MaxError
2D/3D

• MaxError is an option for SolveFree and SolvePath that is used to specify the maximum allowable estimated error in any dependent variable in a single time step.

• If MaxError is exceeded at any time step, the time step size is halved.
• The default error estimation algorithm used by SolveFree assumes that the average magnitude of each of the dependent variables is near to one (which is clearly true for the Euler parameters or angular variables in units of radians).
• Setting MaxError to a larger value effectively relaxes the error estimation criteria that are used to adjust the time step size.
• The MaxError option can also be used to explicitly specify the error limits at which the time step size is halved or doubled, or when a corrective versus predictive step is taken.
• MaxError->{a, b, c} specifies that the time step size will be halved if the maximum error in the current step is greater than a. The step size will be doubled if the maximum error in any of the last 2n steps drops below c, where n is the degree of the integrand fitting polynomial. An Adams-Moulton corrective step will be taken if the error in the current step is between a and b.
• The default setting is MaxError->0.0001.
• MaxError->{0.0001, 0.00001, 0.000001} is equivalent to the default setting.
• See also: Diagnostics.