MDP FANode issues

I'm trying to perform factorial analysis on a distance matrix (made of distances between about 1700 points, all ranging between 0.0 and 1.0, inclusively). I'm a total FA newbie.

Anyways, this code:

fan=mdp.nodes.FANode()
far=fan.execute(a)
# a is a numpy.array, size 1780x1780

Gives me:

Traceback (most recent call last):
  File "<pyshell#29>", line 1, in <mod开发者_StackOverflowule>
    far=fan.execute(a)
  File "/usr/lib/pymodules/python2.7/mdp/signal_node.py", line 575, in execute
    self._pre_execution_checks(x)
  File "/usr/lib/pymodules/python2.7/mdp/signal_node.py", line 451, in _pre_execution_checks
    self._if_training_stop_training()
  File "/usr/lib/pymodules/python2.7/mdp/signal_node.py", line 431, in _if_training_stop_training
    self.stop_training()
  File "/usr/lib/pymodules/python2.7/mdp/signal_node.py", line 556, in stop_training
    self._train_seq[self._train_phase][1](*args, **kwargs)
  File "/usr/lib/pymodules/python2.7/mdp/nodes/em_nodes.py", line 93, in _stop_training
    A = normal(0., sqrt(scale/k), size=(d, k)).astype(typ)
  File "mtrand.pyx", line 1279, in mtrand.RandomState.normal (numpy/random/mtrand/mtrand.c:6943)
ValueError: scale <= 0

I tried replacing 0 values with 0.00001, to no avail. Any idea what this might mean?