{"id":8115,"date":"2026-10-26T01:59:59","date_gmt":"2026-10-26T01:59:59","guid":{"rendered":"https:\/\/chinametalfoundry.com\/?p=8115"},"modified":"2026-10-26T01:59:59","modified_gmt":"2026-10-26T01:59:59","slug":"casting-dual-sourcing","status":"publish","type":"post","link":"https:\/\/chinametalfoundry.com\/pt\/blog\/casting-dual-sourcing\/","title":{"rendered":"Casting Dual Sourcing: How to Qualify a Second Foundry Without Doubling Cost"},"content":{"rendered":"
\n
\n

Single-source buying works right up to the day it doesn’t. A furnace transformer fails, a grid curtailment cuts melt capacity in half, or a quality hold freezes your releases \u2014 and your line stops because nobody else on earth can pour that part. Casting dual sourcing<\/strong> is the obvious answer, and the obvious version of it is expensive: a second tooling set, a second first-article package, a second process audit, and a permanent volume penalty. This guide turns the decision into arithmetic instead of anxiety.<\/p>\n<\/div>\n

What follows is the model we use with buyers who run 500 to 20,000 castings a year: what a second foundry genuinely costs to keep alive, what a disruption genuinely costs when it lands, and the break-even probability that tells you which parts deserve casting dual sourcing and which ones just need deeper safety stock. Cost bands and process capability data below come from shop-floor quoting records and industry references including the American Foundry Society<\/a> e NADCA<\/a>; every number in the worked examples is shown with its inputs so you can substitute your own.<\/p>\n

\"casting
Two pump housings, one drawing, two foundries \u2014 the surface texture and riser scars already tell you the tooling is not identical.<\/figcaption><\/figure>\n

What Casting Dual Sourcing Actually Buys You<\/h2>\n

Most buyers say “dual sourcing” and mean three very different things. The three models carry wildly different costs and deliver wildly different protection, and mixing them up is the single most common reason a second-source programme gets cancelled after year one.<\/p>\n\n\n\n\n\n\n\n
Model<\/th>\nVolume to source B<\/th>\nRamp to full rate<\/th>\nTypical annual carry<\/th>\nBest for<\/th>\n<\/tr>\n<\/thead>\n
Paper alternate<\/strong> \u2014 drawings, material spec and tooling drawings on file, no tooling built<\/td>\n0%<\/td>\n12\u201320 weeks<\/td>\n$400\u2013$900<\/td>\nLow-consequence parts with deep safety stock<\/td>\n<\/tr>\n
Warm backup<\/strong> \u2014 full tooling built, B runs 1\u20132 keep-alive batches a year<\/td>\n10\u201315%<\/td>\n3\u20135 weeks<\/td>\n$5,000\u2013$8,000<\/td>\nCritical parts where you cannot tolerate a 12-week gap<\/td>\n<\/tr>\n
Parallel split<\/strong> \u2014 B runs continuously, typically 70\/30<\/td>\n25\u201340%<\/td>\n3\u20135 days<\/td>\n$7,000\u2013$11,000<\/td>\nGrowth programmes, customer-mandated second sources<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The ramp column is the one that gets ignored, and it is the one that decides everything in casting dual sourcing. A paper alternate does not protect you from a three-week outage \u2014 by the time its tooling is cut, the outage is over. A warm backup still needs tooling pulled from storage and refurbished, a sand system re-set, operators re-briefed, and first-off parts re-qualified before rate climbs. Only a source that has poured your part in the last 30 days can surge inside a week. If your part is a green-sand casting<\/a> or fundi\u00e7\u00e3o \u00e0 cera perdida<\/a> design, the tooling step dominates that ramp; for high-pressure die casting<\/a> the die exists at B only if you paid for a second one.<\/p>\n

The Real Cost of a Second Foundry<\/h2>\n

Take a concrete part: an ASTM A536 65-45-12 ductile iron pump housing, 18.0 kg as-cast, 2,400 pieces a year (46 per week), FOB $41.80 per piece \u2014 roughly $100,320 of annual buy. Here is what a casting dual sourcing programme actually costs to stand up properly, not on paper.<\/p>\n

One-time qualification bill<\/h3>\n
\n\n\n\n\n\n\n\n\n\n\n\n\n
Line item<\/th>\nCost<\/th>\nWhy it cannot be skipped<\/th>\n<\/tr>\n<\/thead>\n
Duplicate tooling set \u2014 matchplate plus core boxes, aluminium<\/td>\n$6,800<\/td>\nYour existing pattern is sized to foundry A’s flask and moulding line \u2014 see how casting moulds are made<\/a> and why a second set is rarely a copy. Budget 60\u2013100% of original tooling for a true second set; see our casting tooling cost<\/a> breakdown for the band by process.<\/td>\n<\/tr>\n
First-article inspection, full dimensional layout<\/td>\n$950<\/td>\nDifferent gating means different as-cast geometry. Cost structure per our FAI cost pack<\/a>.<\/td>\n<\/tr>\n
Material verification \u2014 chemistry, tensile, hardness, nodule count<\/td>\n$640<\/td>\nSame spec on paper is not the same microstructure in the ladle.<\/td>\n<\/tr>\n
Machining trial plus capability study, 30 pieces, 5 CTQ features<\/td>\n$2,150<\/td>\nThe only way to know B holds your datums is to cut it on the same Maquina\u00e7\u00e3o CNC<\/a> routing you run in production.<\/td>\n<\/tr>\n
On-site process audit, 2 auditors, 1.5 days plus travel<\/td>\n$3,400<\/td>\nScope per our casting supplier audit<\/a> checklist.<\/td>\n<\/tr>\n
PPAP Level 3 \/ control plan documentation<\/td>\n$1,200<\/td>\nWithout a control plan, the second source is a rumour.<\/td>\n<\/tr>\n
Internal engineering time, 40 h at $75\/h<\/td>\n$3,000<\/td>\nDrawing review, tooling reviews, FAI adjudication, PPAP sign-off.<\/td>\n<\/tr>\n
Total one-time<\/strong><\/td>\n$18,140<\/strong><\/td>\nAmortise over the 5-year tooling life \u2192 $3,628\/year.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\"casting
Duplicate tooling is the largest single line in the qualification bill, and the one most often forgotten until the outage arrives.<\/figcaption><\/figure>\n

Annual carry for a warm backup<\/h3>\n
\n\n\n\n\n\n\n\n\n\n
Annual line<\/th>\nAmount<\/th>\n<\/tr>\n<\/thead>\n
Amortised qualification, 5-year tooling life<\/td>\n$3,628<\/td>\n<\/tr>\n
Keep-alive volume: 300 pieces at a +$3.10\/pc small-batch premium<\/td>\n$930<\/td>\n<\/tr>\n
Second tooling set: storage, annual condition inspection, minor refurb<\/td>\n$420<\/td>\n<\/tr>\n
Audit refresh plus dual control-plan maintenance, 1 day\/year<\/td>\n$1,400<\/td>\n<\/tr>\n
Total annual carry (C)<\/strong><\/td>\n$6,378<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

That $3.10 keep-alive premium is not a penalty the foundry invents \u2014 it is the standing cost of any casting dual sourcing arrangement \u2014 it is the arithmetic of batch size. A 150-piece batch still eats a full moulding-line set-up and a full melt, so the fixed cost is spread over fewer pieces. The mechanics are laid out in our piece on casting minimum order quantity<\/a>.<\/p>\n

What a Disruption Actually Costs<\/h2>\n

Now the other side of the casting dual sourcing business case. Assume the buyer holds two weeks of safety stock (92 pieces) and the incumbent suffers a six-week outage. Two weeks are covered by stock; four weeks are not.<\/p>\n\n\n\n\n\n\n\n\n\n\n
Loss line, single source, 6-week outage<\/th>\nAmount<\/th>\n<\/tr>\n<\/thead>\n
Air freight differential on the 4-week gap: 184 pcs \u00d7 18 kg = 3,312 kg at $5.50\/kg = $18,216, less ocean $828<\/td>\n$17,388<\/td>\n<\/tr>\n
Assembly line stoppage, 4 days at $4,600\/day<\/td>\n$18,400<\/td>\n<\/tr>\n
Expedite \/ re-sequencing premium at the foundry<\/td>\n$2,800<\/td>\n<\/tr>\n
Late-delivery exposure to your own customer<\/td>\n$6,000<\/td>\n<\/tr>\n
Internal firefighting, 60 h at $75\/h<\/td>\n$4,500<\/td>\n<\/tr>\n
Total loss per event<\/strong><\/td>\n$49,088<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

With a warm backup, B ramps over four weeks. Safety stock covers the first two; the remaining two weeks run short. The residual: air freight on roughly 92 pieces (1,656 kg at $5.50 = $9,108, less ocean $414) = $8,694, plus two days of line stoppage $9,200, plus $2,000 expedite, $3,000 customer exposure and $3,000 internal time \u2014 $25,894<\/strong>. So the loss avoided per event is $23,194<\/strong>.<\/p>\n

The Break-Even Probability: Casting Dual Sourcing as a Number<\/h2>\n

One more input into the casting dual sourcing decision. A live second source changes the annual negotiation: with credible competitive tension, the incumbent’s increase typically lands near +1.8% instead of +4.2%. On 2,100 pieces at $41.80 that 2.4-point gap is worth $2,107 a year<\/strong> \u2014 call it the leverage credit V.<\/p>\n

Net annual value of the programme = P \u00d7 L + V \u2212 C, where P is your honest annual probability of a disruption longer than four weeks.<\/p>\n

Break-even: P* = (C \u2212 V) \/ L = ($6,378 \u2212 $2,107) \/ $23,194 = 18.4%<\/strong>.<\/p>\n

Read that carefully. On pure outage arithmetic, a warm backup only pays if you genuinely believe a four-week-plus outage is more likely than roughly one year in five. For a single-furnace foundry on an unreliable grid, that is plausible \u2014 and for a ductile iron casting<\/a> programme with no second melt route, casting dual sourcing is cheap insurance against a melt outage. For a modern multi-line foundry with redundant melt capacity, it is not \u2014 and paying $6,378 a year to hedge an 8% risk is a losing trade.<\/p>\n

Why the parallel split beats the warm backup<\/h3>\n
\n

Upgrade from warm backup to a 70\/30 live split and two things change. Cost rises: 720 pieces move to a smaller volume tier ($44.30 vs $41.80 = +$2.50\/pc = $1,800), offset against the $930 keep-alive premium you were already paying, so +$870; plus ten extra receiving and incoming-inspection events at $90 each, +$900. Extra carry $1,770<\/strong>, total C = $8,148.<\/p>\n

But B is now pouring every month, so ramp collapses from four weeks to three to five days. Safety stock covers it entirely: no line stoppage, no air freight, only a $700 expedite and $750 of coordination \u2014 residual $3,340<\/strong>, so L rises to $45,748.<\/p>\n\n\n\n\n\n\n\n
Model<\/th>\nAnnual carry C<\/th>\nLoss avoided L<\/th>\nBreak-even P*<\/th>\n<\/tr>\n<\/thead>\n
Single source, do nothing<\/td>\n$0<\/td>\n\u2014<\/td>\n\u2014<\/td>\n<\/tr>\n
Warm backup (85\/15)<\/td>\n$6,378<\/td>\n$23,194<\/td>\n18.4%<\/td>\n<\/tr>\n
Parallel split (70\/30)<\/td>\n$8,148<\/td>\n$45,748<\/td>\n13.2%<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The counterintuitive result: the more expensive model has the lower<\/em> break-even. A warm backup carries about 78% of the split’s cost while capturing only 51% of its benefit, because the expensive part of a disruption is the four-week ramp, not the tooling. If you are going to do casting dual sourcing at all, do it as a live split or do not do it.<\/p>\n

When Casting Dual Sourcing Is the Wrong Answer<\/h2>\n

Same method, different part. A 6.5 kg grey iron counterweight \u2014 a textbook case where casting dual sourcing is the wrong answer: 900 pieces a year, FOB $9.40, no machining, no pressure boundary, and the buyer already holds six weeks of stock.<\/p>\n