and the ejection, wipe and lube the moving parts. Catches the small stuff before it makes scrap.
  • Technician, at shot milestones: pull the tool, clean it properly (dry-ice blasting cleans without scratching the steel), inspect and replace wear parts, check the parting line. This is where flash and drag actually get fixed.
  • Engineer, monthly: read the defect trend on each tool and decide what to rebuild, what to replace, and where the intervals are wrong.

The history record ties it together. Once you know a tool's shot count and what it throws, you can set intervals from data instead of guesswork, and make a rebuild-versus-replace call before the tool fails, not after.

D. Why this is the precondition, spelled out

You cannot scientifically mold or shorten a cycle on a worn, flashing tool. A cavity-pressure sensor (which reads the plastic pressure inside the mold) only means something if the same tool makes the same shot twice, and a worn tool does not. That is why this card lands before scientific molding and cycle time: it is the cheapest one here and it is what makes the expensive ones possible. And restoring a tool to its approved condition is maintenance, not a process change, so it does not trip a customer re-validation (PPAP) and can start immediately, on every tool, without waiting on anyone's sign-off.

The records we did not have for this read and would ask for on day one. Each one replaces an estimate on this card with your own number.

  • Owner
    Quality manager and tool room lead
    Why

    To confirm how much of the scrap and unplanned downtime is tool-driven (flash, drag, dropped parts) versus process-driven, and to rank which tools carry the pain. This one dataset decides whether the card is worth building and which tools to start on.

  • Owner
    Tool room
    Why

    To know which molds have working counters and which need one, and to set the first service intervals. Without shot counts the program has nothing to trigger off.

  • Owner
    Tool room and purchasing
    Why

    To size the starter spares shelf and see how much is going to emergency repairs today, which is the money the program takes back.

  • Owner
    Whoever owns the ERP/MES
    Why

    To find out whether the tool-history and PM triggers can run on software already paid for, the cheapest path, before buying a CMMS or ProShop.