Borui Academy

Chapter 3

Convergence Tests Toolkit

Decision tree + 7 tests

Chapter 3 — The Convergence-Test Toolkit

Why this chapter matters: this is the heart of BC series. Roughly 80% of AP convergence problems boil down to "pick the right test from this menu." Practice until you can identify the right test from a one-second glance at the general term.

By the end of this chapter you will:

  1. Master 7 major tests (n-th term, integral, p-series, comparison, limit comparison, ratio, root)
  2. Use a decision tree to pick a test
  3. Handle every standard AP convergence question

3.1 The Decision Tree — Spirit of the Chapter

Given a series an\sum a_n, ask these questions in order:

Look at a_n
│
├─ lim a_n ≠ 0?              → Divergent (n-th term test, §3.2)
│
├─ Form a·r^n?               → Geometric series (Ch 2)
│
├─ Form 1n\frac{1}{n}^p?               → p-series (Ch 2)
│
├─ Has (-1)^n alternation?   → Alternating Series Test (Ch 4)
│
├─ Has n!, r^n, or n^n?      → Ratio test or Root test (§3.6, 3.7)
│
├─ a_n = f(n) with f         → Integral test (§3.3)
│   positive & decreasing?
│
├─ Looks like a known        → Comparison or Limit Comparison (§3.4, 3.5)
│   conv/div series?
│
└─ None of the above?        → Try telescoping (Ch 2.5)

🔑 Meta-rule: scan the shape of ana_n. Factorials/exponentials → ratio/root. Rational expressions → comparison/limit comparison. Logs/trig → integral test.

Convergence test decision tree

🎵 Audio 1 — Decision-tree overviewaudio/ch3_01_decision_tree.mp3


3.2 n-th Term Test (review)

If limnan0\lim_{n\to\infty} a_n \ne 0, then an\sum a_n diverges. The converse is false.

Already covered in §2.4. Use as your first filter.


3.3 Integral Test

Hypotheses (all three required):

  • an=f(n)a_n = f(n) where ff is positive, continuous, and monotonically decreasing on [1,)[1, \infty)

Conclusion:
n=1an and 1f(x)dx converge or diverge together\sum_{n=1}^{\infty} a_n \text{ and } \int_1^{\infty} f(x)\,dx \text{ converge or diverge together}

Intuition: integrals are the continuous analog of sums. Picture rectangles of width 1 and height ana_n; monotone decrease ensures the rectangles' total area is sandwiched between 1Nf\int_1^N f and 0N1f\int_0^{N-1} f.

🎵 Audio 2 — Integral testaudio/ch3_02_integral.mp3

Example

Decide n=21nlnn\displaystyle\sum_{n=2}^{\infty} \frac{1}{n \ln n}.

Step 1 — Verify hypotheses: f(x)=1xlnxf(x) = \frac{1}{x \ln x} is positive, continuous, decreasing on [2,)[2, \infty). ✓

Step 2 — Evaluate the integral:
21xlnxdxu=lnxln2duu=\int_2^{\infty} \frac{1}{x \ln x} dx \xrightarrow{u = \ln x} \int_{\ln 2}^{\infty} \frac{du}{u} = \infty

Integral diverges → series diverges.

🔑 Strategy: when lnn\ln n appears in the denominator, substitute u=lnxu = \ln x.

⚠️ Trap: the integral test cannot give you the value of the sum. Only convergence/divergence. (For example, 1n2=π26\sum \frac{1}{n^2} = \frac{\pi^2}{6} but 1x2dx=1\int_1^\infty x^{-2}\,dx = 1.)


3.4 Direct Comparison Test (DCT)

Hypotheses: 0anbn0 \le a_n \le b_n for all sufficiently large nn (i.e. positive-term series).

Conclusion:

  • bn\sum b_n converges an\Rightarrow \sum a_n converges
  • an\sum a_n diverges bn\Rightarrow \sum b_n diverges

Analogy: if your wallet contains less than a billionaire's, and the billionaire is finite, then so is yours. The reverse argument is for divergence.

🎵 Audio 3 — Direct comparisonaudio/ch3_03_compare.mp3

Example

Decide n=11n2+n\displaystyle\sum_{n=1}^{\infty} \frac{1}{n^2 + n}.

Reasoning: 1n2+n1n2\frac{1}{n^2 + n} \le \frac{1}{n^2}, and 1n2\sum \frac{1}{n^2} converges (p-series, p=2p=2). DCT → converges.

🔑 Strategy: pick a known yardstick — usually a geometric series or p-series.

⚠️ Trap: direction matters. To prove convergence you must dominate by a larger convergent series. To prove divergence you must be dominated below by a divergent series.


3.5 Limit Comparison Test (LCT)

Hypotheses: an,bn>0a_n, b_n > 0.

Test: compute L=limnanbn\displaystyle L = \lim_{n\to\infty} \frac{a_n}{b_n}.

Conclusion:

  • 0<L<0 < L < \inftyan\sum a_n and bn\sum b_n agree (both converge or both diverge)
  • L=0L = 0 and bn\sum b_n converges → an\sum a_n converges
  • L=L = \infty and bn\sum b_n diverges → an\sum a_n diverges

Why LCT often beats DCT: DCT requires a strict inequality you can prove; LCT only needs the same asymptotic order.

🎵 Audio 4 — Limit comparisonaudio/ch3_04_limit_compare.mp3

Example

Decide n=12n+1n3+5\displaystyle\sum_{n=1}^{\infty} \frac{2n+1}{n^3 + 5}.

Step 1 — Pick bnb_n: leading-term analysis says annn3=1n2a_n \sim \frac{n}{n^3} = \frac{1}{n^2}. Let bn=1n2b_n = \frac{1}{n^2}.

Step 2 — Compute the limit:
limn(2n+1)/(n3+5)1/n2=limn2(2n+1)n3+5=2\lim_{n\to\infty} \frac{(2n+1)/(n^3+5)}{1/n^2} = \lim \frac{n^2(2n+1)}{n^3+5} = 2

0<2<0 < 2 < \infty. ✓

Step 3: 1n2\sum \frac{1}{n^2} converges → original converges.

🔑 How to choose bnb_n: keep only the dominant term. For rational expressions, match the highest-power ratio.


3.6 Ratio Test (a must when factorials or exponentials appear)

When to use: terms contain n!n!, rnr^n, nnn^n, etc.

Compute:
L=limnan+1anL = \lim_{n\to\infty} \left|\frac{a_{n+1}}{a_n}\right|

Conclusion:

  • L<1L < 1absolutely convergent
  • L>1L > 1 (including L=L = \infty) → divergent
  • L=1L = 1inconclusive, switch to a different test

🎵 Audio 5 — Ratio test + factorial cancellationaudio/ch3_05_ratio.mp3

Example 1 (factorial classic)

Decide n=1n!nn\displaystyle\sum_{n=1}^{\infty} \frac{n!}{n^n}.

L=lim(n+1)!/(n+1)n+1n!/nn=lim(n+1)nn(n+1)n+1=limnn(n+1)n=lim1(1+1/n)n=1eL = \lim \left|\frac{(n+1)!/(n+1)^{n+1}}{n!/n^n}\right| = \lim \frac{(n+1) \cdot n^n}{(n+1)^{n+1}} = \lim \frac{n^n}{(n+1)^n} = \lim \frac{1}{(1+1/n)^n} = \frac{1}{e}

L=1e<1L = \frac{1}{e} < 1converges.

Example 2 (when L=1L=1 fails)

Decide 1n\sum \frac{1}{n} (we already know it diverges).

L=lim1/(n+1)1/n=limnn+1=1L = \lim \frac{1/(n+1)}{1/n} = \lim \frac{n}{n+1} = 1

The test is inconclusive — use the integral test or the p-series rule instead.

🔑 Factorial cancellation: (n+1)!n!=n+1\frac{(n+1)!}{n!} = n+1 and n!(n1)!=n\frac{n!}{(n-1)!} = n.

⚠️ Trap: ratio test always gives L=1L=1 on a p-series. Don't waste time — use the p-series rule directly.


3.7 Root Test

When to use: the entire general term is raised to the nn-th power, e.g. (nn+1)n\left(\frac{n}{n+1}\right)^n.

Compute:
L=limnannL = \lim_{n\to\infty} \sqrt[n]{|a_n|}

Conclusion: identical to ratio test (<1<1 converges, >1>1 diverges, =1=1 inconclusive).

🎵 Audio 6 — Root testaudio/ch3_06_root.mp3

Example

Decide n=1(n2n+1)n\displaystyle\sum_{n=1}^{\infty} \left(\frac{n}{2n+1}\right)^n.

L=lim(n2n+1)nn=limn2n+1=12L = \lim \sqrt[n]{\left(\frac{n}{2n+1}\right)^n} = \lim \frac{n}{2n+1} = \frac{1}{2}

L=12<1L = \frac{1}{2} < 1converges.

🔑 Ratio vs. root: an outer nn-th power cries out for the root test; factorials cry out for the ratio test.


3.8 Quick-Reference Table

Shape of ana_n Best test
Obvious liman0\lim a_n \ne 0 n-th term (rule out divergence first)
arna r^n Geometric formula
1/np1/n^p p-series rule
Has n!n! or rnr^n Ratio
Outer ()n(\cdots)^n Root
f(n)f(n) pos./cont./decreasing, easy to integrate Integral
Looks like a known conv/div series Comparison or Limit comparison
Has (1)n(-1)^n alternation AST (Ch 4)

3.9 Mixed Examples

Decide convergence:

(a) n=1n22n\displaystyle\sum_{n=1}^{\infty} \frac{n^2}{2^n}

(b) n=11n2+cosn\displaystyle\sum_{n=1}^{\infty} \frac{1}{n^2 + \cos n}

(c) n=21n(lnn)2\displaystyle\sum_{n=2}^{\infty} \frac{1}{n (\ln n)^2}

(d) n=1n+1n3+2n\displaystyle\sum_{n=1}^{\infty} \frac{n+1}{n^3+2n}

Solutions:

(a) Ratio test (because 2n2^n):
L=lim(n+1)2/2n+1n2/2n=12lim(n+1n)2=12<1L = \lim \frac{(n+1)^2/2^{n+1}}{n^2/2^n} = \frac{1}{2} \lim \left(\frac{n+1}{n}\right)^2 = \frac{1}{2} < 1
Converges.

(b) Direct comparison or limit comparison: for large nn, n2+cosnn21n^2 + \cos n \ge n^2 - 1, so 1n2+cosn1n211n2\frac{1}{n^2 + \cos n} \le \frac{1}{n^2 - 1} \sim \frac{1}{n^2}. Or: limit-compare with 1n2\frac{1}{n^2}, ratio 1\to 1. Converges.

(c) Integral test (because lnn\ln n in the denominator):
21x(lnx)2dxu=lnxln2duu2=1ln2\int_2^{\infty} \frac{1}{x (\ln x)^2} dx \xrightarrow{u = \ln x} \int_{\ln 2}^{\infty} \frac{du}{u^2} = \frac{1}{\ln 2}
Integral finite → converges.

(d) Limit comparison with 1n2\frac{1}{n^2}: leading term is nn3=1n2\frac{n}{n^3} = \frac{1}{n^2}. Converges.

🎵 Audio 7 — Walkthrough of (a)–(d)audio/ch3_07_examples.mp3


3.10 Trap Alerts ⚠️

  1. Don't push past L=1L=1 in ratio/root — switch tests immediately.
  2. Comparison/LCT only work for positive-term series. Take absolute values first if signs are mixed.
  3. Integral test gives convergence, not the sum.
  4. DCT direction matters: dominate above for convergence, dominate below for divergence.
  5. LCT requires 0<L<0 < L < \infty for the two-way conclusion. L=0L = 0 or \infty gives a weaker one-way conclusion.
  6. Memorize p-series and geometric series: nearly every comparison rests on them.

🎵 Audio 8 — Trap roundupaudio/ch3_08_traps.mp3


3.11 30-Second Decision Algorithm

1. Does a_n → 0? No → diverges (done)
2. Geometric or p-series? → apply formula
3. Has (-1)^n? → jump to Ch 4
4. Has n! or r^n? → Ratio test
5. Outer (...)^n? → Root test
6. Looks like 1n\frac{1}{n}^p or r^n? → Comparison or LCT
7. f(n) positive, decreasing, easily integrable? → Integral test
8. None of the above? → try telescoping

🎵 Audio 9 — 30-second algorithmaudio/ch3_09_algorithm.mp3


3.12 Standard Yardsticks (memorize)

Common comparison yardsticks

Yardstick Conv/Div
1n\sum \frac{1}{n} Diverges
1np,p>1\sum \frac{1}{n^p}, p>1 Converges
rn,r<1\sum r^n, |r|<1 Converges
1n!\sum \frac{1}{n!} Converges fast (sum =e= e)
1nlnn\sum \frac{1}{n \ln n} Diverges
1n(lnn)2\sum \frac{1}{n (\ln n)^2} Converges
1n2+1\sum \frac{1}{n^2 + 1} Converges

3.13 Mnemonic

"Recognize the shape, walk the decision tree."

  • factorial / exponential → ratio
  • outer nn-th power → root
  • easy-to-integrate function → integral
  • looks like p-series or geometric → comparison / LCT
  • L=1L = 1 → switch tests immediately
  • comparison only on positive terms

🎵 Audio 10 — Recap + mnemonicaudio/ch3_10_recap.mp3


Media Inventory

File Purpose
audio/ch3_01_decision_tree.mp3 Decision-tree overview
audio/ch3_02_integral.mp3 Integral test
audio/ch3_03_compare.mp3 Direct comparison
audio/ch3_04_limit_compare.mp3 Limit comparison
audio/ch3_05_ratio.mp3 Ratio test
audio/ch3_06_root.mp3 Root test
audio/ch3_07_examples.mp3 Mixed examples
audio/ch3_08_traps.mp3 Six traps
audio/ch3_09_algorithm.mp3 30-second algorithm
audio/ch3_10_recap.mp3 Chapter recap
images/ch3_decision_tree.png Decision flowchart
images/ch3_benchmarks.png Yardstick reference

End of chapter. Next: alternating series and absolute/conditional convergence.