Troubleshooting

No results are drawn

Results appear only when the model can be solved. Check, in this order:

1. Is there a red alert in the viewer? No materials defined / No cross sections defined → add one. Model has N error(s) → click Show details and fix each item (see the table below).
2. Enough supports? The solver needs at least three restrained DOFs and no mechanism. A beam on two rollers (Dz + Dz) has only two and slides away; a frame with all pins and no bracing may be a mechanism even with many supports. Add a Dx somewhere, or an Ry at one support.
3. Are the elements connected? Two nodes at the same coordinates are still two separate nodes. Delete the stray one and reconnect, or place nodes on elements with Connect to structure.
4. Hinged everywhere? A node whose every element is hinged and that has no Ry support has an undefined rotation. Untick one hinge or restrain Ry at that node.
5. Absurd numbers (displacements in the millions) mean the structure is almost a mechanism; EduBeam hides such results. Look for a missing support or a near-zero stiffness (E, A or Iy accidentally entered as 0 or in the wrong unit).

Error messages

Messages from Show details (the Cannot solve model dialog):

Message	Meaning / fix
Model needs at least 3 constrained DOFs to be stable in 2D analysis.	Add supports until at least three DOFs are restrained in total.
Element X references missing node / material / cross section Y.	The referenced entity was deleted (usually via a hand-edited JSON file). Reassign it in the Elements table.
Element X must reference exactly 2 nodes.	Corrupt element in an imported file; delete and recreate it.
Nodal load / Prescribed displacement #n references missing node Y.	Delete the load or re-target it.
Element load #n references missing element Y.	Delete the load.
Solver failed due to an internal model inconsistency…	A generic failure; undo the last step or save the file and report it.

Warnings (Model warnings dialog) do not stop the solve: Element X references the same node at both ends (zero-length element—delete it) and … contains invalid values (a load with a non-numeric component—edit it).

Results look wrong

Symptom	Likely cause
Loads act upward	Global z points down, so positive Fz/fz is downward. Negative values point up. See conventions.
Deflection is 1000× too big or small	Unit mix-up—E entered in Pa while the unit is MPa, or Iy in cm⁴ while the unit is m⁴. Check the units chip at the bottom-right of the viewer.
Deflection slightly larger than the textbook formula	Timoshenko shear deformation. Increase the section's shear coefficient (or use a slender member) to approach Euler–Bernoulli values. See Checking results by hand.
Moment diagram is on the "wrong" side	The side is a drawing convention only; read the sign from the labels—positive means sagging (tension in the bottom fibre).
Diagrams are huge / tiny	Purely visual—adjust Results scale in Settings → Viewer settings.
Local-coordinate load points the wrong way	The element's local x runs from its initial to its end node. Use Swap nodes or flip the sign.
Temperature load does nothing	Determinate structures deform freely under temperature without internal forces. Check that α ≠ 0 and, for the gradient part, that the section height h is set.

Interface issues

Symptom	Fix
Shortcuts do nothing	Click on the canvas first—keys are ignored while a text field has focus.
Cannot pan	Panning uses the middle/right mouse button by default; change it in Settings → Controls & Shortcuts. On a trackpad, use two fingers or switch the setting to Right button.
My model disappeared after an update	An update that resets storage is announced in a dialog first; choose to cancel it and save the project before updating.
Wrong language	Settings → Language & Locale, or add ?lang=en to the URL.
Settings changes do not stick	Local storage is blocked (private window, strict privacy mode). Settings and autosave need it.

Reporting a bug

Open a GitHub issue with the browser and OS, what you expected, and—most useful of all—a share link or the project JSON reproducing the problem.