evan's notes
cal 3discreteodeenv systems

On this page

  • 1. Why we need new tests
  • 2. Alternating Series & the Alternating Series Test (AST)
  • Example A (Alternating harmonic series)
  • Example B (Fails AST)
  • 3. Absolute vs Conditional Convergence
  • 4. Ratio Test (great when factorials or exponentials appear)
  • Example C 
  • Example D 
  • 5. Root Test (handy when the whole term is raised to the nnn-th power)
  • Example E 
  • 6. Quick Strategy Cheat‑Sheet
  • 7. Practice – Try these quick checks
  • 8. Key Take‑aways

Integral Calculus

§8.4 Other Convergence Tests

Evan Luo · May 9, 2025

Integral Calculus

§8.4 Other Convergence Tests

Evan LuoMay 9, 2025

3 min read

1. Why we need new tests

All the tests you learned earlier (Integral, Comparison, ‑p‑series, etc.) assume every term is positive. But many useful series zig‑zag above and below zero, for example

1−12+13−14+…1-\tfrac12+\tfrac13-\tfrac14+\dots1−21​+31​−41​+…

So we need tests that handle sign changes.


2. Alternating Series & the Alternating Series Test (AST)

Key ideaPlain‑English explanation
Alternating seriesTerms switch sign: an=(−1)n−1bna_n = (-1)^{n-1} b_nan​=(−1)n−1bn​ with bn>0b_n>0bn​>0.
AST conditions1. The positive part bnb_nbn​ gets smaller: bn+1≤bnb_{n+1}\le b_nbn+1​≤bn​.
2. bnb_nbn​ shrinks to 0: lim⁡n→∞bn=0\displaystyle\lim_{n\to\infty} b_n = 0n→∞lim​bn​=0.
ConclusionIf both hold, the series converges.

Why it works (picture) – partial sums bounce left‑right but the “zig‑zag” steps keep shrinking, so they squeeze toward a single number.

Example A (Alternating harmonic series)

∑n=1∞(−1)n−11n\sum_{n=1}^{\infty}(-1)^{n-1}\frac1nn=1∑∞​(−1)n−1n1​

bn=1nb_n=\tfrac1nbn​=n1​ is decreasing and →0, so it converges by AST.

Example B (Fails AST)

∑n=1∞(−1)n1−3nn+7\sum_{n=1}^{\infty}(-1)^{n}\frac{1-3n}{n+7}n=1∑∞​(−1)nn+71−3n​

Limit of bnb_nbn​ is 3 ≠ 0, so AST cannot help; the series actually diverges by the basic “limit of a_n” test (often called Test for Divergence).


3. Absolute vs Conditional Convergence

TermMeaning
Absolute convergenceThe series made from (ana_nan​) converges.
Conditional convergenceThe original series converges but (ana_nan​) diverges.

Big fact: Absolute   ⟹  \implies⟹ Convergent (proof uses Comparison Test with ∣an∣|a_n|∣an​∣).

  • Alternating harmonic series is conditionally convergent – its absolute version is the divergent harmonic series.
  • ∑n=1∞cos⁡nn2\displaystyle\sum_{n=1}^{\infty}\frac{\cos n}{n^{2}}n=1∑∞​n2cosn​ is absolutely convergent because ∣cos⁡n∣≤1|\cos n|\le1∣cosn∣≤1 and ∑1/n2\sum 1/n^{2}∑1/n2 converges (‑p‑series with p>1p>1p>1).

4. Ratio Test (great when factorials or exponentials appear)

Compute L=lim⁡n→∞∣an+1an∣L=\displaystyle\lim_{n\to\infty}\Bigl|\frac{a_{n+1}}{a_n}\Bigr|L=n→∞lim​​an​an+1​​​.

ResultVerdict
L<1L<1L<1Absolutely (hence totally) convergent
L>1L>1L>1 or L=∞L=\inftyL=∞Divergent
L=1L=1L=1No information – try another test

Example C 

∑n=1∞(−1)nn22n\sum_{n=1}^{\infty}(-1)^n\frac{n^2}{2^n}n=1∑∞​(−1)n2nn2​ ∣an+1an∣=(n+1)22 n+1⋅2 nn2→12<1\left|\frac{a_{n+1}}{a_n}\right| =\frac{(n+1)^2}{2^{\,n+1}}\cdot\frac{2^{\,n}}{n^{2}} \to\frac12<1​an​an+1​​​=2n+1(n+1)2​⋅n22n​→21​<1

Series converges absolutely. (Your handwritten page marks it “AC” ✔.)

Example D 

∑n=1∞nnn!\sum_{n=1}^{\infty}\frac{n^n}{n!}n=1∑∞​n!nn​ an+1an→e>1\frac{a_{n+1}}{a_n}\to e>1an​an+1​​→e>1

Series diverges. (Matches your slide with result “D” for divergent.)


5. Root Test (handy when the whole term is raised to the nnn-th power)

Compute L=lim⁡n→∞∣an∣nL=\displaystyle\lim_{n\to\infty}\sqrt[n]{|a_n|}L=n→∞lim​n∣an​∣​.

Same conclusions as the Ratio Test: L<1L<1L<1   ⟹  \implies⟹ converge, L>1L>1L>1   ⟹  \implies⟹ diverge, L=1L=1L=1   ⟹  \implies⟹ inconclusive.

Example E 

∑n=1∞(2n−33n−2)n\sum_{n=1}^{\infty}\Bigl(\frac{2n-3}{3n-2}\Bigr)^{n}n=1∑∞​(3n−22n−3​)n ∣an∣n  =  2n−33n−2  ⟶  23<1\sqrt[n]{|a_n|}\;=\;\frac{2n-3}{3n-2}\;\longrightarrow\;\frac23<1n∣an​∣​=3n−22n−3​⟶32​<1

So the series converges.


6. Quick Strategy Cheat‑Sheet

Situation you noticeTest to try first
Terms alternate cleanly & bnb_nbn​ ↓ 0AST
Factorials, knk^nkn, or many productsRatio test
Something like (expression)n(\text{expression})^{n}(expression)nRoot test
Positive terms & resembles 1/np1/n^p1/npComparison or p‑series
Integral of f(n)f(n)f(n) easyIntegral test
Signs irregular → check absolute value firstAbsolute/conditional approach

7. Practice – Try these quick checks

  1. ∑n=1∞(−1)n+1n(n+2)\displaystyle \sum_{n=1}^{\infty}\frac{(-1)^{n+1}}{n(n+2)}n=1∑∞​n(n+2)(−1)n+1​
  2. ∑n=1∞(2n)!4 n(n!)2\displaystyle \sum_{n=1}^{\infty}\frac{(2n)!}{4^{\,n}(n!)^{2}}n=1∑∞​4n(n!)2(2n)!​
  3. ∑n=1∞sin⁡nn3/2\displaystyle \sum_{n=1}^{\infty}\frac{\sin n}{n^{3/2}}n=1∑∞​n3/2sinn​

(Hint: 1 → AST then absolute test, 2 → Ratio, 3 → compare |sin n| to 1 and use p‑series.)


8. Key Take‑aways

  • Alternating signs + shrinking size   ⟹  \implies⟹ convergence (AST).
  • Absolute convergence is stronger; if it happens, you’re done.
  • Ratio & Root tests are powerhouse tools whenever factorials or nth‑powers show up.
  • Always check the limit of ana_nan​ first: if it isn’t zero, the series diverges instantly.

Keep this roadmap handy and series tests will feel a lot less mysterious!

Source: https://notes.ohevan.com/notes/integral-calculus/8-4-other-convergence-tests

© 2026 Evan Luo. All rights reserved.

Back to Integral Calculus

0 reads

·Last edited May 22, 2025
  • newsletter

  • about me

  • sponsor

© 2026 Evan Luo. All rights reserved.