diff --git a/content/content/articles/rgb-driver.md b/content/content/articles/rgb-driver.md
index cb5f3b8..19ae137 100644
--- a/content/content/articles/rgb-driver.md
+++ b/content/content/articles/rgb-driver.md
@@ -18,21 +18,21 @@ These signals are related to code under tag `cycler_works_include_output_stage`.
 
 First octets:
 
-![](./docs/cycler_working_first_octets.png)
+![](/cycler_working_first_octets.png)
 
 Last octets:
 
-![](./docs/cycler_working_last_octets.png)
+![](/cycler_working_last_octets.png)
 
 Schematics and legend for signals:
 
-![](./docs/schematics.jpeg)
+![](/schematics.jpeg)
 
 #### Some more explanations
 
 Consider above schematics and the screen shot "Last octets" from the oscilloscope.
 
-![](./docs/timing.png)
+![](/timing.png)
 
 Timer TA1 is running in "up mode" to the value 45 set in compare register `TA1CCR0`. The compare registers `TA1CCR1` is set to 10, `TA1CCR2` is set to 22.
 The output mode of the timer is set to "Reset/Set", which means the GPIO associated with `TA1CCR1` (P2.1) and `TA1CCR2` (P2.4) are set at the overflow and 
@@ -40,7 +40,7 @@ restart of the counter and reset when the counter matches the associated compare
 
 So, on P2.1 (D1 on the oscilloscope) we have a long pulse and at P2.4 (D0 on the oscilloscope) we have a short pulse, with synchronous raising edge.
 
-![](./docs/74hc74-function-table.png)
+![](/74hc74-function-table.png)
 
 The inverted signal P2.4 is connected to the Clock input of a 74HC74 D-flipflop, the data input of the flipflop is connected to GPIO P1.0 (D2 on the oscilloscope).
 
@@ -64,26 +64,26 @@ Additionally, when the first bit of a full draw screen cycle is presented at P1.
 
 Complete cycle: 2.48us
 
-![](./docs/pulse_complete.png)
+![](/pulse_complete.png)
 
 Short pulse: 550ns
 
-![](./docs/pulse_short.png)
+![](/pulse_short.png)
 
 Long pulse: 1.18us
 
-![](./docs/pulse_long.png)
+![](/pulse_long.png)
 
 
 ### Load Time
 
 During of loading data into five LEDs: 297us
 
-![](./docs/five_leds.png)
+![](/five_leds.png)
 
 During of loading data into six LEDs: 297us
 
-![](./docs/six_leds.png)
+![](/six_leds.png)
 
 
 | # of LEDs | Load Time measured | calculated |
@@ -103,11 +103,11 @@ will not handle the reset correctly.
 The following circuitry should generate a valid reset signal far enough from the raise 
 of the supply voltage:
 
-![](./docs/reset-circuit.jpeg)
+![](/reset-circuit.jpeg)
 
 The circuit generates the following signals:
 
-![](./docs/reset-signal.png)
+![](/reset-signal.png)
 
 ##### Reference voltage (green):
 
diff --git a/content/content/articles/tetris.md b/content/content/articles/tetris.md
index 7ef9c7f..085b241 100644
--- a/content/content/articles/tetris.md
+++ b/content/content/articles/tetris.md
@@ -4,11 +4,11 @@
 
 Update Amplifier (separate input circuitry per PSG, it appears, that a silent PSG has a DC level on its output which is summarized to the AC output of the working PSG, so two input circuits with individual couping capacitor):
 
-![](IMG_4941.jpg)
+![](/IMG_4941.jpg)
 
 Update of the power switch of the amplifier (at appears, that the small transistor couldn't deliver enough current):
 
-![](./docs/IMG_4958.jpeg)
+![](/IMG_4958.jpeg)
 
 This Tetris implementation consists of a hardware and a software (running on that hardware).
 
@@ -26,7 +26,7 @@ The buttons are debounced using RC circuitry and Schmitt triggers and connected
 
 The peripherial microcontrollers and the EEPROM are connected via SPI including individual chip select lines.
 
-![](./docs/game-ctrl.jpg)
+![](/game-ctrl.jpg)
 
 
 ## Play Ground Canvas
@@ -37,7 +37,7 @@ The play ground is implemented using a 10 * 20 matrix of PL9823 RGB LEDs which a
 
 The communcation with the game play controller is implemented as a sequences of tuples of LED address (0 to 211) and color code. A single octet of 253 where the LED address is expected is taken as the end-of-telegram mark. Readiness to receive a telegram is signaled to the game play controller via a single line connected to a GPIO of the game play controller.
 
-![](./docs/rgb-driver.jpg)
+![](/rgb-driver.jpg)
 
 [Details are here]({{< ref "rgb-driver.md" >}} "Details are here")
 
@@ -50,7 +50,7 @@ In the first place, a MAX7221 was meant to be used for connecting a multiple dig
 
 Communication with the game play controller is just a 16 bit number to be displayed.
 
-![](./docs/display-driver.jpg)
+![](/display-driver.jpg)
 
 
 ## Sound Effects
@@ -63,8 +63,8 @@ An amplifier following the proposal of the AY-3-8913 datasheet is implemented us
 
 The clock generator proposed by the AY-3-8913 does not work reliably, so an alternative design from "The Art of Electronics" has been used.
 
-![](./docs/sound-driver-1.jpg) 
-![](./docs/sound-driver-2.png) 
-![](./docs/sound-driver-3.jpg) 
-![](./docs/sound-driver-4.jpg) 
+![](/sound-driver-1.jpg) 
+![](/sound-driver-2.png) 
+![](/sound-driver-3.jpg) 
+![](/sound-driver-4.jpg)