Open yenpomao opened 5 months ago
Problem can be solve by using more states/more accurate simulation.
During the graph calculation (graph = mr0.compute_graph(seq0, A, 200, 1e-4)) the first value (200 here) is the maximal number of states and 1e-4 is the threshold. Use a higher value (e.g. 20000) and more accurate simulation (e.g. 1e-6), then the signal is lower than the FID signal.
graph = mr0.compute_graph(seq0, A, 200, 1e-4)
graph = mr0.compute_graph(seq0, A, 20000, 1e-6)
We will look into possible solutions to estimate and report the achieved precision with the used settings. However, this might proof to be difficult or even impossible without actually computing with higher precision.
https://colab.research.google.com/drive/1EckZcHbvGZ5XBMlLJZIYTla9muIbzZhP?authuser=1#scrollTo=xrlenUgfDCYP I have a simple FID readout in the first part of colab. In the second part of code I have a prepare readout and the signal is higher than FID which is unexpected. furthermore, when I rounded digit into 1(origin is 4) the issue disappeared.
![image](https://github.com/MRsources/MRzero-Core/assets/141250768/299bbf0c-6a12-4379-b0c5-dec7d976636f)