diff --git a/src/bech32/bech32.zig b/src/bech32/bech32.zig
index 3216482..2136c3a 100644
--- a/src/bech32/bech32.zig
+++ b/src/bech32/bech32.zig
@@ -153,7 +153,7 @@ pub const Variant = enum {
 fn splitAndDecode(allocator: std.mem.Allocator, s: []const u8) Error!struct { std.ArrayList(u8), std.ArrayList(u5) } {
     // Split at separator and check for two pieces
 
-    const raw_hrp, const raw_data = if (std.mem.indexOfScalar(u8, s, SEP)) |sep| .{
+    const raw_hrp, const raw_data = if (std.mem.lastIndexOfScalar(u8, s, SEP)) |sep| .{
         s[0..sep], s[sep + 1 ..],
     } else return Error.MissingSeparator;
 
@@ -177,20 +177,27 @@ fn splitAndDecode(allocator: std.mem.Allocator, s: []const u8) Error!struct { st
     var data = std.ArrayList(u5).init(allocator);
     errdefer data.deinit();
 
+    var it = std.unicode.Utf8View.initUnchecked(raw_data).iterator();
     // Check data payload
-    for (raw_data) |c| {
+    while (it.nextCodepoint()) |c| {
         // Only check if c is in the ASCII range, all invalid ASCII
         // characters have the value -1 in CHARSET_REV (which covers
         // the whole ASCII range) and will be filtered out later.
-        if (!std.ascii.isAscii(c)) return error.InvalidChar;
+        if (c >= 128) return error.InvalidChar;
 
-        if (std.ascii.isLower(c)) {
+        if (switch (c) {
+            'a'...'z' => true,
+            else => false,
+        }) {
             switch (case) {
                 .upper => return Error.MixedCase,
                 .none => case = .lower,
                 .lower => {},
             }
-        } else if (std.ascii.isUpper(c)) {
+        } else if (switch (c) {
+            'A'...'Z' => true,
+            else => false,
+        }) {
             switch (case) {
                 .lower => return Error.MixedCase,
                 .none => case = .upper,
@@ -531,6 +538,15 @@ test "roundtrip_without_checksum" {
     try std.testing.expectEqualSlices(u5, data.items, decoded_data.items);
 }
 
+test "decode ln" {
+    const str = "lnbc10n1pnw2hkzdqqpp5ve584t0cv27hwmy0cx9ca8uwyqyfw9y9dm3r8vus9fv36r2l9yjssp59g4z52329g4z52329g4z52329g4z52329g4z52329g4z52329g4qcqzysc39n3w6zq50y3775yc66mmvt2fe5aa9mzzdnvq5palgyw6j9uu04qnm0g4ftm5ehtm5aulwtuy0hfrtqcdaxcl2r7wz2x4503levlacqswkkem";
+
+    const h, const d, const v = try decode(std.testing.allocator, str);
+    _ = v; // autofix
+    defer h.deinit();
+    defer d.deinit();
+}
+
 test "test_hrp_case_decode" {
     const hrp, const data, const variant = try decode(std.testing.allocator, "hrp1qqqq40atq3");
     defer hrp.deinit();