velxio/frontend/src/simulation/AVRSimulator.ts

986 lines
36 KiB
TypeScript
Raw Normal View History

import {
CPU,
AVRTimer,
timer0Config,
timer1Config,
timer2Config,
AVRUSART,
usart0Config,
AVRIOPort,
portAConfig,
portBConfig,
portCConfig,
portDConfig,
portEConfig,
portFConfig,
portGConfig,
portHConfig,
portJConfig,
portKConfig,
portLConfig,
avrInstruction,
AVRADC,
adcConfig,
AVRSPI,
spiConfig,
AVRTWI,
twiConfig,
ATtinyTimer1,
attinyTimer1Config,
} from 'avr8js';
feat(attiny85+customchip): full ATtiny85 ADC/Timer0 + custom-chip pipeline fixes ATtiny85 (AVRSimulator + collectPinStates + connectAnalogInputsToMcu + SimulatorCanvas + Attiny85Element + examples): - Add attiny85AdcConfig with correct register addresses (ADMUX=0x27, ADCSRA=0x26, ADCSRB=0x23, ADCL=0x24, ADCH=0x25, DIDR0=0x34, adcInterrupt=0x08). Without this, analogRead() polled the wrong address forever and the firmware hung on first ADC read. - Add attiny85Timer0Config + instantiate AVRTimer so OVF fires at the ATTinyCore-expected ~1.024 ms cadence. delay() advance is still blocked on avr8js TIFR auto-clear semantics (separate upstream issue, see ATTINY85_TIMER0_UPSTREAM_ISSUE.md in velxio-prod test plan). - Map ATtiny85 ADC channels to PB-style pin names (PB5/PB2/PB4/PB3 -> 0..3) in connectAnalogInputsToMcu so SPICE node voltages reach the right ADC channel. - Recognise /^PB\d+$/ in collectPinStates.pinNameToArduinoPin so wires named "PB1" emit v_attiny85_pb1 V-source and the LED responds to MCU writes. Previously every PB-wire returned -1 and SPICE saw no source. - SimulatorCanvas: subscribe pin 1 (PB1) for the built-in LED on the attiny85 board kind (Digispark convention), instead of falling through to the pin-13 default. - Attiny85Element: remove the hand-drawn "yellow LED" circle that was floating above the chip. The bare DIP-8 has no on-board LED; examples wire a real wokwi-led + resistor instead. - examples.ts: add a real wokwi-led + 220 Ohm wokwi-resistor + wires to attiny85-blink, and add missing series resistors to attiny85-button-led and attiny85-ntc-sensor. attiny85-pwm-fade was already correct. Custom-chip pipeline (CustomChipPart + simulatorBridges): - Add a requestAnimationFrame loop that calls instance.tickTimers() every frame in CustomChipPart. Chips that register vx_timer_create (e.g. an i8080 stepping its core, or a sensor publishing samples) had timers added to the queue but nothing fired them; tickTimers was dead code. - Gate the ESP32 backend path with detectSimulatorKind(sim)==='esp32'. The previous `typeof sim.registerSensor === 'function'` check matched AVR and RP2040 simulators too (they expose registerSensor for I2C sensor proxies), routing client-side chips to a non-existent ESP32 worker on those boards. - Replace direct simulator.usart.writeByte calls in avrUartTx with a JS-level FIFO + setTimeout(1ms) drainer. avr8js writeByte drops bytes under burst load (a chip emitting print_string lost ~99% of bytes via non-immediate, or kept only the last byte via immediate). The drainer attempts one non-immediate write per tick and retries on RXC busy / RXEN off. Added a guard for ATtiny85 (no USART -> would queue forever). End-to-end verified: i8080-banner-streamer now prints the boot banner followed by "uptime ticks: 0xNN" lines stepping every ~50 ms, executing real Intel 8080 instructions inside the WASM chip. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-20 04:10:08 +07:00
import type { AVRTimerConfig } from 'avr8js/dist/esm/peripherals/timer';
import type { ADCConfig, ADCMuxConfiguration } from 'avr8js/dist/esm/peripherals/adc';
import { ADCMuxInputType, ADCReference } from 'avr8js/dist/esm/peripherals/adc';
import { PinManager } from './PinManager';
import { hexToUint8Array } from '../utils/hexParser';
import { I2CBusManager, nullI2CMaster } from './I2CBusManager';
import type { I2CDevice } from './I2CBusManager';
/**
* AVRSimulator - Emulates Arduino Uno (ATmega328p) using avr8js
*
* Features:
* - CPU emulation at 16MHz
* - Timer0/Timer1/Timer2 support (enables millis(), delay(), PWM)
* - USART support (Serial)
* - GPIO ports (PORTB, PORTC, PORTD)
* - ADC support (analogRead())
* - PWM monitoring via OCR register polling
* - Pin state tracking via PinManager
*/
// OCR register addresses → Arduino pin mapping for PWM (ATmega328P / Uno / Nano)
const PWM_PINS_UNO = [
{ ocrAddr: 0x47, pin: 6, label: 'OCR0A' }, // Timer0A → D6
{ ocrAddr: 0x48, pin: 5, label: 'OCR0B' }, // Timer0B → D5
{ ocrAddr: 0x88, pin: 9, label: 'OCR1AL' }, // Timer1A low byte → D9
{ ocrAddr: 0x8a, pin: 10, label: 'OCR1BL' }, // Timer1B low byte → D10
{ ocrAddr: 0xb3, pin: 11, label: 'OCR2A' }, // Timer2A → D11
{ ocrAddr: 0xb4, pin: 3, label: 'OCR2B' }, // Timer2B → D3
];
// OCR register addresses → Arduino Mega pin mapping for PWM (ATmega2560)
// Timers 0/1/2 same addresses; Timers 3/4/5 at higher addresses.
const PWM_PINS_MEGA = [
{ ocrAddr: 0x47, pin: 13, label: 'OCR0A' }, // Timer0A → D13
{ ocrAddr: 0x48, pin: 4, label: 'OCR0B' }, // Timer0B → D4
{ ocrAddr: 0x88, pin: 11, label: 'OCR1AL' }, // Timer1A → D11
{ ocrAddr: 0x8a, pin: 12, label: 'OCR1BL' }, // Timer1B → D12
{ ocrAddr: 0xb3, pin: 10, label: 'OCR2A' }, // Timer2A → D10
{ ocrAddr: 0xb4, pin: 9, label: 'OCR2B' }, // Timer2B → D9
// Timer3 (0x800x8D, but OCR3A/B/C at 0x98/0x9A/0x9C)
{ ocrAddr: 0x98, pin: 5, label: 'OCR3AL' }, // Timer3A → D5
{ ocrAddr: 0x9a, pin: 2, label: 'OCR3BL' }, // Timer3B → D2
{ ocrAddr: 0x9c, pin: 3, label: 'OCR3CL' }, // Timer3C → D3
// Timer4 (OCR4A/B/C at 0xA8/0xAA/0xAC)
{ ocrAddr: 0xa8, pin: 6, label: 'OCR4AL' }, // Timer4A → D6
{ ocrAddr: 0xaa, pin: 7, label: 'OCR4BL' }, // Timer4B → D7
{ ocrAddr: 0xac, pin: 8, label: 'OCR4CL' }, // Timer4C → D8
// Timer5 (OCR5A/B/C at 0x128/0x12A/0x12C — extended I/O)
{ ocrAddr: 0x128, pin: 46, label: 'OCR5AL' }, // Timer5A → D46
{ ocrAddr: 0x12a, pin: 45, label: 'OCR5BL' }, // Timer5B → D45
{ ocrAddr: 0x12c, pin: 44, label: 'OCR5CL' }, // Timer5C → D44
];
/**
* ATmega2560 port-bit Arduino Mega pin mapping.
* Index = bit position (07). -1 = not exposed on the Arduino Mega header.
*/
const MEGA_PORT_BIT_MAP: Record<string, number[]> = {
// PA0-PA7 → D22-D29
PORTA: [22, 23, 24, 25, 26, 27, 28, 29],
// PB0=D53(SS), PB1=D52(SCK), PB2=D51(MOSI), PB3=D50(MISO), PB4-PB7=D10-D13
PORTB: [53, 52, 51, 50, 10, 11, 12, 13],
// PC0-PC7 → D37, D36, D35, D34, D33, D32, D31, D30 (reversed)
PORTC: [37, 36, 35, 34, 33, 32, 31, 30],
// PD0=D21(SCL), PD1=D20(SDA), PD2=D19(RX1), PD3=D18(TX1), PD7=D38
PORTD: [21, 20, 19, 18, -1, -1, -1, 38],
// PE0=D0(RX0), PE1=D1(TX0), PE3=D5, PE4=D2, PE5=D3
PORTE: [0, 1, -1, 5, 2, 3, -1, -1],
// PF0-PF7 → A0-A7 (pin numbers 54-61)
PORTF: [54, 55, 56, 57, 58, 59, 60, 61],
// PG0=D41, PG1=D40, PG2=D39, PG5=D4
PORTG: [41, 40, 39, -1, -1, 4, -1, -1],
// PH0=D17(RX2), PH1=D16(TX2), PH3=D6, PH4=D7, PH5=D8, PH6=D9
PORTH: [17, 16, -1, 6, 7, 8, 9, -1],
// PJ0=D15(RX3), PJ1=D14(TX3)
PORTJ: [15, 14, -1, -1, -1, -1, -1, -1],
// PK0-PK7 → A8-A15 (pin numbers 62-69)
PORTK: [62, 63, 64, 65, 66, 67, 68, 69],
// PL0=D49, PL1=D48, PL2=D47, PL3=D46, PL4=D45, PL5=D44, PL6=D43, PL7=D42
PORTL: [49, 48, 47, 46, 45, 44, 43, 42],
};
/**
* Reverse of MEGA_PORT_BIT_MAP: Arduino Mega pin { portName, bit }.
* Pre-built for fast setPinState() lookups.
*/
const MEGA_PIN_TO_PORT = (() => {
const map: Record<number, { portName: string; bit: number; port?: AVRIOPort }> = {};
for (const [portName, pins] of Object.entries(MEGA_PORT_BIT_MAP)) {
pins.forEach((pin, bit) => {
if (pin >= 0) map[pin] = { portName, bit };
});
}
return map;
})();
// OCR register addresses → ATtiny85 pin mapping for PWM
// Timer0: OC0A→PB0, OC0B→PB1 (ATtiny85 Timer0 OCR regs at 0x56, 0x5C)
// Timer1: OC1A→PB1, OC1B→PB4 (ATtinyTimer1 OCR regs from attinyTimer1Config)
const PWM_PINS_TINY85 = [
{ ocrAddr: 0x56, pin: 0, label: 'OCR0A' }, // Timer0A → PB0
{ ocrAddr: 0x5c, pin: 1, label: 'OCR0B' }, // Timer0B → PB1
{ ocrAddr: 0x4e, pin: 1, label: 'OCR1A' }, // Timer1A → PB1 (attinyTimer1Config.OCR1A)
{ ocrAddr: 0x4b, pin: 4, label: 'OCR1B' }, // Timer1B → PB4 (attinyTimer1Config.OCR1B)
];
/**
* ATtiny85 PORTB config registers are at different addresses than ATmega328P.
* ATtiny85: PINB=0x36, DDRB=0x37, PORTB=0x38 (vs ATmega: 0x23/0x24/0x25)
*/
const attiny85PortBConfig = {
PIN: 0x36,
DDR: 0x37,
PORT: 0x38,
externalInterrupts: [] as never[],
};
feat(attiny85+customchip): full ATtiny85 ADC/Timer0 + custom-chip pipeline fixes ATtiny85 (AVRSimulator + collectPinStates + connectAnalogInputsToMcu + SimulatorCanvas + Attiny85Element + examples): - Add attiny85AdcConfig with correct register addresses (ADMUX=0x27, ADCSRA=0x26, ADCSRB=0x23, ADCL=0x24, ADCH=0x25, DIDR0=0x34, adcInterrupt=0x08). Without this, analogRead() polled the wrong address forever and the firmware hung on first ADC read. - Add attiny85Timer0Config + instantiate AVRTimer so OVF fires at the ATTinyCore-expected ~1.024 ms cadence. delay() advance is still blocked on avr8js TIFR auto-clear semantics (separate upstream issue, see ATTINY85_TIMER0_UPSTREAM_ISSUE.md in velxio-prod test plan). - Map ATtiny85 ADC channels to PB-style pin names (PB5/PB2/PB4/PB3 -> 0..3) in connectAnalogInputsToMcu so SPICE node voltages reach the right ADC channel. - Recognise /^PB\d+$/ in collectPinStates.pinNameToArduinoPin so wires named "PB1" emit v_attiny85_pb1 V-source and the LED responds to MCU writes. Previously every PB-wire returned -1 and SPICE saw no source. - SimulatorCanvas: subscribe pin 1 (PB1) for the built-in LED on the attiny85 board kind (Digispark convention), instead of falling through to the pin-13 default. - Attiny85Element: remove the hand-drawn "yellow LED" circle that was floating above the chip. The bare DIP-8 has no on-board LED; examples wire a real wokwi-led + resistor instead. - examples.ts: add a real wokwi-led + 220 Ohm wokwi-resistor + wires to attiny85-blink, and add missing series resistors to attiny85-button-led and attiny85-ntc-sensor. attiny85-pwm-fade was already correct. Custom-chip pipeline (CustomChipPart + simulatorBridges): - Add a requestAnimationFrame loop that calls instance.tickTimers() every frame in CustomChipPart. Chips that register vx_timer_create (e.g. an i8080 stepping its core, or a sensor publishing samples) had timers added to the queue but nothing fired them; tickTimers was dead code. - Gate the ESP32 backend path with detectSimulatorKind(sim)==='esp32'. The previous `typeof sim.registerSensor === 'function'` check matched AVR and RP2040 simulators too (they expose registerSensor for I2C sensor proxies), routing client-side chips to a non-existent ESP32 worker on those boards. - Replace direct simulator.usart.writeByte calls in avrUartTx with a JS-level FIFO + setTimeout(1ms) drainer. avr8js writeByte drops bytes under burst load (a chip emitting print_string lost ~99% of bytes via non-immediate, or kept only the last byte via immediate). The drainer attempts one non-immediate write per tick and retries on RXC busy / RXEN off. Added a guard for ATtiny85 (no USART -> would queue forever). End-to-end verified: i8080-banner-streamer now prints the boot banner followed by "uptime ticks: 0xNN" lines stepping every ~50 ms, executing real Intel 8080 instructions inside the WASM chip. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-20 04:10:08 +07:00
/**
* ATtiny85 Timer0 config Arduino `millis()` / `delay()` rely on the
* TIMER0_OVF interrupt to tick the millisecond counter. avr8js's generic
* `AVRTimer` is fully data-driven, so we just supply ATtiny85's register
* addresses (different from the ATmega328P defaults in `timer0Config`)
* and the right interrupt vector offsets.
*
* Refs: <avr/iotnx5.h> for register addresses; ATtiny25/45/85 datasheet
* (Atmel-2586) for vector indices.
* _VECTOR(5) TIMER0_OVF word 0x0A
* _VECTOR(10) TIMER0_COMPA word 0x14
* _VECTOR(11) TIMER0_COMPB word 0x16
*/
/**
* ATtiny85 ADC config required because the chip's ADC registers live at
* completely different memory addresses than the ATmega328P defaults that
* avr8js's `adcConfig` ships with. Without this, `analogRead()` writes
* ADSC at ATtiny85's ADCSRA (0x26) and polls forever because avr8js is
* listening at 0x7A instead.
*
* Refs: <avr/iotnx5.h>; ATtiny25/45/85 datasheet (Atmel-2586) sec. 17.
* ADMUX = 0x07 (I/O) -> 0x27 (mem)
* ADCSRA = 0x06 -> 0x26
* ADCSRB = 0x03 -> 0x23
* ADCL = 0x04 -> 0x24
* ADCH = 0x05 -> 0x25
* DIDR0 = 0x14 -> 0x34
* ADC_vect = _VECTOR(8) -> word 0x10
*
* MUX field is 4 bits (bits 3:0). Single-ended channels 0..3 = PB5/PB2/PB4/PB3.
* Reference bits REFS1:REFS0 at ADMUX[7:6] select VCC/AREF/Internal1V1 by default;
* full REFS2 extension lives at ADMUX[4] but the avr8js helper checks bit 3,
* so the rare 2.56 V internal reference is currently unsupported every
* default-ref sketch (`analogReference(DEFAULT)`) works fine.
*/
const attiny85AdcChannels: ADCMuxConfiguration = {
0: { type: ADCMuxInputType.SingleEnded, channel: 0 }, // PB5
1: { type: ADCMuxInputType.SingleEnded, channel: 1 }, // PB2
2: { type: ADCMuxInputType.SingleEnded, channel: 2 }, // PB4
3: { type: ADCMuxInputType.SingleEnded, channel: 3 }, // PB3
12: { type: ADCMuxInputType.Constant, voltage: 1.1 }, // VBG
13: { type: ADCMuxInputType.Constant, voltage: 0 }, // GND
15: { type: ADCMuxInputType.Temperature },
};
const attiny85AdcConfig: ADCConfig = {
ADMUX: 0x27,
ADCSRA: 0x26,
ADCSRB: 0x23,
ADCL: 0x24,
ADCH: 0x25,
DIDR0: 0x34,
// ATtiny85 vectors are 1-word RJMP (vs ATmega328P's 2-word JMP) so the
// avr8js "address" field is the raw vector index, not vector*2.
adcInterrupt: 0x08, // _VECTOR(8) ADC_vect
numChannels: 4,
muxInputMask: 0xf,
muxChannels: attiny85AdcChannels,
adcReferences: [
ADCReference.AVCC, // 00 = VCC
ADCReference.AREF, // 01 = external AREF (PB0)
ADCReference.Internal1V1, // 10 = internal 1.1 V
ADCReference.Reserved, // 11 = reserved
],
};
const attiny85Timer0Config: AVRTimerConfig = {
bits: 8,
captureInterrupt: 0,
// ATtiny85 vectors are 1-word RJMP (vs ATmega328P's 2-word JMP) so the
// avr8js "address" field is the raw vector index, not vector*2.
compAInterrupt: 0x0a, // _VECTOR(10) TIMER0_COMPA_vect
compBInterrupt: 0x0b, // _VECTOR(11) TIMER0_COMPB_vect
compCInterrupt: 0,
ovfInterrupt: 0x05, // _VECTOR(5) TIMER0_OVF_vect
TIFR: 0x58,
OCRA: 0x56,
OCRB: 0x5c,
OCRC: 0,
ICR: 0,
TCNT: 0x52,
TCCRA: 0x4f,
TCCRB: 0x53,
TCCRC: 0,
TIMSK: 0x59,
TOV: 0b00000010,
OCFA: 0b00010000,
OCFB: 0b00001000,
OCFC: 0,
TOIE: 0b00000010,
OCIEA: 0b00010000,
OCIEB: 0b00001000,
OCIEC: 0,
compPortA: 0x38,
compPinA: 0,
compPortB: 0x38,
compPinB: 1,
compPortC: 0,
compPinC: 0,
externalClockPort: 0x36,
externalClockPin: 2,
dividers: { 0: 0, 1: 1, 2: 8, 3: 64, 4: 256, 5: 1024, 6: 0, 7: 0 },
};
/** Ordered list of Mega ports with their avr8js configs */
const MEGA_PORT_CONFIGS = [
{ name: 'PORTA', config: portAConfig },
{ name: 'PORTB', config: portBConfig },
{ name: 'PORTC', config: portCConfig },
{ name: 'PORTD', config: portDConfig },
{ name: 'PORTE', config: portEConfig },
{ name: 'PORTF', config: portFConfig },
{ name: 'PORTG', config: portGConfig },
{ name: 'PORTH', config: portHConfig },
{ name: 'PORTJ', config: portJConfig },
{ name: 'PORTK', config: portKConfig },
{ name: 'PORTL', config: portLConfig },
];
export class AVRSimulator {
private cpu: CPU | null = null;
/** Peripherals kept alive by reference so GC doesn't collect their CPU hooks */
private peripherals: unknown[] = [];
private portB: AVRIOPort | null = null;
private portC: AVRIOPort | null = null;
private portD: AVRIOPort | null = null;
/** Extra ports used by the Mega (A, EL); keyed by port name */
private megaPorts: Map<string, AVRIOPort> = new Map();
private megaPortValues: Map<string, number> = new Map();
private adc: AVRADC | null = null;
public spi: AVRSPI | null = null;
public usart: AVRUSART | null = null;
public twi: AVRTWI | null = null;
public i2cBus!: I2CBusManager;
private program: Uint16Array | null = null;
private running = false;
private animationFrame: number | null = null;
public pinManager: PinManager;
private speed = 1.0;
/** 'uno' for ATmega328P boards (Uno, Nano); 'mega' for ATmega2560; 'tiny85' for ATtiny85 */
private boardVariant: 'uno' | 'mega' | 'tiny85';
/** Cycle-accurate pin change queue — used by timing-sensitive peripherals (e.g. DHT22). */
private scheduledPinChanges: Array<{ cycle: number; pin: number; state: boolean }> = [];
/** Serial output buffer — subscribers receive each byte or line */
public onSerialData: ((char: string) => void) | null = null;
/** Fires whenever the sketch changes Serial baud rate (Serial.begin) */
public onBaudRateChange: ((baudRate: number) => void) | null = null;
/**
* Fires for every digital pin transition with a millisecond timestamp
* derived from the CPU cycle counter (cycles / CPU_HZ * 1000).
* Used by the oscilloscope / logic analyzer.
*/
public onPinChangeWithTime: ((pin: number, state: boolean, timeMs: number) => void) | null = null;
private lastPortBValue = 0;
private lastPortCValue = 0;
private lastPortDValue = 0;
private lastOcrValues: number[] = [];
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
/**
* Last known TXEN bit value, used to detect 01 transitions and seed the
* TX pin baseline at idle HIGH the moment the firmware enables the USART.
* Without this seed the oscilloscope shows a floating/LOW baseline until
* the first byte transmits, which doesn't match real hardware.
*/
private lastTxEnable = false;
constructor(pinManager: PinManager, boardVariant: 'uno' | 'mega' | 'tiny85' = 'uno') {
this.pinManager = pinManager;
this.boardVariant = boardVariant;
// Create the bus up-front with a placeholder master so that
// Interconnect can install cross-board bridges and parts can
// register devices BEFORE the firmware loads. The real AVRTWI
// takes over via `i2cBus.attachMaster(twi)` inside loadHex.
this.i2cBus = new I2CBusManager(nullI2CMaster());
}
private get pwmPins() {
if (this.boardVariant === 'mega') return PWM_PINS_MEGA;
if (this.boardVariant === 'tiny85') return PWM_PINS_TINY85;
return PWM_PINS_UNO;
}
/**
* Load compiled hex file into simulator
*/
loadHex(hexContent: string): void {
console.log('Loading HEX file...');
const bytes = hexToUint8Array(hexContent);
// ATmega328P: 32 KB = 16 384 words. ATmega2560: 256 KB = 131 072 words.
// ATtiny85: 8 KB = 4 096 words, 512 bytes SRAM.
const progWords =
this.boardVariant === 'mega' ? 131072 : this.boardVariant === 'tiny85' ? 4096 : 16384;
// ATmega2560 data space: 0x00000x21FF = 8704 bytes total.
// avr8js: data.length = sramBytes + registerSpace (0x100 = 256).
// So sramBytes must be >= 8704 256 = 8448 to fit RAMEND=0x21FF on the stack.
// ATmega328P RAMEND = 0x08FF; default 8192 is already a safe over-alloc.
// ATtiny85 RAMEND = 0x025F; 512 bytes SRAM.
const sramBytes =
this.boardVariant === 'mega' ? 8448 : this.boardVariant === 'tiny85' ? 512 : 8192;
this.program = new Uint16Array(progWords);
for (let i = 0; i < bytes.length; i += 2) {
this.program[i >> 1] = (bytes[i] || 0) | ((bytes[i + 1] || 0) << 8);
}
console.log(`Loaded ${bytes.length} bytes into program memory`);
this.cpu = new CPU(this.program, sramBytes);
if (this.boardVariant === 'tiny85') {
feat(attiny85+customchip): full ATtiny85 ADC/Timer0 + custom-chip pipeline fixes ATtiny85 (AVRSimulator + collectPinStates + connectAnalogInputsToMcu + SimulatorCanvas + Attiny85Element + examples): - Add attiny85AdcConfig with correct register addresses (ADMUX=0x27, ADCSRA=0x26, ADCSRB=0x23, ADCL=0x24, ADCH=0x25, DIDR0=0x34, adcInterrupt=0x08). Without this, analogRead() polled the wrong address forever and the firmware hung on first ADC read. - Add attiny85Timer0Config + instantiate AVRTimer so OVF fires at the ATTinyCore-expected ~1.024 ms cadence. delay() advance is still blocked on avr8js TIFR auto-clear semantics (separate upstream issue, see ATTINY85_TIMER0_UPSTREAM_ISSUE.md in velxio-prod test plan). - Map ATtiny85 ADC channels to PB-style pin names (PB5/PB2/PB4/PB3 -> 0..3) in connectAnalogInputsToMcu so SPICE node voltages reach the right ADC channel. - Recognise /^PB\d+$/ in collectPinStates.pinNameToArduinoPin so wires named "PB1" emit v_attiny85_pb1 V-source and the LED responds to MCU writes. Previously every PB-wire returned -1 and SPICE saw no source. - SimulatorCanvas: subscribe pin 1 (PB1) for the built-in LED on the attiny85 board kind (Digispark convention), instead of falling through to the pin-13 default. - Attiny85Element: remove the hand-drawn "yellow LED" circle that was floating above the chip. The bare DIP-8 has no on-board LED; examples wire a real wokwi-led + resistor instead. - examples.ts: add a real wokwi-led + 220 Ohm wokwi-resistor + wires to attiny85-blink, and add missing series resistors to attiny85-button-led and attiny85-ntc-sensor. attiny85-pwm-fade was already correct. Custom-chip pipeline (CustomChipPart + simulatorBridges): - Add a requestAnimationFrame loop that calls instance.tickTimers() every frame in CustomChipPart. Chips that register vx_timer_create (e.g. an i8080 stepping its core, or a sensor publishing samples) had timers added to the queue but nothing fired them; tickTimers was dead code. - Gate the ESP32 backend path with detectSimulatorKind(sim)==='esp32'. The previous `typeof sim.registerSensor === 'function'` check matched AVR and RP2040 simulators too (they expose registerSensor for I2C sensor proxies), routing client-side chips to a non-existent ESP32 worker on those boards. - Replace direct simulator.usart.writeByte calls in avrUartTx with a JS-level FIFO + setTimeout(1ms) drainer. avr8js writeByte drops bytes under burst load (a chip emitting print_string lost ~99% of bytes via non-immediate, or kept only the last byte via immediate). The drainer attempts one non-immediate write per tick and retries on RXC busy / RXEN off. Added a guard for ATtiny85 (no USART -> would queue forever). End-to-end verified: i8080-banner-streamer now prints the boot banner followed by "uptime ticks: 0xNN" lines stepping every ~50 ms, executing real Intel 8080 instructions inside the WASM chip. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-20 04:10:08 +07:00
// ATtiny85: PORTB only (PB0-PB5). Timer0 powers millis()/delay() in
// ATTinyCore via TIMER0_OVF. Timer1 is the high-speed 8-bit PWM
// timer (PLL clock). No hardware USART on this chip.
//
// Known limitation (task #116): the Timer0 OVF interrupt does fire at
// the correct cadence (1.024 ms simulated, verified via debug
// instrumentation), but real ATTinyCore-compiled `delay()` does not
// observably advance — the LED stays stuck either HIGH or LOW
// depending on which phase the firmware was in when the first OVF
// hit. Likely a subtle interaction between the avr8js clearInterrupt
// semantics (only clears the pending queue entry, leaves TIFR bit
// set) and ATTinyCore's ISR relying on hardware auto-clear of TOV0.
// Workaround attempts (manual TIFR clear after ISR entry) did not
// change the visible behavior. Needs a deeper avr8js dive.
this.portB = new AVRIOPort(this.cpu, attiny85PortBConfig as typeof portBConfig);
feat(attiny85+customchip): full ATtiny85 ADC/Timer0 + custom-chip pipeline fixes ATtiny85 (AVRSimulator + collectPinStates + connectAnalogInputsToMcu + SimulatorCanvas + Attiny85Element + examples): - Add attiny85AdcConfig with correct register addresses (ADMUX=0x27, ADCSRA=0x26, ADCSRB=0x23, ADCL=0x24, ADCH=0x25, DIDR0=0x34, adcInterrupt=0x08). Without this, analogRead() polled the wrong address forever and the firmware hung on first ADC read. - Add attiny85Timer0Config + instantiate AVRTimer so OVF fires at the ATTinyCore-expected ~1.024 ms cadence. delay() advance is still blocked on avr8js TIFR auto-clear semantics (separate upstream issue, see ATTINY85_TIMER0_UPSTREAM_ISSUE.md in velxio-prod test plan). - Map ATtiny85 ADC channels to PB-style pin names (PB5/PB2/PB4/PB3 -> 0..3) in connectAnalogInputsToMcu so SPICE node voltages reach the right ADC channel. - Recognise /^PB\d+$/ in collectPinStates.pinNameToArduinoPin so wires named "PB1" emit v_attiny85_pb1 V-source and the LED responds to MCU writes. Previously every PB-wire returned -1 and SPICE saw no source. - SimulatorCanvas: subscribe pin 1 (PB1) for the built-in LED on the attiny85 board kind (Digispark convention), instead of falling through to the pin-13 default. - Attiny85Element: remove the hand-drawn "yellow LED" circle that was floating above the chip. The bare DIP-8 has no on-board LED; examples wire a real wokwi-led + resistor instead. - examples.ts: add a real wokwi-led + 220 Ohm wokwi-resistor + wires to attiny85-blink, and add missing series resistors to attiny85-button-led and attiny85-ntc-sensor. attiny85-pwm-fade was already correct. Custom-chip pipeline (CustomChipPart + simulatorBridges): - Add a requestAnimationFrame loop that calls instance.tickTimers() every frame in CustomChipPart. Chips that register vx_timer_create (e.g. an i8080 stepping its core, or a sensor publishing samples) had timers added to the queue but nothing fired them; tickTimers was dead code. - Gate the ESP32 backend path with detectSimulatorKind(sim)==='esp32'. The previous `typeof sim.registerSensor === 'function'` check matched AVR and RP2040 simulators too (they expose registerSensor for I2C sensor proxies), routing client-side chips to a non-existent ESP32 worker on those boards. - Replace direct simulator.usart.writeByte calls in avrUartTx with a JS-level FIFO + setTimeout(1ms) drainer. avr8js writeByte drops bytes under burst load (a chip emitting print_string lost ~99% of bytes via non-immediate, or kept only the last byte via immediate). The drainer attempts one non-immediate write per tick and retries on RXC busy / RXEN off. Added a guard for ATtiny85 (no USART -> would queue forever). End-to-end verified: i8080-banner-streamer now prints the boot banner followed by "uptime ticks: 0xNN" lines stepping every ~50 ms, executing real Intel 8080 instructions inside the WASM chip. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-20 04:10:08 +07:00
this.adc = new AVRADC(this.cpu, attiny85AdcConfig);
this.peripherals = [
new AVRTimer(this.cpu, attiny85Timer0Config),
new ATtinyTimer1(this.cpu, attinyTimer1Config),
];
// usart stays null — ATtiny85 has no hardware USART
} else {
// ATmega2560 has more vectors before the timers/USART (8 external INTs, etc.),
// so the interrupt WORD addresses differ from ATmega328P.
//
// avr8js config values are WORD addresses = _VECTOR(N) * 2
// (each JMP vector = 4 bytes = 2 words; cpu.pc * 2 == byte address).
//
// ATmega2560 word addresses (_VECTOR(N) → N * 2):
// TIMER2_COMPA=_V(13)→0x1A TIMER2_COMPB=_V(14)→0x1C TIMER2_OVF=_V(15)→0x1E
// TIMER1_CAPT=_V(16)→0x20 TIMER1_COMPA=_V(17)→0x22 TIMER1_COMPB=_V(18)→0x24
// TIMER1_COMPC=_V(19)→0x26 TIMER1_OVF=_V(20)→0x28
// TIMER0_COMPA=_V(21)→0x2A TIMER0_COMPB=_V(22)→0x2C TIMER0_OVF=_V(23)→0x2E
// SPI_STC=_V(24)→0x30 USART0_RX=_V(25)→0x32
// USART0_UDRE=_V(26)→0x34 USART0_TX=_V(27)→0x36
// TWI=_V(39)→0x4E
const isMega = this.boardVariant === 'mega';
const activeTimer0Config = isMega
? { ...timer0Config, compAInterrupt: 0x2a, compBInterrupt: 0x2c, ovfInterrupt: 0x2e }
: timer0Config;
const activeTimer1Config = isMega
? {
...timer1Config,
captureInterrupt: 0x20,
compAInterrupt: 0x22,
compBInterrupt: 0x24,
ovfInterrupt: 0x28,
}
: timer1Config;
const activeTimer2Config = isMega
? { ...timer2Config, compAInterrupt: 0x1a, compBInterrupt: 0x1c, ovfInterrupt: 0x1e }
: timer2Config;
const activeUsart0Config = isMega
? {
...usart0Config,
rxCompleteInterrupt: 0x32,
dataRegisterEmptyInterrupt: 0x34,
txCompleteInterrupt: 0x36,
}
: usart0Config;
const activeSpiConfig = isMega ? { ...spiConfig, spiInterrupt: 0x30 } : spiConfig;
const activeTwiConfig = isMega ? { ...twiConfig, twiInterrupt: 0x4e } : twiConfig;
this.spi = new AVRSPI(this.cpu, activeSpiConfig, 16000000);
this.spi.onByte = (value) => {
this.spi!.completeTransfer(value);
};
this.usart = new AVRUSART(this.cpu, activeUsart0Config, 16000000);
this.usart.onByteTransmit = (value: number) => {
if (this.onSerialData) this.onSerialData(String.fromCharCode(value));
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
// Synthesize the UART frame on PD1 so the oscilloscope sees a real
// waveform during Serial.print. See emitUartTxFrame() for details.
this.emitUartTxFrame(value);
};
this.usart.onConfigurationChange = () => {
if (this.onBaudRateChange && this.usart) this.onBaudRateChange(this.usart.baudRate);
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
// Seed idle HIGH on the TX pin the first time TXEN flips on.
this.handleUartConfigChange();
};
this.twi = new AVRTWI(this.cpu, activeTwiConfig, 16000000);
// Attach the real AVRTWI to the bus created in the constructor;
// any devices already registered + bridges already installed are
// preserved across firmware (re)loads.
this.i2cBus.attachMaster(this.twi);
this.peripherals = [
new AVRTimer(this.cpu, activeTimer0Config),
new AVRTimer(this.cpu, activeTimer1Config),
new AVRTimer(this.cpu, activeTimer2Config),
this.usart,
this.spi,
this.twi,
];
this.adc = new AVRADC(this.cpu, adcConfig);
// ── GPIO ports ──────────────────────────────────────────────────────
this.portB = new AVRIOPort(this.cpu, portBConfig);
this.portC = new AVRIOPort(this.cpu, portCConfig);
this.portD = new AVRIOPort(this.cpu, portDConfig);
if (this.boardVariant === 'mega') {
this.megaPorts.clear();
this.megaPortValues.clear();
for (const { name, config } of MEGA_PORT_CONFIGS) {
this.megaPorts.set(name, new AVRIOPort(this.cpu, config));
this.megaPortValues.set(name, 0);
}
}
}
this.lastPortBValue = 0;
this.lastPortCValue = 0;
this.lastPortDValue = 0;
this.lastOcrValues = new Array(this.pwmPins.length).fill(0);
this.setupPinHooks();
const boardName =
this.boardVariant === 'mega'
? 'ATmega2560'
: this.boardVariant === 'tiny85'
? 'ATtiny85'
: 'ATmega328P';
console.log(`AVR CPU initialized (${boardName}, ${this.peripherals.length} peripherals)`);
}
/**
* Expose ADC instance so components (potentiometer, etc.) can inject voltages
*/
getADC(): AVRADC | null {
return this.adc;
}
/** Returns the CPU clock frequency in Hz (16 MHz for AVR). */
getClockHz(): number {
return 16_000_000;
}
/**
* Returns the current CPU cycle count.
* Used by timing-sensitive peripherals to schedule future pin changes.
*/
getCurrentCycles(): number {
return this.cpu?.cycles ?? 0;
}
/**
* Schedule a pin state change at a specific future CPU cycle count.
* The change fires between AVR instructions, enabling cycle-accurate protocol simulation.
* Used by DHT22 and other timing-sensitive single-wire peripherals.
*/
schedulePinChange(pin: number, state: boolean, atCycle: number): void {
// Callers are expected to push entries in ascending cycle order.
// Insert at the correct position to maintain sort (linear scan from end, O(1) for ordered pushes).
let i = this.scheduledPinChanges.length;
while (i > 0 && this.scheduledPinChanges[i - 1].cycle > atCycle) i--;
this.scheduledPinChanges.splice(i, 0, { cycle: atCycle, pin, state });
}
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
/**
* Synthesize a real bit-level UART frame on the TX pin so an oscilloscope
* sees a waveform during Serial.print, matching real ATmega328P / ATmega2560
* behavior. avr8js's USART only intercepts the byte at the UDR0 register
* level it never toggles PD1 (Uno/Nano) / PE1 (Mega), so without this
* shim the TX pin is flat in the scope while real hardware would show the
* UART frame at the configured baud rate.
*
* Frame layout (8N1, the Arduino default):
* [start LOW] [data LSB ... data MSB] [parity?] [stop1] [stop2?]
*
* We honour avr8js's USART configuration getters (bitsPerChar, parityEnabled,
* parityOdd, stopBits, baudRate) so unusual configurations stay accurate.
*
* Each transition is emitted via onPinChangeWithTime so the oscilloscope
* stamps it with simulator time (cpu.cycles / 16_000 ms), giving bit-level
* timing that holds at any sweep speed.
*/
private emitUartTxFrame(byte: number): void {
const usart = this.usart;
if (!usart || !this.cpu || !this.onPinChangeWithTime) return;
if (!usart.txEnable) return;
const baud = usart.baudRate;
if (!baud || baud <= 0) return;
// ATmega328P (Uno/Nano) UART0: TX = PD1 → Arduino pin 1
// ATmega2560 (Mega) UART0: TX = PE1 → Arduino pin 1 (Mega TX0)
// ATtiny85 has no hardware USART so this method is never called.
const txPin = 1;
const freqHz = 16_000_000;
const cyclesPerBit = freqHz / baud;
const startCycle = this.cpu.cycles;
// Build the frame bit-by-bit. UART idles HIGH; start = LOW; data LSB first;
// optional parity; stop bit(s) HIGH. Idle->start gives the first transition.
const dataBits = usart.bitsPerChar; // typically 8
const bits: boolean[] = [false]; // start bit
let onesCount = 0;
for (let i = 0; i < dataBits; i++) {
const b = (byte >> i) & 1;
bits.push(b !== 0);
onesCount += b;
}
if (usart.parityEnabled) {
// Even parity = bit that makes total ones even; odd = total ones odd.
const parity = usart.parityOdd ? (onesCount % 2 === 0) : (onesCount % 2 !== 0);
bits.push(parity);
}
for (let i = 0; i < usart.stopBits; i++) bits.push(true);
// Emit only the bits that change state to keep buffer churn minimal.
// The "previous" state at startCycle is idle HIGH.
let prevState = true;
for (let i = 0; i < bits.length; i++) {
if (bits[i] !== prevState) {
const timeMs = (startCycle + i * cyclesPerBit) / 16_000;
this.onPinChangeWithTime(txPin, bits[i], timeMs);
prevState = bits[i];
}
}
// After the stop bit(s) the line is already HIGH (idle) so no trailing
// transition is needed — the next byte will start from HIGH automatically.
}
/**
* Seed the TX pin at idle HIGH when the firmware sets TXEN for the first
* time (typically inside Serial.begin). Without this seed the scope's
* "initial state before the first byte" defaults to LOW, hiding the start
* bit transition of the very first byte sent.
*/
private handleUartConfigChange(): void {
if (!this.usart || !this.cpu) return;
const tx = this.usart.txEnable;
if (tx && !this.lastTxEnable && this.onPinChangeWithTime) {
const timeMs = this.cpu.cycles / 16_000;
this.onPinChangeWithTime(1, true, timeMs);
}
this.lastTxEnable = tx;
}
/** Flush all scheduled pin changes whose target cycle has been reached. */
private flushScheduledPinChanges(): void {
if (this.scheduledPinChanges.length === 0 || !this.cpu) return;
const now = this.cpu.cycles;
while (this.scheduledPinChanges.length > 0 && this.scheduledPinChanges[0].cycle <= now) {
const { pin, state } = this.scheduledPinChanges.shift()!;
this.setPinState(pin, state);
}
}
/**
* Fire onPinChangeWithTime for every bit that differs between newVal and oldVal.
* @param pinMap Optional explicit per-bit Arduino pin numbers (Mega).
* @param offset Legacy pin offset (Uno/Nano): PORTB8, PORTC14, PORTD0.
*/
private firePinChangeWithTime(
newVal: number,
oldVal: number,
pinMap: number[] | null,
offset = 0,
): void {
if (!this.onPinChangeWithTime || !this.cpu) return;
const timeMs = this.cpu.cycles / 16_000;
const changed = newVal ^ oldVal;
for (let bit = 0; bit < 8; bit++) {
if (changed & (1 << bit)) {
const pin = pinMap ? pinMap[bit] : offset + bit;
if (pin < 0) continue;
const state = (newVal & (1 << bit)) !== 0;
this.onPinChangeWithTime(pin, state, timeMs);
}
}
}
/**
* Monitor pin changes and update component states
*/
private setupPinHooks(): void {
if (!this.cpu) return;
console.log('Setting up pin hooks...');
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
// DDR register addresses (used to distinguish OUTPUT pins from
// INPUT_PULLUP — see PinManager.updatePort ddrMask param).
// ATmega328P/Uno/Nano: DDRB=0x24, DDRC=0x27, DDRD=0x2A
// ATtiny85: DDRB=0x37
// ATmega2560: per-port table below
const cpu = this.cpu;
const readDdr = (addr: number) => cpu.data[addr] ?? 0;
if (this.boardVariant === 'tiny85') {
// ATtiny85: PORTB only, PB0-PB5 → pins 0-5
// Must pass an explicit pinMap so updatePort uses offset 0 instead of the
// legacy PORTB offset (8) which would map PB1 → pin 9, etc.
const TINY85_PIN_MAP = [0, 1, 2, 3, 4, 5, -1, -1];
this.portB!.addListener((value) => {
if (value !== this.lastPortBValue) {
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
this.pinManager.updatePort('PORTB', value, this.lastPortBValue, TINY85_PIN_MAP, readDdr(0x37));
this.firePinChangeWithTime(value, this.lastPortBValue, null, 0);
this.lastPortBValue = value;
}
});
} else if (this.boardVariant === 'mega') {
// Mega: use explicit per-bit pin maps for all 11 ports
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
const MEGA_DDR_ADDRS: Record<string, number> = {
PORTA: 0x21, PORTB: 0x24, PORTC: 0x27, PORTD: 0x2A,
PORTE: 0x2D, PORTF: 0x30, PORTG: 0x33, PORTH: 0x101,
PORTJ: 0x104, PORTK: 0x107, PORTL: 0x10A,
};
for (const [portName, port] of this.megaPorts) {
const pinMap = MEGA_PORT_BIT_MAP[portName];
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
const ddrAddr = MEGA_DDR_ADDRS[portName];
this.megaPortValues.set(portName, 0);
port.addListener((value) => {
const old = this.megaPortValues.get(portName) ?? 0;
if (value !== old) {
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
this.pinManager.updatePort(portName, value, old, pinMap, ddrAddr ? readDdr(ddrAddr) : undefined);
this.firePinChangeWithTime(value, old, pinMap);
this.megaPortValues.set(portName, value);
}
});
}
} else {
// Uno / Nano: simple 3-port setup
this.portB!.addListener((value) => {
if (value !== this.lastPortBValue) {
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
this.pinManager.updatePort('PORTB', value, this.lastPortBValue, undefined, readDdr(0x24));
this.firePinChangeWithTime(value, this.lastPortBValue, null, 8);
this.lastPortBValue = value;
}
});
this.portC!.addListener((value) => {
if (value !== this.lastPortCValue) {
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
this.pinManager.updatePort('PORTC', value, this.lastPortCValue, undefined, readDdr(0x27));
this.firePinChangeWithTime(value, this.lastPortCValue, null, 14);
this.lastPortCValue = value;
}
});
this.portD!.addListener((value) => {
if (value !== this.lastPortDValue) {
fix(spice+pipeline): LED visualization, INPUT_PULLUP, ESP32-C3, PWM fade, examples End-to-end pipeline fixes uncovered while auditing the /examples gallery. Each bug shipped past green unit + snapshot tests because none of those run firmware + render LEDs. Added scripts/visual-led-test.mjs as a CDP-driven visual harness that loads each example, runs the simulator, samples `wokwi-led.brightness`, and asserts toggle / gradient / initial-off invariants — exits non-zero on any regression. Frontend simulator - PinManager.updatePort: new optional ddrMask param. A pin is added to `outputPins` only if the DDR bit is set, so the PORTx write that enables INPUT_PULLUP (DDR=0, PORT=1) no longer falsely marks the pin as MCU output. AVRSimulator now reads DDRB/C/D (0x24/0x27/0x2A on Uno/Nano, 0x37 on ATtiny85, per-port table on Mega) and forwards it. - AVRSimulator: pass DDR mask alongside every port-listener fire. - BasicParts pushbutton{,-6mm}: seed pin HIGH in attachEvents so `digitalRead()` returns HIGH while idle. avr8js doesn't auto-simulate INPUT_PULLUP — without this the firmware reads LOW from boot and thinks the button is permanently pressed (the "LED is always on, pressing does nothing" UX bug). - connectMcuEdgesToService: suppress synthetic digital edges on pins with active PWM, AND subscribe to onPwmChange to re-tick the netlist on duty changes. Fade-LED now produces a true gradient (6 distinct brightness levels across a fade cycle) instead of a binary 0/full toggle. - CircuitSimulationService.handleMcuEdge: replace single-slot pendingMcuEdge with a per-pin Map. Multiple pins toggling during the same in-flight tick used to overwrite each other; now every pin's most-recent edge replays after the tick. Fixes Traffic-Light RED→ YELLOW→GREEN sequencing. - NetlistBuilder: new sanitizeSpiceId() helper replaces hyphens with underscores in V-source names. ngspice's interactive `alter` command treats `-` as an operator and silently no-ops on hyphenated source names, so mid-simulation MCU pin transitions stopped propagating after the first solve. MixedModeScheduler.onMcuPinChange and CircuitSimulationService self-heal use the same sanitizer so names stay consistent across emit/alter/lookup. Also added a regex-based fallback in step 2 so any board pin matching `GND.\d+` canonicalises to net "0" — ESP32-C3 dev kits expose up to 10 GND pins and the per-board `groundPinNames` list missed several, leaving wires floating instead of grounded. - collectPinStates: emit V-sources only for pins in `outputPins`, not every wired board pin. Leaves INPUT pins (analog sensors on A0, pull-down dividers, etc.) free for the SPICE solver instead of being shorted to 0 V by an ideal MCU V-source. - start.ts: extended __spiceDebug to also expose outputPinsByBoard + nodeVoltages + pinNetMapEntries for the visual harness. - ESP32 / RP2040 / RISC-V / C3 simulators: pass `'mcu'` source flag to triggerPinChange / setPinState so the new outputPins tracking fires on those boards too (was AVR-only before). - useSimulatorStore: stopBoard/resetBoard call pm.resetPinStates() so outputPins clears between runs; Esp32Bridge.onPinChange passes the `'mcu'` flag in all three places it's wired. - types/board.ts: ATtiny85 FQBN `clock=internal16mhz` → `clock=16pll` (ATTinyCore 1.5.2 renamed the option). Backend - esp-idf-template/main/CMakeLists.txt: skip the `-DLED_BUILTIN=2` fallback for esp32c3 and esp32s3 targets. Both variants already define LED_BUILTIN in pins_arduino.h via a self-define macro (`#define LED_BUILTIN LED_BUILTIN` + `static const uint8_t LED_BUILTIN = ...;`). Pre-defining the symbol from the command line expanded the static-const declaration to `static const uint8_t 2 = ...;` — a syntax error that broke every ESP32-C3 / S3 build (`expected unqualified-id before numeric constant`). Examples - examples.ts: bulk-fix 72 wire endpoints that referenced `componentId: 'nano-rp2040'` / `'esp32-c3'` etc. (boards that don't exist on the canvas). Replaced with `'arduino-uno'` (the canvas board-id convention) and converted `D<n>` pin names to `GP<n>` for Pico-style boards. Affects pico-blink, pico-i2c-scanner, pico-i2c-rtc-read, pico-spi-loopback, c3-blink and others. Tests - scripts/visual-led-test.mjs: CDP-driven harness. Default suite covers Blink (single-pin), Button (idle-OFF invariant — catches the INPUT_PULLUP regression), Traffic-Light (multi-pin sequencing), Fade-LED (PWM gradient — ≥3 distinct levels), RGB-LED (≥3 PWM pins driven). Run via `npm --prefix frontend run test:visual` against a Chrome on `:9222` + vite on `:5174` + backend on `:8001`. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 07:47:45 +07:00
this.pinManager.updatePort('PORTD', value, this.lastPortDValue, undefined, readDdr(0x2A));
this.firePinChangeWithTime(value, this.lastPortDValue, null, 0);
this.lastPortDValue = value;
}
});
}
console.log('Pin hooks configured successfully');
}
/**
* Poll OCR registers and notify PinManager of PWM duty cycle changes
*/
private pollPwmRegisters(): void {
if (!this.cpu) return;
const pins = this.pwmPins;
for (let i = 0; i < pins.length; i++) {
const { ocrAddr, pin } = pins[i];
const ocrValue = this.cpu.data[ocrAddr];
if (ocrValue !== this.lastOcrValues[i]) {
this.lastOcrValues[i] = ocrValue;
this.pinManager.updatePwm(pin, ocrValue / 255);
}
}
}
/**
* Start simulation loop
*/
start(): void {
if (this.running || !this.cpu) {
console.warn('Simulator already running or not initialized');
return;
}
this.running = true;
console.log('Starting AVR simulation...');
// Browser-only debug hook. Guarded so node-side vitest runs don't
// ReferenceError on `window` and spam stderr.
if (typeof window !== 'undefined') {
const dbg = (window as unknown as { __spiceDebug?: () => void }).__spiceDebug;
if (typeof dbg === 'function') dbg();
feat(sim): Phase 1c G+F3 — retire legacy CircuitScheduler / eecircuit-engine The mixed-mode migration's endgame. After this commit there is ONE SPICE solver path in the codebase — the vendored ngspice WASM via SolverPort, behind both NgSpiceWorkerAdapter (production browser) and NgSpiceNodeAdapter (Vitest Node). Zero hybrids; zero legacy left to maintain. Deleted production files: • simulation/spice/CircuitScheduler.ts (200ms-poll legacy) • simulation/spice/SpiceEngine.ts (eecircuit-engine wrap) • simulation/spice/SpiceEngine.lazy.ts (lazy code-split) • simulation/spice/subscribeToStore.ts (legacy solve loop) • simulation/spice/connectLegacySolverToMixedMode.ts (bridge) • simulation/spice/connectMixedModeSchedulerToStore.ts (feature flag) Deleted tests (no longer cover any live code): • connect-legacy-solver-to-mixed-mode.test.ts • connect-mixed-mode-scheduler-to-store.test.ts • spice-rectifier-live-bootstrap.test.ts Migrated 6 tests off the deleted `circuitScheduler.solveNow` API to the new `__tests__/helpers/solveInput.ts` (same shape, backed by NgSpiceNodeAdapter). `useElectricalStore` rewritten as a pure state container: • setSolveResult(snapshot) — atomic publish from the service • paused / setPaused — UI control unchanged • reset — project unload • REMOVED: triggerSolve, solveNow, setDebounceMs, scheduler hook • REMOVED: dependency on SpiceEngine.lazy preload EditorPage now mounts a single `startSimulation()` from `simulation/spice/start.ts`, which constructs CircuitSimulationService + ADC bridge + MCU edge bridge. Four useEffect calls collapsed to one. `circuitVerifier.ts` (production) and `runNetlist.ts` use an environment-aware factory: Web Worker in browser, in-proc WASM in Node tests. `/* @vite-ignore */` keeps the Node adapter chain (node:fs, node:url) out of the browser bundle while still letting Node resolve it dynamically. Removed `eecircuit-engine` from package.json dependencies. `collectPinStates` extracted to its own module so the service doesn't depend on the (now deleted) subscribeToStore.ts. Verification: • 1392/1392 tests pass across 103 files (28 pre-existing skips). • `tsc --noEmit` clean. • `vite build` succeeds (27 s, only the existing chunk-size warning that pre-dates this work). Phase 1c — COMPLETE. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-16 02:46:34 +07:00
else console.warn('[spice] __spiceDebug not attached — startSimulation never called');
}
// ATmega328p @ 16MHz
const CPU_HZ = 16_000_000;
const CYCLES_PER_MS = CPU_HZ / 1000;
// Cap: never execute more than 50ms worth of cycles in one frame.
// This prevents a runaway burst when the tab was backgrounded and
// then becomes visible again (browser may deliver a huge delta).
const MAX_DELTA_MS = 50;
let lastTimestamp = 0;
let frameCount = 0;
const execute = (timestamp: number) => {
if (!this.running || !this.cpu) return;
// Clamp delta so we never overshoot after a paused/backgrounded tab.
// MAX_DELTA_MS already handles large initial deltas (e.g. first frame),
// so no separate first-frame guard is needed.
const rawDelta = timestamp - lastTimestamp;
const deltaMs = Math.min(rawDelta, MAX_DELTA_MS);
lastTimestamp = timestamp;
const cyclesPerFrame = Math.floor(CYCLES_PER_MS * deltaMs * this.speed);
try {
for (let i = 0; i < cyclesPerFrame; i++) {
avrInstruction(this.cpu); // Execute the AVR instruction
this.cpu.tick(); // Update peripheral timers and cycles
if (this.scheduledPinChanges.length > 0) this.flushScheduledPinChanges();
}
// Poll PWM registers every frame
this.pollPwmRegisters();
frameCount++;
if (frameCount % 60 === 0) {
console.log(`[CPU] Frame ${frameCount}, PC: ${this.cpu.pc}, Cycles: ${this.cpu.cycles}`);
}
} catch (error) {
console.error('Simulation error:', error);
this.stop();
return;
}
this.animationFrame = requestAnimationFrame(execute);
};
this.animationFrame = requestAnimationFrame(execute);
}
/**
* Stop simulation
*/
stop(): void {
if (!this.running) return;
this.running = false;
if (this.animationFrame !== null) {
cancelAnimationFrame(this.animationFrame);
this.animationFrame = null;
}
this.scheduledPinChanges = [];
console.log('AVR simulation stopped');
}
/**
* Reset simulator (re-run program from scratch without recompiling)
*/
reset(): void {
this.stop();
if (this.program) {
// Re-use the stored hex content path: just reload
const sramBytes =
this.boardVariant === 'mega' ? 8448 : this.boardVariant === 'tiny85' ? 512 : 8192;
console.log('Resetting AVR CPU...');
this.cpu = new CPU(this.program, sramBytes);
if (this.boardVariant === 'tiny85') {
this.portB = new AVRIOPort(this.cpu, attiny85PortBConfig as typeof portBConfig);
feat(attiny85+customchip): full ATtiny85 ADC/Timer0 + custom-chip pipeline fixes ATtiny85 (AVRSimulator + collectPinStates + connectAnalogInputsToMcu + SimulatorCanvas + Attiny85Element + examples): - Add attiny85AdcConfig with correct register addresses (ADMUX=0x27, ADCSRA=0x26, ADCSRB=0x23, ADCL=0x24, ADCH=0x25, DIDR0=0x34, adcInterrupt=0x08). Without this, analogRead() polled the wrong address forever and the firmware hung on first ADC read. - Add attiny85Timer0Config + instantiate AVRTimer so OVF fires at the ATTinyCore-expected ~1.024 ms cadence. delay() advance is still blocked on avr8js TIFR auto-clear semantics (separate upstream issue, see ATTINY85_TIMER0_UPSTREAM_ISSUE.md in velxio-prod test plan). - Map ATtiny85 ADC channels to PB-style pin names (PB5/PB2/PB4/PB3 -> 0..3) in connectAnalogInputsToMcu so SPICE node voltages reach the right ADC channel. - Recognise /^PB\d+$/ in collectPinStates.pinNameToArduinoPin so wires named "PB1" emit v_attiny85_pb1 V-source and the LED responds to MCU writes. Previously every PB-wire returned -1 and SPICE saw no source. - SimulatorCanvas: subscribe pin 1 (PB1) for the built-in LED on the attiny85 board kind (Digispark convention), instead of falling through to the pin-13 default. - Attiny85Element: remove the hand-drawn "yellow LED" circle that was floating above the chip. The bare DIP-8 has no on-board LED; examples wire a real wokwi-led + resistor instead. - examples.ts: add a real wokwi-led + 220 Ohm wokwi-resistor + wires to attiny85-blink, and add missing series resistors to attiny85-button-led and attiny85-ntc-sensor. attiny85-pwm-fade was already correct. Custom-chip pipeline (CustomChipPart + simulatorBridges): - Add a requestAnimationFrame loop that calls instance.tickTimers() every frame in CustomChipPart. Chips that register vx_timer_create (e.g. an i8080 stepping its core, or a sensor publishing samples) had timers added to the queue but nothing fired them; tickTimers was dead code. - Gate the ESP32 backend path with detectSimulatorKind(sim)==='esp32'. The previous `typeof sim.registerSensor === 'function'` check matched AVR and RP2040 simulators too (they expose registerSensor for I2C sensor proxies), routing client-side chips to a non-existent ESP32 worker on those boards. - Replace direct simulator.usart.writeByte calls in avrUartTx with a JS-level FIFO + setTimeout(1ms) drainer. avr8js writeByte drops bytes under burst load (a chip emitting print_string lost ~99% of bytes via non-immediate, or kept only the last byte via immediate). The drainer attempts one non-immediate write per tick and retries on RXC busy / RXEN off. Added a guard for ATtiny85 (no USART -> would queue forever). End-to-end verified: i8080-banner-streamer now prints the boot banner followed by "uptime ticks: 0xNN" lines stepping every ~50 ms, executing real Intel 8080 instructions inside the WASM chip. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-20 04:10:08 +07:00
this.adc = new AVRADC(this.cpu, attiny85AdcConfig);
this.peripherals = [
new AVRTimer(this.cpu, attiny85Timer0Config),
new ATtinyTimer1(this.cpu, attinyTimer1Config),
];
this.usart = null;
} else {
this.spi = new AVRSPI(this.cpu, spiConfig, 16000000);
this.spi.onByte = (value) => {
this.spi!.completeTransfer(value);
};
this.usart = new AVRUSART(this.cpu, usart0Config, 16000000);
this.usart.onByteTransmit = (value: number) => {
if (this.onSerialData) this.onSerialData(String.fromCharCode(value));
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
this.emitUartTxFrame(value);
};
this.usart.onConfigurationChange = () => {
if (this.onBaudRateChange && this.usart) this.onBaudRateChange(this.usart.baudRate);
feat(avr/uart): synthesize bit-level TX waveform on PD1/PE1 avr8js intercepts the transmitted byte at the UDR0 register and never toggles the corresponding GPIO. Real ATmega328P / ATmega2560 hardware drives PD1 / PE1 with a start bit, 8 data bits LSB-first, and a stop bit at the configured baud rate the moment TXEN is set. An oscilloscope probe on D1 therefore showed nothing in Velxio while the same probe in the real world would resolve the UART frame. Synthesize the frame from the inside of `onByteTransmit`: * Read `usart.baudRate`, `usart.bitsPerChar`, `usart.parityEnabled`, `usart.parityOdd`, `usart.stopBits` so unusual configurations stay accurate (avr8js already exposes these as public getters). * Build the bit list start + data(LSB first) + parity? + stopBit(s). * For each transition vs. previous state (initial = idle HIGH), call `onPinChangeWithTime(1, state, timeMs)` where `timeMs = (cpu.cycles + i * cyclesPerBit) / 16_000`. Same simulator-time clock the existing port-listener path uses, so the scope draws the UART waveform cycle-accurately alongside other GPIO activity. Also hook `onConfigurationChange` to detect TXEN flipping 0→1 and seed the scope baseline at idle HIGH; without that, the very first byte's start bit transition would be invisible because the scope's pre-first- sample default is LOW. Both USART construction sites (initial setupSimulation around line 423, re-init after stop around line 749) get the same hook. Covered by `__tests__/avr-uart-tx-waveform.test.ts` (5 cases): idle seed, byte with internal transitions, 0xFF edge case, TXEN-disabled no-op, bit-period timing.
2026-05-23 00:49:43 +07:00
this.handleUartConfigChange();
};
this.twi = new AVRTWI(this.cpu, twiConfig, 16000000);
this.i2cBus.attachMaster(this.twi);
this.peripherals = [
new AVRTimer(this.cpu, timer0Config),
new AVRTimer(this.cpu, timer1Config),
new AVRTimer(this.cpu, timer2Config),
this.usart,
this.spi,
this.twi,
];
this.adc = new AVRADC(this.cpu, adcConfig);
this.portB = new AVRIOPort(this.cpu, portBConfig);
this.portC = new AVRIOPort(this.cpu, portCConfig);
this.portD = new AVRIOPort(this.cpu, portDConfig);
if (this.boardVariant === 'mega') {
this.megaPorts.clear();
this.megaPortValues.clear();
for (const { name, config } of MEGA_PORT_CONFIGS) {
this.megaPorts.set(name, new AVRIOPort(this.cpu, config));
this.megaPortValues.set(name, 0);
}
}
}
this.lastPortBValue = 0;
this.lastPortCValue = 0;
this.lastPortDValue = 0;
this.lastOcrValues = new Array(this.pwmPins.length).fill(0);
this.setupPinHooks();
console.log('AVR CPU reset complete');
}
}
isRunning(): boolean {
return this.running;
}
setSpeed(speed: number): void {
this.speed = Math.max(0.1, Math.min(10.0, speed));
console.log(`Simulation speed set to ${this.speed}x`);
}
getSpeed(): number {
return this.speed;
}
step(): void {
if (!this.cpu) return;
avrInstruction(this.cpu);
this.cpu.tick();
}
/**
* Set the state of an Arduino pin externally (e.g. from a UI button)
*/
setPinState(arduinoPin: number, state: boolean): void {
if (this.boardVariant === 'mega') {
const entry = MEGA_PIN_TO_PORT[arduinoPin];
if (entry) {
const port = this.megaPorts.get(entry.portName);
port?.setPin(entry.bit, state);
}
return;
}
if (this.boardVariant === 'tiny85') {
// ATtiny85: PB0-PB5 = pins 0-5
if (arduinoPin >= 0 && arduinoPin <= 5 && this.portB) {
this.portB.setPin(arduinoPin, state);
}
return;
}
// Uno / Nano
if (arduinoPin >= 0 && arduinoPin <= 7 && this.portD) {
this.portD.setPin(arduinoPin, state);
} else if (arduinoPin >= 8 && arduinoPin <= 13 && this.portB) {
this.portB.setPin(arduinoPin - 8, state);
} else if (arduinoPin >= 14 && arduinoPin <= 19 && this.portC) {
this.portC.setPin(arduinoPin - 14, state);
}
}
/**
* Send a byte to the Arduino serial port (RX) as if typed in the Serial Monitor.
*/
serialWrite(text: string): void {
if (!this.usart) return;
for (let i = 0; i < text.length; i++) {
this.usart.writeByte(text.charCodeAt(i));
}
}
/**
* Register a virtual I2C device on the bus (e.g. RTC, sensor).
*/
addI2CDevice(device: I2CDevice): void {
if (this.i2cBus) {
this.i2cBus.addDevice(device);
}
}
/**
* Remove a virtual I2C device by address. Mirrors RP2040Simulator's
* `removeI2CDevice(addr, bus)` shape so Interconnect / parts can use
* the same uniform API across boards.
*/
removeI2CDevice(address: number, _bus: 0 | 1 = 0): void {
this.i2cBus?.removeDevice(address);
}
/**
* Get the I2CBusManager for a given hardware I2C bus. AVR has only
* one TWI so `bus` is ignored. Available from construction time so
* Interconnect can install cross-board I2C bridges immediately
* (the bus's master peripheral is swapped in later by `loadHex`).
*/
getI2CBus(_bus: 0 | 1 = 0): I2CBusManager {
return this.i2cBus;
}
// ── Generic sensor registration (board-agnostic API) ──────────────────────
// AVR handles all sensor protocols locally via schedulePinChange,
// so these return false / no-op — the sensor runs its own frontend logic.
registerSensor(_type: string, _pin: number, _props: Record<string, unknown>): boolean {
return false;
}
updateSensor(_pin: number, _props: Record<string, unknown>): void {}
unregisterSensor(_pin: number): void {}
}