diff --git a/App/main.c b/App/main.c
index 771a9ff..a4c4dc1 100644
--- a/App/main.c
+++ b/App/main.c
@@ -145,43 +145,38 @@ void WS2812_EncodeByte(u8 v, u8 *buf)
 void WS2812_Write24Bit(u8 g, u8 r, u8 b)
 {
     u8 data buf[10];
-    u16 timeout;
+    u8 b0, b1, b2, b3, b4, b5, b6, b7, b8, b9;
     
     buf[0] = 0x00; // Prepend dummy 0x00 to absorb SPI hardware startup delay/glitch
     WS2812_EncodeByte(g, &buf[1]);
     WS2812_EncodeByte(r, &buf[4]);
     WS2812_EncodeByte(b, &buf[7]);
     
+    b0 = buf[0]; b1 = buf[1]; b2 = buf[2]; b3 = buf[3];
+    b4 = buf[4]; b5 = buf[5]; b6 = buf[6]; b7 = buf[7];
+    b8 = buf[8]; b9 = buf[9];
+    
     EA = 0;
     
     // Ensure P2.3 (MOSI) and P2.5 (SCLK) are configured as push-pull output
     P2M1 &= ~((1 << 3) | (1 << 5));
     P2M0 |=  ((1 << 3) | (1 << 5));
     
-    SPSTAT = 0xC0; // Clear flags
-    SPCTL = 0xD0;  // Enable SPI at SYSCLK/4 (3.0 MHz actual clock on board)
+    SPSTAT = 0xC0; // Clear flags first
+    SPDAT = b0;    // Write data first!
+    SPCTL = 0xD0;  // Enable SPI after writing data. Should start shifting b0 immediately!
     
-    SPDAT = buf[0]; // Start transmitting dummy 0x00 (absorbs startup delay at MOSI=0)
-    // Wait 30 NOPs to ensure buf[0] has moved to the shifter
-    _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_();
-    _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_();
-    _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_();
-    _nop_(); _nop_(); _nop_(); _nop_(); _nop_(); _nop_();
-    
-    SPDAT = buf[1]; // Buffer buf[1] (zero gap!)
-    
-    // Transmission loop unrolled with double-buffering and timeout protection
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[2];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[3];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[4];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[5];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[6];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[7];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[8];
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; SPDAT = buf[9];
-    
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; // Wait for buf[8] to finish shifting, buf[9] moves to shifter
-    timeout = 2000; while (!(SPSTAT & 0x80) && --timeout); SPSTAT = 0xC0; // Wait for buf[9] to finish shifting
+    // Unrolled 10-byte transmission with minimal gap (~300ns)
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b1;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b2;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b3;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b4;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b5;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b6;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b7;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b8;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0; SPDAT = b9;
+    while (!(SPSTAT & 0x80)); SPSTAT = 0xC0;
     
     SPCTL = 0x90; // Disable SPI
     WS2812_DI = 0;