Samtech-Instruments
diff --git a/‎README.md‎
Lines changed: 24 additions & 93 deletions b/‎README.md‎
Lines changed: 24 additions & 93 deletions
diff --git a/‎docs/hardware/esp14.md‎
Lines changed: 85 additions & 4 deletions b/‎docs/hardware/esp14.md‎
Lines changed: 85 additions & 4 deletions
@@ -23,7 +23,7 @@ couple of days.
 
 Find more in detail information about the supported boards, needed tools and
 the library APIs on the
-[project website](https://tenbaht.github.io/sduino/).
+[project documentation files](https://github.com/tenbaht/sduino/blob/master/docs/index.md).
 
 All you need to get stated is a simple STM8S103F breakout board and a
 ST-Link V2 compatible flash programmer. Three boards and one flash
@@ -103,7 +103,7 @@ improves significantly from version to version. Be sure to use
 version that might be included in your distribution. Version 3.5.0 as
 included with ubuntu 16.04 is definitly too old and compilation will fail
 due to some compiler errors.
-[More information on installing SDCC](sdcc.md)
+[More information on installing SDCC](blob/master/docs/sdcc.md)
 
 Support for the Cosmic compiler under Windows and integration into the ST
 visual developer IDE might be possible, but is not done (yet?).
@@ -112,13 +112,13 @@ visual developer IDE might be possible, but is not done (yet?).
 
 ## Supported hardware
 
-[STM8S103 breakout board](https://tenbaht.github.io/sduino/hardware/stm8blue.html):
+[STM8S103 breakout board](blob/master/docs/hardware/stm8blue.md):
 works  
 
-[ESP14 Wifi board](https://tenbaht.github.io/sduino/hardware/esp14.html):
+[ESP14 Wifi board](blob/master/docs/hardware/esp14.md):
 untested, but expected to work.
 
-[STM8S105 Discovery board](https://tenbaht.github.io/sduino/hardware/stm8disco.html):
+[STM8S105 Discovery board](blob/master/docs/hardware/stm8disco.md):
 not supported yet, but work in progress.
 
 
@@ -147,6 +147,8 @@ an option.
 
 
 
+## Current status and to-do list
+
 ## Current status and to-do list
 
 #### tested and working
@@ -157,16 +159,23 @@ an option.
 `analogWrite()`  
 `ShiftOut()`  
 WMath: `map()`  
-HardwareSerial  
+[HardwareSerial]  
 Print (without float)  
-SPI: working, no interrupt support  
-LiquidCrystal (for text LCD based on the HD44780 controller)  
-PCD8544 (for Nokia 5110 type displays)  
-[Stepper](Stepper.md) (multi-instance design for more than one stepper at a time)  
-ssd1306 (for monochrome OLED displays based on SSD1306 controller)  
+[SPI](https://github.com/tenbaht/sduino/blob/master/docs/SPI.md):
+  working, no interrupt support  
+[LiquidCrystal](https://github.com/tenbaht/sduino/blob/master/docs/api/LiquidCrystal.md):
+  Text LCD based on the HD44780 controller  
+[PCD8544](https://github.com/tenbaht/sduino/blob/master/docs/api/PCD8544.md):
+  Nokia 5110 type displays  
+[Mini_SSD1306](https://github.com/tenbaht/sduino/blob/master/docs/api/Mini_SSD1306.md):
+  Monochrome OLED displays based on the SSD1306 controller
+[Stepper](https://github.com/tenbaht/sduino/blob/master/docs/api/Stepper.md):
+  Multi-instance design for more than one stepper at a time  
+[Servo](https://github.com/tenbaht/sduino/blob/master/docs/api/Servo.md):
+  Multi-instance design for more than one servo at a time)  
 
 #### implemented and partly working
-Wire/I2C: no deadlock protection yet  
+Wire/I2C  
 
 #### tested, but not working
 `alternateFunctions()`  
@@ -186,6 +195,7 @@ module WCharacter
 module WString  
 
 
+
 The compile environment needs to detect which interrupts are actively used
 and link only the needed ones into the binary. See test/digitalWrite:
 Compiling with the straight Makefile.classic does not add UART interrupt
@@ -227,7 +237,7 @@ $1.60-$3.00 for an ATmega8.
 
 ## Further reading and application notes
 
-[sduino project website](https://tenbaht.github.io/sduino/):
+[project documentation files](https://github.com/tenbaht/sduino/blob/master/docs/index.md).
 More in detail information about supported boards, tools and the API.
 
 [PM0051](http://www.st.com/resource/en/programming_manual/cd00191343.pdf):
@@ -261,87 +271,8 @@ hard-coded assumptions about the number of pins with special functions.
 Ideally, all these numbers would be the same and all programs could be
 compiled without changes.
 
-[Here](https://tenbaht.github.io/sduino/pin_mapping.md)
+[Here](blob/master/docs/pin_mapping.md)
 I discuss some possible pin mapping strategies and check how close we could
 get the the ideal mapping. Unfortunatly, it turns out that a perfect mapping
 is not possible.
 
-In the end I chose a simple geometric numbering for the square UFQFPN20
-package starting with port pin PA1 and counting up from 0. This results in
-this mapping:
-
-sduino pin	| STM8S103 CPU port pin
-----------	| ---------------------
- 0-2		| PA1-PA3
- 3-4		| PB5-PB4 (reverse order)
- 5-9		| PC3-PC7
-10-15		| PD1-PD6
-
-serial: 14,15  
-SPI: 2,7,8,9  
-I2C: 3,4  
-Analog: 6,11,12,14,15  
-PWM: 2,5,6,12 plus either only 13 or 7-9 but not 13 (via alternate mapping)  
-
- + Easy and logical for use on a breadboard
- + Very clear and logical port pin ordering
- - Analog pins are still scattered around
- + TX and RX would be the rarely used analog pin numbers A3/A4 at
-   the end of the analog pin number list
- + At least the analog pins are in data sheet order
- - All functions use totally different pin numbers than Arduino
-
-I am still not really happy with this mapping. Instead of simplifing things
-it only adds another layer of abstraction and confusion. To avoid this I
-added definitions for the regular CPU pin names like `PA1` and `PD2`. In the
-end, this notation seem a lot easier to me. I am open for suggestions for a
-better pin number mapping.
-
-The chosen pin mapping for the STM8S103 (possible alternate function in
-paratheses):
-
-|Phys. STM8 pin|Name	|Functions	|Geometrical mapping|special funcion
-|---:	|---	|---			| ---	|---
-|1	|PD4	|UART_CLK/T2-1/beep	|13	|PWM
-|2	|PD5	|TX/Ain5		|14	|Analog A3
-|3	|PD6	|RX/Ain6		|15	|Analog A4
-|5	|PA1	|(OscIn, kein HS)	|0	|
-|6	|PA2	|(OscIn, kein HS)	|1	|
-|10	|PA3	|SS/T2-3		|2	|PWM
-|11	|PB5	|SDA	LED		|3	|
-|12	|PB4	|SCL			|4	|
-|13	|PC3	|T1-3/[T1-n1]		|5	|PWM, (n~)
-|14	|PC4	|T1-4/Ain2/[T1-n2]	|6	|PWM, Analog A0, (n~)
-|15	|PC5	|SCK/[T2-1]		|7	|(~)
-|16	|PC6	|MOSI/[T1-1]		|8	|(~)
-|17	|PC7	|MISO/[T1-2]		|9	|(~)
-|18	|PD1	|(SWIM)			|10	|
-|19	|PD2	|Ain3/[T2-3]		|11	|Analog A1, (~~)
-|20	|PD3	|Ain4/T2-2		|12	|PWM, Analog A2
-
-
-
-
-
-### Performance compared with the original Arduino environment
-
-Benchmarking the original Arduino examples from Arduino 1.0.5. The simple
-Blinky compiles to 57 bytes of code, the total binary including the sduino
-libraries is 1868 Bytes (0x74c).
-
-So far, wiring_analog depends on wiring_digital, even when analogWrite is not
-used. This could be solved by compiling the sduino functions separately into
-a library.
-
-|Name			|Code	|Total	|Linked files other than main and wiring
-|---------------------	|------	|-----	|-----------------------------
-|01. Basics/		|
-|BareMinimum		|2	|1238	|-
-|Blink			|57	|1870	|wiring_digital
-|AnalogReadSerial	|205	|3452	|digital, analog, serial, print
-|DigitalReadSerial	|57	|3160	|digital, serial, print
-|Fade			|226	|2189	|digital, analog
-|ReadAnalogVoltage	|	|	|float not yet implemented
-|02. Digital/		|
-|Debounce		|192	|2016	|digital
-
@@ -1,9 +1,90 @@
 # ESP-14
 
 An ESP-8266 Wifi-Controller together with a STM8S003 CPU for about $2 -
-incredible. I am still waiting for mine to be delivered, so nothing is
-tested yet. But the STM8S003 is very similar to the STM8S103 used on the
-[STM8S103 breakout boards](stm8blue.html) and I expect all software to run
-here as well.
+incredible.  Since the STM8S003 is almost identical to the STM8S103 used on
+the [STM8S103 breakout boards](stm8blue.html) all software runs here as well.
+
+
+Using the same pin numbering scheme as with the stm8blue boards it looks
+like this:
 
 ![ESP14 sduino pin numbering](esp14.png)
+
+
+
+## Differences between the '003 and the '103 CPUs
+
+Very slim. All differences that I could find so far:
+
+Feature		| STM8S003	| STM8S103
+------------	| -------------	| ------------
+EEPROM size	| 128 Bytes	| 640 Bytes
+
+I am not sure why the '003 is sold for half the price, or, to but it the
+other way around, why anybody should choose the '103 at all. Did I miss an
+important difference?
+
+
+
+## Using the Wifi connection
+
+The ESP-14 is basically an ESP-01 Wifi-module together with a STM8S breakout
+board combined into one PCB. The ESP part comes preprogrammed with the
+usual ESP AT-Firmware and is connected to the RX/TX pins of the STM8S.
+
+It feels like having a classic ATmega8-Arduino with a build-in ESP-01
+module. Using the [WiFiEsp library](https://github.com/bportaluri/WiFiEsp)
+we could have a programming experience very similar to using an Arduino with
+a Wifi-shield.
+
+Problem is the limited flash space of the '003. Even the very basic
+WebServer example that comes with the WiFiEsp library needs more then 14kB
+flash and more than 1kB RAM when compiled for an Arduino Uno. Even if the
+library could be carefully optimized to fit into the 8kB flash of an '003 -
+there would be no space left for the actual user code.
+
+It is still possible to do very simple task by just sending AT-strings to
+the ESP. But here we are definitly leaving Ardunio territory. Here we have
+to start from scratch and do it all by ourself. This is not going to be
+beginner-friendly anymore.
+
+The major remaining advantage of the STM8S over the ESP is power efficiency.
+The STM8S runs on almost nothing, while the ESP draws anything between 50
+and 200mA even for the simplest tasks.
+
+In most cases you might want to implement the external communication part of
+the application on the ESP, using the
+[ESP8266 Arduino port](https://github.com/esp8266/Arduino).
+
+The STM8S is better suited to local control and measurement and other
+energy- and time-critical tasks. Both CPUs use separate power supply inputs.
+Using an additional PMOS transistor the STM8S could control the power supply
+for the ESP and help extending battery life by activating the big brother
+only when its services are actually required.
+
+
+
+## Poor design choice
+
+Overall, the '003 is a very questionable CPU choice by AI Thinker. The '003
+is clearly one of the cheapest CPUs on the market, but the applications for
+this module are heavily restricted the the small flash memory.
+
+Why didn't they use an ATmega328? Just one dollar more and the whole Arduino
+universe opens up? Or at least an STM8S005K6T6C, offering 32kB flash and 2kB
+RAM for only 40 cents more, if you want to go really cheap.
+
+And why would you need an additional PMOS to control the ESP power supply?
+Why is this not included on the PCB already? This could the big killer
+feature of the current combination.
+
+
+
+## The upside
+
+Having a beefy 32 bit CPU and using it as an I/O-coprocessor while running
+the main application on an resticted 8-bitter is a weired concept. Things
+like the Arduino Yun or the Arduino WiFi-shield don't make too much sense.
+
+It is much more sensible to use the big CPU to control the small one than
+the other way around. Maybe this board can help to promote that concept.