On This Page
- 10 min read
Using Commercial Vacuum Flanges on ASME Section VIII Code Vessels
Vacuum‑tight is easy; code‑stamped pressure isn’t. Now what?
CF (ConFlat, knife‑edge metal gasket) flanges are engineer candy: ultra‑tight, clean, cryo‑friendly, and off the shelf. Then ASME (American Society of Mechanical Engineers) Section VIII (the pressure‑vessel code) walks in and asks: where’s the math, the weld rationale, the leak criteria? Last quarter we watched a 6‑inch CF assembly hold 1×10^-8 mbar·L/s helium, yet the stamp review paused—no recognized design basis, no traceable torque plan. Sound familiar?
We’ve been in that room with your Authorized Inspector (AI, the third‑party code inspector). The fix wasn’t magic: treat the CF as a custom flange via Appendix 2 (ASME’s flange math), run local FEA (finite‑element analysis) on the knife‑edge, or prove it with UG‑101 (code proof test). We’ll show the paths, a worked 6‑inch, 75‑psig (gauge pressure) example, and a CRN (Canadian Registration Number)‑ready plan. Talk to us early and save weeks. So why does the code push back on CF geometry?
Why CF Flanges Shine—and Why ASME Reviewers Push Back
You’re right: CF seals excel, but compliance demands traceable design math, sound welds, and proof. CF (ConFlat, knife‑edge metal gasket) flanges are outstanding for cryogenic cycling, UHV (ultra‑high vacuum) leak tightness, and clean, demountable joints. We see them daily on beamlines, cryomodules, and high‑purity gas skids. Code‑stamped service changes the game: positive pressure, energy storage, and safety margins. Think 15–150 psig with −196 to 200 °C. Different stakes.
Why you like CF is obvious: tighten once, helium‑tight for years, and no elastomers to age. But your Authorized Inspector (AI, the third‑party code inspector) won’t accept vacuum pedigree as proof for pressure. They’ll ask how the joint behaves under pressure, temperature, and external loads from piping and supports. Example: a 6‑inch CF on a 75 psig nitrogen purifier sees nozzle moments at startup. That rotation matters.
Here’s the rulebook in plain English. ASME Section VIII Division 1 (pressure‑vessel code) points you to UG‑4 (materials and traceability—MTRs, heat numbers), Appendix 2 (bolted flange design—bolt loads, rotation, stresses), and UG‑34 (flat heads/covers—thickness formulas). Because CF geometry isn’t in ASME B16.5/B16.47 (listed flange standards), you must treat it as an unlisted, custom flange and qualify it.
Three approval paths exist. Appendix 2 math, UG‑101 proof testing (code pressure test to demonstrate strength), or Division 2 design‑by‑analysis (FEA, finite‑element analysis). Reviewers then check gasket seating stress, bolt stresses and relaxation, flange rotation/leak risk, welding details, and MDMT (minimum design metal temperature) with impact testing where required. Skipping any of these slows or stops your stamp.
Can your catalog CF parts meet UG‑4 traceability?
UG‑4 (materials acceptance in ASME Section VIII) expects listed materials with traceable certification. CF flanges lack an ASME or Manufacturer’s Standard with a pressure‑temperature rating, so the part itself isn’t “pre‑approved.” That means most catalog items are out unless they arrive with proper MTRs (mill test reports), heat numbers, and unique serialization that ties each piece to the MTR. Bolts, ferrules, and hubs are all in scope.
Because there’s no B16.5/B16.47‑style listing, you must prove material pedigree part‑by‑part. Serialization (individual ID numbers stamped or etched) lets you link the delivered flange, bolt set, and ferrule to specific heats on the MTRs. Without that chain, your AI can’t verify the material in service equals the material in the paperwork. Many vacuum suppliers ship great hardware, but only provide generic “304L” statements. That won’t pass UG‑4.
Procurement impact: if you buy anonymous catalog parts, you risk a last‑minute rejection, rework, and schedule slip. We’ve seen teams scramble for PMI (positive material identification) at receiving and still miss heat links. The fix is simple but must start at the PO (purchase order): specify ASME‑listed specs, full MTRs for all pressure‑retaining items, unique serialization on each piece, and bolt/ferrule certificates that match. Your future self will thank you.
You’ve got two clean paths. One: source commercial CF components that come with full MTRs, heat numbers, and part‑level serialization. Fast and familiar, but you’re dependent on the vendor’s documentation quality. Two: machine CF‑compatible geometry from certified plate/forging with MTRs. You control dimensions, tolerances, and marking, and we can stamp heat numbers as we go. It takes more shop time, but it removes ambiguity.
Tradeoffs to watch: lead time and bolt compatibility. Commercial parts may arrive in days, but missing bolt certs stall you. Custom machining adds 1–3 weeks, yet gives perfect documentation and knife‑edge control. Whichever route, plan PMI at receiving, record serial‑to‑heat links, and store photos of markings in your MDR (Manufacturer’s Data Report package). This avoids painful backfills later.
- Pitfall 1: Missing heat numbers or partial MTRs for bolting/ferrules
- Pitfall 2: Supplier markings not traceable to a Section II Part D listed material
- Pitfall 3: Mixed heats in the bolt set without full documentation
- Pitfall 4: No positive material identification (PMI) record during receiving
Appendix 2 Math: Manageable, if you feed it good inputs
Appendix 2 (the bolted‑flange design rules in ASME Section VIII Div. 1) needs solid inputs: design pressure/temperature, bolt pattern and material, flange dimensions, and gasket geometry. You’ll also need gasket m and y factors (seating and operating factors per Tables 2‑5.1 and 2‑5.2). CF copper gaskets aren’t directly listed, so we justify equivalents or test‑derived values. Using a cover or blind? UG‑34 (flat heads/covers) checks thickness separately.
Don’t forget the realities that move the numbers: external nozzle loads from B31.3/B31.1 piping (process piping codes), MDMT (minimum design metal temperature), and bolt allowables from Section II‑D (material properties tables). Tolerances on b, h, and t (gasket seating width, hub height, and thickness) matter because small changes swing rotation. We capture all this in a one‑page worksheet before running calcs or FEA (finite‑element analysis).
Practical rule: many commercial CF flanges clear Appendix 2 at lower pressures—often below about 50 psig. Above that, expect tweaks: thicker hubs, altered bolt patterns, verified m,y, and tighter tolerances. Budget time to iterate once, then prove with a leak/pressure test. Next, protect the knife‑edge and control weld effects.
| Input/Assumption | Appendix Ref. | Validate This | Common Pitfall | Impact |
|---|---|---|---|---|
| Design pressure and temperature, including hydro/pneumatic test and upset cases. | Appendix 2‑1 (scope, symbols, fundamental flange relations). | Use correct load cases, including gasket seating and test conditions. | Ignoring upset, test, thermal, or transient operating conditions. | Under‑designed flange; leakage, yielding, or nonconformance at test. |
| Bolt pattern, size, count, and ASME‑listed material grade. | Appendix 2‑5 (bolt load and allowable stress checks). | Bolt stress within allowable; creep/relaxation at service temperature verified. | Using non‑listed bolt materials or mixed property classes. | Bolt failure or preload loss during operation or test. |
| Gasket type and m,y values for seating and operating. | Tables 2‑5.1 and 2‑5.2 (gasket factors and seating stresses). | m,y compatible with knife‑edge copper gasket behavior and seat width. | Guessing m,y values from vendor marketing without validation. | Leakage at seat‑up or inadequate preload window. |
| Flange dimensions b, h, t, bore, hub, and seating width. | Appendix 2‑7 through 2‑14 (geometry, stresses, rotation, moments). | Moments, rotation, and stresses within Appendix 2 limits. | Using catalog OD/ID without tolerances or corrosion allowance. | Excess rotation, leak risk, or knife‑edge damage. |
| Cover or blind flange subject to UG‑34 checks. | UG‑34 (flat heads and covers thickness formulas). | Thickness and corrosion allowance per formulas; stay or reinforcement if needed. | Treating CF cap as un‑stayed cover. | Buckling, yielding, or excessive deflection at pressure test. |
Weld distortion can ruin knife-edge seals fast
If buckling, yielding, or excessive deflection shows at the pressure test, weld distortion likely set the trap. Code welds are bigger—3/16 to 1/4 inch fillets or full‑pen grooves—while vacuum shops favor tiny 1/16 inch TIG beads (GTAW, gas tungsten arc welding). More heat input means more shrinkage. Distortion shows up as angular pull, peaking at the hub, or flange ovality. Even 0.1 mm runout or a 20 microinch scuff can blunt the knife‑edge, shift gasket seating, and spike required bolt preload.
After welding, the heat‑affected zone (HAZ, metal near the weld) hardens and holds residual stress that creeps into the knife‑edge. Expect diameter growth, face tilt, and concentricity drift—small numbers, big consequences. We lock surface finish with post‑weld machining and verify RA (roughness average) ≤ 32 microinch, knife‑edge angle 20° ± 1°, radius 0.05–0.10 mm. QA needs PT (penetrant testing, dye check) on the seal face, UT/RT (ultrasonic/radiographic testing) on welds, and a WPS/PQR (welding procedure specification/procedure qualification record) that limits heat input.
- Mitigation: Use a pre-machined oversize boss and finish-machine after welding
- Mitigation: Control WPS/PQR heat input and sequence to minimize distortion
- Mitigation: Consider alternate joint prep that isolates the knife-edge
Three practical paths to compliance
You’ve got options, and each one fits a different mix of pressure, size, temperature, schedule, and paperwork maturity. We group them into three paths so you can pick quickly. If you’re at 15–75 psig with good supplier certs, one path works. If pressure climbs or geometry shifts, another path wins. Tight schedule, limited documentation, or aggressive MDMT (minimum design metal temperature)? We’ll steer you to the route that gets approval without rework.
Start by ranking constraints: required pressure rating, available MTRs (mill test reports), access to machining, and reviewer expectations. Smaller diameters and lower PSIG often clear with calculations and targeted testing. Larger diameters, high external piping loads, or cryo cycling benefit from design‑by‑analysis and a proof test. When in doubt, we pre‑brief the Authorized Inspector and pick the path they’ll sign off on fastest.
- Path 1: Cert-backed commercial CF flange + full Appendix 2/UG-34 calculations
- Path 2: Machine CF-compatible flange from MTR-backed stock to code geometry
- Path 3: Custom Appendix 2 flange that retains knife-edge and uses standard CF gasket
Use this comparison to decide fast: pressure envelope, speed to buy or build, engineering effort, and best fit.
| Option | Pressure Range | Pros | Cons | Best For |
|---|---|---|---|---|
| Cert-backed commercial CF | Low to moderate pressure | Fast procurement, minimal machining | Limited pressure envelope; documentation varies | Small, low-PSIG retrofits |
| Machined from certified stock | Moderate | Full material control; precise dimensions | Longer lead time; higher shop cost | Mid-range pressures/diameters |
| Custom Appendix 2 flange | Moderate to high | Optimized for code and sealing | Highest engineering effort; analysis required | New builds or major redesigns |
| Post-weld finish machining | Any | Protects knife-edge quality | Requires precision machining; fixturing | Cryogenic and ultra-high vacuum |
| Bolt/material optimization | Any | Improves allowable loads | May need alternate alloys | Temperature and creep-sensitive services |
Worked example: 6-inch CF-style flange at 75 psig
Because we’re in temperature- and creep-sensitive service, let’s put real numbers on it. Example: 6-inch CF-style flange at 75 psig nitrogen, −196 to +50 °C. Materials: 316L body, OFHC (oxygen-free high-conductivity) copper gasket, bolts ASTM A193 B8M Class 2 (strain-hardened 316 stainless). Joint: hub-to-nozzle full-penetration groove weld, post-weld face finish to Ra (roughness average) ≤ 32 microinch. This is illustrative, not a blanket approval. A licensed PE (professional engineer) must review and seal.
Assume zero internal corrosion allowance, external nozzle loads per ASME B31.3 (process piping code) translated into Appendix 2 rotation and bolt checks, and gasket factors m and y based on copper test data for a 0.5–1.0 mm knife-edge seat. We supplement Appendix 2 with local FEA (finite-element analysis) around the knife-edge and set a torque window that keeps operating bolt stress within Section II-D allowables. This guide isn’t approval; your AI (Authorized Inspector) may require UG-101 proof.
Here’s the traceable, reviewer-friendly sequence we run, with inputs and checkpoints captured.
- Step 1: Gather pressure/temperature envelopes, test conditions, corrosion allowance
- Step 2: Confirm MTRs for flange, bolts, and any cover; verify Section II listings
- Step 3: Select gasket and m,y values appropriate to knife-edge geometry
- Step 4: Run Appendix 2 flange checks (rotation, stress) and UG-34 if cover applies
- Step 5: Iterate thickness/bolt pattern until margins meet targets
- Step 6: Freeze WPS/PQR with heat input limits and plan post-weld machining
- Step 7: Compile design report with calcs, drawings, and test plan
Now prove the numbers. Plan a hydro at 1.3–1.5× design with calibrated gauges and a timed hold. Our guide to ASME hydrostatic pressure testing covers setup, safety, and acceptance. Next, we’ll layer pneumatic and helium checks in the validation plan.
Validation roadmap to prove sealing and compliance
You just saw the math; now we prove it. We validate in stages so each step de-risks the next, catches damage early, and builds a regulator-ready record. Simple, sequential, no surprises.
- Stage 1: Visual/PMI and dimensional inspection to baseline the knife-edge — verify markings and materials with PMI (positive material identification), measure flatness/ovality, and photograph the edge as-received for traceability.
- Stage 2: Leak-check at sub-atmospheric (helium mass spec) to validate sealing surfaces — helium mass spectrometer (sniff or vacuum mode); target ≤1E‑6 mbar·L/s and document dwell and background.
- Stage 3: ASME (American Society of Mechanical Engineers) hydrostatic test at code-required factor with instrumentation — 1.3–1.5× design; calibrated gauge, independent verification, timed hold, and metal temperature logged.
- Stage 4: Pneumatic or functional test at service pressure, where appropriate — control stored energy, use pressure-decay or soap/helium sniff, verify instrumentation, and record leak acceptance and valve/function behavior.
- Stage 5: Post-test NDE and dimensional re-verification to confirm no knife-edge damage — NDE (non-destructive examination) with penetrant; re-measure flatness, edge height/radius, and compare to as-received photos.
If you need ultimate strength data beyond code factors, we can characterize failure pressure with controlled burst testing and fold results into your safety margins and CRN package.
CRN expectations for CF-style flanges
Since we’re folding your test results into a CRN package, here’s how reviews work. A CRN (Canadian Registration Number) is provincial approval under CSA B51 (the national pressure equipment standard) to use a design in that jurisdiction. Because CF-style flanges are unlisted (not ASME B16.5/B16.47), authorities treat them as non-standard parts that must be justified. You can qualify with Appendix 2 calculations, Division 2 FEA (finite-element analysis), or UG-101 proof tests (demonstrated strength). Clean math plus traceable tests usually gets traction. Example: a 6-inch, 75 psig copper-sealed joint cleared on first review with both.
Documentation wins. Provincial reviewers focus on MDMT (minimum design metal temperature) and impact testing basis, welding procedure records, and whether your leak and proof tests match the acceptance criteria you cite. Package signed calculations, drawings, MTRs (mill test reports), PMI (positive material identification), and NDE (non-destructive examination) maps cleanly. When we pre-brief an Authorized Inspector (third-party code inspector) and align on acceptance thresholds—like ≤1E‑6 mbar·L/s helium—reviews move faster. One design can register in multiple provinces with the same evidence, but formatting and cover letters matter.
New to CRN? Start with our Canadian Registration Number explainer—scope, steps, and common pitfalls.
- Include: Signed calculations (Appendix 2/UG-34) by qualified engineer
- Include: Drawings with weld details, tolerances, and materials
- Include: MTRs and PMI records for pressure parts and bolts
- Include: Test plans and results (hydro, leak, NDE)
- Include: Quality system references (WPS/PQR, welder quals)
Design-to-CRN (Canadian Registration Number) timeline, step by step
With WPS/PQR (welding procedures) and welder qualifications set, what happens next? A clear, shared timeline cuts iteration, locks procurement lead times, and keeps reviewers aligned. Use it to trigger POs, book machining, and reserve test slots before bottlenecks hit.
| Phase | Primary Owner | Core Activities | Main Deliverables | Notes |
|---|---|---|---|---|
| Concept and feasibility | Engineering | Select path (Appendix 2/UG‑101/Div. 2), gather service data | Feasibility memo with scope and constraints | List risks, assumptions, and required approvals |
| Detailed design | Engineering + QA/Welding | Appendix 2/UG‑34 calcs, gasket factors, WPS/PQR | Design report, drawings, torque table, WPS | Freeze heat input limits; align Authorized Inspector expectations |
| Fabrication | Fabricator | Controlled welding, in‑process quality checks, traceability | Traveler, weld maps, heat links, photo records | Prepare for post‑weld machining and protection |
| Post‑weld and machining | Machining shop | Finish‑machine knife‑edge, verify final dimensions | Dimensional report, surface finish verification | Ra 32 microinch max; protect edges |
| Testing and CRN submission | Engineering + QA | Hydro, pneumatic, helium tests; compile dossier | Complete CRN package and cover letter | Expect provincial questions; pre‑draft responses |
Our boilers and pressure vessels CRN registration guide outlines documents, timelines, and province-specific quirks. Next, we translate nozzle loads into flange checks.
Designing the vessel–piping interface that actually seals
So how do we translate nozzle loads into flange checks at the vessel–piping interface? Your external piping moments and shear (from weight, thermal growth, and startup forces) drive flange rotation and squeeze the bolt preload window. Even 1 mm misalignment or a 0.5° skew at assembly can shift the knife‑edge seating and demand 10–20% more torque. Appendix 2 (ASME’s flange rules) assumes clean alignment and controlled loads; field realities can blow those assumptions.
Plan the piping like it’s part of the seal. Use guides and springs so hot‑cold movement doesn’t dump a 1.5 kN·m moment into a 6‑inch CF. During bolt‑up, lubricate consistently, cross‑pattern in stages, and verify torque scatter (±10%) with a calibrated wrench. Then translate measured nozzle loads per WRC (Welding Research Council) methods into Appendix 2 rotation checks, and adjust torque or supports before you ever see pressure.
Validate the whole loop with pressure piping testing so loads match the model, then we’ll lock materials and finishes that keep preload and sealing stable across temperature.
- Check: External loads at the nozzle (WRC) vs. flange rotation
- Check: Bolt-up sequence and torque control in procedures
- Check: Field surface protection of the knife-edge during assembly
Materials and finishes that hold seals through thermal cycles
Protecting the knife‑edge in the field is half the battle; the other half is choosing materials and finishes that don’t drift under thermal cycles. Start with 316L or 304L stainless for the flange; they weld cleanly and play with OFHC (oxygen‑free high‑conductivity) copper gaskets. At MDMT (minimum design metal temperature) near −196 °C, we favor 316L for toughness. Bolting: A193 B8M Class 2 (strain‑hardened 316) or A320 B8M; pair with silver‑plated nuts to tame galling. Keep corrosion allowance small (0–1.5 mm) to preserve stiffness for Appendix 2 (ASME flange rules).
Knife‑edge behavior depends on hardness and finish. Target Ra (roughness average) 16–32 microinch on the seat, knife‑edge angle ~20° and radius 0.05–0.10 mm. Don’t coat the sealing land; passivate only—coatings can flake and leak. Duplex 2205 or 347 stainless can work for higher strength or sensitization control, but confirm impact toughness at MDMT (minimum design metal temperature) and galvanic compatibility with copper. For galling, use silver plating 8–25 micrometers on nuts or oxygen‑compatible lubricants (Krytox) for O2 (oxygen) service. Example: that switch cut torque scatter from ~25% to ~10%.
Here’s a quick surface and finish checklist you can use on every CF (ConFlat) joint. It keeps sealing predictable across hot‑cold cycles and makes reviewers comfortable.
- Target: Controlled Ra on knife-edge seat per gasket supplier
- Target: Post-weld finish machining to restore geometry
- Target: Protective caps and cleanliness controls
- Target: Verified flatness/runout of mating faces
Cryogenic CF flange: red flags to CRN in 10 weeks
With flatness and runout verified, we turned that checklist into results on a real build. A team wanted to reuse commercial CF (ConFlat, knife‑edge metal gasket) hardware for 75 psig cryogenic nitrogen. The snags were familiar: thin Appendix 2 (ASME flange math) margins, missing MTRs (mill test reports) on bolts, and weld distortion risk near the seal. We aligned scope with their Authorized Inspector in week one. Clear goals. No surprises.
We chose a custom Appendix 2 flange in 316L with an OFHC (oxygen‑free high‑conductivity) copper gasket, backed by local FEA (finite‑element analysis) at the knife‑edge. Joint detail: full‑pen weld, then post‑weld finish machining to restore Ra 16–32 microinch and edge radius 0.05–0.10 mm. We set a torque window (about 55–70 N·m, verified, not assumed) to keep bolt stress within allowables. Proof came fast: UG‑101 (code proof test) at 1.3–1.5× pressure passed, then helium leak ≤1E‑6 mbar·L/s in vacuum mode.
After five thermal cycles from −196 to +40 °C, leakage held at ~1E‑7 mbar·L/s and torque loss stayed under 10%. The CRN (Canadian Registration Number) landed in two provinces in roughly 10 weeks with one RFI (request for information) on impact data, closed in 48 hours using our pre‑planned MDMT package. Net result: stable seal, reviewer‑friendly math, and a repeatable torque table the maintenance team trusts. Want the full report outline and torque template? Ask and we’ll share a redacted sample.
Expert resources to de-risk CF flange compliance
If that redacted sample would help, we also support life‑science skids—our ASME BPE code engineering practice bridges CF (ConFlat) flanges with sanitary design so your quality assurance (QA) team and CRN (Canadian Registration Number) reviewers stay aligned.
Want a shortcut? Get our CF (ConFlat) flange checklist—MTRs (mill test reports), Appendix 2 inputs, gasket factors, weld procedures, post‑weld finish targets, and leak/pressure acceptance. We’ll email the PDF after a short form so your team can use it as an MDR (Manufacturer’s Data Report) index.
Quick answers to your CF flange questions
If the checklist sparked questions, start here. These answers are guidance only—final designs need a qualified engineer’s review and alignment with your Authorized Inspector (AI, code inspector) and CRN (Canadian Registration Number) expectations.
- Q: Can I use catalog CF flanges as‑is on a code vessel? A: Only with traceability (MTRs, mill test reports, serialization) and passing Appendix 2/UG‑34 checks; many need thicker hubs, verified bolt grades, or post‑weld machining to restore sealing faces.
- Q: What pressure can CF flanges typically handle? A: Many commercial sizes work at low‑to‑moderate pressure; above about 50 psig, Appendix 2 often flags rotation/bolt limits, so geometry, thickness, or bolt pattern changes are common to keep margins.
- Q: Do I need to post-weld machine the knife-edge? A: Often yes. Code‑sized welds add distortion and heat‑affected zone (HAZ) hardening; finish‑machining restores flatness, Ra, edge angle/radius, and concentricity so gasket seating and torque windows match your calculations.
- Q: Which gasket m,y should I use? A: Use m,y (operating and seating factors) consistent with ring‑type metal gaskets; confirm with vendor data or test, and ensure compatibility with your knife‑edge seat width and finish.
- Q: Will CRN reviewers accept CF‑style ports? A: Yes, with a complete package: signed calculations, drawings, MTRs (mill test reports), weld and NDE (non‑destructive examination) records, plus test evidence. Acceptance varies by province; documentation clarity often decides outcomes.
- Q: Is pneumatic testing required? A: It depends on service and risk. Hydrostatic per ASME UG‑99 (hydro rules) is typical; pneumatic is used selectively with controls, stored‑energy safeguards, and leak criteria matched to service, often followed by helium mass‑spectrometer testing.
CRN‑ready CF flanges—approved without leaks or delays
If you’re weighing hydro vs. pneumatic tests and leak thresholds after those FAQs, let’s map your approval path in one call. We’ll review service, materials, and loads, then outline the fastest route to a stamp—Appendix 2 math, targeted FEA (finite‑element analysis), or a proof test. You leave with risks ranked, a torque window, and a checklist to brief your inspector. 30 minutes. Clear next steps.
Prefer proof over debate? Our compliance‑by‑design review packages the worked example, acceptance criteria, and validation plan your reviewer expects. We’ve supported 120+ pressure equipment projects across multiple provinces, including 30+ CF‑style flange approvals. Typical rhythm: week 1 alignment, week 2 calculations and torque table, week 3 test plan ready. Your schedule—and your seal—stay intact.
For a fast primer, visit CRN help Canada and get steps, documents, and timelines.
Get Expert Help
Let's discuss your test plan