commit 996abe6efb4e97ba0f9f26d51fad54466110a1bd
parent 6439bdb3df95a382896716e0de2cf35d631e8327
Author: Daniel Barlow <dan@telent.net>
Date: Sun, 16 Nov 2025 21:18:16 +0000
some notes
Diffstat:
| A | README.md | | | 4 | ++++ |
| A | notes.md | | | 178 | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ |
2 files changed, 182 insertions(+), 0 deletions(-)
diff --git a/README.md b/README.md
@@ -0,0 +1,4 @@
+# ECU-logical
+
+Read data from the K-line of a Honda motorbike ECU and publish it
+using Bluetooth LE
diff --git a/notes.md b/notes.md
@@ -0,0 +1,178 @@
+# Resources
+
+https://github.com/aeaphichart/HondaECU-1
+
+ - interface with optocouplers
+ - table of "keihin codes" (38770-foo model numbers)
+ - "If you make a bench harness to work with your ECU you will need a 2.5 amp power supply."
+ - fork https://github.com/Roxxas21/HondaEcuTest/tree/Roxxas21-patch-1
+
+https://github.com/sophienyaa/Honda-Motorcycle-ECU-Tools/blob/master/scan.ino
+
+ - for CRF250, scan all tables
+ - scanning the code it looks like it's mostly 11 and D1
+
+https://gonzos.net/projects/ctx-obd/
+
+ - Honda CTX-700
+ - explains checksum algo
+
+https://github.com/aster94/Keyword-Protocol-2000
+
+ - Suzuki and Kawasaki, untested on Hondas
+ - circuit diagram for L9637
+ - needs 510R resistor, 100n capacitor
+
+https://github.com/martinhol221/ISO14230-4-KWP
+
+ - Russian
+ - citroen c3
+ - https://github-com.translate.goog/martinhol221/ISO14230-4-KWP/blob/master/README.md?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en-US&_x_tr_pto=wapp
+
+https://rust-classes.com/chapter_embedded_stmf401
+
+ - Rust for embedded STM32F401CCU6 boards
+
+https://sigrok.org/wiki/Main_Page
+
+ - software for usb signal analyser
+
+https://netcult.ch/elmue/HUD%20ECU%20Hacker/
+
+ - more ecu hacking software, charityware
+ - describes honda init sequence
+
+https://www.liwebikes.com/pgm_fi-90.html
+ - cbr125rw ecm pinout
+ - this may be an earlier year
+
+https://www.manualslib.com/manual/1244336/Honda-Cbr125r.html?page=99#manual
+
+https://themotorbikeforum.co.uk/uploads/monthly_2018_05/7.jpg.e9632588ec1d17c0f74223b06cb81b0f.jpg
+https://themotorbikeforum.co.uk/topic/30235-cbr125-electrical-fault/
+
+ - photo of wiring diagram for CBR125R 2012, looks
+ to be the same ecu connection to DLC as whatever year is the
+ carlsalter.com source
+
+https://www.carlsalter.com/mcpdf/Honda%20-%20cbr125rw-service-manual.pdf
+ - wirig diagram, yay. !Caution: this is 2007 model!
+ pin 1 Bl/W IGP "ignition power" (also seen on DLC)
+ pin 15 SCS
+ pin 30 k-line
+ pin 9 and 10 ground, so maybe that helps us find which way the pins
+ are numbered
+
+service manual p 17-6 has a diagram of the ecu harness connector
+(female side) showing blue/yellow and white/yellow wires in the right
+hand side (CKP sensor) (pins 12, 23). this implies that pin 1 is top right
+when looking at the female side => top left looking at the ecu itself
+
+additionally, p5-66 labels pin 1 at top right of "wire side/female
+connector". It says pins 2, 9, 10 sohuld be ground
+
+as a side note, there is no permanent power to the ecu, only ignition
+power
+
+
+https://publicism.info/engineering/penetration/8.html
+
+ - build an ecu test bench
+
+https://github.com/AutotronicCommunity/Honda_Keihin_KLine_Protocol
+
+ - has a different init sequence which supposedly sends a response to
+ the first byte sequence
+
+ > Wake Up with send to ECU code = FE 04 72 8C
+ > ECU Response with code = 0E 04 72 7C
+
+ vs gonzo.net which says
+ > Send “Wakeup” message: FE 04 FF FF. No response is expected
+
+ (The second req/resp is the same)
+
+
+
+# Data available
+
+What BLE shall we publish?
+
+From https://gonzos.net/projects/wp-content/uploads/2015/09/Honda-data-tables.pdf
+
+Not all attributes are present for all bikes.
+
+RPM
+TPS volt/%
+ECT volt/deg C
+IAT volt/deg C
+MAP volt/kPa
+Batt volt
+Speed
+Fuel inj (?)
+Engine on/off
+[Neutral/clutch, in gear, kickstand]
+
+Should we have one service for all the things, or separate services
+for each? I assume that separate services make it easier to add more
+later without having to change clients
+
+
++ rpm
++ tps
+ - voltage
+ - percentage
++ ect
+ - voltage
+ - degrees
+
+... etc
+
+
+
+
+We could independently source bank angle and acceleration/deceleration
+if we use a board with an IMU, so consider adding them to the profile
+(or have two profiles).
+
+there don't appear to be services for this kind of thing
+some characteristics that exist:
+ - uuid: 0x2C09
+ name: Rotational Speed
+ id: org.bluetooth.characteristic.rotational_speed
+ (32 bits, signed => -2^31 < rpm < 2^31)
+ - uuid: 0x2A6E
+ name: Temperature
+ id: org.bluetooth.characteristic.temperature
+ (16 bits, 2dp => 0 <= t < 655.36)
+ - uuid: 0x2B18
+ name: Voltage
+ id: org.bluetooth.characteristic.voltage
+ (16 bits, V/64 => 0 < v < 1022
+ - uuid: 0x2A6D
+ name: Pressure
+ id: org.bluetooth.characteristic.pressure
+ - id: org.bluetooth.characteristic.percentage_8
+ (8 bits, 0.5% resolution)
+ - id: org.bluetooth.characteristic.boolean
+ 0 = false, 1 = true
+
+is there any characteristic here that
+
+
+specs of each characteristic are at
+https://bitbucket.org/bluetooth-SIG/public/src/main/gss/
+
+A common practice is to choose a base UUID for the custom service and then increment the 3rd and 4th Most Significant Bytes (MSB) within the UUID of each included characteristic
+
+
+
+# System architecture
+
+So far there is none. This is my first Rust program
+
+We'd like shared "scoreboard" of all the values which is mutated by
+the ecu-polling code and read by the BLE publishing code.
+
+It would be good to know whether the data was stale (have we lost ECU
+connection? some other problem?)
