How to Compute RREF on a TI-84 Calculator
The TI-84 Plus, TI-84 Plus Silver Edition, and TI-84 Plus CE all include a built-in rref( function under the Matrix Math menu. Here is the complete step-by-step procedure:
- Store your matrix. Press
2nd→MATRIX(orNAMES). Navigate right toEDIT. Select matrix [A]. Enter the number of rows and columns, then type each entry. Press2nd→QUITwhen done. - Open the Matrix Math menu. Press
2nd→MATRIX. Navigate right toMATH. - Select rref(. Scroll down to
B: rref((it may be labeled differently on older OS versions). PressENTER. - Enter the matrix name. Press
2nd→MATRIX, select[A]from NAMES, pressENTER. - Close the parenthesis. Type
)and pressENTER. - Read the RREF. The result is displayed as a matrix. You may need to scroll right to see columns beyond the screen width.
For an augmented matrix [A|b], store all columns (including the b column) as a single matrix. For example, a 3-equation, 3-variable augmented matrix becomes a 3×4 matrix in [A].
The Floating-Point Problem on TI-84
The TI-84 stores all numbers in floating-point format with 14 significant decimal digits of precision. Most fractions — including 1/3, 1/7, 2/11, and many others — cannot be represented exactly in this format. They are rounded to the nearest representable value.
In a single arithmetic operation, this rounding error is negligible (roughly 10⁻¹⁴). But Gauss-Jordan elimination performs many operations, each using the results of previous ones. Errors accumulate. For a 3×3 matrix with fraction entries, the final RREF may contain values like:
Expected: 0 Got: 9.9999e-13
Expected: 1 Got: 0.9999999998
Expected: 1/3 Got: 0.3333333334
For homework where you must show exact fraction answers, these floating-point residuals are problematic. A student who gets 9.9999e-13 cannot tell whether this is "effectively zero" or whether they made an error. The TI-84 has a Frac function (press MATH → 1: ►Frac) that attempts to convert decimals back to fractions, but it only works for simple fractions and fails for more complex cases.
How the Online Calculator Avoids This Problem
The RREF calculator on this site represents every number as an exact rational number: a pair of integers (numerator, denominator). The fraction 1/3 is stored as the pair (1, 3) — not as a decimal approximation. All arithmetic (addition, subtraction, multiplication, division) on these fraction pairs produces another exact fraction pair, following the standard rules for rational arithmetic.
This means there is no rounding at any stage. A value that should be zero is stored as the integer 0, not as 10⁻¹³. A value that should be 7/12 is stored as the pair (7, 12), not as 0.58333.... The output matches exactly what you would compute by hand with exact fraction arithmetic.
The tradeoff: the online calculator requires an internet connection and cannot be used on a closed-book exam. For exams and on-the-fly calculation, the TI-84 is the right tool — but for checking homework and understanding fraction results, the online calculator gives verifiably correct answers.
RREF on Casio Graphing Calculators
Casio calculators (fx-9750GIII, fx-9860GIII, fx-CG50) also include RREF functionality, accessed differently:
- From the Main Menu, enter the Run-Matrix mode.
- Press
F1(MAT/VCT) to open the matrix editor. - Enter your matrix dimensions and values, then press
EXIT. - In the Run-Matrix editor, press
OPTN→F2(MAT) →F6(more) →F4(Rref). - Select your matrix and press
EXE.
Like the TI-84, the Casio uses floating-point internally. The same rounding issues apply. Casio does have a fraction display mode (MathI/MathO settings), which helps for simple results, but does not eliminate the underlying floating-point arithmetic.
Side-by-Side Comparison
| Feature | TI-84 | This Online Calculator |
|---|---|---|
| Arithmetic type | Floating-point (14 digits) | Exact rational (BigInt) |
| Fraction inputs | Stored as decimals | Stored exactly as a/b |
| Zero values | May show 9.99e-13 | Always exactly 0 |
| Fraction outputs | Decimals (use ►Frac to convert) | Exact fractions always |
| Step-by-step output | Final result only | Every row operation named |
| Works offline/exam | Yes | No (browser required) |
| Max matrix size | Up to 99×99 | Up to 6×6 on this site |
When to Use the TI-84
The TI-84 is the right choice when:
- Taking a closed-book exam that permits graphing calculators but not phones or laptops.
- Working with large matrices (more than 6 rows or columns) where the online calculator's size limit is reached.
- Doing quick approximate checks where floating-point results are close enough for your purpose.
- Chaining calculations using stored matrices in the calculator's memory for a sequence of operations.
When to Use the Online RREF Calculator
The online calculator is the right choice when:
- Your matrix contains fractions (1/3, 5/7, etc.) and you need exact fraction answers in the output.
- You need step-by-step explanations for each row operation — essential for learning and homework verification.
- You are checking your hand computation and need to know which specific step is wrong.
- You got a suspicious TI-84 result (like 9.9999e-13) and want to know if it should be zero.
- You need LaTeX output to paste into an assignment or paper.
For a complete worked tutorial using the online calculator, see our RREF step-by-step tutorial. For a comparison of Gauss-Jordan and Gaussian elimination methods, see the Gauss-Jordan vs Gaussian elimination guide.
Frequently Asked Questions
How do I use rref( on a TI-84?
Press 2nd → MATRIX → MATH → scroll to rref(. After selecting it, press 2nd → MATRIX → NAMES → [A] → ) → ENTER. The matrix must be stored in [A] (or another slot) beforehand using MATRIX → EDIT.
Why does my TI-84 show 9.99999e-13 instead of 0?
The TI-84 uses floating-point arithmetic. Fractions like 1/3 cannot be stored exactly in binary, so rounding errors accumulate across many row operations, producing tiny near-zero values. An online calculator with exact rational arithmetic shows exactly 0.
Does the Casio fx-9750GIII have RREF?
Yes — accessible via Run-Matrix mode, OPTN → MAT → Rref. Casio calculators also use floating-point arithmetic, so the same rounding caveats apply for fraction inputs.