Finding Limits Using Partial Sums- Techniques and Examples
What Partial Sums Actually Are
A partial sum is exactly what it sounds like—you sum up the first n terms of a sequence and call that your partial sum Sn. Nothing fancy. You're not adding infinity terms. You're adding a finite chunk and seeing what happens as that chunk gets bigger.
The limit of a sequence of partial sums is the foundation for understanding infinite series. If you can't get this part right, you're going to struggle with everything that comes after. No way around it.
Why Partial Sums Matter for Finding Limits
When you need to find the limit of an infinite series, you're really asking: "What does this sum approach as I add more and more terms?" The sequence of partial sums gives you a way to answer that question.
Instead of jumping straight to "infinite," you build up gradually. Each partial sum is a checkpoint. If those checkpoints approach a specific number, that's your limit. If they don't, the series diverges and you're done—there's no hidden answer waiting for you.
The Basic Framework
For a series Σak, the partial sum is:
Sn = a1 + a2 + a3 + ... + an
You're finding lim Sn as n→∞. That's the entire game.
Techniques for Common Series Types
Arithmetic Series
These are the easiest. If you have a sequence with a constant difference, there's a direct formula:
Sn = n/2 × (first term + last term)
Or equivalently: Sn = n/2 × (2a1 + (n-1)d)
Then take the limit as n approaches infinity. If the common difference is nonzero, the series diverges. Simple as that.
Geometric Series
Geometric series are where partial sums actually become useful. The formula for a finite geometric series is:
Sn = a1 × (1 - rn) / (1 - r) when r ≠ 1
Here's where it gets interesting. To find the limit, you need to examine what happens to rn as n grows:
- If |r| < 1, then rn → 0, and the limit is a1 / (1 - r)
- If |r| ≥ 1, the series diverges—no sum exists
This is the most common application you'll encounter. Memorize it or know how to derive it quickly.
Telescoping Series
Telescoping series collapse beautifully when you write out the partial sums. Terms cancel with each other, leaving just a few stragglers.
Example: Σ(1/k - 1/(k+1))
Your partial sum is:
Sn = (1 - 1/2) + (1/2 - 1/3) + (1/3 - 1/4) + ... + (1/n - 1/(n+1))
Everything cancels. You get Sn = 1 - 1/(n+1)
Now take the limit: lim Sn = 1 - 0 = 1
The trick is writing out enough terms to see the cancellation pattern. Don't try to do it all in your head.
Comparison: Partial Sums vs. Direct Limit Tests
There are other methods for testing convergence. Here's how partial sums stack up:
| Method | Best For | Limitation |
|---|---|---|
| Partial Sums | Geometric, arithmetic, telescoping series | Often requires finding closed form for Sn |
| Ratio Test | Factorials, exponential patterns | Doesn't give the actual sum, just convergence |
| Integral Test | p-series, continuous decreasing functions | Only gives approximation, not exact value |
| Comparison Test | Series with known benchmarks | Only determines convergence, not sum value |
Partial sums give you the actual limit value when they work. That's their advantage. The downside is they're not always easy to compute.
Getting Started: A Practical How-To
Here's how to approach any limit-finding problem using partial sums:
- Identify the series — Write out the first 3-5 terms explicitly. This shows you the pattern and helps you spot telescoping or geometric structures.
- Write the partial sum formula — Express Sn in terms of n. If you can't find a pattern from the first few terms, this step will expose that you need a different method.
- Simplify Sn — Combine like terms, look for cancellations, reduce fractions. For geometric series, use the standard formula. For telescoping, write out enough terms to see what survives.
- Take the limit — Examine what happens to each term involving n as n→∞. Zero terms go to zero. Constant terms stay. Growing terms mean divergence.
- State the result — Either give the limiting value or state that the series diverges. Don't hedge. Don't add caveats. If the limit exists, state it. If it doesn't, say so.
Common Mistakes That Kill Your Answer
People lose points on partial sum problems for predictable reasons:
- Forgetting the convergence condition for geometric series — Students apply the sum formula a/(1-r) without checking that |r| < 1. The formula is wrong if the series diverges.
- Not simplifying the partial sum first — Trying to take the limit of a messy Sn before simplifying it. Clean it up first.
- Misidentifying the series type — A telescoping series that isn't recognized as telescoping will look impossible. Write out terms. Always write out terms.
- Assuming convergence without proof — Just because a series "looks like" it should converge doesn't mean it does. The partial sum analysis tells you the truth.
Example Walkthrough
Find the limit of the series: Σk=1∞ (2/3)k
Step 1: This is geometric with a1 = 2/3 and r = 2/3
Step 2: Partial sum formula: Sn = (2/3)(1 - (2/3)n) / (1 - 2/3)
Step 3: Simplify: Sn = (2/3)(1 - (2/3)n) / (1/3) = 2(1 - (2/3)n)
Step 4: Take the limit: |r| = 2/3 < 1, so (2/3)n → 0
Step 5: Result: lim Sn = 2(1 - 0) = 2
The series converges to 2.
When Partial Sums Won't Save You
Some series don't have nice partial sum formulas. The harmonic series Σ 1/k is a famous example. You can show it diverges by comparing partial sums to an integral or by grouping terms, but there's no closed form for Sn that gives you the limit directly.
In those cases, partial sum analysis tells you whether a limit exists, even if you can't compute it exactly. That's still valuable information.
Know when to use the method and when to switch to another test. Partial sums work beautifully when the series has a recognizable structure. They fail when the structure is too messy to sum.