Key takeaways
- Matching a 2 mm CRCA bracket in brass needs about 2.5 mm of brass and ends up 36% heavier, because brass modulus is roughly 110 GPa against 200 GPa for steel.
- At 28% IACS, brass needs about 3.6 times the copper section for the same resistance, so any part carrying more than roughly 100 A belongs in copper.
- Brass cannot be arc welded in production: zinc boils at 907 °C and leaves porosity, so a welded frame is a steel or stainless part by definition.
- At 0.5–3% lead, free-machining brass sits outside the 0.1% RoHS threshold unless an exemption applies to the application.
- Brass earns its place when a part must conduct, resist corrosion bare, draw deeply and stay non-magnetic at once; drop any one of those and a cheaper material usually wins.
Brass material selection usually starts with a part that worked, in a material nobody questioned, and a new drawing that copies it. The quotation then comes back several times the steel price and the meeting turns into an argument about the supplier rather than the alloy.
The rule is short. Brass earns its place when a part has to conduct, resist corrosion without a coating, draw deeper than steel will, and stay non-magnetic — ideally two of those at once. Drop any one of them and there is usually a cheaper material that does the remaining job better. What follows is the arithmetic behind that statement, and the point at which each alternative takes over.
Brass material selection: the decision rule, part by part
Read the requirement in the first column, not the material in the drawing title. The fourth column is what we would quote at Ganesh, which is a narrower set of processes than a specialist brass house would offer.
| Requirement | Brass verdict | Better route | Crossover number and Ganesh note |
|---|---|---|---|
| Welded into a frame or bracket | Wrong | CRCA or 304 stainless | Zinc boils at 907 °C; our robotic cells run steel MIG |
| Stiffness at the drawn thickness | Wrong | CRCA, 200 GPa | Brass at 110 GPa needs 1.26× the thickness for equal stiffness |
| Carries more than about 100 A | Wrong | C11000 copper | 28% IACS against 101% IACS, so 3.6× the section |
| Weight-critical, outdoor | Wrong | 5052 aluminium | 8.5 g/cm³ against 2.68 g/cm³ |
| Corrosion resistance with no coating | Marginal | 304 or 316L | Brass tarnishes; 63/37 dezincifies in chloride |
| Deep draw plus conductivity | Right | C26000 brass | Draw ratio to about 2.0 in multiple passes |
| Decorative hardware, plated | Right | C27000 or C27200 | Plating adhesion is good; plating is bought in |
| Food or potable-water contact | Conditional | Lead-free brass or 316L | Lead must be declared on the mill certificate |
| High-volume flat cover | Wrong | CRCA with powder coat | Tooling pays back above roughly 5,000 parts a year |
| Threaded boss in a thin panel | Wrong | Free-cutting brass rod or a welded nut | Sheet brass strips at M6 in 1.5 mm |
| Marine bracket | Conditional | Naval brass, DZR brass or 316L | Sheet availability decides; confirm before design |
The two rows people argue about are stiffness and current. Both have a number, and both numbers are worse for brass than the drawing usually assumes.
The stiffness calculation that ends the substitution argument
Brass modulus is about 110 GPa; CRCA is about 200 GPa. Bending stiffness goes with the cube of thickness, so a brass part of the same thickness is 45% as stiff as the steel part it replaces.
Matching the steel stiffness means scaling thickness by the cube root of the modulus ratio: (200/110)^(1/3) = 1.22. So a 2 mm steel bracket becomes a 2.44 mm brass bracket, which is not a stock thickness, so it becomes 2.5 mm. Mass per square metre then goes from 2.0 × 7.85 = 15.7 kg/m² for steel to 2.5 × 8.5 = 21.25 kg/m² for brass. The brass part is 35% heavier and, at the same section, no stronger.
That is the number to put in front of the person who asked for brass because it "feels more solid". It also explains why brass covers on lift cars and genset canopies are rare: the mass lands on the door gear and the hinges, and every hinge then has to be re-rated.
Where deflection does not matter — a decorative trim, a blanking plate, a shroud with no load — the stiffness argument disappears and the decision moves to corrosion and appearance. That is the honest boundary.
Where the conductivity argument breaks
Brass is a conductor, and that is often the reason it was chosen. At 28% IACS for C26000, it is a poor one.
Take a 1,000 A busbar. A 50 × 10 mm copper bar has 500 mm² of section and a resistance of about 34 µΩ per metre, so it dissipates 34 W per metre at full load and drops 34 mV. The same bar in brass has 3.6 times the resistance, dissipates about 123 W per metre and drops 123 mV. To match the copper loss the brass bar needs 1,790 mm², which is 50 × 36 mm: a bar that no longer fits the enclosure and weighs 7.6 kg per metre against 4.5 kg for the copper it replaced.
That is why brass appears in terminals, contact pins and small current-carrying brackets rather than in busbars. It machines well, takes a thread, and holds a contact force. Above roughly 100 A, copper wins on both loss and volume, and the enclosure design should follow the copper.
The joining limit is absolute
Brass cannot be arc welded in production. Zinc boils at 907 °C, below the 900–940 °C melting range of the alloy, so the weld pool loses its zinc as fume before the copper ever freezes. The result is a porous, weak joint, a plume of metal oxide over the operator, and a part that fails a bend test.
Brazing works, at 600–800 °C with a copper-zinc or silver-bearing filler, and soldering works for electrical joints. Both are slower than MIG welding and both leave a heat-affected zone that softens a hard temper. Spot welding fails for a different reason: at 28% IACS the sheet conducts the current away faster than a nugget can form.
The procurement consequence is simple. If the assembly drawing shows a brass bracket welded to a steel frame, the drawing is asking for two incompatible processes joined by a joint that will not exist. Split it into a steel bracket with a brass insert, or change the bracket to steel and plate it.
What brass costs you after the quote
The piece price is only the first line. Brass carries four supply-chain costs that steel does not.
Material segregation. Brass swarf and offcuts cannot go into the steel scrap bin. Once mixed, the whole load is downgraded, and a shop that runs brass occasionally will charge for the segregation whether it says so or not.
Grade availability. C26000 in 0.5–3 mm is a stock item in most Indian markets. A dezincification-resistant grade or naval brass in the same thickness is often a mill order, and the lead time is measured in weeks. A design that needs DZR and a six-week material lead is a design that should have been 316L.
Tooling. Press tools for brass need different clearances and a polished die face, because the alloy galling on a rough die is worse than steel. The tool is quoted separately from the piece price, and it is not transferable to a steel part of the same shape.
Compliance. Free-machining brass carries 0.5–3% lead for chip control. That is fine for a machined fitting and a problem for a sheet part sold into electrical equipment, where the RoHS limit is 0.1% unless a specific exemption covers the application. The alloy certificate has to state lead, and the exemption position has to be confirmed rather than assumed.
Failure modes that arrive after the material is chosen
Season cracking. Cold-worked brass under residual tensile stress cracks in ammonia-bearing air. The crack starts at a bend or a punched edge and runs across the part, and it looks like a manufacturing defect until the environment is checked. Stress relief at 250–325 °C for about an hour removes most of the risk.
Dezincification. Warm chloride water leaches zinc and leaves a copper sponge. Thin sheet loses section fast: 0.1–0.5 mm a year in aggressive service is enough to turn a 1.5 mm bracket into a 0.5 mm one inside three years. Coatings slow it; they do not stop it.
Stress relaxation. A brass spring contact at 120 °C loses 10–20% of its contact force over 1,000 hours. If the part is a spring, the temper and the temperature belong on the drawing, and phosphor bronze is usually the better alloy.
Thread stripping. A 1.5 mm brass panel will not hold an M6 thread under vibration. The fix is a welded nut, a rivet nut or a thicker insert, and it is cheaper to design it in than to repair it in the field.
The cheaper alternative, by part shape
A flat cover with no duty: CRCA at 1.5–2 mm with powder coating. The material is stiffer, the coating is a known process, and the part welds to the frame.
A bracket with a load path: CRCA or 304, depending on corrosion. If the environment is coastal, 304 is often enough and 316L is the safe answer.
A conductor above 100 A: C11000 copper, with the section calculated from the allowable temperature rise rather than copied from the old drawing.
A decorative part that must look like brass: brass, because a brass-look coating on steel is a different product with a different failure mode, and the film will wear through at the touch points.
A high-volume flat part: CRCA stamping, with the tool amortised. The press side of that route is covered in CRCA stamping for OEM brackets and panels.
Three substitutions, worked
An earth strap. A 20 mm × 1 mm brass strap has the conductance of a 20 mm × 0.28 mm copper strap, because 28% IACS scales linearly with section. If the drawing only needs the strap to carry fault current for a few seconds, copper at 0.3 mm does the same job at 30% of the mass and a fraction of the material volume. If the strap also has to survive being trodden on, the brass thickness is a mechanical decision and should be labelled as one.
A cover panel. A 600 × 400 mm cover in 1.5 mm brass weighs 3.06 kg; the same cover in 1.5 mm CRCA weighs 2.83 kg. The 230 g difference is invisible on one part and adds 23 kg across a hundred panels, which is the point at which the packing case and the installer's lift both change.
A mounting bracket. A 2 mm brass bracket that must match a 2 mm steel bracket in deflection becomes 2.5 mm brass at 35% more mass, or stays at 2 mm and accepts 45% more deflection. Neither is a small change, and both are visible on the first page of a calculation that the drawing never asks for.
What Ganesh does with a brass enquiry
We cut brass on the 12 kW fibre laser, form it on the 200 T, 5-axis press brake with a 4,450 mm bed, and draw it in press tools built in our own tool room. Our custom part range covers 0.8–40 mm, and brass sits alongside copper, stainless and aluminium in the same plant, which is why we can quote the alternative rather than only the drawing.
That is the useful part of the conversation. A brass drawing that arrives with a weld symbol, a 3 mm thickness and a 1,000 A duty will get a reply that says so, with the steel or copper route priced beside it. The custom parts page lists the material and process range, and the pressed-part side is in copper and brass pressed components.
What to put in the RFQ so the comparison is real
- Duty, in one line: what the part carries, what it is bolted to, and what environment it sees.
- Alloy and temper if it is fixed, or the property that matters if it is not.
- Joining method, with the weld symbols removed if the material stays brass.
- Finish, including whether bare metal is acceptable.
- Annual quantity and release size, because that decides laser, press or a bought-in machined part.
- Compliance, including the RoHS position and any lead limit.
- Mass target, if the assembly has one, because it usually rules brass out before the quote.
The DXF and RFQ checklist covers the file and drawing fields, and the materials hub has the equivalent decision rules for stainless, aluminium and copper. If the duty is still ambiguous, request a quote with the load, the joint and the environment stated, and the material comparison comes back with the price.
When brass is right
Keep brass for the part that must conduct, resist corrosion bare, form deeply and stay non-magnetic, or for the decorative part where the metal is the finish. In those cases the higher material cost buys out a coating line, a plating operation or a corrosion allowance, and the total cost can be lower than the steel route at low volume.
Everything else is a substitution. The drawing is the cheapest place to make it, and the RFQ is the second cheapest.
Frequently asked questions
Is brass stronger than steel?
No. Annealed C26000 runs 300–370 MPa tensile against 350–450 MPa for CRCA, and its modulus is about 110 GPa against 200 GPa, so a brass part of the same thickness is both weaker and roughly half as stiff in bending.
Can brass replace copper for a busbar?
Only at low current. Brass conducts at about 28% IACS against 101% for C11000 copper, so the same current needs 3.6 times the cross-section. Above roughly 100 A the bar becomes impractical and copper is the correct material.
Why can't brass be welded?
Zinc boils at 907 °C, below the melting range of the alloy, so the weld pool loses zinc as fume and finishes porous and weak. Brass is brazed, soldered or mechanically joined; a welded assembly has to be steel or stainless.
Is brass good for outdoor parts?
Only in the right grade. Standard 63/37 brass dezincifies in chloride and ammonia, and thin sheet can lose 0.1–0.5 mm a year in warm coastal service. Naval brass or a dezincification-resistant grade works, but 316L is usually easier to buy.
When is brass actually the cheaper choice?
When it removes a finishing operation. A bare brass part that needs no plating, no paint and no corrosion allowance can beat a coated steel part on total cost, particularly at low volume where the coating setup is charged to the batch.
Sources and further reading
- ASTM B36/B36M — Brass plate, sheet, strip and rolled bar · reviewed 3 Nov 2026
- Copper Development Association — conductivity, modulus and alloy data for copper and brass · reviewed 3 Nov 2026
- Directive 2011/65/EU (RoHS) — lead restriction and the copper-alloy exemption · reviewed 3 Nov 2026
Not sure brass is the right call?
Send the drawing and the duty. We will tell you whether brass, CRCA, stainless, aluminium or copper is the better route before you commit to a tool.
Request a quote