A test bar cast with the bracket answers one question: what did this heat do? It shares the pour and the furnace cycle, so the pull test and the parts carry the same heat number. A bar cut from the finished part answers a different question, and costs a bracket. That is why investment casting shops pour test bars with the part.
The bar is not a separate order. It is poured from the same ladle, tied into the same gating, and cut off after heat treatment. The tag carries the heat number of the brackets beside it. That link turns a tensile result into batch evidence.
Running with the furnace load is written into the route, and quench and temper come before the bar step. The order reads backwards until you see it: brackets and bar are tempered together, and only then does the bar go to the saw.
What the arrangement does not give you is a statement about any single bracket. The bar speaks for the heat.
The bracket is not one section. It runs 8–12 mm at the wall and 25–30 mm through the boss where the load lands, both design targets. The boss solidifies last. A bar near the thin wall cools fast, pulls clean, and has never seen the middle of the heavy section.
Position matters more than count. One bar in the right place beats three on the runner. The process sheet should name where it sits and which section it represents; when it is silent, settle that before the pattern is approved, because the mold fixes the position, not the purchase order.
Quench at 840–860 °C, temper at 550–600 °C, hold the body to HB 220–250. Those windows are narrow. A bar tempered alone at 550 °C and a bracket load at 600 °C leave the furnace as two different materials, 50 °C apart on one chart.
Running with the load removes that gap: one ramp, one hold, one line in the batch record. The material behind the record is 45 steel to GB/T 699-2015 in the quenched and tempered condition, and the grade alone does not set the result.
Ask for the furnace chart with the load number and the bar tag on it. A certificate that says quenched and tempered, with no load identity, cannot be matched to the bar that was pulled.
ASTM A370 is the method, not the acceptance number. It fixes how the specimen is pulled and what the report carries: yield strength, tensile strength, elongation, reduction of area. The drawing supplies the figure those results are checked against, 355 MPa as a design target rather than a property of the grade.
A report with a tensile value and no yield value leaves the target unchecked. A number above 355 MPa does not stand in for it. A bar pulled to another method is a different test.
Three fields decide whether a certificate is usable: the heat number on the bar tag, the method cited as ASTM A370, and a yield result in MPa. The first two let you defend the third years later.
The bar and the body cool at different rates. The bar is smaller in section than the 25–30 mm boss, so it hardens through while the boss center may not. That gap has a cost: a bar can pass while the body lands under 355 MPa, and nothing on the certificate flags it.
A larger bar does not close the gap. The section has to be fed, and the body has to be read. Hardness inside HB 220–250, taken on the bracket and not on the bar, is that second record. It is not a yield value and does not convert into one. What it catches is the heavy section that did not harden, the one failure the bar cannot report.
Record | Reading and unit | What it covers | Blind spot |
|---|---|---|---|
Cast-on bar, per heat | ASTM A370 tensile, MPa | The heat and its temper cycle | Core of the 25–30 mm boss |
Body hardness | HB 220–250, on the bracket | The section actually read | A yield value in MPa |
Bar position | Named in the process sheet | Which section the bar stands for | Body core away from the reading |
Put three items in the purchase order: a cast-on bar pulled to ASTM A370 with a report per heat, a bar position on the process sheet, and body hardness read at named points against HB 220–250. Settle all three before the pattern is approved.