A red-and-blue overlay can look like a major design change when one file was simply exported in a different coordinate system. Before reviewing the colored regions, establish that the two models are the intended revisions and that the comparison conditions make sense.
Open the matching tool. Viewing is available without an account; sign in when the conversion workflow requires it.
Run Solid Differences with its actual constraints
In Compare, choose fixed model A and B. Start with the same units and original coordinates; CADProps does not automatically align the files. Side-by-side, overlay and A/B modes are visual aids. Solid Differences is available for supported STEP pairs with valid closed solids, with at most 100 solids per file. It shows A-only material in red, B-only in blue and common material in gray. Each input is fused first, so overlapping material is counted once for this operation. The Boolean tolerance is a calculation setting, not manufacturing acceptance. Preserve the coordinate frame, scope and warnings with the result.
A useful comparison starts with a question: are you checking placement, shape or both? Automatic alignment can help inspect shape, but it can also hide a placement change that was important to the assembly.
Fix the two inputs first
Record the exact earlier and later files. If the system supports saved versions, compare fixed snapshots rather than two links that always resolve to “latest.” Onshape's version documentation provides an example of immutable design snapshots, distinct from an active workspace. Versions and history.
Keep conversion outputs attached to their source revision. A STEP exported from V2 is not V3 simply because a new file appeared. If you compare a native file with a derivative, include that relationship in the record.
For assemblies, ensure that each revision includes its own dependencies. Loading an old top-level file with newer component files is not a reliable reconstruction of the earlier assembly.
Start with the original coordinates
Inspect units, orientation and placement before making an alignment choice. If both models use the same frame, a direct overlay preserves meaningful positional differences.
When the purpose is shape comparison, a deliberate rigid alignment may help. Record the transform and let the reviewer confirm it. Do not allow automatic scaling or mirroring to make different designs appear alike.
Two models can have the same envelope and volume yet differ locally. Those properties are useful warning signals, not proofs of geometric identity.
Use visual modes for what they actually show
| Mode | Good use | Limitation |
|---|---|---|
| Side by side | Overall form and component context | Small changes can be hard to locate |
| Overlay | Position and outline comparison | Overlapping colors are not a computed difference |
| A/B toggle | Checking one region in a fixed view | Relies on human inspection |
| Property table | Extents, volume and other summary differences | Equal totals do not prove equal shapes |
Label overlay colors as Model A and Model B unless a difference algorithm has actually classified added and removed material. The legend should describe the result that exists, not the result you hope to add later.
Separate a display threshold from engineering acceptance
A geometric comparison may use a distance threshold. Record its value, unit and calculation method. That parameter is a detection setting, not a tolerance specification for manufacturing.
“No difference detected within this threshold and scope” is more precise than “identical.” The former leaves room for sampling limits, untested components and algorithm constraints. A coarse preview should never be the hidden basis of a confident dimensional claim.
When a reviewer needs to know whether a hole moved, identify and measure that region with an appropriate method. Do not invent a feature-level explanation from a color patch alone.
Keep human notes separate from computed observations
A note saying “increase the bore to 8 mm” is a request. A user comment saying “the bore has changed” is an observation. A measured diameter difference is another kind of result. Give each an explicit label.
Attach comments to a version pair and view. Store the camera and section state so that the next person can inspect the same area. Avoid automatically attaching an old face marker to a new face unless the relationship has been verified.
Share the comparison, not just two filenames
A review package should include the version pair, configuration or body scope, unit choice, alignment method and saved observations. Fix that package when sharing it. A later upload should not silently replace one side.
CADProps processes self-contained STEP/IGES geometry, supported static meshes and bounded ZIP packages. STEP pairs support solid differences; other 3D inputs support visual comparison. SolidWorks, Siemens NX / UG NX PRT and Inventor IPT/IAM support validated STEP conversion and viewing of the converted geometry. Package assembly references in ZIP; partial geometry is flagged. DWG and DXF use the separate 2D viewer for bounded preview and basic measurement. Open the STEP workspace or open the DWG viewer.
Diagnose equal totals and a changed local feature
Two versions can have the same volume after one pocket becomes deeper and another becomes shallower. Equal volume does not establish geometric identity. Start with a fixed view of the relevant interface, then use supported Solid Differences only when both STEP inputs satisfy its solid and count limits.
If the overlay shows widespread separation, verify units and original placement before interpreting red or blue as changed material. CADProps does not automatically register the inputs. Perform any required alignment in an appropriate source workflow and document the transformation; do not scale a model merely to make its silhouette agree.
A computed difference that fails is an unavailable comparison, not evidence of zero difference. Preserve that state and use visual review for its stated limits, or obtain valid solid inputs. The comparison workspace is also not a shareable single-model snapshot in the current release. Deliver a review record with the exact version pair, scope, coordinate frame and observations through the agreed project process.
Reproduce the observed result
We opened block.step and a byte-identical block.stp copy in the same local workspace. Both displayed 40 × 30 × 20 mm. The computed solid comparison found A-only 0 mm³, B-only 0 mm³ and common 24000 mm³ at 0.001 mm Boolean tolerance, in original coordinates. The same checksum proves identical input bytes; this single pair does not establish that unrelated STEP files are equivalent.
Choose two files in the comparison tool, wait for both inputs, then inspect Differences and its Properties. Side by side, Overlay and A/B are separate visual modes. A tolerance is a Boolean calculation setting; a successful comparison is not manufacturing approval.
Download original sample · block.stp · Verification record
Open the matching tool. Viewing is available without an account; sign in when the conversion workflow requires it.
Original article: CADProps · https://www.cadprops.com/guides/compare-cad-revisions/