diff --git a/docs/usage.rst b/docs/usage.rst
index 3414a7f8..acaf020b 100644
--- a/docs/usage.rst
+++ b/docs/usage.rst
@@ -1057,7 +1057,7 @@ HDF5 Files
 
 The :mod:`unyt` library provides a hook for writing data both to a new HDF5 file and an existing file and then subsequently reading that data back in to restore the array. This works via the :meth:`unyt_array.write_hdf5 <unyt.array.unyt_array.write_hdf5>` and :meth:`unyt_array.from_hdf5 <unyt.array.unyt_array.from_hdf5>` methods. The simplest way to use these functions is to write data to a file that does not exist yet:
 
-  >>> from unyt import cm, unyt_array
+  >>> from unyt import cm
   >>> import os
   >>> data = [1, 2, 3]*cm
   >>> data.write_hdf5('my_data.h5')
diff --git a/unyt/array.py b/unyt/array.py
index b92b2e06..a58f90fd 100644
--- a/unyt/array.py
+++ b/unyt/array.py
@@ -127,8 +127,13 @@
 from unyt.equivalencies import equivalence_registry
 from unyt._on_demand_imports import _astropy, _pint
 from unyt._pint_conversions import convert_pint_units
+from unyt._unit_lookup_table import default_unit_symbol_lut
 from unyt.unit_object import _check_em_conversion, _em_conversion, Unit
-from unyt.unit_registry import _sanitize_unit_system, UnitRegistry
+from unyt.unit_registry import (
+    _sanitize_unit_system,
+    UnitRegistry,
+    default_unit_registry,
+)
 
 NULL_UNIT = Unit()
 POWER_SIGN_MAPPING = {multiply: 1, divide: -1}
@@ -1319,7 +1324,11 @@ def write_hdf5(self, filename, dataset_name=None, info=None, group_name=None):
             info = {}
 
         info["units"] = str(self.units)
-        info["unit_registry"] = np.void(pickle.dumps(self.units.registry.lut))
+        lut = {}
+        for k, v in self.units.registry.lut.items():
+            if k not in default_unit_registry.lut:
+                lut[k] = v
+        info["unit_registry"] = np.void(pickle.dumps(lut))
 
         if dataset_name is None:
             dataset_name = "array_data"
@@ -1382,7 +1391,9 @@ def from_hdf5(cls, filename, dataset_name=None, group_name=None):
         dataset = g[dataset_name]
         data = dataset[:]
         units = dataset.attrs.get("units", "")
-        unit_lut = pickle.loads(dataset.attrs["unit_registry"].tostring())
+        unit_lut = default_unit_symbol_lut.copy()
+        unit_lut_load = pickle.loads(dataset.attrs["unit_registry"].tostring())
+        unit_lut.update(unit_lut_load)
         f.close()
         registry = UnitRegistry(lut=unit_lut, add_default_symbols=False)
         return cls(data, units, registry=registry)