Troubleshooting · Cutting
Table Saw Repeat Cuts Drift: Find the Cause
Stop the batch and compare the first, middle, and last pieces. Then check fence locking, stop blocks, reference edges, feed pressure, blade condition, and measurement method before producing a new first article.
The answer first
What should you do right now?
Stop the batch and compare the first, middle, and last pieces. Then check fence locking, stop blocks, reference edges, feed pressure, blade condition, and measurement method before producing a new first article.
The important distinction is between a calculation problem, where the inputs or arithmetic are wrong; a fabrication problem, where the setup or material does not match the plan; and a design problem, where the intended part cannot satisfy the available geometry or clearance. The checks below separate those paths before you consume more stock or make a compensating edit that spreads the error.
Symptom fingerprint
A run of nominally identical rips or crosscuts becomes progressively wider, narrower, or less square.
Photograph the condition, record the active revision, and write down the measurement method before touching the setup. A useful symptom is specific: which part, which face or edge, which direction, how many pieces, and whether the error is consistent or changing. That evidence helps distinguish one incorrect input from a process that is drifting.
Do not begin by trimming, forcing, or globally changing dimensions. Those actions may make one assembly look better while destroying interchangeability, grain intent, hardware clearance, or the ability to explain the final material count.
Most likely causes
Use these as testable hypotheses, not automatic conclusions. More than one cause can exist, but changing several variables together makes it impossible to know which correction worked.
Fence movement
A fence can lock out of parallel or shift under side pressure. This cause becomes more likely when it matches the physical pattern instead of only the expected answer.
Changing reference face
Warp, debris, or flipping parts changes how stock contacts the fence. This cause becomes more likely when it matches the physical pattern instead of only the expected answer.
Blade and feed effects
Burning, deflection, and inconsistent pressure can alter the edge. This cause becomes more likely when it matches the physical pattern instead of only the expected answer.
10-minute diagnostic matrix
Work down the rows in order. The “signal” column names the question, the test creates evidence, and the last two columns tell you what the result means and what to do next. Keep the original list or setup unchanged in a saved baseline so the comparison remains reversible.
| Signal | Test | What the result means | Next action |
|---|---|---|---|
| Sequence sample | Stack and measure the first, middle, and last parts. | Progressive change confirms drift rather than a single bad blank. | Quarantine the batch and inspect the setup. |
| Fence lock test | Measure front and rear fence distance before and after locking. | Different readings reveal alignment or locking error. | Adjust or service the fence. |
| Contact-surface check | Inspect edges for bow and remove dust at the fence and table. | Intermittent gaps create inconsistent widths. | Joint or select a reliable reference edge. |
| Blade-path review | Check blade cleanliness, runout indicators, feed rate, and burning. | Changing cutting force suggests tool or technique instability. | Correct the blade setup and make a test cut. |
If a check fails, correct that source and repeat the same check before moving on. If every check passes but the symptom remains, widen the investigation to the machine manual, hardware instructions, material supplier, field conditions, or a qualified trade as appropriate. A calculator is useful evidence, but it cannot observe physical setup.
Fix sequence: make the smallest verified correction
- 1. Stop and label the batch. Record the starting condition and result so the next step tests one variable instead of changing several assumptions at once.
- 2. Compare first, middle, and last. Record the starting condition and result so the next step tests one variable instead of changing several assumptions at once.
- 3. Verify fence lock and parallelism. Record the starting condition and result so the next step tests one variable instead of changing several assumptions at once.
- 4. Clean and confirm the reference edge. Record the starting condition and result so the next step tests one variable instead of changing several assumptions at once.
- 5. Approve a new first article. Record the starting condition and result so the next step tests one variable instead of changing several assumptions at once.
After step five, create one controlled first article or dry-fit. Compare it with the released dimension, orientation, clearance, and visible-face requirements. Only then should the corrected process be applied to repeated parts, the complete cabinet run, or the purchasing plan.
Stop condition
When not to continue
Stop when the fence does not lock repeatably, stock loses full contact, or sample parts exceed tolerance.
A stop condition protects more than material. It prevents a questionable assumption from reaching machining, edge treatment, finishing, hardware installation, or site work where recovery becomes slower and more expensive. Mark the affected parts or documents so another person cannot resume from the unresolved setup.
Prevention checklist for the next release
- Check fence lock before batches. Add the decision to the project record when it affects dimensions, material, orientation, hardware, or purchasing.
- Use consistent feed pressure. Add the decision to the project record when it affects dimensions, material, orientation, hardware, or purchasing.
- Clean reference surfaces. Add the decision to the project record when it affects dimensions, material, orientation, hardware, or purchasing.
- Measure samples through the run. Add the decision to the project record when it affects dimensions, material, orientation, hardware, or purchasing.
- Do not use a drifting part as the new master. Add the decision to the project record when it affects dimensions, material, orientation, hardware, or purchasing.
Prevention works best when it lives in the source project, not in memory. Put critical unit, material, grain, edge, hardware, trim, kerf, and revision decisions next to the affected part. When a change alters sheet count or assembly clearance, regenerate the downstream layout, labels, and purchase total together.
Use the right next tool
Group repeat dimensions
Use your measured project inputs to verify the relevant dimensions, quantities, or layout.
Open tool →LearnControl repeat cuts
Read the deeper method and understand which assumptions must be recorded.
Read guide →ApplyInspect repeated parts
Compare a related example, dataset, template, or project workflow before release.
Open resource →Method and limits
This page uses a fault-isolation method: define the observed failure, preserve a baseline, test one likely cause, record the result, make the smallest justified change, and verify a first article. It deliberately avoids universal tolerance claims because acceptable error depends on material, joinery, hardware, finish, machine capability, installation conditions, and the project drawing.
WoodCutTool calculators can help check arithmetic, usable stock, kerf, repeated quantities, or layout geometry. They do not certify structural design, machine condition, electrical or plumbing work, product installation, or building-code compliance. Follow manufacturer instructions and engage a qualified professional when the diagnosis crosses those boundaries.
Frequently asked questions
What should I check first when dealing with table saw repeated cuts that drift in size?
Begin with the observable symptom and the first row of the diagnostic matrix. Record the result before changing the design or machine setup. That keeps one correction from hiding another cause and gives you a repeatable explanation for the final decision.
Should I change the cut list to make the problem disappear?
Only after measurements show the source dimensions or approved design must change. Do not edit a finished part merely to force a layout, conceal a setup error, or recover an installation conflict. Preserve the original revision and document the reason for every changed field.
When is it safe to continue cutting or assembly?
Continue only after one controlled test or first-article part passes the relevant dimension, fit, orientation, and safety checks. Stop when the fence does not lock repeatably, stock loses full contact, or sample parts exceed tolerance.
Can an optimizer or calculator guarantee the fix?
No. A calculator can expose geometry, quantities, kerf, sheet size, or clearance assumptions, but it cannot inspect the material, machine, wall, hardware, or physical assembly. Use the linked tool to verify arithmetic, then confirm the result against the actual project.
Turn the diagnosis into a verified plan
Group repeat dimensions Browse all troubleshooting topics
Save the corrected inputs, regenerate the layout or list, and release one revision that matches the shop and installation documents.
Preserve the corrected revision
Preserve the corrected revision after fixing table Saw Repeat Cuts Drift: Find the Cause
Keep the baseline, corrected parts, regenerated plywood layout, and approved cutting sequence together so the same failure does not return through an older note or screenshot.
- Free basic layouts
- No login
- No cloud upload
- Works offline
On a computer? Scan with your iPhone or iPad.