changed all the internal links/anchors

keypad_to_microcontroller.md

@@ -1,10 +1,10 @@
 # Keypad Scanning
 
-There's a dozen ways to connect a keyboard to a microcontroller, and this paper attempts to enumerate many of them. [test link](#one-per-pin)
-There are similar articles on the internet [[see below](#ext_ref)], 
-but this paper introduces some new techniques: [row-grounded](#3x4-plus-ground-[new-design]), 
-[row-grounded-with-diodes](#3x4diodes), 
-[row-grounded-with-4-diodes](#3x4double_diodes)
+There's a dozen ways to connect a keyboard to a microcontroller, and this paper attempts to enumerate many of them.
+There are similar articles on the internet [[see below](#external-references)], 
+but this paper introduces some new techniques: [row-grounded](#3x4-plus-ground-new-design), 
+[row-grounded-with-diodes](#3x4-plus-ground-and-2-diodes-new-design), 
+[row-grounded-with-4-diodes](#3x4-plus-ground-and-4-diodes-new-design)
 
 Some articles focus on a full-function, unambiguous keyboard, 
 where multiple key-presses can be detected. Something
@@ -23,7 +23,7 @@ In this document, *n* refers to the number of keys that are in the keyboard. The
 that is being pressed during an explanation.
 
 
-## <a name="one_per_pin" /> One per Pin 
+## One per Pin 
 
 The simplest way to connect keys to a microcontroller is one key to one pin. 
 Connect the other side of the key to ground. 
@@ -40,7 +40,7 @@ This configuration has a few advantages:
 - there is no keyboard scan....which helps keep scan-noise out of the power supply; helpful in an audio environment
 - can easily be set to wake-on-interrupt, which means that you can power down between keystrokes
 
-See [KeyBounce](#bounce) below.
+See [KeyBounce](#keyboard-bounce) below.
 
 Of course, this solution does not apply to a keyboard which has *rows* and *columns*.
 
@@ -56,7 +56,6 @@ microcontrollers have an A/D converter, and that can be used to determine which
 
 A naive approach would be to simply use equal-valued resistors, and then each tap would be *k*/(*n*+1) of the supply voltage.
 
- <a name="3x4chain" />
 ### 3x4 Chain
 
 Sometimes a keyboard will be pre-wired as a grid, so you don't have access to the individual keys. You can still use 
@@ -109,7 +108,7 @@ See also [this article](https://github.com/sgmne/AnalogKeypad)
 
 ## Scanning
 
-### 3x4 Scanning <a name="3x4scan" />
+### 3x4 Scanning
 
 The simplest connection of a 3x4 keyboard is to simply use 7 GPIO pins, 4 for the rows, and 3 for the columns.
 Enable input & pullups on the columns and rows. At scan time, assert each row pin, one at a time, as a low, and measure
@@ -136,10 +135,10 @@ This scheme can be configured for interrupt triggering
 columns are set to input+pullup+interrupt. 
 Any keypress will then pull at least one column line low.  Once the 
 interrupt is triggered, change to the logic described above, scan the keyboard 
-[over the [debounce period](#bounce) and then restore they keyboard 
+[over the [debounce period](#keyboard-bounce) and then restore they keyboard 
 to the "ready-for-interrupt" configuration.
 
-### 3x4 plus Ground [new design] <a name="3x4gnd" />
+### 3x4 plus Ground [new design]
 
 If you are really tight on pins, you can ground one of the rows and still get a functioning keyboard.
 
@@ -161,7 +160,6 @@ The resulting 4 groups of 3 bits indicate the state of the keyboard; 1=open, 0=k
 
 Again, the same logic will allow you to manage a 4x4 keyboard with 7 GPIO pins.
 
-<a name="3x4diodes" />
 ### 3x4 plus Ground and 2 Diodes [new design]
 
 Again, we can scrape off one more GPIO if we can add a pair of diodes to the row pins.
@@ -215,7 +213,6 @@ look for the symbol **FIVE_PIN**.
 This configuration is also interrupt-capable, by setting the row drivers to output-low, and the columns as 
 input+pullup+interrupt.
 
-<a name="3x4double_diodes" />
 ### 3x4 plus Ground and 4 Diodes [new design]
 
 As an extreme, you can decode a 3x4 keypad with only 4 GPIO pins, 
@@ -268,7 +265,7 @@ can happen.
 look for the symbol **FOUR_PIN**. Here is a working prototype. ![working](keypad.jpg)
 
 
-### Keyboard Debounce <a name="bounce" />
+### Keyboard Debounce 
 
 When a key is pressed, there is a short time during closure when the 
 physical elements almost-touch, and the 
@@ -301,8 +298,12 @@ One involves using an SPI output and a shift register
 <https://hackaday.com/2015/04/15/simple-keypad-scanning-with-spi-and-some-hardware/>   
 This design requires an external chip ($0.10), 8 diodes and 4 resistors.
 
+### Interrupts
 
-### Reading DIP switches <a name="dip_switches" />
+
+
+
+### Reading DIP switches
 
 In some situations, it's helpful to read an array of DIP switches, 
 perhaps to read a configuration at power-on. In this case, 
@@ -325,7 +326,7 @@ and it's impossible to determine if the "5" key is pressed or not. Diodes are
 required at each junction to disambiguate.
 
 
-### External keyboard processor <a name="external_hardware" />
+### External keyboard processor
 
 With the current cost of processors, consider adding an external coprocessor to your project. 
 
@@ -364,7 +365,7 @@ It's U$0.11, 16k flash, 2k ram, 6gpio, sop8.
 | scan with diodes   | 6 for 4x4, 5 for 3x4     | 1 | 2 diodes | 
 | scan with 4 diodes | 5 for 4x4, 4 for 3x4     | 1 | 4 diodes | 
 
-### External References <a name="ext_ref" />
+### External References
 
 An excellent overview of keyboard scanning directly from a microcontroller.  
 <https://ww1.microchip.com/downloads/aemDocuments/documents/MCU08/ApplicationNotes/ApplicationNotes/00003407A.pdf>