Yield Strength vs Tensile Strength: How to Read Them for CNC Part Selection

Yield strength is the stress where a metal starts to deform permanently, and it is the value you design to. Tensile strength (UTS) is the highest stress the material reaches before it necks and breaks, and it sits well above yield. For a CNC part that has to hold its shape under load, size the part against yield strength divided by a safety factor, not against UTS. A 6061-T6 bracket with a published yield near 276 MPa and a tensile strength near 310 MPa bends permanently long before it reaches that 310 MPa number. This guide covers how to read both values off a datasheet, how they sit on the stress-strain curve, what the yield-to-tensile ratio says about margin, and where the two numbers stop behaving the way a simple table implies. Strength values here are published typical figures for the stated temper and product form; definitions follow ISO 6892-1:2019 and ASTM E8/E8M-25

1. Yield strength vs tensile strength: the one-line answer for part selection

Yield strength answers one question: at what stress does this part stop springing back? Below yield, load the part and it returns to shape when you remove the load. Push past yield and some deformation stays. That permanent set is usually the point where a machined part stops doing its job, because a shaft that bends 0.3 mm or a flange that no longer sits flat is scrap even if it never cracked.

Tensile strength answers a different question: what is the highest stress the material can take before it starts to fail? On a tensile test the load climbs past yield, the material work-hardens, reaches a peak, then the cross-section necks down and the part breaks. That peak stress is the tensile strength, also written UTS. It tells you how far the material is from separating, not how far it is from becoming useless.

For part selection the split is clean. Use yield strength to set the working stress. Use tensile strength to understand the failure margin above yield and to read ductility. A 304 stainless part with a published yield near 215 MPa and tensile strength near 505 MPa has a large gap between the two, so it deforms a lot before it breaks. A 7075-T6 part with yield near 503 MPa and tensile near 572 MPa has a narrow gap, so it carries more load before yielding but gives far less warning before fracture. Cross-check these against the tensile and yield strength values for common CNC metals when you are picking stock.

EPOC CRAFTER engineering note: On an RFQ, the yield number drives the go/no-go on whether a part survives its service load. When a customer sends only a tensile strength target, we ask for the yield value or the load case, because a part sized on UTS alone is sized on the wrong number.

1.1 What yield strength tells you (the design boundary)

Yield strength is the ceiling on stress you let a working part see. Divide it by a safety factor to get the allowable stress, then keep the calculated stress under that. Published yield values assume a specific temper and product form, so a 6061 value only holds if your stock is actually T6 and the section matches the test bar. Heat treatment moves this number, which is why how heat treatment shifts yield and hardness matters before you lock a material callout.

1.2 What tensile strength tells you (the failure ceiling)

Tensile strength marks the top of the curve, and the distance from yield up to that peak is the plastic reserve. A wide gap means the part deforms visibly before it fails, which gives an inspector or an operator a warning. A narrow gap means the part holds more load but jumps from fine to fractured with little visible bending. Tensile strength is not the fracture stress; the material breaks at a lower apparent stress after necking.

2. Why the datasheet UTS is not the load your part can carry

A datasheet lists tensile strength as the headline number because it is the biggest one, and that number misleads on load capacity. The part fails as a functional part at yield, not at UTS, so the working load has to stay below yield with margin to spare. The gap between the two values is plastic deformation you are not allowed to use on a real part.

Run the 6061-T6 case. Published yield sits near 276 MPa, tensile near 310 MPa. Design the bracket to the tensile number and you have signed off on a part that permanently bends at 276 MPa, roughly 11 percent below the stress you thought was your limit. By the time the material is near 310 MPa it has already yielded, stretched, and lost the dimensions on the drawing. The part failed inspection long before it failed the tensile test.

The safety factor makes the split wider on purpose. Take that 276 MPa yield and a safety factor of 2, and the allowable working stress is about 138 MPa. That is the real ceiling for the working load, less than half the tensile strength on the datasheet. Anyone sizing the part off the 310 MPa figure is off by more than a factor of two.

Two numbers off the same datasheet, two very different limits:

Yield strength (Rp0.2)	~276 MPa	Onset of permanent set	Sizing the working load
Allowable stress (SF = 2)	~138 MPa	Design ceiling with margin	The number the part lives under
Tensile strength (UTS, Rm)	~310 MPa	Onset of necking before fracture	Reading plastic reserve and ductility

Values are published typical figures for 6061-T6 bar; confirm against the certified mill test report for your stock. Definitions per ISO 6892-1:2019.

The safety factor itself is not a fixed number. It comes from the load case, how well the load is known, the failure consequence, and whether the part sees fatigue or a single static load. A lightly loaded cover and a rotating shaft under cyclic load do not get the same factor. What holds across all of them is the base: the factor divides yield, not tensile.

EPOC CRAFTER engineering note: The common RFQ mistake is a customer citing a high UTS to argue a thin section is strong enough. The section can hit that UTS in a lab and still be wrong for the part, because the part yields first and the drawing dimensions are already gone. When the load case is real, we size against yield over the safety factor and treat UTS as the fracture-warning read, not the load number.

There is one honest exception. A brittle material with almost no plastic region, like gray cast iron, does not give a useful yield-based design margin the same way, because it fractures with little warning near its ultimate stress. That case is handled in section 7, and the split between how cast iron and steel differ under load shows up whenever you compare them for load-bearing work.

3. How the two values sit on the stress-strain curve

Pull a metal test bar in tension, plot stress against strain, and both numbers show up as specific points on one curve. Reading them off the curve is what keeps yield, tensile, and fracture from getting mixed up. ISO 6892-1:2019 defines each point and gives it a symbol, so the terms below are the standard ones, not shop shorthand.

The curve starts as a straight line. Stress and strain climb together, the metal is elastic, and it springs back if you release it. That straight part ends at yield. Past yield the curve bends into the plastic region, the metal keeps stretching, work-hardens, and the stress climbs more slowly to a peak. After the peak the bar necks, the apparent stress drops, and the bar breaks.

Stress on this curve is engineering stress: force divided by the original cross-section, written R = F/S0 in the standard. That detail is the reason tensile strength is not the fracture stress.

Where each value lands:

Yield, well-defined case (ReH and ReL). Some metals, many carbon steels among them, show a distinct yield point. The stress rises, drops at the first yield, then wobbles along a plateau. ISO 6892-1 calls the peak before that first drop the upper yield strength, ReH, and the lowest stress during the plateau the lower yield strength, ReL. Ignore the initial transient when reading ReL.

Yield, no clear point (Rp0.2). Aluminum, most stainless, titanium, and cold-worked metals bend smoothly out of the elastic line with no visible yield point. Here the standard uses proof strength: draw a line parallel to the elastic slope, offset it by 0.2 percent strain, and the stress where it crosses the curve is Rp0.2. Section 5 covers this in full.

Tensile strength (Rm). Rm is the stress at maximum force, Rm = Fm/S0. It is the top of the engineering curve, the point where necking begins, not the point where the bar separates.

That last point is where a lot of writing gets it wrong. Tensile strength is not the stress at fracture. The bar reaches maximum force at Rm, then necks down. Because engineering stress still divides by the original area S0, the plotted stress falls after Rm even though the metal in the neck is being pulled harder over a smaller area. The bar finally breaks at an apparent stress below Rm. So the order on the curve is fixed: yield first, then Rm at the peak, then fracture lower and to the right. Calling Rm the breaking stress reads the curve backward.

EPOC CRAFTER engineering note: For part design you never operate near Rm or ReH. Both live deep in territory where the part has already deformed. The working stress sits on the elastic line, below yield, with the safety factor holding the gap. The upper reaches of the curve are for reading how the material fails, not for setting the load.

One reminder the standard is strict about: these values are test outputs tied to conditions. ISO 6892-1 runs at room temperature, defined as 10 °C to 35 °C, or 23 °C ± 5 °C under controlled conditions. A published yield or tensile number carries the material state, product form, and test method behind it, which is why the same alloy in a different temper gives a different number.

4. Yield-to-tensile ratio and how much margin you have

Divide yield strength by tensile strength and you get the yield-to-tensile ratio. It tells you how much of the material’s total strength you are already using at the point where design stops. A low ratio means yield sits far below tensile, so the part deforms a lot before it breaks. A high ratio means yield sits close to tensile, so the part carries more load before yielding but gives little warning before fracture.

The ratio is a read on plastic reserve, not on which material is stronger. A part that has to bend, absorb a crash, or show visible distortion before failure wants a low ratio. A part where you need maximum load in a fixed section and can accept a brittle-style failure tolerates a high ratio, as long as fatigue and notch sensitivity are handled separately.

Eight common CNC bar metals, published typical values, ranked by what the ratio means for part behavior:

304 stainless (annealed)	215	505	0.43	Huge plastic reserve; visible warning before failure
316L stainless (annealed)	170	485	0.35	More reserve than 304; low yield, size sections generously
C360 brass (free-machining)	125	340	0.37	Wide reserve, easy to machine; low yield, not structural
1045 carbon steel (cold drawn)	450	585	0.77	Moderate reserve; higher yield with some warning left
6061-T6 aluminum	276	310	0.89	Narrow reserve; yields close to its tensile limit
4140 alloy steel (normalized)	655	1020	0.64	High yield with real reserve; heat treatable to shift both
Ti-6Al-4V (annealed)	880	950	0.93	Very high yield, narrow reserve; abrupt failure
7075-T6 aluminum	503	572	0.88	High aluminum yield, narrow reserve; low fracture warning

Published typical figures for the stated temper and product form, cross-checked across supplier and materials-reference datasheets. Confirm against the certified mill test report for your stock. Ratios rounded.

Read the table by what the part has to do, not by the biggest number. 304 and 316L give up yield strength to gain a large safe-deformation window, which is why they show up where a bend is recoverable but a crack is not. Ti-6Al-4V and 7075-T6 do the opposite: high yield in a small section, but the gap to fracture is thin, so once they yield they are close to done. 4140 is the middle ground, high yield with reserve still in hand, and both numbers move with heat treatment, which is where choosing between 4140, 1018 and 1045 steel turns into a real decision.

EPOC CRAFTER engineering note: A high yield-to-tensile ratio is not a defect, it is a trade. On a 7075-T6 or titanium part we treat the narrow gap as a design fact: less forgiveness for overload, so the load case has to be well known before the section is trusted. On 304 or 316L the wide gap buys tolerance for an uncertain load, at the cost of a lower yield you size around.

One limit on the ratio: it describes static tensile behavior only. It says nothing about fatigue life, notch sensitivity, or impact toughness, and a good static ratio can sit next to poor fatigue performance. Those are separate reads.

5. Metals with no clear yield point: how Rp0.2 is defined

Most non-ferrous metals and many stainless grades never show a sharp yield point. The stress-strain line curves out of the elastic region gradually, with no drop and no plateau to read. Aluminum, copper alloys, titanium, and cold-worked metals all behave this way. You cannot point at a single yield spot on the curve, so ISO 6892-1 defines one by construction.

The value is proof strength, and the common one is Rp0.2. The construction the standard lays out:

1. Plot the stress-extension curve and find its straight elastic portion.

2. Draw a line parallel to that elastic portion.

3. Offset it along the strain axis by a set plastic extension. For Rp0.2 that offset is 0.2 percent.

4. Where the offset line crosses the curve, read the stress. That stress is Rp0.2.

That is the number datasheets print as the yield strength for aluminum and stainless. When a 6061-T6 sheet lists 276 MPa yield, that figure is Rp0.2 by the 0.2 percent offset method, not a natural yield point the metal shows.

Two points the standard is careful about, and where casual writing slips:

0.2 percent is a chosen value, not a rule for all metals. ISO 6892-1 does not say every metal without a clear yield point must use 0.2 percent. It defines the proof-strength method and lets the offset be specified. Rp0.2 is the widely used case, written into most material specs, but the standard also defines Rp with other offsets and defines a second proof strength, Rt, based on total extension rather than plastic offset. So 0.2 percent proof is a specification choice that happens to be near-universal, not a law of the material.

Rp0.2 is not the upper or lower yield strength. ReH and ReL only exist for metals that show a real yield point. Rp0.2 is the offset construction for metals that do not. Writing that a standard defines the 0.2 percent proof strength as the upper or lower yield strength is wrong on the standard’s own terms; these are separate definitions for different curve shapes. A datasheet gives you one or the other depending on the metal, not both for the same value.

EPOC CRAFTER engineering note: For part design the distinction rarely changes the working stress, since you divide the listed yield by a safety factor either way. It changes how you compare datasheets. A carbon steel yield read off a real yield point and an aluminum yield read off a 0.2 percent offset are both yield on paper but come from different constructions, so treat the aluminum number as the offset value it is when you cross-compare materials.

One caution the standard flags on the low side of the same test: for a metal that does show a yield point, ISO 6892-1 allows a shortened procedure to read ReL when the yield point extension does not need to be determined, and it requires that shortcut to be recorded in the test report. It is a defined, conditional procedure, not a default you assume. If a report’s yield number matters to your part, the method behind it is worth knowing.

6. Reading a strength table for CNC part selection

A strength table is only useful if you read it in the right order. The load case sets the yield floor, the yield-to-tensile ratio tells you how much warning you get, and the temper line tells you whether the numbers even apply to your stock. Skip that order and you pick a material off the biggest number, which is how UTS ends up driving a decision it should not.

Work a table like this:

1. Start from the working load, not the material. Get the stress the part sees in service, apply the safety factor to yield, and set the allowable stress. That gives the minimum yield you can accept.

2. Filter by yield first. Drop every material whose listed yield, minus its margin, sits below your allowable stress. This is where the part survives or bends, so it comes before anything else.

3. Read the ratio for warning behavior. Among the materials that pass on yield, use the yield-to-tensile ratio to sort by how the part fails. Low ratio for parts that must show distortion before failure, high ratio where you need load per section and can accept abrupt failure.

4. Check the temper and product form on every value. A 6061 number is a T6 number. A 4140 number is a normalized or a quenched-and-tempered number, and they are far apart. If your stock condition does not match the table’s condition, the number does not transfer. Bar, plate, and extrusion of the same alloy can also differ.

5. Bring in the non-strength reads before you commit. Machinability, heat treatability, corrosion resistance, and cost decide between two materials that both clear the yield filter.

That last step is where a strength-only pick goes wrong. 316L and Ti-6Al-4V both survive plenty of load cases on yield, but one machines easily and cheaply and one is slow and tool-hungry. 4140 clears high loads but its listed strength assumes a heat-treat step that changes your process route and cost. The strength table narrows the field; it does not finish the choice. For the properties that sit next to strength, the material property references and the ASTM-traced sheet metal material mechanical data carry the machinability, thermal, and forming numbers that decide the tie.

EPOC CRAFTER engineering note: On a real RFQ the yield filter usually leaves two or three candidates, and the winner is decided by machinability and heat-treat route, not by another 30 MPa of yield. A part that needs 4140 quenched and tempered is a different quote from the same part in 1045 cold drawn, because the heat-treat sequence changes tolerance control and lead time, not just the strength on paper. That is a process decision as much as a material one, which is why the heat treatment sequence around machining belongs in the material call.

One thing a strength table never shows: how tight the part has to be after it is made. Strength picks the metal; the drawing tolerances pick the process and the cost, and the two decisions run in parallel. Where those tolerances need to be tight is its own call, covered in where tighter CNC tolerances are worth the cost.

7. Boundaries and exceptions

The yield-first rule holds for ductile metals under a known static load. Push outside that box and the rule bends. Four cases matter for CNC part selection.

Brittle materials break near their ultimate stress, not at a useful yield. Gray cast iron has almost no plastic region. It does not stretch and warn; it reaches its ultimate stress and fractures. Designing it on a yield-minus-safety-factor basis the way you would 6061 does not fit, because the plastic reserve that makes yield a safe working limit is not there. Brittle materials get sized on their ultimate strength with a larger safety factor and careful attention to stress concentrations, which is why the design approach for cast iron differs from steel, not just the number.

Slender parts can buckle before they yield. A long thin column or a thin wall under compression can fail by buckling at a stress well below its yield strength. That failure is geometric, not a material limit, so a high-yield alloy does not save it. When a part is slender and loaded in compression, check buckling separately; yield strength alone does not predict it.

Temperature and heat treatment move the numbers. Every value here is a room-temperature figure, 10 °C to 35 °C per ISO 6892-1. Yield and tensile strength both drop as temperature climbs, so a part running hot does not have its datasheet strength. Heat treatment moves the numbers the other way and by a lot: 4140 normalized and 4140 quenched-and-tempered are different materials on a strength table even though the bar stock is the same alloy. A yield number without its temper and temperature attached is not usable.

Engineering stress hides what happens after necking. The values here are engineering stress, force over original area. True stress, which uses the actual necked area, keeps rising after Rm even though the plotted engineering curve falls. For part design this stays in the background, because you work below yield and never reach necking. It matters when someone reads a fracture point off an engineering curve and concludes the material got weaker at the end, which is a plotting artifact, not the metal losing strength.

These exceptions tell you when to stop applying the yield-first rule, not that it was wrong. For a ductile CNC part under a known static load, size against yield over a safety factor and read tensile strength for failure warning.

8. FAQ

Is tensile strength the same as yield strength?

No. Yield strength is the stress where permanent deformation starts. Tensile strength is the highest stress the material reaches before necking and fracture, and it sits above yield. On a datasheet they are two separate numbers with two separate jobs.

Why is yield strength lower than tensile strength?

Yield marks where the metal starts to deform permanently, which happens partway up the stress-strain curve, while tensile strength marks the peak of that same curve after the metal work-hardens. The metal keeps carrying more load after it yields, up to the tensile peak, so tensile is the higher number for nearly all engineering metals.

Which one do I design a CNC part to?

Yield strength, divided by a safety factor. The part stops doing its job when it deforms permanently, and that is yield. Tensile strength tells you how much reserve sits above yield and how much warning you get before fracture, but it is not the number you size the load against.

What does the yield-to-tensile ratio tell me?

How much plastic reserve the material has. A low ratio, like 304 stainless near 0.43, means the part deforms a lot before it breaks and gives visible warning. A high ratio, like Ti-6Al-4V near 0.93, means high load capacity in a small section but little warning before fracture.

Is Rp0.2 the yield strength for aluminum?

Yes, in effect. Aluminum has no sharp yield point, so ISO 6892-1 defines its yield by the 0.2 percent proof method, Rp0.2. The yield value on an aluminum datasheet is that offset construction, not a natural yield point the metal shows on the curve.

The yield value carries the part; the tensile value tells you how it lets go. Size the load against yield, read the ratio for how much warning you get, and confirm both against the mill test report for the exact temper and form you are buying. When you have the load case and the material shortlist, send your part drawing for a quote and we will confirm the material call against the working load.

Related Resources

Related Material Guide: tensile and yield strength values for common CNC metals. Full strength, hardness, and density table for eight CNC bar metals.

Related Article: choosing between 4140, 1018 and 1045 steel. How yield and hardenability decide between three common steels.

Related Capability: CNC-machined load-bearing parts. The process behind the parts this material call feeds.

Related Standards Reference: design-for-manufacturing guidance. How material and tolerance decisions land on a drawing.