NivaarExam PrepOfficial exam papers ↗

22-Mec-A4 Design and Manufacture of Machine Elements · Undated paper

Question 3 of 6: Springback control on a precision bent lever

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Notes on this paper 

Paper format. National Examinations, May 2019 — 16-Mec-A4 Design and Manufacture of Machine Elements. Three hours, open book, any non-communicating calculator. Six questions in two parts: Part A (Q1–Q3, manufacturing processes) and Part B (Q4–Q6, machine-element analysis). The rubric asks for two questions from each part; all six are solved here. All questions carry equal value (25 %).

Reference texts. S. Kalpakjian and S. Schmid, Manufacturing Engineering and Technology, 7th ed. (Part A: Ch. 16 sheet-metal forming, Ch. 26 grinding); M. Groover, Fundamentals of Modern Manufacturing, 6th ed. (Ch. 20, 25); R. Budynas and K. Nisbett, Shigley's Mechanical Engineering Design, 10th ed. (Ch. 3 stress, Ch. 6 fatigue, Ch. 7 shafts and keys, Ch. 16 brakes); R. Hibbeler, Mechanics of Materials, 10th ed. (Ch. 6, 7, 9 transverse shear and stress transformation).

Question 3: Springback control on a precision bent lever (25 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

(a) What varies in a production batch of strip. Sheet and strip are sold to tolerances on dimensions and to a specification band on properties, not to a single value, so several quantities scatter from coil to coil and, more slowly, along a single coil:

These matter here because springback in bending is governed by the elastic-recovery group $S_y/E$ together with the ratio of bend radius to thickness. Kalpakjian gives the recovery of a bend as

$$\frac{R_i}{R_f} \;=\; 4\left(\frac{R_i S_y}{E T}\right)^{3} - 3\left(\frac{R_i S_y}{E T}\right) + 1$$

with $R_i$ the initial (die) radius, $R_f$ the radius after release, $T$ the strip thickness, $S_y$ the yield strength and $E$ the modulus. Since $E$ is essentially constant for a given alloy family while $S_y$ and $T$ are precisely the two quantities that scatter, the batch variation in the as-formed angle is inherited directly from the batch variation in the strip.

Given. Low-carbon strip, nominal $S_y = 250\ \text{MPa}$ with a $\pm 10\,\%$ coil-to-coil band, $T = 1.0\ \text{mm}$, die radius $R_i = 5.0\ \text{mm}$, $E = 207\ \text{GPa}$.

Find. The spread in the finished included angle if a single fixed overbend is set from the nominal properties.

  1. Evaluate the recovery group at the nominal condition. With $R_iS_y/(ET) = (5)(250)/[(207000)(1)] = 6.039\times10^{-3}$, the formula gives $R_i/R_f = 0.98188$.
  2. Convert to an overbend. The bend arc length is conserved on release, so the angle scales with the same ratio, $\theta_f = \theta_i\,(R_i/R_f)$. To land on 90° at nominal properties the tool must be set to $\theta_i = 90/0.98188 = 91.66^\circ$ — that is, an overbend of about $1.66^\circ$.
  3. Apply the same fixed tool to the extremes of the property band. At $S_y = 225\ \text{MPa}$ the ratio is $0.98370$ and the part closes to $90.17^\circ$; at $S_y = 275\ \text{MPa}$ it is $0.98007$ and the part opens to $89.83^\circ$.
  4. Quote the residual spread. The finished angle therefore wanders across $$\boxed{\Delta\theta \approx 0.33^\circ \ \text{(about } \pm 0.17^\circ \text{)}}$$ from coil variation alone, before thickness scatter, tool wear or lubrication are counted.

(b) Partly — overbending is necessary but it cannot "always assure" the angle. Overbending is the right first move: it removes the mean springback, and no bending process for a precision part should be set up without it. What it cannot do is remove the variance. The overbend is fixed in the tool steel; the springback it is compensating is a function of each coil's own $S_y$ and $T$. When the property band shifts, the compensation is no longer matched, and the part comes off at the wrong angle in a perfectly repeatable way. The calculation above puts a number on that: roughly a third of a degree of drift, which for a camera linkage held to a tenth of a degree is a scrap-generating amount.

The correct answer is therefore to overbend and to add a mechanism that is insensitive to the strip:

The recommendation for this part is to overbend as a first-order correction, bottom the bend to kill the residual sensitivity, and specify strip with a tightened yield range — with in-process angle checks at coil changes to catch anything the first two miss.

Springback of the 90° camera lever with a fixed overbend
Condition$S_y$ (MPa)$R_i/R_f$Finished angle
Nominal (tool set here)2500.9818890.00°
Low end of the band2250.9837090.17°
High end of the band2750.9800789.83°
Residual spread—0.33°