[Bug]: pybamm.ScipySolver fails to run experiment

[martinjrobins]

PyBaMM Version

develop

Python Version

3.10

Describe the bug

the pybamm.ScipySolver aborts early on the experiment given below, whereas the casadi and idaklu solver solve it correctly

Steps to Reproduce

import pybamm
import numpy as np
import matplotlib.pyplot as plt

experiment_2step = pybamm.Experiment(
    [
        (
            "Discharge at C/20 for 1 hour",
            "Charge at 1 A until 4.1 V",
            "Hold at 4.1 V until C/2",
            "Discharge at 2 W for 30 min",
            "Discharge at 2 W for 30 min",  # repeat to cover this case (changes initialisation)
        ),
    ]
    * 2,
)

solutions = []
for solver in [
    pybamm.CasadiSolver(),
    pybamm.IDAKLUSolver(),
    pybamm.ScipySolver(),
]:
      model = pybamm.lithium_ion.SPM()
      param = model.default_parameter_values
      input_param_name = "Negative electrode active material volume fraction"
      input_param_value = param[input_param_name]
      param.update({input_param_name: "[input]"})
      sim = pybamm.Simulation(
          model,
          experiment=experiment_2step,
          solver=solver,
          parameter_values=param,
      )
      solution = sim.solve(
          inputs={input_param_name: input_param_value},
      )
      solutions.append(solution)

for i, s in enumerate(solutions):
    plt.plot(s.t, s["Voltage [V]"](s.t) + float(i), label=f"Solver {i}")
plt.savefig("test.png")


for i in range(1, len(solutions)):
    np.testing.assert_allclose(
        solutions[0]["Voltage [V]"].data[:-1],
        solutions[i]["Voltage [V]"](solutions[0].t[:-1]),
        rtol=5e-2,
        equal_nan=True,
    )

Relevant log output

No response

[rtimms]

This is because voltage control turns the model from an ODE to a DAE. Surprised we don't get a better error message here?

[martinjrobins]

ahh yes, and the scipy solver doesn't support DAEs. Ok, so the main thing here is raising an error (and updating the relevent test in #4415)

[martinjrobins]

my bad. All the noise from my assertions hid the error, which was perfectly fine. Closing this one as done