Simulating power-on with a TLV3691 in TINA or P-SPICE produces a 2V 360us output glitch. I assume this is a problem with the macromodel- can you confirm that the actual device does not exhibit this behavior? If it is a problem with this model you might want to fix it as a customer is already questioning the selection of a TLV3691 for his circuit design based on a simulation showing this output glitch.