diff --git a/nametag8_CH592/user/comm/i2c.h b/nametag8_CH592/user/comm/i2c.h
index 4476725..54269aa 100644
--- a/nametag8_CH592/user/comm/i2c.h
+++ b/nametag8_CH592/user/comm/i2c.h
@@ -18,7 +18,9 @@
 
 #include "soft_i2c_master.h"
 
-
+/* known bugs / limitations:
+ * - there is no ack on address. code always assumes slave is responding.
+ */
 #define i2c_init()          i2cm_init()
 #define i2c_start()         SetSysClock(SYSCLK_FREQ_USEI2C); i2cm_start()
 #define i2c_restart()       i2cm_restart()
diff --git a/nametag8_CH592/user/comm/soft_i2c_master.c b/nametag8_CH592/user/comm/soft_i2c_master.c
index 329d405..3476518 100644
--- a/nametag8_CH592/user/comm/soft_i2c_master.c
+++ b/nametag8_CH592/user/comm/soft_i2c_master.c
@@ -19,12 +19,12 @@
 #include <stdint.h>
 
 
-#define CYCLES_TO_HI        16
-#define CYCLES_TO_LO        16
+#define CYCLES_TO_HI    5   // these settings will result in ~500KHz (32MHz HCLK)
+#define CYCLES_TO_LO    2
 
-#define CYCLES_RD           2   // cycles spent in read routine
-#define CYCLES_EXTRA_WR_HI  2   // extra cycles spent in write routine
-#define CYCLES_EXTRA_WR_LO  4
+#define CYCLES_RD       2   // cycles spent in read routine
+#define CYCLES_WR_HI    2   // extra cycles spent in write routine
+#define CYCLES_WR_LO    4
 
 
 #define SDA_PIN         4
@@ -62,18 +62,18 @@ static volatile uint16_t spin;
 #define bit_delay_hi()      spin = delay_hi; while(spin--)
 #define bit_delay_lo()      spin = delay_lo; while(spin--)
 
-#define rd_delay()          spin = delay_hi - CYCLES_RD; while(spin--)
+#define rd_delay()          bit_delay_hi()
 
-#define wr_delay_hi()       spin = delay_hi - CYCLES_EXTRA_WR_HI; while(spin--)
-#define wr_delay_lo()       spin = delay_lo - CYCLES_EXTRA_WR_LO; while(spin--)
+#define wr_delay_hi()       bit_delay_hi()
+#define wr_delay_lo()       bit_delay_hi()      // spin = delay_hi - CYCLES_EXTRA_WR_HI; while(spin--)
 
 
 
 // re-run init any time the clock speed changes to recalculate delays.
 void i2cm_init()
 {
-    uint32_t sysclk;
-    uint32_t cycles;
+    //uint32_t sysclk;
+    //uint32_t cycles;
 
     // configure GPIO
     SCL_IN_HI();
@@ -83,13 +83,14 @@ void i2cm_init()
     SDA_SET_LO();
 
     // configure timer
-    sysclk = GetSysClock();
-    cycles = sysclk / 500000;
+    //sysclk = GetSysClock();
+    //cycles = sysclk / 500000;
 
-    delay_hi = (cycles - CYCLES_TO_HI - 2) / 4;
-    delay_lo = (cycles - CYCLES_TO_LO - 2) / 4;
+    delay_hi = CYCLES_TO_HI;
+    delay_lo = CYCLES_TO_LO;
 }
 
+__attribute__((section(".ramfunc")))
 void i2cm_start()
 {
     SDA_IN_HI(); bit_delay_hi();
@@ -98,6 +99,7 @@ void i2cm_start()
     SCL_OUTLO(); bit_delay_lo();
 }
 
+__attribute__((section(".ramfunc")))
 void i2cm_restart()
 {
     SDA_IN_HI(); bit_delay_hi();
@@ -105,6 +107,7 @@ void i2cm_restart()
     i2cm_start();
 }
 
+__attribute__((section(".ramfunc")))
 void i2cm_stop()
 {
     SDA_OUTLO(); bit_delay_lo();
@@ -118,11 +121,13 @@ void i2cm_stop()
 }
 
 // returns: data byte
+__attribute__((section(".ramfunc")))
 uint8_t i2cm_rd(uint8_t ack)
 {
     uint8_t x, in = 0;
 
     SDA_IN_HI();
+
     for (x = 8; x; x--) {
         in <<= 1;               // clock next bit
 
@@ -142,29 +147,30 @@ uint8_t i2cm_rd(uint8_t ack)
     SCL_IN_HI(); bit_delay_hi();
 
     SDA_IN_HI();
-    SCL_OUTLO();
+    SCL_OUTLO(); bit_delay_lo();
 
     return in;
 }
 
 // returns: possible ack from target
+__attribute__((section(".ramfunc")))
 uint8_t i2cm_wr(uint8_t dat)
 {
     uint8_t x;
     uint8_t ack;
 
-    SCL_IN_HI();
-    while (!SCL_GET());     // clock stretch
+    SCL_OUTLO();
 
     for (x = 8; x; x--) {
         if (dat & 0x80) { SDA_IN_HI(); }
         else            { SDA_OUTLO(); }
 
         SCL_IN_HI(); wr_delay_hi();
+        while (!SCL_GET());     // clock stretch
 
         dat <<= 1;
 
-        SCL_OUTLO();
+        SCL_OUTLO(); bit_delay_lo();
     }
 
     SDA_IN_HI();
@@ -172,11 +178,12 @@ uint8_t i2cm_wr(uint8_t dat)
 
     ack = SDA_GET();
 
-    SCL_OUTLO();
+    SCL_OUTLO(); bit_delay_lo();
     return ack;
 }
 
 // use a left-aligned address with this implementation.
+__attribute__((section(".ramfunc")))
 uint8_t i2cm_addr(uint8_t addr, uint8_t reading_bit)
 {
     addr &= ~0x1;
@@ -185,13 +192,13 @@ uint8_t i2cm_addr(uint8_t addr, uint8_t reading_bit)
 }
 
 
-void i2cm_rdbuf(uint8_t *dat, uint8_t len)
+void i2cm_rdbuf(uint8_t *dat, uint16_t len)
 {
     while (len--) *dat++ = i2cm_rd(len > 0);
     // i2cm_stop();
 }
 
-void i2cm_wrbuf(const uint8_t *dat, uint8_t len)
+void i2cm_wrbuf(const uint8_t *dat, uint16_t len)
 {
     uint8_t nack;
 
diff --git a/nametag8_CH592/user/comm/soft_i2c_master.h b/nametag8_CH592/user/comm/soft_i2c_master.h
index 80da7b6..0f67499 100644
--- a/nametag8_CH592/user/comm/soft_i2c_master.h
+++ b/nametag8_CH592/user/comm/soft_i2c_master.h
@@ -20,8 +20,8 @@ uint8_t i2cm_rd(uint8_t ack);
 uint8_t i2cm_wr(uint8_t dat);
 uint8_t i2cm_addr(uint8_t addr, uint8_t reading_bit);
 
-void i2cm_rdbuf(uint8_t *dat, uint8_t len);
-void i2cm_wrbuf(const uint8_t *dat, uint8_t len);
+void i2cm_rdbuf(uint8_t *dat, uint16_t len);
+void i2cm_wrbuf(const uint8_t *dat, uint16_t len);
 
 
 
diff --git a/nametag8_CH592/user/hw/ssd1306.h b/nametag8_CH592/user/hw/ssd1306.h
index 89e2330..9600c45 100644
--- a/nametag8_CH592/user/hw/ssd1306.h
+++ b/nametag8_CH592/user/hw/ssd1306.h
@@ -82,7 +82,7 @@ extern uint8_t oled_fb[((SSD1306_HEIGHT >> 3) * SSD1306_WIDTH) + 1];
 
 
 
-#define ssd1306_write(dat, len, cb)     i2c_addr(SSD1306_I2C_ADDR, 1); i2c_wrbuf(dat, len); if (cb) ssd1306_cb(0)
+#define ssd1306_write(dat, len, cb)     i2c_addr((SSD1306_I2C_ADDR << 1), 0); i2c_wrbuf(dat, len); if (cb) ssd1306_cb(0)
 #define ssd1306_idle()                  0
 
 
diff --git a/nametag8_CH592/user/main.c b/nametag8_CH592/user/main.c
index df91198..f5cfbb6 100644
--- a/nametag8_CH592/user/main.c
+++ b/nametag8_CH592/user/main.c
@@ -43,7 +43,7 @@
 
 
 // global settings
-#define OLED_UPDATE_RATE        64      // framerate of OLED; (256*0.75) / OLED_UPDATE_RATE
+#define OLED_UPDATE_RATE        8      // framerate of OLED; (256*(6/8)) / OLED_UPDATE_RATE; 8 = 24FPS
 
 // flags
 #define FLAG_OLED_UPDATE        (1 << 0)
@@ -87,6 +87,7 @@ void ch59x_xtal_conf()
  * trigger an interrupt when bits match a mask. at this time we're not using the RTC
  * for anything else so putting RTC should suffice for our main tick interrupt.
  */
+__HIGH_CODE
 void rtc_reset_trig()
 {
     sys_safe_access_enable();
@@ -113,6 +114,7 @@ void rtcisr_init()
     PFIC_EnableIRQ(RTC_IRQn);
 }
 
+__HIGH_CODE
 void oled_update_done()
 {
     int16_t a;
@@ -228,9 +230,10 @@ int main()
     // configure port-based interrupts (used for ch32sub interrupt)
     port_intr_init();
 
-    // start up the OLED
+    // start up the OLED and menu system
     ssd1306fb_set_target(&oled);
     ssd1306_init(1);    // we'll try to init later too, since sometimes they fail
+    menu_stop(0);       // lol. yes, this "starts" the "menu" in nametag mode
 
     // configure main program tick interrupt
     rtcisr_init();
@@ -289,31 +292,37 @@ void RTC_IRQHandler(void)
         uptime_sec  = (uint8_t)((uptime       ) % 60);
     }
 
-    // render and update RGBLED at 64Hz
-    if ((st_tick & 0x3) == 0x3) {
-        // make sure a valid program is selected
-        if (uconf.ledprog_rgb_idx > (sizeof(rgb_pgm) / sizeof(rgb_pgm[0]))) {
-            uconf.ledprog_rgb_idx = 0;
+    // operations
+    switch (st_tick & 0x7) {
+        case 0:
+        case 3: {
+            // make sure a valid program is selected
+            if (uconf.ledprog_rgb_idx > (sizeof(rgb_pgm) / sizeof(rgb_pgm[0]))) {
+                uconf.ledprog_rgb_idx = 0;
+            }
+
+            // send any rendered data now
+            rgbled_send();
+
+            // defer rendering
+            flags |= FLAG_RGBLED_RUN_PROG;
+            break;
         }
-
-        // send any rendered data now
-        rgbled_send();
-
-        // defer rendering
-        flags |= FLAG_RGBLED_RUN_PROG;
-    }
-
-    // render and update the oled during non-rgbled frames
-    else {
-        oled_tick++;
-        if (oled_tick >= OLED_UPDATE_RATE) {
-            oled_tick = 0;
-            flags |= FLAG_OLED_UPDATE;
+        case 1: {
+            accel_poll();
+            // no break
+        }
+        default: {
+            oled_tick++;
+            if (oled_tick >= OLED_UPDATE_RATE) {
+                oled_tick = 0;
+                flags |= FLAG_OLED_UPDATE;
+            }
         }
     }
 
     // read accelerometer data at 32Hz
     if ((st_tick & 0x7) == 0x7) {
-        accel_poll();
+
     }
 }
diff --git a/nametag8_CH592/user/render/draw_ssd1306.c b/nametag8_CH592/user/render/draw_ssd1306.c
index afab911..a9cb899 100644
--- a/nametag8_CH592/user/render/draw_ssd1306.c
+++ b/nametag8_CH592/user/render/draw_ssd1306.c
@@ -4,6 +4,11 @@
  *
  * mostly implemented by true
  * some functions shamelessly copied and modified from interwebs
+ *
+ * todo:
+ * - new MCU is fast enough to convert entire framebuffer quickly.
+ *   it may save time and power to render a standard pixel framebuffer,
+ *   then convert. would also make code much cleaner. need to test.
  */
 
 #include "draw_ssd1306.h"
@@ -47,6 +52,7 @@ void ssd1306fb_set_color(uint8_t color)
  * there are assumptions about pages as well; Y pages will always
  * start on a boundary.
  */
+__HIGH_CODE
 void ssd1306fb_rotate(SSD1306 *src, SSD1306 *dst, int8_t rot)
 {
     int8_t x, y, nx, ny;
@@ -100,6 +106,7 @@ void ssd1306fb_rotate(SSD1306 *src, SSD1306 *dst, int8_t rot)
  * draws from data stored in horizontal priority page byte order
  * (useful for copying between buffers)
  */
+__HIGH_CODE
 void ssd1306fb_copy(int8_t x, int8_t y, uint8_t width, uint8_t height, const uint8_t *dat)
 {
     uint16_t i;
@@ -189,6 +196,7 @@ void ssd1306fb_copy(int8_t x, int8_t y, uint8_t width, uint8_t height, const uin
 /*
  * draws from data stored in vertical priority page byte order (why? why not just horizontal? ...)
  */
+__HIGH_CODE
 void ssd1306fb_draw_raw(int8_t x, int8_t y, uint8_t width, uint8_t height, const uint8_t *dat, uint16_t dat_offset, uint16_t bytes)
 {
     uint16_t i;
@@ -280,6 +288,7 @@ void ssd1306fb_draw_raw(int8_t x, int8_t y, uint8_t width, uint8_t height, const
     }
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_pix(uint8_t x, uint8_t y)
 {
     uint8_t *fb;
@@ -302,6 +311,7 @@ void ssd1306fb_draw_pix(uint8_t x, uint8_t y)
     }
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_hline(uint8_t x1, uint8_t x2, uint8_t y)
 {
     uint8_t t;
@@ -332,6 +342,7 @@ void ssd1306fb_draw_hline(uint8_t x1, uint8_t x2, uint8_t y)
     }
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_vline(uint8_t x, uint8_t y1, uint8_t y2)
 {
     uint8_t t;
@@ -407,6 +418,7 @@ void ssd1306fb_draw_vline(uint8_t x, uint8_t y1, uint8_t y2)
     }
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_line(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
 {
     int8_t t;
@@ -453,6 +465,7 @@ void ssd1306fb_draw_line(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
     }
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_rect(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
 {
     ssd1306fb_draw_vline(x1, y1 + 1, y2 - 1);
@@ -463,6 +476,7 @@ void ssd1306fb_draw_rect(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
 
 
 // note: rect fill not working right now, lol
+__HIGH_CODE
 void ssd1306fb_draw_rect_fill(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
 {
     uint8_t t;
@@ -476,6 +490,7 @@ void ssd1306fb_draw_rect_fill(uint8_t x1, uint8_t y1, uint8_t x2, uint8_t y2)
     } while (y1++ < y2);
 }
 
+__HIGH_CODE
 void ssd1306fb_draw_circle(int8_t x, int8_t y, uint8_t radius)
 {
     int8_t xs = 0;
@@ -508,6 +523,7 @@ void ssd1306fb_draw_circle(int8_t x, int8_t y, uint8_t radius)
 /*
  * character string drawing functions draw at the cursor position.
  */
+__HIGH_CODE
 uint8_t ssd1306fb_get_str_width(const uint8_t *font, const char *str, uint8_t len, int8_t extra_spacing)
 {
     uint8_t first = font[FONT_FIRST_CHAR_POS];
@@ -537,6 +553,7 @@ uint8_t ssd1306fb_get_font_height(const uint8_t *font)
     return font[FONT_HEIGHT_POS];
 }
 
+__HIGH_CODE
 void ssd1306fb_internal_str(const uint8_t *font, const char *str, uint8_t len, int8_t spacing)
 {
     const uint8_t height        = font[FONT_HEIGHT_POS];