update FAQ on how to register pandas converters

`pandas.tseries.converter` no longer exists, and the individual converters are
not part of the public API so this change updates the documentation to reflect
how to properly register the pandas converters with matplotlib.

Author: terencehonles

doc/faq/howto_faq.rst

@@ -21,20 +21,12 @@ Plot `numpy.datetime64` values
 As of Matplotlib 2.2, `numpy.datetime64` objects are handled the same way
 as `datetime.datetime` objects.
 
-If you prefer the pandas converters and locators, you can register their
-converter with the `matplotlib.units` module::
-
-  from pandas.tseries import converter as pdtc
-  pdtc.register()
-
-If you only want to use the `pandas` converter for `numpy.datetime64` values ::
-
-  from pandas.tseries import converter as pdtc
-  import matplotlib.units as munits
-  import numpy as np
-
-  munits.registry[np.datetime64] = pdtc.DatetimeConverter()
+If you prefer the pandas converters and locators, you can register them.  This
+is done automatically when calling a pandas plot function and may be
+unnecessary when using pandas instead of Matplotlib directly. ::
 
+  from pandas.plotting import register_matplotlib_converters
+  register_matplotlib_converters()
 
 
 .. _howto-figure-empty: