{"id":1689,"date":"2026-11-12T13:17:00","date_gmt":"2026-11-12T05:17:00","guid":{"rendered":"https:\/\/www.bosssealing.com\/?p=1689"},"modified":"2026-11-10T12:00:13","modified_gmt":"2026-11-10T04:00:13","slug":"molded-vs-machined-seals","status":"publish","type":"post","link":"https:\/\/www.bosssealing.com\/ja\/molded-vs-machined-seals\/","title":{"rendered":"Molded vs Machined Seals: Cost, Lead Time and Performance Compared"},"content":{"rendered":"

Two quotes for the same 45 mm polyurethane rod seal. Same drawing, same quantity, 800 pieces a year. One supplier comes back at US$6.10 per piece with a six-week lead time because the seal will be molded. Another comes back at US$13.80 with a two-week lead time because the seal will be turned from PU rod stock. Neither supplier is wrong. They are selling different manufacturing routes, and the route decides tooling, lead time, tolerance, and \u2014 in ways that rarely appear on a quotation \u2014 how the part behaves after 200,000 strokes.<\/p>\n

We quote both routes on the same drawing most weeks, and the conversation usually starts with price and ends with something else: a field failure that traced back to a flash line, or a revision that cost three weeks because the cavity was already cut. This article compares the two routes the way a buyer and a design engineer actually need them compared.<\/p>\n

\"Molded
Molded elastomer rings (left) and machined PTFE rings (right), both to the same nominal drawing, made by different routes.<\/figcaption><\/figure>\n

The Two Processes, Stated Plainly<\/h2>\n

Molded<\/strong> means the geometry is created by a tool. Compound, or PTFE powder in the case of PTFE parts, goes into a cavity, is heated under pressure until it cures or sinters, and is ejected. The cavity is the specification. Change nothing and part 50,000 is nominally identical to part 1.<\/p>\n

Machined<\/strong> means the geometry is cut out of stock: rod, tube, sheet, or a molded blank. A CNC lathe or mill removes material until the drawing is satisfied. No tool is dedicated to the part, so nothing has to be paid for before the first piece exists.<\/p>\n

Almost every seal in a hydraulic cylinder could be made either way. What decides it in practice is quantity, section thickness, and how many dimensions genuinely have to be right. A 12 mm O-ring is molded because nobody would machine one. A 400 mm PTFE wear ring is machined because a tool for it would cost more than a decade of parts.<\/p>\n

What Each Process Does to the Material<\/h2>\n

The differences that survive a drawing review are microstructural. They show up later as differences in compression set, wear rate, and dimensional drift.<\/p>\n

Cure state<\/h3>\n

Molded elastomer cures inside the tool under pressure. A well-run press holds cavity temperature within \u00b12-3 \u00b0C and clamps at 100-200 bar, which suppresses porosity and drives crosslink density to target. A poorly run press does neither. Undercure leaves a soft part with high set and tacky surfaces. Overcure hardens the skin, drops elongation, and raises set. Neither problem shows on a dimensional report, and both are common in cheap tooling with uneven heating.<\/p>\n

Machined elastomer is already cured before it reaches the lathe. The billet or slab was vulcanized in one piece, so cure state is uniform through the section \u2014 often better than a thick molded part, where the core lags the skin by minutes. The trade-off is that slab stock is produced to a durometer and a hardness tolerance, not to a compression set target for your gland.<\/p>\n

Filler orientation<\/h3>\n

This is the one that catches engineers who switch routes without requalifying. Filled PTFE \u2014 glass, carbon, bronze, graphite, moly \u2014 develops filler orientation along the flow direction during molding and extrusion. A ring machined from a molded billet keeps the filler in planes that may or may not align with the sliding direction. A compression-molded ring forms its own flow pattern in the cavity, which can be better or worse depending on preform placement and where the material has to travel.<\/p>\n

Practical consequence: wear rate on a filled PTFE guide ring can shift 30-50% between a machined and a molded version of the same drawing, with the same compound name on both certificates. If a wear part was qualified as machined, requalify it as molded. Do not assume the material name carries the performance.<\/p>\n

Residual stress<\/h3>\n

Machining introduces stress. PTFE is the extreme case: cut a 300 mm sealing ring from sheet and it will move 0.2-0.5 mm out of round within 24 hours as stress redistributes. Shops that understand this rough-machine, stress-relieve, rest the part, then finish. Shops that do not will ship a part that measured perfectly at the machine and will not sit flat a week later.<\/p>\n

Molded parts carry residual stress from cure shrinkage too, but it is repeatable and it is the same on piece 1 and piece 50,000, which makes it possible to compensate in the cavity dimensions.<\/p>\n

Skin, flash and tool marks<\/h3>\n

A molded part has a skin formed against polished steel: smoother than any machined rubber surface, and slightly harder, because mold-side cure is more complete. It also has a flash line, and that line is a geometric feature whether or not the drawing mentions it. Where a parting line crosses a dynamic lip, it is both a leak path and a wear edge. On a static gasket it is cosmetic, and specifying zero flash there adds cost for nothing.<\/p>\n

A machined part has no flash line, but it has tool marks, and mark direction matters as much as the roughness number. Circumferential turning marks on a rod sealing bore work like a screw thread and pump oil along the shaft. The same Ra produced by grinding, with a random lay, behaves completely differently. That is why a finish callout without a lay direction is only half a specification.<\/p>\n

Tolerance: What Each Route Can Actually Hold<\/h2>\n

Rubber does not hold metal tolerances, and the standard that formalizes this is ISO 3302-1. It defines four classes applied to nominal dimensions, with separate values for fixed dimensions and closure dimensions \u2014 anything measured across the parting line.<\/p>\n\n\n\n\n\n\n\n\n\n
Table 1 \u2014 ISO 3302-1 fixed-dimension tolerances vs machining capability<\/caption>\n
Nominal size<\/th>\nM1 precision<\/th>\nM2 high quality<\/th>\nM3 commercial<\/th>\nM4 non-critical<\/th>\nMachined PTFE<\/th>\n<\/tr>\n<\/thead>\n
0-4 mm<\/td>\n\u00b10.08 mm<\/td>\n\u00b10.10 mm<\/td>\n\u00b10.25 mm<\/td>\n\u00b10.50 mm<\/td>\n\u00b10.02 mm routine<\/td>\n<\/tr>\n
10-16 mm<\/td>\n\u00b10.15 mm<\/td>\n\u00b10.20 mm<\/td>\n\u00b10.40 mm<\/td>\n\u00b10.80 mm<\/td>\n\u00b10.02 mm; \u00b10.005 mm achievable<\/td>\n<\/tr>\n
25-40 mm<\/td>\n\u00b10.25 mm<\/td>\n\u00b10.45 mm<\/td>\n\u00b10.70 mm<\/td>\n\u00b11.00 mm<\/td>\n\u00b10.02-0.05 mm<\/td>\n<\/tr>\n
63-100 mm<\/td>\n\u00b10.35 mm<\/td>\n\u00b10.50 mm<\/td>\n\u00b11.00 mm<\/td>\n\u00b11.30 mm<\/td>\n\u00b10.05 mm, geometry dependent<\/td>\n<\/tr>\n
Hardness<\/td>\n\u00b15 Shore A across all classes<\/td>\nBillet tolerance, typically \u00b15<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Two facts in that table decide most quotations. First, closure dimensions run 20-30% looser than fixed dimensions at the same class, so a 25-40 mm dimension molded across the parting line is closer to \u00b10.9 mm at M3 than to \u00b10.7 mm. Second, compression molding defaults to M3. Reaching M2 needs a better tool and tighter process control. M1 on a compression-molded part is a specialist job with ground cavities and 100% inspection, and the price reflects all three.<\/p>\n

Machining is not automatically tighter in the way buyers assume. On stable geometry in PTFE, \u00b10.02 mm is routine and \u00b10.005 mm is achievable with the right setup and a 20 \u00b0C inspection room. On a thin-walled ring, or on 70 Shore A rubber, the material moves more than the tolerance: a soft ring reads 0.1-0.2 mm differently depending on how hard the caliper closes. Measuring rubber with a micrometer and calling it a dimension is a common source of arguments that are not about the part at all.<\/p>\n

The useful rule: if the critical dimension is formed by a metal surface \u2014 a lip diameter, a bonded insert, a groove on a molded diaphragm \u2014 molding holds it better, because the tool defines it and the cutter never touches it. If the critical dimension is a section thickness on a short run, machining gives you room to iterate without cutting steel.<\/p>\n

Where the Cost Sits<\/h2>\n

Tooling is the visible difference and the least interesting one. A simple single-cavity compression tool for a 60 mm seal runs US$4,000-8,000. A multi-cavity production tool in hardened steel for a large FKM part runs US$25,000-50,000. Machining has no tool cost at all, which is why it wins every comparison below 200 pieces.<\/p>\n

The per-piece gap is what compounds. Machined parts carry machine time, stock removal and scrap. Molded parts carry cycle time divided across cavities, and the cavity count is the lever: going from 4 to 16 cavities cuts unit price roughly in half and doubles tool cost.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n
Table 2 \u2014 Worked cost model, 45 mm PU rod seal, 30% glass-filled<\/caption>\n
Line item<\/th>\nMachined from PU rod<\/th>\nCompression molded<\/th>\n<\/tr>\n<\/thead>\n
Tooling<\/td>\nUS$0<\/td>\nUS$7,500 (8-cavity)<\/td>\n<\/tr>\n
Unit price at 500 pcs<\/td>\nUS$14.20<\/td>\nUS$7.40<\/td>\n<\/tr>\n
Unit price at 800 pcs<\/td>\nUS$13.80<\/td>\nUS$6.10<\/td>\n<\/tr>\n
Break-even volume<\/td>\n7,500 \u00f7 (13.80 \u2212 6.10) = 974 pieces<\/td>\n<\/tr>\n
Minimum order<\/td>\n20 pieces<\/td>\nOne press run, ~400 pieces<\/td>\n<\/tr>\n
Drawing change before first article<\/td>\nProgramming only<\/td>\nUS$800-3,000 and 1-3 weeks<\/td>\n<\/tr>\n
Drawing change after tool exists<\/td>\nProgramming only<\/td>\nNew cavity or new tool<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Three cost lines get left out of comparisons and should not be:<\/p>\n