diff --git a/scripts/legacy_xrd/convert_legacy_xrd.py b/scripts/legacy_xrd/convert_legacy_xrd.py
index b87014a..afd3f94 100644
--- a/scripts/legacy_xrd/convert_legacy_xrd.py
+++ b/scripts/legacy_xrd/convert_legacy_xrd.py
@@ -66,14 +66,18 @@
     "CW": "Caleb Wasserbeck",
 }
 
-# Scan settings by step size (degrees 2-theta). Times are as reported by the
-# lab (effective time per step); quick scans are routine, long ones are the
-# ~2 h scans used for Rietveld refinement.
+# Scan settings by step size (degrees 2-theta), with the time per step in
+# seconds confirmed by Tufail Hassan (email, 2026-10-12). Quick scans are
+# routine; long ones are used for Rietveld refinement.
 SCAN_TYPES = {
-    Fraction("0.02"): ("quick", 49.0),
-    Fraction("0.015"): ("long", 361.24),
+    Fraction("0.02"): ("quick", 0.191),
+    Fraction("0.015"): ("long", 1.41),
 }
 
+# Column for the pattern file: the output name on the lab's registered
+# diffractometer.
+PATTERN_FIELD = "XRD pattern file"
+
 # Conditions the lab reports as constant for its MAX synthesis runs.
 LEGACY_CONSTANTS = {
     "Atmosphere": "Ar",
@@ -129,7 +133,7 @@
     "Two Theta Max",
     "Points",
     "Researcher",
-    "XRD Pattern",
+    PATTERN_FIELD,
     "Source File",
     "Tags",
 ]
@@ -499,7 +503,7 @@ def convert(
                 "Two Theta Max": _fmt_number(max(pattern.two_theta)),
                 "Points": str(len(pattern.two_theta)),
                 "Researcher": researcher,
-                "XRD Pattern": f"patterns/{pattern_name}",
+                PATTERN_FIELD: f"patterns/{pattern_name}",
                 "Source File": path.name,
                 "Tags": " ".join(parsed.tags),
             }
diff --git a/scripts/legacy_xrd/test_convert_legacy_xrd.py b/scripts/legacy_xrd/test_convert_legacy_xrd.py
index 59c4316..529c210 100644
--- a/scripts/legacy_xrd/test_convert_legacy_xrd.py
+++ b/scripts/legacy_xrd/test_convert_legacy_xrd.py
@@ -11,7 +11,13 @@
 
 sys.path.insert(0, str(Path(__file__).parent))
 
-from convert_legacy_xrd import FormulaError, convert, parse_filename, parse_formula  # noqa: E402
+from convert_legacy_xrd import (  # noqa: E402
+    FormulaError,
+    classify_scan,
+    convert,
+    parse_filename,
+    parse_formula,
+)
 
 RAW_LINES = [f'"{3 + i * 0.02:.16g}\t{100 + i}"\r\n' for i in range(50)]
 
@@ -98,6 +104,13 @@ def test_bad_date_is_an_error(self):
         self.assertTrue(parse_filename("133024_Ti3AlC2_1500_4h_TH").problems)
 
 
+class ClassifyScanTests(unittest.TestCase):
+    def test_lab_confirmed_times_per_step(self):
+        self.assertEqual(classify_scan(0.02), ("quick", 0.191))
+        self.assertEqual(classify_scan(0.015), ("long", 1.41))
+        self.assertEqual(classify_scan(0.05), ("unknown", None))
+
+
 class ConvertTests(unittest.TestCase):
     def test_splits_into_synthesis_and_xrd_linked_by_sample_id(self):
         with tempfile.TemporaryDirectory() as tmp:
@@ -124,18 +137,19 @@ def test_splits_into_synthesis_and_xrd_linked_by_sample_id(self):
             self.assertRegex(sample["Sample ID"], r"^PUDT[0-9A-Z]{10}$")
             self.assertEqual({row["Sample ID"] for row in xrd}, {sample["Sample ID"]})
             self.assertEqual(xrd[0]["Scan Type"], "quick")
+            self.assertEqual((xrd[0]["Step Size"], xrd[0]["Time per Step"]), ("0.02", "0.191"))
             self.assertEqual(xrd[1]["Tags"], "rescan")
             self.assertEqual((xrd[0]["Loading"], xrd[0]["Sample Preparation"]), ("top", "ground powder"))
             self.assertIn("not stripped", xrd[0]["Radiation"])
             self.assertEqual(sample["Precursors"], "")
 
-            pattern = (out / xrd[0]["XRD Pattern"]).read_text().splitlines()
+            pattern = (out / xrd[0]["XRD pattern file"]).read_text().splitlines()
             self.assertEqual(pattern[0], "two_theta_deg,intensity_counts")
             self.assertEqual(pattern[1], "3.0,100")
             self.assertEqual(len(pattern), 51)
             with zipfile.ZipFile(out / "patterns.zip") as archive:
                 self.assertEqual(
-                    sorted(archive.namelist()), sorted(row["XRD Pattern"] for row in xrd)
+                    sorted(archive.namelist()), sorted(row["XRD pattern file"] for row in xrd)
                 )
 
             # It's exactly the ID the platform issues for that synthesis run.