запущен проект motor identification c терминалкой

This commit is contained in:
2026-06-05 12:15:36 +03:00
commit 177431f3d2
1383 changed files with 840275 additions and 0 deletions

95
docs/MD_DRIVER_MAP.md Normal file
View File

@@ -0,0 +1,95 @@
# MD / Motor Driver Map
## Status
No STM32G474CEU6 motor-driver files such as `md.c`, `motor_driver.c`, `gate_driver.c`, `drv.c`, `inverter.c`, `pwm.c`, or STM32 `tim.c/adc.c/gpio.c` were found in the current `AD` workspace.
The available motor-driver evidence is a C2000/TMS320 INU implementation under `Inu_im_1wnd_3lvl/Inu`.
A sibling STM32G474 candidate was found at `F:\set\workspace\setcorp\set506\git_project\stm32g474\stm32g474_IHM08M\IHM08M`, but it targets `STM32G474RETx`/NUCLEO-G474RE, not the requested `STM32G474CEU6`.
## Driver pins
| Function | STM32G474 status | Existing source evidence | Notes |
|----------|------------------|--------------------------|-------|
| Gate/PWM outputs | UNKNOWN/TODO | GPIO0..GPIO11 mapped to ePWM1A/B..ePWM6A/B in `Inu/init28335.c:177..188` | STM32 TIM channels must be extracted from `.ioc` when available |
| Driver enable | UNKNOWN/TODO | No explicit `EN_GATE`/`DRV_EN` STM32 pin found | Existing C2000 enables pulses by clearing Trip-Zone latch in `pwm()` |
| Driver fault | UNKNOWN/TODO | `DI_24V_SOURCE_FAULT` on C2000 GPIO50, active high | Real STM32 fault input must be documented from schematic/CubeMX |
| EEPROM/SPI CS | UNKNOWN/TODO | C2000 GPIO57 `CS_SET/CS_CLEAR`, SPIA on GPIO54..56 | Exact external device UNKNOWN |
| Status LEDs | UNKNOWN/TODO | GPIO59..61 active-low LEDs | Board-specific, not motor-driver critical |
## PWM generation
Existing C2000 implementation:
- ePWM1..ePWM6 are initialized in `SetupEpwm()` in `Inu/init28335.c`.
- ePWM period uses `T1_PRD` derived from `T_PWM`.
- PWM compare registers are written in `pwm()` in `Inu/upr.c:537..558`.
- Dead-time uses `DT` in `Inu/def.h`.
- Emergency PWM disable uses Trip-Zone force `TZFRC.all = 0x0004` in `stop_inu()` in `Inu/isr.c:300..328`.
STM32G474 TODO:
- Identify TIM1/TIM8 or other advanced timer instance.
- Identify complementary channels, dead-time, break input, BKIN polarity, off-state behavior.
- Verify that all PWM outputs remain disabled after reset and during ParamID default state.
Candidate STM32G474RETx project:
- TIM1 PWM is configured only for `TIM_CHANNEL_1` with output pin `PA7/TIM1_CH1N`.
- `MX_TIM1_Init()` has `DeadTime = 0`, `BreakState = TIM_BREAK_DISABLE`, `Break2State = TIM_BREAK2_DISABLE`, `AutomaticOutput = TIM_AUTOMATICOUTPUT_DISABLE`.
- Calls to `HAL_Init()`, `SystemClock_Config()`, `MX_GPIO_Init()`, and `MX_TIM1_Init()` are commented out in `Core/Src/main.c`, so the candidate firmware currently does not initialize peripherals at runtime.
## ADC/current sense
Existing C2000 implementation:
| Measurement | Existing source | Notes |
|-------------|-----------------|-------|
| DC-link voltage `udc1` | `AdcMirror.ADCRESULT0`, `Inu/isr.c:37` | Offset `offset.Udc1`, scale `GAIN_UDC` |
| Current `ic1` | `AdcMirror.ADCRESULT2`, `Inu/isr.c:38` | Offset `offset.Ic1`, scale `GAIN_IAC` |
| Current `ia1` | `AdcMirror.ADCRESULT4`, `Inu/isr.c:39` | Offset `offset.Ia1`, scale `GAIN_IAC` |
| Current `ib1` | `AdcMirror.ADCRESULT6`, `Inu/isr.c:40` | Offset `offset.Ib1`, scale `GAIN_IAC` |
STM32G474 TODO:
- Identify ADC instances and channels from `.ioc`.
- Identify injected/regular conversion, DMA buffers, trigger timer, sampling time.
- Map ADC raw values into `AD_Measurements_t`.
Candidate STM32G474RETx project:
- No ADC module is enabled in `stm32g4xx_hal_conf.h`.
- No ADC pins/channels are configured in `IHM08M.ioc`.
## Protection logic found
| Protection | Existing implementation | Fault |
|------------|-------------------------|-------|
| Overcurrent | `result.ia1/ib1/ic1` checked against `protect.IacMax/IacMin` in `Inu/isr.c:97..101` | `faultNo = 22` |
| DC-link undervoltage | `udc1Nf < protect.UdcMin` with debounce in `Inu/isr.c:104..115` | `faultNo = 30` |
| DC-link overvoltage | `udc1Nf > protect.UdcMax` in `Inu/isr.c:121..123` | `faultNo = 24` |
| Overspeed | `wm > protect.WmMax` with debounce in `Inu/isr.c:126..137` | `faultNo = 32` |
| +24 V supply fault | `DI_24V_SOURCE_FAULT == 1` with debounce in `Inu/isr.c:140..151` | `faultNo = 7` |
| PWM shutdown | `stop_inu()` forces Trip-Zone and clears compare registers | Safe stop path |
## Missing or unverified protections
- STM32G474 hardware break input: UNKNOWN.
- Gate-driver fault polarity and latch behavior: UNKNOWN.
- Overtemperature source: UNKNOWN.
- Emergency stop pin: UNKNOWN.
- DC-link discharge/ready signal: UNKNOWN.
- Hardware overcurrent comparator/OPAMP/COMP chain: UNKNOWN.
- Watchdog strategy for STM32 project: UNKNOWN.
## New safe software interlock
`Core/Src/ad_parameter_identification.c` adds:
- Compile-time default `AD_PARAM_ID_ENABLE_POWER_TESTS = 0`.
- Software enable gate via `AD_ParamID_SetSoftwareEnable`.
- Fault latch via `AD_PARAM_ID_STATUS_FAULT_LATCHED`.
- Emergency stop API `AD_ParamID_EmergencyStop`.
- Checks for measurement status flags and configured current/voltage/temperature/speed limits.
- Locked-rotor mode blocked unless `AD_ParamID_SetLockedRotorAllowed(1)` is called and power tests are compiled in.