Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.
1.1Vector scan paths for exposure
What ifWhat if each cross section uses a vector contour and hatch paths to drive a scanning exposure beam?
Claim language“a mask image of each printing cross section”
A slicer intersects the model with each layer plane and connects the resulting segments into closed contours. The controller fills those contours with vector hatch segments and sends the contour and hatch coordinates to two galvanometer mirrors after the pre-exposure dwell.
1.2Projected area and coverage change
What ifWhat if the controller uses projected area and changes in pixel coverage to characterize each layer?
Claim language“determining edge angles in the mask image”
The controller multiplies the active pixel count by the projected pixel area and counts the pixels that switch state relative to the preceding layer. A resin-specific table maps those two quantities to a dwell selected from measured settling times.
1.3Analytical resin refill timing
What ifWhat if the controller calculates waiting time from a thin-gap flow equation and measured resin properties?
Claim language“a pre-trained prediction model”
For each corner region, the controller derives a characteristic refill distance from the local contour geometry. A thin-gap flow calculation estimates refill time from that distance, the resin viscosity at its measured temperature, the layer gap, and a specified pressure difference.
1.3Force based exposure timing
What ifWhat if a load sensor determines when to start exposure by measuring settling after platform motion?
Claim language“a pre-trained prediction model”
A load cell on the build-platform carriage measures the axial force after the platform reaches the next layer position. The controller starts exposure when the filtered force stays within a specified band around the stationary baseline for a set number of samples.
1.4Resin temperature dwell table
What ifWhat if the layer dwell comes from a table indexed by resin temperature and layer gap?
Claim language“according to the waiting time corresponding to the edge angles”
A vat thermistor supplies resin temperature, and the motion controller supplies the commanded layer gap. The controller interpolates between table entries populated from refill measurements and holds the platform at the exposure position for the resulting interval.
Section 2 · machine-generated
Variations
Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.
Other field of use
Exposure timing for ceramic slurry printing
A vat photopolymerization printer forms ceramic parts from a suspension of ceramic powder in a photocurable binder. The slicer generates a binary mask for each layer and measures the opening angle, adjacent edge lengths, and local gap width at each corner. A model trained on optical measurements of slurry settling predicts a delay for each corner after the build platform stops moving. The exposure controller uses the 95th percentile of the corner delays for the layer and uses a calibrated baseline delay for masks without corners.
Component substitution
Gradient boosted trees for delay prediction
A gradient boosted regression tree ensemble predicts a settling delay for each corner in the layer mask. The input vector contains the interior angle, the lengths of the two adjoining edges, and the distance from the corner to the nearest nonadjacent edge. Training labels record the time from platform stop until local surface motion stays below a calibrated threshold for a fixed interval. The printer computes the root mean square of the predicted delays and applies a fixed margin calibrated from complete layer masks.
Parameter or scale variation
Corner sampling at two physical scales
A slicer traces the mask boundary at a projected pixel pitch of 50 micrometers. At each candidate corner, it fits boundary segments in windows extending 0.25 millimeters and 1.0 millimeter along each adjoining edge. A trained model receives both angle estimates and the corresponding line fit residuals and outputs a delay for each corner. The controller takes the 90th percentile of the outputs, rounds up to the next 50 milliseconds, and loads the result into the exposure timer.
Combination with existing systems
Delay prediction with resin temperature control
A heated resin vat uses a thermistor and a feedback controller to regulate bulk resin temperature. The slicer derives corner angles, adjoining edge lengths, and gap widths from each layer mask. A prediction model combines those features with the measured vat temperature to estimate a delay for each corner. The exposure controller starts its timer when the platform stops and sets the delay to the 95th percentile of the corner estimates plus a calibrated allowance for temperature measurement lag.
Automated or data-driven variant
Camera verification of predicted settling
A camera images the resin surface under illumination outside the resin's curing band, and a calibration maps image coordinates to the layer mask. A trained model uses mask corner geometry to predict local delays and supplies their 90th percentile as the first time to check for settling. The controller computes frame-to-frame intensity changes in image windows around the mapped corners. Exposure starts after all windows stay below a calibrated change threshold for five consecutive frames.
Delay estimation from successive layer masks
The slicer aligns the current and preceding layer masks on the printer's pixel grid and computes their pixelwise difference. At each corner of the current mask, the software measures the interior angle, adjoining edge lengths, and fraction of changed pixels in a fixed physical window. A trained regression model uses these features to predict a delay for each corner from measured settling times for successive layer pairs. The controller weights each predicted delay by the length of the adjoining edges, computes the weighted mean, and adds a fixed margin calibrated from layer settling measurements.
Specific version of a part
Independent regional exposure timing
The controller assigns each region a start time based on the longest predicted dwell among its associated corners. A spatial light modulator activates each region at its scheduled time for the prescribed exposure duration, and the platform moves after all regions finish.
Bit packed layer masks
The slicer rasterizes each cross section onto the exposure pixel grid and stores it as a one-bit-per-pixel image. The print controller reads the image for the current layer and unpacks its rows into an analysis buffer.
Corner angles from fitted lines
The controller traces the boundary between active and inactive pixels and fits a line to a short contour segment on each side of a candidate corner. It computes the interior angle from the oriented lines and records the lengths of the fitted segments in millimeters.
Neural network corner timing
A pre-trained feedforward neural network receives the interior angle and the two adjacent segment lengths for each corner. Its output gives the estimated dwell in milliseconds for the resin and layer gap represented in the training measurements.
Maximum corner dwell per layer
The controller selects the largest predicted corner dwell for the current mask and rounds it up to the next timer tick. The timer starts when the platform reaches its exposure position, and the light source turns on when the timer expires.
Section 3 · quoted from USPTO records
Other work
Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.
1.1acquiring a mask image of each printing cross section in a 3D printing model;
To do this, first, a 3D image is sliced to produce 2D masking images.
B · US 2022/0266515 A1 · ¶ 28
1.2determining edge angles in the mask image, and determining geometric parameters of the edge angles;
Derivation of the corner locations (e.g., corner radials and corner location angles) from the determined dimensions and offset of the load may then be performed in the manner discussed above, such that an effective consumption rate and/or effective circumference/radius-based wrap speed model may be employed to control the dispense rate during a wrapping operation.
A · US 2014/0116008 A1 · ¶ 134
1.3inputting the geometric parameters of the edge angles into a pre-trained prediction model for processing to generate waiting time corresponding to the edge angles; and
1.4according to the waiting time corresponding to the edge angles, determining waiting time of the printing cross section corresponding to the mask image.
An empty cell means only that no passage was reproduced there.
Statements in these documents
In particular, wrappers have historically suffered from packaging material breaks and limitations on the amount of wrap force applied to the load (as determined in part by the amount of pre-stretch used) due to erratic speed changes required to wrap loads.
A · US 2014/0116008 A1 · ¶ 5
Stretch wrapping, whether accomplished by a turntable, rotating arm, vertical rotating ring, or horizontal rotating ring, typically covers the four vertical sides of the load with a stretchable packaging material such as polyethylene packaging material.
A · US 2014/0116008 A1 · ¶ 3
Through the use of the 3D printer, advantageously, manufacturing cost and time may be substantially reduced, individual customized objects may be made, and solid objects complicated in shape may be simply manufactured.
B · US 2022/0266515 A1 · ¶ 6
The 3D printer having such many advantages is applied in a variety of fields such as automobiles, flights, construction, medicine, home appliances, toys, and the like.
B · US 2022/0266515 A1 · ¶ 7
Cite aspatent.fyi, US 2026/0356043 A1, claim 1, first published 2026-11-19T19:14:50-05:00, https://patent.fyi/US20260356043/, SHA-256 035988f617b3aa79e326b604c93ec78ff5fbfe7fa440b830317485910ea17268