Mold Flow Analysis: When It Pays in Mold Design

Trial 6 of a 16-cavity connector program stopped at 97% fill with cavities 9 and 10 still short-shot. The toolmaker had re-cut two gates, the buyer had spent $4,600 on press time, and no one had modeled the runner before steel.

Mold flow analysis is computer-aided engineering that simulates how resin fills, packs and cools inside a cavity. Software in the Autodesk Moldflow, Moldex3D and Cadmould class solves the flow equations on a mesh of the part. Output shows where the melt goes, where pressure peaks, where weld lines form, and how the part will shrink and warp. This guide is for the buyer who must decide whether a $500 to $3,000 analysis line item earns its place in the tooling budget.

It covers what each analysis type predicts, when the fee pays for itself, and what accuracy you can trust. It also covers when to skip CAE altogether. Values are typical industry figures from toolmaker and software vendor literature. Where a number depends on your resin or machine, verify it against your own datasheet. No vendor benchmark survives contact with your real material card.

The Snapshot

• A fill or flow study on an 8 to 16 cavity tool typically costs $500 to $1,500 and returns in 3 to 5 days.
• Programs that run CAE before steel cutting report 1 to 3 fewer mold trials, with each trial worth $800 to $2,000 in press time and labor.
• Weld line and air trap positions usually land within 5 to 10 mm of the simulation on unfilled engineering resins.
• Packing pressure available at the last-filled cavity runs 30 to 60% of peak injection pressure, and that window is the real budget for sink control.
• Cooling analysis targets ±3 to 5 °C cavity uniformity, and every 5 to 10 °C recovered from the cooling layout shortens cycle time.

When Mold Flow Analysis Pays for Itself

Analysis earns its fee when a wrong guess is expensive and hard to reverse. Four program profiles qualify. The first is a multi-cavity tool above 8 cavities, where one imbalance starves every cavity fed by the same runner. The second is an engineering resin such as PA66, PBT or LCP, which shrinks 0.5 to 2.0% and forms weak weld lines.

The third profile is a warpage-critical part holding flatness under 0.10 mm across a large area. The fourth is thin-wall geometry below 1.0 mm, where required fill pressure climbs fast. On those programs one avoided trial covers the study fee. A typical trial costs $800 to $2,000 once you add press time, resin, labor and inspection.

The arithmetic is direct. A fill or flow study runs $500 to $1,500, and avoiding a single $1,500 trial pays for it outright. Avoiding two pays twice over. Steel changes made after the trial cost more still.

Welding a new gate into a hardened cavity runs 2 to 4 times the analysis fee, plus 3 to 7 days of schedule. CAE therefore pays twice: once by removing trials, once by keeping gate decisions on the drawing board. Gate position decides weld line location, sink marks and warpage direction. That is why our guide on 7 gate design best practices starts with fill logic rather than gate size.

Connector housings illustrate the payoff clearly. An LCP 16-cavity tool runs melt at 320 to 350 °C and fills in 0.5 to 1.5 seconds. Fill errors in that regime cost real money, because LCP weld lines and flash windows are narrow. A flow study on such a program is routine among automotive tier suppliers. Buyers on 64-cavity tools report the same pattern: gate symmetry decided on screen, not on the press.

What Simulation Predicts: Fill, Pack, Cool, Warp

A fill analysis shows the melt front advancing through the cavity as colored time steps. You see which region fills last, where flow hesitates, and where two fronts meet to form a weld line. You also see where air has no escape path. Fill time on a typical housing lands within 5 to 15% of the measured trial value for an unfilled resin.

A flow and pack study adds pressure, shear and shrinkage data. Injection pressure climbs through fill and peaks near the end of the shot. Packing then holds pressure while the gate freezes. Useful packing pressure at the cavity runs 30 to 60% of peak injection pressure at the screw. That window decides how much sink and shrink you can correct.

A cooling analysis models the mold steel and water circuits, not just the part. It predicts cavity surface temperature, hot spots and the temperature spread across the part. Well-designed circuits hold uniformity to ±3 to 5 °C. Each 5 to 10 °C you recover from the cooling layout shortens cooling time by 5 to 15%.

Warpage analysis couples fill, pack and cooling results, then predicts the shape after ejection. Warpage comes mainly from shrinkage mismatch. Thick and thin sections shrink at different rates, and flow direction shrinks differently from cross-flow. Simulation ranks each contributor and tests fixes such as moving the gate or re-routing a circuit.

Filled resins add orientation effects that unfilled grades skip. Glass fibers align with flow during fill, so shrinkage differs between flow and cross-flow directions by 0.1 to 0.4% in common grades. Warpage results that ignore orientation under-predict twist in long flat parts. Ask whether the material card includes fiber orientation data before you trust a deflection map.

Analysis Types: Fill, Flow, Cooling, Warpage, Shrinkage

The five analysis types build like a pyramid, and you can stop at any level. Fill analysis is the cheapest and answers gate count and position. Flow and pack analysis answers pressure, shear and shrinkage distribution. Cooling and warpage analysis sit on top and need the lower results first.

← swipe to scroll →

Table: typical published fee ranges for a single iteration on a mid-size part. Multi-cavity tools and 3D meshes run higher. Verify against your supplier's quote.

Resin shear rate limits deserve their own line. Every grade has a window, commonly 40,000 to 100,000 s-1, above which the polymer degrades and burns. Gate size is often set by this limit, since a gate that is too small pushes shear past the ceiling. Flow analysis shows shear rate along the whole runner, gate and cavity.

Pick the analysis type by risk, not by vendor menu. A thin-wall connector housing gains most from fill and flow studies. A flat cover with a sealing bead needs cooling and warpage work. Requesting full warpage on a simple tray often overspends by $1,000 or more.

Model and Inputs: Mesh, Material Card, Process Settings

Three inputs set simulation quality before the first run: the mesh, the material card and the process settings. A mesh divides the part geometry into elements for the solver. Mesh density decides what the solver can see. A coarse mesh misses thin ribs and hesitation effects. A fine mesh resolves them but multiplies run time.

Two mesh families dominate practice. Dual-domain and midplane meshes treat the wall as a shell, which suits parts up to roughly 3 to 4 mm wall thickness. 3D tetrahedral meshes fill the volume and fit thick walls, bosses and flow around inserts. A connector housing typically runs 200,000 to 2 million elements depending on mesh type.

The material card often matters more than the mesh. Cards come from two sources: the resin datasheet, published for marketing, and measured rheology, run on a specific lot. Measured viscosity curves track real behavior far better than datasheet estimates. Resin lots vary, and regrind shifts viscosity by 10 to 30% at common regrind levels.

Ask which card version was used and when it was built. An old card can misstate pressure by 10 to 20% on a sensitive grade. Process settings are the third input: melt temperature, mold temperature, injection time and packing profile. PA66 housings run melt at 270 to 300 °C and steel at 60 to 90 °C. Feed the analysis the window you will actually run, not the widest datasheet range.

Runner and gate geometry is meshed with the part in the same model. A hot runner system adds manifold and nozzle drops to the pressure budget. Each drop adds 50 to 300 bar of pressure loss depending on size and flow length. Cold runners add sprue and runner volume that must freeze before ejection, which lengthens cycle time.

Reading Results: Gate Imbalance, Pressure Drop, Air Traps, Weld Lines

The report arrives as color plots plus a written summary. Four plots decide most gate and runner decisions. The fill time plot shows last-filled regions in red. The pressure plot at end of fill shows total drop from screw tip to cavity. The weld line plot marks every meeting front, and the air trap plot flags volumes with no escape.

Runner balance is the most common multi-cavity question. A naturally balanced runner feeds equal flow length to every cavity. An artificially balanced runner tunes diameters instead. Simulation reports the residual spread as a fill time difference. Good practice holds cavity-to-cavity imbalance within 2 to 3%.

Beyond 3%, the fastest cavity overpacks while the slowest one short-shots. Runner geometry is the usual fix, and our guide to balancing multi-cavity molds walks through the options. Pressure drop tells you whether the mold fits the machine. Add sprue, runner and gate losses, then compare with the press. Typical injection machines deliver 1,400 to 2,500 bar of maximum injection pressure.

If the model needs more than 80 to 90% of that pressure at end of fill, the gate, runner or wall must change. Weld lines are the result buyers care about most. Simulation places them where fronts meet, and on unfilled resins the position usually lands within 5 to 10 mm of reality. Glass-filled grades keep only 30 to 50% of base strength at the weld. Air traps flagged by the model show where vents must go, and our guide to mold venting covers vent depth rules for each resin family.

Sink marks hide in thick sections even when fill looks clean. The pack plot shows volumetric shrinkage by region. Rib bases and boss walls above 1.5 to 2.0 times nominal wall are the usual suspects. Simulation shows whether packing pressure at 30 to 60% of injection can feed them before gate freeze. Gate freeze timing decides how much packing reaches the far cavity.

Accuracy Limits: What Simulation Cannot Promise

Simulation predicts trends reliably and absolute values loosely. Treat pressure, temperature and deflection outputs as comparisons between options. Deviation between predicted and measured deflection of 20 to 50% is common on complex parts. The ranking of two gate options usually survives even when the numbers do not.

Four limits explain most of the gap. Real molds deflect under pressure, while many models treat steel as rigid. Cooling continues after the modeled cycle stops. The viscosity in the card represents one lot, while the press runs another. Venting, gas, mold surface roughness and small geometry are simplified or ignored.

Shrinkage deserves special caution. Material cards give a shrink range, typically 0.3 to 2.0% depending on grade. Real shrinkage depends on orientation, wall thickness and packing history in your specific part. Two parts in the same resin can shrink differently. Treat the shrinkage map as a starting point for cavity compensation, then verify on trial 1.

Mesh type sets one more ceiling. Shell models cannot resolve thick-to-thin transitions or flow around ribs. 3D models capture them but depend on mesh density and material data. Thin-wall parts below 1.0 mm carry wider error bands because shear heating and hesitation dominate. Our guide to avoiding warpage in thin-wall molds shows the practical fixes that follow a study.

Validation is how good shops close the accuracy gap. They compare predicted fill time and pressure against trial 1 data, then recalibrate the card for the next run. Published case studies show prediction error falling from 15 to 25% down to 5 to 10% after one calibration cycle. Ask your mold maker for their calibration history before you pay for iteration 2.

When to Skip Analysis

CAE is a tool, not a ritual, and several program profiles gain little from it. A single-cavity or 2-cavity mold in a commodity resin such as PP, PE or ABS rarely needs simulation. Simple two-plate tools in those resins commonly reach a usable part within 1 to 2 trials. Spend the money on steel quality or a spare cavity instead.

Tolerance is the second filter. Parts holding ±0.2 mm or looser, with no sealing or fit surface tied to flatness, rarely repay warpage analysis. At ±0.1 mm or tighter the study starts paying. Flatness-critical surfaces shift the balance fastest.

A moving design kills the value of any run. Each gate move or wall change voids the previous result, and each re-run costs $500 to $3,000 plus 3 to 10 days. Wait until DFM settles the geometry before spending. Also skip when the material is missing from the database. Output quality drops with input quality.

Experience is the final filter. If your mold maker built the identical family tool last year with the same resin, the new tool is a copy exercise. Ask one question before approving any add-on. What decision will this analysis actually change? With no answer, skip.

There is a middle path between full CAE and blind trial. A DFM review with a molding-risk register catches most geometry traps in 1 to 3 days and often at no fee. Scientific molding at trial, with pressure curves read against predicted values, substitutes for a full study on simple tools. Both options cost less than the $500 to $3,000 study and cover most low-risk programs.

How to Request CAE on Your RFQ

When analysis is justified, specify it like any other deliverable. Name the acceptable software class: Autodesk Moldflow, Moldex3D or Cadmould. Name the material card source and ask for the card vintage. Name the process window in melt and mold temperatures. Ask for a stated number of iterations, since the first run always forces a change.

List the deliverables you want to see. Typical requests cover the fill time plot, end-of-fill pressure, weld line and air trap plots, cooling map and deflection map. Ask for a one-page summary stating the recommended gate position, expected fill pressure, imbalance percentage and top three risks. You are paying for a decision aid, not a decoration.

Schedule the study before steel cutting. A typical cycle runs 3 to 10 days from CAD to final report. On a 2 to 5 week mold build the slot fits between DFM sign-off and cavity steel start. If the mold maker proposes analysis at the same time as the trial, the timing is already wrong.

Who runs the study matters as much as which software. Ask how many analyses the engineer completed in the last year. A mold maker running CAE weekly reads results differently from one opening the software twice a year. DieStrike reviews gate and fill risk inside its mold design and DFM service before your RFQ becomes steel.

Plan for the result to change the design. If the study moves a gate or adds a vent, the change needs a revision loop in the contract. State who approves gate moves, who pays for the re-run, and who owns the updated drawing. A change clause of one included iteration plus paid re-runs keeps the budget honest. Without it, iteration 2 arrives as a surprise invoice.

Cost and Schedule Impact

Fee ranges are stable across toolmakers because the work is similar. A fill-only study on a mid-size part runs $500 to $1,200. Adding flow and pack takes it to $800 to $1,800. Cooling and warpage push a full study to $1,500 to $3,000. 3D meshes on multi-cavity tools run higher.

← swipe to scroll →

Table: typical published ranges for a single mid-size part study with two to three iterations. Verify fees against your supplier's quote.

Duration depends on mesh, iterations and queue. Model cleanup and meshing take 1 to 3 days on typical CAD. Each iteration adds 1 to 2 days, and useful studies run 2 to 3 iterations. Confirm the delivery date in writing before you pay.

Compare the fee against the alternative. A trial day runs $800 to $2,000 counting press, resin and labor. A troublesome new tool burns 3 to 6 trials. Analysis that removes 1 to 3 of those trials returns 1.5 to 4 times its fee. That ratio explains why 16-cavity engineering-resin programs rarely skip the study.

Schedule cost cuts both ways. Analysis requested after steel cutting cannot change a gate without welding, so its recommendations lose value fast. Ask the mold maker to state the analysis slot inside the program plan. Keep the 3 to 10 day window visible when you commit to launch dates.

Cost drivers follow a short list. 3D solid meshes run longer than shell meshes. Parts over 300 mm in length need larger models and more compute time. Iterations multiply the fee almost linearly, and tools above 32 cavities complicate the runner model. Budget the study at roughly 1 to 3% of mold cost on multi-cavity tools, and it rarely breaks a program.

FAQ: Mold Flow Analysis

Q1. How much does mold flow analysis cost?

A single study runs $500 to $3,000 depending on scope. Fill-only work lands at $500 to $1,200. Full cooling and warpage runs $1,500 to $3,000. Duration is 3 to 10 days including two or three iterations.

Q2. Will analysis guarantee a first-shot mold?

No. Predicted deflection can deviate 20 to 50% from measured parts on complex geometry, so the trial still verifies reality. Analysis reliably ranks options and cuts 1 to 3 trials on multi-cavity or engineering-resin programs. Simple commodity tools often run fine without it.

Q3. What must I send the mold maker for a study?

Send the final part CAD in STEP format, the exact resin grade and supplier, and gate preferences or none. Add the process window: melt temperature, mold temperature and injection time. Include target cycle, annual volume, machine tonnage and maximum injection pressure. Measured rheology on your actual lot beats a datasheet card.

Q4. When should I refuse to pay for analysis?

Skip single-cavity and 2-cavity commodity tools holding ±0.2 mm or looser. Skip while the design is still moving, because each change voids the run and adds $500 to $3,000. Skip when no material card exists, since output quality follows input quality.

Q5. Is analysis included in the mold quote?

Some toolmakers bundle a fill check into the DFM quote at no extra fee. Deep flow, cooling and warpage studies are almost always separate line items. Ask in the RFQ which analyses are included. Request the report before steel cutting starts.

The Bottom Line

Mold flow analysis is an insurance policy priced at $500 to $3,000 with a 3 to 10 day term. It pays when a wrong gate or runner is expensive: multi-cavity tools, engineering resins and tight tolerances. It earns little on simple commodity tools where trial and error is cheap.

Read the report for rankings, not prophecies. Compare gate A against gate B, one cooling layout against another, imbalance under 2 to 3% or over. Verify the prediction on trial 1, then feed the measured result into the next study. Each cycle makes the next mold cheaper.

DieStrike builds connector programs for TE Connectivity, Amphenol and Luxshare under IATF 16949 with 120+ machines and hardened cavity steel to HRC 62. Production molds ship in 2 to 5 weeks, standard components in 3 to 7 days, at ±0.005 mm standard precision or ±0.002 mm on request. Put the CAE study in your RFQ, then read it before the first trial, not after the third.