diff --git a/config/example-extras.cfg b/config/example-extras.cfg
index 1398f992..81231f75 100644
--- a/config/example-extras.cfg
+++ b/config/example-extras.cfg
@@ -539,10 +539,6 @@
 #sid_pin:
 #   The pins connected to an st7920 type lcd. These parameters must be
 #   provided when using an st7920 display.
-#chip_delay:
-#   This parameter specifies the internal command delay (in seconds)
-#   on st7920 displays. It may be necessary to increase this if the
-#   display shows garbled data. The default is 0.000020.
 
 
 # Custom thermistors (one may define any number of sections with a
diff --git a/config/generic-re-arm.cfg b/config/generic-re-arm.cfg
index 4a9c9fab..13a2c792 100644
--- a/config/generic-re-arm.cfg
+++ b/config/generic-re-arm.cfg
@@ -93,8 +93,3 @@ max_z_accel: 100
 #cs_pin: P0.16
 #sclk_pin: P0.15
 #sid_pin: P0.18
-#chip_delay needs to be set to the below for the LCD to initialise correctly.
-#chip_delay: .000040
-
-
-
diff --git a/klippy/extras/display.py b/klippy/extras/display.py
index e68c3db7..00a4777d 100644
--- a/klippy/extras/display.py
+++ b/klippy/extras/display.py
@@ -161,7 +161,9 @@ HD44780_chars = [
 # ST7920 (128x64 graphics) lcd chip
 ######################################################################
 
-ST7920_DELAY = .000020 # Spec says 72us, but faster is possible in practice
+# Spec says 72us, but faster is possible in practice
+ST7920_CMD_DELAY  = .000020
+ST7920_SYNC_DELAY = .000045
 
 class ST7920:
     char_right_arrow = '\x1a'
@@ -182,7 +184,6 @@ class ST7920:
         self.mcu = mcu
         self.oid = self.mcu.create_oid()
         self.mcu.add_config_object(self)
-        self.chip_delay = config.getfloat('chip_delay', ST7920_DELAY, minval=0.)
         self.send_data_cmd = self.send_cmds_cmd = None
         self.is_extended = False
         # framebuffers
@@ -195,9 +196,10 @@ class ST7920:
     def build_config(self):
         self.mcu.add_config_cmd(
             "config_st7920 oid=%u cs_pin=%s sclk_pin=%s sid_pin=%s"
-            " delay_ticks=%d" % (
+            " sync_delay_ticks=%d cmd_delay_ticks=%d" % (
                 self.oid, self.pins[0], self.pins[1], self.pins[2],
-                self.mcu.seconds_to_clock(self.chip_delay)))
+                self.mcu.seconds_to_clock(ST7920_SYNC_DELAY),
+                self.mcu.seconds_to_clock(ST7920_CMD_DELAY)))
         cmd_queue = self.mcu.alloc_command_queue()
         self.send_cmds_cmd = self.mcu.lookup_command(
             "st7920_send_cmds oid=%c cmds=%*s", cq=cmd_queue)
diff --git a/src/lcd_st7920.c b/src/lcd_st7920.c
index 4351d6ac..8cd76c50 100644
--- a/src/lcd_st7920.c
+++ b/src/lcd_st7920.c
@@ -12,7 +12,7 @@
 #include "sched.h" // DECL_SHUTDOWN
 
 struct st7920 {
-    uint32_t last_cmd_time, cmd_wait_ticks;
+    uint32_t last_cmd_time, sync_wait_ticks, cmd_wait_ticks;
     struct gpio_out sclk, sid;
 };
 
@@ -45,16 +45,31 @@ st7920_xmit_byte(struct st7920 *s, uint8_t data)
 static void
 st7920_xmit(struct st7920 *s, uint8_t count, uint8_t *cmds)
 {
-    uint32_t last_cmd_time=s->last_cmd_time, cmd_wait_ticks=s->cmd_wait_ticks;
-    while (count--) {
+    if (!count)
+        return;
+
+    // Send first byte (with longer delay)
+    uint32_t last_cmd_time = s->last_cmd_time, wait_ticks = s->sync_wait_ticks;
+    uint8_t cmd = *cmds++;
+    st7920_xmit_byte(s, cmd & 0xf0);
+    while (timer_read_time() - last_cmd_time < wait_ticks)
+        // Can't complete transfer until delay complete
+        irq_poll();
+    st7920_xmit_byte(s, cmd << 4);
+    last_cmd_time = timer_read_time();
+
+    // Send subsequent bytes
+    wait_ticks = s->cmd_wait_ticks;
+    while (--count) {
         uint8_t cmd = *cmds++;
         st7920_xmit_byte(s, cmd & 0xf0);
-        // Can't complete transfer until delay complete
-        while (timer_read_time() - last_cmd_time < cmd_wait_ticks)
+        while (timer_read_time() - last_cmd_time < wait_ticks)
+            // Can't complete transfer until delay complete
             irq_poll();
         st7920_xmit_byte(s, cmd << 4);
         last_cmd_time = timer_read_time();
     }
+
     s->last_cmd_time = last_cmd_time;
 }
 
@@ -80,13 +95,16 @@ command_config_st7920(uint32_t *args)
     uint32_t end = timer_read_time();
     irq_enable();
     s->last_cmd_time = end;
-    uint32_t diff = end - start, delay_ticks = args[4];
-    if (delay_ticks > diff)
-        s->cmd_wait_ticks = delay_ticks - diff;
+    uint32_t diff = end - start, sync_delay_ticks = args[4];
+    if (sync_delay_ticks > diff)
+        s->sync_wait_ticks = sync_delay_ticks - diff;
+    uint32_t cmd_delay_ticks = args[5];
+    if (cmd_delay_ticks > diff)
+        s->cmd_wait_ticks = cmd_delay_ticks - diff;
 }
 DECL_COMMAND(command_config_st7920,
              "config_st7920 oid=%c cs_pin=%u sclk_pin=%u sid_pin=%u"
-             " delay_ticks=%u");
+             " sync_delay_ticks=%u cmd_delay_ticks=%u");
 
 void
 command_st7920_send_cmds(uint32_t *args)