A new counter-example to Kelvin's conjecture on minimal surfaces