​‌​​‌​

AISI 8620 / (1.6523) / UNS G86200 / AMS 6274

​‌​​‌​

AISI 8620 is a low-carbon nickel-chromium-molybdenum alloyed case hardening (carburising) steel. What separates it from through-hardening grades such as 4140 and 4340 in the alloy steel group is that it gains its hardness not through its bulk but by carbon being diffused into its surface.

Keeping the carbon content low allows the core to remain tough and impact resistant. After the part has been machined, carbon is diffused into its surface in a carburising furnace, then it is quenched and tempered. The result is a hard, wear resistant outer layer over a tough core able to absorb impact — exactly the structure sought for gears and rolling contact surfaces.​‌​​‌​

Nickel raises core toughness, while chromium and molybdenum increase hardenability and the strength of the carburised case. This composition makes 8620 the most widely used and most readily available grade among the case hardening steels.

It is used in aerospace and defence for gearbox components, transmission gears, cams, pins and bushings, and in general machine building for shafts, sprockets and bearing housing parts. It can be supplied as round bar and forgings.​‌​​‌​

​‌​​‌​

What AISI 8620 Is — and Why You Cannot Order a Hardness Number

​‌​​‌​

AISI/SAE 8620 (UNS G86200 / W.Nr. 1.6523 / EN 20NiCrMo2-2 / former DIN designation 21NiCrMo2 / BS 805M20 / AFNOR 20NCD2 / JIS SNCM220) is a low-carbon, triple-alloyed (Ni-Cr-Mo) CASE-HARDENING — that is, carburizing — steel. Its single distinguishing sentence is this: it is the most widely used carburizing steel in the world, because it carries small amounts of all three alloying elements, and those small amounts buy a core toughness that single-element budget grades cannot deliver at the same price.​‌​​‌​

And the first thing that has to be said: when you buy 8620, you are NOT buying a hardness. Carbon is nominally 0.20 %. The maximum martensitic hardness achievable with that carbon is roughly 45 HRC — so 8620 can never reach 60 HRC on its own. The 58–62 HRC figure you see on datasheets is not the hardness of the steel; it is the hardness of carbon that a carburizing furnace loaded into the surface afterwards. The steel is only a carrier; the hardness is produced by the heat treater. That is the single most important sentence on this page, and everything below is built on it.

The engineering sits in three elements present together in small amounts. Chromium (0.40–0.60 %) and molybdenum (0.15–0.25 %) delay the pearlite and bainite transformations — that is, they provide hardenability. Nickel (0.40–0.70 %) contributes to hardenability too, but does its real work elsewhere: it solid-solution strengthens the ferrite matrix and pushes the ductile-to-brittle transition temperature DOWN — it keeps the core tough in the cold. Chromium-manganese carburizing steels (16MnCr5, 20MnCr5) have no nickel, and that is exactly where the difference shows up.​‌​​‌​

8620, 8620H, 8620RH and “carburized 8620” — four different purchase-order lines​‌​​‌​

8620 is a chemistry specification (ASTM A29 / A322, SAE J404). 8620H is a hardenability specification: the chemistry band widens slightly (C 0.17–0.23 %, Mn 0.60–0.95 %, Cr 0.35–0.65 %, Ni 0.35–0.75 %, Mo 0.15–0.25 %) but the mill guarantees the Jominy end-quench band. 8620RH (restricted hardenability, SAE J1868) narrows the band further. “Carburized 8620” is neither a chemistry nor a standard — it is a process outcome, and what defines that outcome is not the steel’s chemistry but the recipe the heat treater writes. Do not put all four on the same purchase-order line.

The practical rule: if you or your subcontractor will do the heat treatment and you want batch-to-batch repeatability, buy 8620H, not 8620. A plain “8620” order does not guarantee hardenability; two heats that both meet the specification can come out of the same furnace on the same cycle several HRC apart in the core.​‌​​‌​

Standards by Product Form

​‌​​‌​

​‌​​‌​

Specification Gaps — What “8620” Does Not Buy You​‌​​‌​

In 8620 the gap is not in the chemistry, it is in the PROCESS OUTCOME. And that gap is the source of nearly every commercial dispute around this alloy: you buy the chemistry, but what does the work is the case — and the case is usually nowhere on the purchase order.

Chemical Composition

The table below puts three specification families side by side. They are close but not identical, and the sentence “8620 = 20NiCrMo2-2” should not be said without reading the mill certificate.​‌​​‌​

Three weak points in the sentence “8620 = 20NiCrMo2-2”​‌​​‌​

First: manganese. SAE asks for 0.70–0.90 %, EN for 0.65–0.95 %. The common band is 0.70–0.90 % — here EN is the wider one, so a European heat at Mn = 0.93 % meets EN but fails SAE 8620. (It does fit the 8620H band: 0.60–0.95 %.)
Second: chromium. SAE 0.40–0.60 %, EN 0.35–0.70 %. Again EN is wider; a heat at Cr = 0.66 % fails SAE 8620.
Third, and most important: aluminium and cleanliness. EN 10084 requires Al 0.020–0.050 % and a cleaner steel at P ≤0.025 %; on the SAE side there is no aluminium requirement and P runs to 0.035 %. Aluminium is not decorative here: in a steel held for hours at 900–955 °C, AlN precipitation is the principal brake on grain growth. A heat without aluminium carries a real risk of abnormal grain growth, and with it distortion and loss of case toughness.
What to do: before issuing a dual certificate, compare the Mn, Cr, P and especially Al lines of the mill certificate one by one. If the Al line is blank, EN 10084 compliance cannot be claimed.

The resulphurised variant: 20NiCrMoS2-2 (1.6526)​‌​​‌​

EN 10084 also lists the same chemistry with sulphur forced into the 0.020–0.040 % band. Sulphur precipitates as manganese sulphide, breaks the chip and markedly improves machinability. The price: MnS inclusions elongate in the rolling direction and reduce transverse ductility and impact toughness. Rule: high-volume small parts on screw machines → 1.6526; parts loaded transversely or subject to impact → stay with 1.6523.

CARBURIZING — the Heart of This Page​‌​​‌​

Carburizing means holding the steel in the austenite range (above Ac3) in a carbon-donating atmosphere so that carbon diffuses into the surface. 8620 was designed for exactly this: the core stays at 0.20 % C and stays tough; the surface rises to 0.80–1.00 % C and, when quenched, transforms to high-carbon martensite and gets hard. You end up with two different steels in one part.

Before and after carburizing — the complete cycle

1. Normalize (BEFORE machining). 870–925 °C (1600–1700 °F), 1 h per 25 mm of section, still air cool. Result 149–179 HB. The purpose is not hardness but erasing the banded structure and residual stresses left by rolling or forging. Skip it and post-carburizing distortion becomes unpredictable.
2. Rough machining. The part is at its most machinable here (149–183 HB).
3. Stress relief (optional, but strongly advised after heavy cuts). 595–650 °C (1100–1200 °F). It removes the stresses roughing left behind, before the part goes into the carburizing furnace.
4. Masking. Surfaces that must not carburize (areas to be threaded, bearing seats, surfaces that will not be ground) are protected with copper plating or stop-off paint. The alternative: leave extra stock and machine the case away afterwards.
5. Carburize + quench + temper. The table above.
6. Cryogenic treatment (if needed). BEFORE tempering — see the next section.
7. Grinding / honing. This is where distortion is taken out. The stock left must not exceed half the case; otherwise you grind away the hardest, highest-carbon layer.​‌​​‌​

RETAINED AUSTENITE and DIMENSIONAL INSTABILITY — the Section Usually Left Out

Retained austenite in a carburized case is UNAVOIDABLE, and it is not a defect — it is physics. Here is why: carbon lowers the martensite start temperature (Ms). In the core, carbon is 0.20 % and Ms is far above room temperature — the core transforms completely. At the surface, carbon has risen to 0.80–1.00 % and Ms has fallen to near or below room temperature. Because the quench stops at room temperature, the transformation cannot finish and some austenite remains in the case.​‌​​‌​

The remedy: cryogenic (deep-freeze) treatment​‌​​‌​

The logic is simple: if the transformation stopped at room temperature, go colder and let it continue. In practice:
Temperature. The published range is wide: cryogenic chambers reach −185 °C (−300 °F), but for carburized 8620 the −68 to −79 °C (−90 to −110 °F) band is often reported as sufficient. Going to liquid-nitrogen temperature is not always necessary and brings thermal-shock risk.
Sequence — this is critical. The cryogenic step goes immediately after the quench and BEFORE tempering. The reason: tempering redistributes carbon in the retained austenite and stabilises it. Once stabilised, that austenite no longer transforms readily on subsequent cooling. Delay between quench and freeze starts the same stabilisation, so keep the wait short.
And you must still temper. The cryogenic step produces fresh martensite, and fresh martensite is stressed and brittle. Tempering after the freeze is not optional. On critical work a temper–freeze–temper double cycle is used.
When it is needed: close-tolerance gears, bearing races, hydraulic valve bodies, long-life transmission parts. When it is not: loose-tolerance wear bushings and pins.

Is retained austenite always the enemy?​‌​​‌​

No — and this is the honest side of the topic. A measured amount (roughly 15–20 % by volume) yields plastically under Hertzian contact and spreads the contact stress, and in some rolling-contact fatigue (RCF) applications it extends life. The effect is well known in the bearing and gear literature. The distinction is this: if the loading is rolling contact, measured retained austenite can help; if the governing requirement is dimensional stability or edge strength, it hurts. So “eliminate retained austenite” is not a universal rule; write an application-specific ceiling instead.

Hardenability and Core Properties​‌​​‌​

A carburized part raises two separate hardenability questions. The first is whether the case hardens — with carbon at 0.80 % that is nearly guaranteed. The second is whether the core hardens, and that is the real engineering question: if the core does not harden, the case sits on a soft foundation and collapses like an eggshell.

Three things can be read from that table.
First — the band is WIDE. At 7 mm the standard +H band is 25–42 HRC: a 17 HRC spread, and both ends fully meet the specification. So two heats both stamped “8620” can genuinely come out of the same furnace on the same cycle with different cores. The batch-to-batch scatter you are seeing is usually not a furnace fault but a scatter the standard permits.
Second — the fix is to order +HH or +HL. These grades split the band in two. At the same distance, +HH gives you 31–42 and +HL 25–36. Narrowing the band is buying repeatability. The US-side equivalent is the 8620RH grades (SAE J1868).
Third — 8620 hardens SHALLOW. At 11 mm, hardness can fall to 20 HRC. That means that a few millimetres inside the centre of an oil-quenched bar, the core will already be ferritic-bainitic. If you need core strength in a heavy section, 8620 is the wrong steel — look at 4320, 18CrNiMo7-6 or 9310.

A single-source US figure: one distributor quotes 8620H at 48 HRC maximum at 1/16″ (1.6 mm) and 32 HRC minimum at 3/16″ (4.8 mm). Those values are consistent with the EN band above but have not been independently verified; check the current SAE J1268 table before making them an acceptance criterion.​‌​​‌​

Mechanical Properties​‌​​‌​

Read the first two rows of the table below carefully: they are the properties of UN-CARBURIZED 8620, and many datasheets publish them on their own as “the mechanical properties of 8620”. They do not describe the behaviour of a carburized part.

One oddity in the table needs explaining: the annealed yield (385 MPa) appears higher than the normalized yield (357–360 MPa), even though normalized tensile strength is higher. This is a known inconsistency in the published ASM-derived data and most likely comes from different specimen diameters and mixed yield definitions (lower yield point versus 0.2 % offset). Do not use either value for design; work from the certificate of your own delivery condition.

Physical Properties​‌​​‌​

Welding — When and Why​‌​​‌​

Because its carbon is low (0.18–0.23 %), 8620 welds MORE EASILY than 4140 or 8740 — but that does not mean “without trouble”. The alloying raises hardenability, so the HAZ turns to martensite if it cools fast. Preheat is mandatory.

The critical question: weld BEFORE or AFTER carburizing?

The answer is almost always “before”. In order:
1. A carburized surface cannot be welded. Surface carbon is 0.80–1.00 %. At that carbon level the HAZ transforms to very hard, very brittle martensite and the weld cracks. The carbon equivalent jumps into through-hardening steel territory.
2. Welding heat destroys the case. Welding takes the adjacent region far above the tempering temperature and locally softens the case. Your hardness map is no longer known.
3. The required PWHT kills the case. PWHT runs at 620–650 °C, which is far above the case tempering temperature of 150–190 °C. Do the PWHT and you lose most of the case hardness; skip it and you accept the cracking risk. There is no way out of that dilemma.
The correct order: weld → stress relieve → machine → carburize → quench → temper → grind. If you are forced to weld in the hardened condition: grind the case off completely in the weld zone, raise the preheat to 260 °C, use a soft filler, cool slowly under ceramic wool, and accept that the zone will no longer be hard.​‌​​‌​

Machining

8620 machines better than 4140 or 8740 thanks to its low carbon — but that same low carbon creates the alloy’s one machining problem: a tendency to built-up edge (BUE). A soft, ductile steel welds itself to the tool nose at low cutting speed and ruins the finish.​‌​​‌​

Planning the grinding stock — the step most often skipped. Carburizing plus quenching inevitably produces distortion, so critical surfaces need grinding stock. But that stock must not exceed half the case. If you target a 1.0 mm effective case and leave 0.6 mm of grinding stock, you grind away the hardest, highest-carbon layer and what remains is no longer at the target hardness. Write the case depth into the specification as a value to be measured AFTER grinding.​‌​​‌​

Corrosion — the “NOT Stainless” Section

AISI 8620 is NOT stainless, and carburizing does not make it stainless. It contains about 0.50 % chromium. The threshold for a steel to behave passively — to build a self-repairing chromium oxide film — is roughly 10.5 % chromium. 8620 has about one twentieth of that. The chromium is there for hardenability, not corrosion resistance.​‌​​‌​

Where it does well

It does well in dry, closed systems running under an oil film. Gearbox internals, sealed reducers, gears and bearing components in an oil bath — in all of these 8620 causes no corrosion trouble at all, because the surface is permanently wetted with oil. Much of the alloy’s industrial success comes from exactly this environment.​‌​​‌​

Where it FAILS

1. Unprotected in the atmosphere. Left unpainted, unplated and unoiled, an 8620 part shows surface rust within days — like plain carbon steel.
2. A carburized surface rusts FASTER. Counter-intuitive but true: high-carbon martensite and its carbides are galvanically more active than low-carbon ferrite-pearlite. “We hardened it, so it is more durable” is wrong.
3. After grinding. A freshly ground case surface is stripped of protective oxide and may be micro-cracked. Oil or passivate immediately after grinding; a ground gear left overnight on the bench comes back with rust staining.
4. Seawater, salt spray, road salt. Not usable. Plating is mandatory (zinc, zinc-nickel, phosphate + oil, cadmium in aerospace).
5. Sour (H₂S) service. Under NACE MR0175 / ISO 15156, the typical limit for carbon and low-alloy steels is 22 HRC. A carburized 8620 surface is 58–62 HRC — far above it. Carburized 8620 is not suitable for sour service. This is almost never stated on datasheets and causes real failures in oilfield equipment.
6. Hydrogen from electroplating. A carburized surface is very hard and very high strength, and therefore susceptible to hydrogen charged in during electroplating or acid pickling. Post-plate baking must be a purchase requirement.
7. Oxidising acids, chloride-bearing aqueous media, food and pharmaceutical cleaning cycles. Unusable in all of them.​‌​​‌​

If you need corrosion resistance as well as a hard case

There are three routes, and none of them is 8620. First, nitriding — it adds some corrosion resistance (especially nitrocarburizing + oxidation cycles) but the layer is shallow. Second, a martensitic stainless (440C, 17-4 PH): corrosion resistance and hardness together, at a cost in fracture toughness and price. Third, stainless carburizing grades (e.g. high-nickel aerospace gear steels): expensive and long lead time. The fourth “route” — plate 8620 and manage it — is often the cheapest, but decide on the basis of plating life and hydrogen risk, not first cost.​‌​​‌​

Frequently Asked Questions

The datasheet says “8620 hardness 58–62 HRC” but the bar we received measured 180 HB. Did we get the wrong material?​‌​​‌​

Most likely the material is right and the datasheet is wrong — or at least badly written.
To be clear: the as-delivered hardness of 8620 bar is typically 149–212 HB, depending on whether it is annealed, normalized or as-rolled. 180 HB is entirely normal and is exactly what a normalized 8620 is expected to be (the published typical value is 183 HB).
58–62 HRC is NOT a delivery condition; it is a PROCESS RESULT. To reach it, the part must be held for hours at 900–955 °C in a carbon-donating atmosphere, then oil quenched and tempered — that is, carbon has to be loaded into the surface afterwards. The steel’s own carbon is 0.20 %, and 58 HRC is physically impossible with it.
The confusion is so common that many distributor pages publish the line “Hardness: 58–62 HRC” with no qualification at all. The result: the buyer expects hard bar, receives soft bar, and a return process begins.
What to do: write two things separately on the order — (a) the delivery condition and hardness band of the bar, and (b) if carburizing is to follow, the target effective case depth, surface hardness and core hardness. They cannot go on the same line because they are not the same thing.

Our gears come out in tolerance after carburizing but seize in the field within months. What causes that?​‌​​‌​

Your first suspect should be RETAINED AUSTENITE. What you describe is its classic signature.
The mechanism: in the carburized surface, carbon rises to 0.80–1.00 % and that carbon drives the martensite start temperature (Ms) to or below room temperature. Because the oil quench stops at room temperature, the transformation does not complete and some austenite remains in the case. In service, stress, cold and time gradually convert it to martensite — and martensite occupies more volume than austenite. So the part GROWS OVER TIME. Backlash closes, bearing clearance shrinks, the part seizes.
Why your inspection does not catch it: the part is in tolerance when it leaves the factory. The transformation takes months. Hardness testing will not show it either — in fact hardness creeps slightly upward.
Confirmation: measure retained austenite by X-ray diffraction (ASTM E975). Test one failed part and one new part. High in the new part and low in the returned one and the diagnosis is certain.
The fix has three steps. First and most effective: add a cryogenic step — immediately after the quench, BEFORE tempering, in the −68 to −79 °C band. The sequence matters: tempering stabilises retained austenite, and freezing afterwards achieves much less. Second: lower the surface carbon — pull the carbon potential back from 1.1 % to the 0.85–0.90 % band; the higher the surface carbon, the more retained austenite. Third: move from direct quenching to reheat quenching — finer grain, less retained austenite. And write a ceiling into your specification (typically 15–20 % by volume), together with the measurement method.

We carburized a heavy shaft; the case is hard but the teeth collapsed under load. Is the case too thin?​‌​​‌​

Most likely the case is fine and the CORE is inadequate. This is 8620’s most typical application error.
The physics: a hard case carries load only if there is a hard foundation beneath it to support it. If the core did not harden, the case collapses like an eggshell — no matter how hard the case itself is.
And 8620 is a shallow-hardening steel. Look at the Jominy band: at 11 mm from the quenched end the standard +H band has fallen to 20–33 HRC. The centre of a heavy oil-quenched bar cools even more slowly than that. In heavy sections the core stays ferritic-bainitic and has no load-carrying capacity.
The numerical check: in EN practice the minimum core Rm after case hardening is 1100 MPa at ⌀≤16 mm, 800 MPa at ⌀16–40 mm and 700 MPa at ⌀40–100 mm. The same steel loses a third of its core strength when the section quadruples. Which number was your design based on?
The fix, in order: (1) Buy a restricted-hardenability grade — +HH (or 8620RH on the US side). That guarantees the upper half of the band and is essentially a free gain. (2) If that is not enough, move to AISI 4320: same carbon, but nickel at 1.65–2.00 % hardens markedly deeper. (3) Still not enough — 18CrNiMo7-6 in Europe, AISI 9310 in aerospace. (4) Increase quench severity (more agitation, faster oil) — but know that distortion and cracking risk rise with it.
Making the case thicker does NOT fix this, and can make it worse: a thicker case means more retained austenite and more distortion.

A customer wants “carburized 8620, 60 HRC, NACE compliant”. Can we supply it?​‌​​‌​

No. That specification contradicts itself and should be rejected.
NACE MR0175 / ISO 15156-2 typically imposes a 22 HRC maximum and a quenched-and-tempered condition on carbon and low-alloy steels in sour (H₂S) service. The reason is sulphide stress cracking (SSC): hard martensite, charged with hydrogen in an H₂S environment, cracks without warning under stress.
A carburized 8620 surface is 58–62 HRC — nearly three times the limit. And that hardness is exactly at the outer surface, the part in contact with the environment. These two requirements cannot be met on the same part.
The NACE limit is also LOCAL, not an average. “The core is 30 HRC, so the average is acceptable” is not a defence: weld metal, HAZ, rolled thread roots and the carburized surface must each individually sit below the limit.
The honest answer to the customer has three options. (1) If sour service is real: drop the carburizing and use a steel quenched and tempered below 22 HRC (the logic of 4140 / A193 B7M) — and obtain wear resistance by another route. (2) If wear resistance is non-negotiable: isolate the hard surface from the environment (closed oil-filled system, coating, sleeve) and describe that in writing. (3) If both are genuinely required: the material class has to change — nickel-base alloys or purpose-built duplex/stainless solutions. Under no circumstances issue a written statement that a carburized part is NACE compliant.

We hold bar certified to 20NiCrMo2-2 and the customer wants SAE 8620. Can we ship it?​‌​​‌​

Not without reading the certificate line by line — and the trap here is DIFFERENT from the 4140/42CrMo4 case.
On chemistry, the EN band is WIDER than the SAE band. On manganese, EN asks for 0.65–0.95 % and SAE for 0.70–0.90 %. On chromium, EN 0.35–0.70 % and SAE 0.40–0.60 %. So not every heat that meets EN meets SAE 8620: a European heat at Mn = 0.93 % or Cr = 0.66 % satisfies EN 10084 perfectly but falls outside the SAE 8620 band. (Interestingly it does largely fit the 8620H band — Mn 0.60–0.95 %, Cr 0.35–0.65 %.)
In the other direction, EN demands a CLEANER steel: P ≤0.025 % (SAE ≤0.035 %) and, crucially, Al 0.020–0.050 % is mandatory. So a US heat certified to SAE 8620 cannot claim EN 10084 compliance if the aluminium line is blank. And that line is not decorative: in a steel held for hours at carburizing temperature, AlN precipitation is the principal brake on grain growth.
What to do: compare the Mn, Cr, P and Al lines of the mill certificate against the target bands one by one. If they all fit, a dual certificate can be issued — and many mills deliberately produce narrow heats that satisfy both bands. But that is a choice, not a rule. If they do not fit, supply the customer material to the specification they asked for; “same steel” is not a defensible sentence in a third-party audit.

Common datasheet errors — check these before you order​‌​​‌​

1. “8620 hardness: 58–62 HRC” — THE MOST COMMON ERROR. That is the hardness of the carburized case, not the steel. As-delivered 8620 bar is 149–212 HB. 58 HRC is physically impossible at 0.20 % carbon.
2. Case hardness quoted as if it were core hardness. A carburized part has two hardnesses: surface 58–62 HRC, core 25–40 HRC. Any line that does not say which one will be measured generates a dispute.
3. Case depth quoted with no definition. Effective (to 50 HRC) and total (to the core structure) differ by 1.5–2×. “0.8 mm case” — by which definition?
4. Time-versus-depth tables conflict. One source gives 18 h for 1.5 mm at 925 °C, another 16 h; one gives 4 h for 0.75 mm while another gives 2 h for 0.5 mm. The difference comes from how the boost/diffusion cycle is built. Make none of them an acceptance criterion.
5. Values like “Rc 90” get published — A UNIT ERROR. The Rockwell C scale does not go above about 70. That number is almost certainly HR15N. The same pages print thermal conductivity as “26” and modulus as “31” with no units at all (Btu and 10⁶ psi).
6. Retained austenite is never mentioned. It is the most common cause of dimensional instability and no base specification sets a ceiling. You have to write one.
7. The SEQUENCE of the cryogenic step is described wrongly. The correct order is quench → cryogenic → temper. Freezing after tempering is far less effective, because tempering stabilises the austenite.
8. A single “core strength” number is published. The same steel has a minimum core Rm of 1100 MPa at ⌀16 mm and 700 MPa at ⌀100 mm. Any value quoted without a section size is meaningless.
9. Jominy values are given as single numbers. The EN +H band at 7 mm is 25–42 HRC — a 17 HRC spread, both ends compliant. The batch-to-batch scatter you see is scatter the standard permits. The cure: order +HH / +HL or 8620RH.
10. “8620 = 20NiCrMo2-2” — INCOMPLETE. The EN band is wider on Mn and Cr, but tighter on P, and Al 0.020–0.050 % is MANDATORY. EN compliance cannot be claimed from a certificate with a blank aluminium line.
11. Nobody questions why annealed yield exceeds normalized yield. The published data give 385 MPa annealed against 357–360 MPa normalized. That inconsistency comes from different specimen diameters and yield definitions; neither is a design value.
12. AMS 6274, 6276 and 6277 are used interchangeably. The difference is melting practice: 6274 vacuum degassed (VD), 6276 CEVM/VAR, 6277 CEM. And 6276/6277 tighten the chemistry to P ≤0.012 %, S ≤0.010 %. Get the number right on aerospace orders.
13. “Carburizing makes 8620 corrosion resistant” — WRONG, AND THE OPPOSITE IS TRUE. High-carbon martensite is more active and rusts faster than the low-carbon structure.
14. NACE compliance is claimed for carburized parts. The limit is 22 HRC; a carburized surface is 58–62 HRC. It cannot be met.
15. Welding AFTER carburizing is recommended. Surface carbon is 0.80–1.00 %; the HAZ cracks. And the required PWHT (620–650 °C) is far above the case tempering temperature and destroys the case. Correct order: weld FIRST.
16. Grinding stock is planned independently of case depth. If the stock exceeds half the case, you grind off the hardest layer. Specify case depth as measured AFTER grinding.
17. “Poor finish — slow down” — BACKWARDS on 8620. Poor finish is usually built-up edge; the cure is a sharp tool and a HIGHER speed.
18. 8620 is compared with 4140/8740 on the “hardness” axis. 8620 is a carburizing steel (0.18–0.23 % C); the others are through-hardening steels (0.38–0.43 % C). They answer different questions.
19. Thermal expansion is given as a single number. It runs from 10.5 ×10⁻⁶ /K (−100 °C) to 14.4 ×10⁻⁶ /K (600 °C). A single figure produces errors in shrink-fit calculations. Likewise modulus falls from 212 GPa to 164 GPa.
20. “8620 was made for carburizing, so it suits every carburizing job” — INCOMPLETE. 8620 hardens shallow. Large-module gears, heavy sections and low-temperature impact call for 4320, 18CrNiMo7-6 or 9310.
21. The normalizing step is skipped. An 8620 that was not normalized before machining distorts unpredictably after carburizing. That step is not a cost, it is insurance.
22. Resulphurised 20NiCrMoS2-2 is offered in place of 20NiCrMo2-2. Sulphur improves machinability but reduces transverse ductility and impact toughness.

​‌​​‌​

​‌​​‌​

​‌​​‌​