SerialPIO on RP235X

[vslinuxdotnet]

Hello,
I test the SerialPIO on RP235X with a ESP8266 over serial and only the Transmit works, I don't receive anything over serial.

some test:

SerialPIO ser6(10,11);//tx/rx

void setup() {
    delay(1000);
    Serial.begin(9600);
    ser6.begin(115200);
    
    Serial.println("\n-----boot--------");
   }

void loop() {

ser6.write("AT+GMR\r\n");
delay(100);
  while (ser6.available()) {
    Serial.println(ser6.readString());
  }
delay(1000);
}

I connect a serial to usb module in 115200 and have data from ESP8266 :

OK
AT+GMR
AT version:2.2.0.0(b097cdf - ESP8266 - Jun 17 2021 12:57:45)
SDK version:v3.4-22-g967752e2
compile time(6800286):Aug 4 2021 17:20:05
Bin version:2.2.0(Cytron_ESP-01S)

Can be some bug in SerialPIO on RP235X ?

the device:

Thanks VS

[earlephilhower]

Ah, very interesting! I have triple-checked the output (which was the one using 32b bit shifts so wasn't working on GPIOS >31 on the 2350B) and it's good, and just re-verified it and it's still good.

Input, however, seems busted and only on the 2350. If you install a loopback from GP4<-GP6> and run this modified test on a 2040

SerialPIO ser6(4, 6); //tx/rx
void setup() {
  delay(5000);
  Serial.begin(9600);
  ser6.begin(115200);
  Serial.println("\n-----boot--------");
}

void loop() {
  ser6.write("AT+GMR\r\n");
  delay(100);
  while (ser6.available()) {
    Serial.printf("%c", ser6.read());
  }
  Serial.printf("----\n----\n");
}

You get the AT+GMR reported back

---
----
AT+GMR
----
----
AT+GMR
----
----
AT+GMR
----
----

But same app on the 2350 doesn't actually notice received data.

Should be a simple pin issue because there's no special casing in SerialPIO per-chip type.

[earlephilhower]

Can you try #2451, @vslinuxdotnet, and report back? Loopback tests now work, and it was pretty obvious the issue once I looked into it.

[vslinuxdotnet]

Hello, yes In two hours.
Thanks

[vslinuxdotnet]

I tested and it works :) but not full working, the read buffer don't catch all text:

I only get "AT version:2.2.0.0(b097c" and must be:

AT version:2.2.0.0(b097cdf - ESP8266 - Jun 17 2021 12:57:45)
SDK version:v3.4-22-g967752e2
compile time(6800286):Aug 4 2021 17:20:05
Bin version:2.2.0(Cytron_ESP-01S)

I tested with no delays:

   while (ser6.available()) {
    Serial.printf("%c", ser6.read());
  }

Thanks

[earlephilhower]

Sorry, can't reproduce that here. If a single character gets read, they all should be deposited into the FIFO. I bet you're reading faster than the 8266 can write chars out and as such the available() goes to 0 while the 8266 is still thinking.

I suggest something simple like

SerialPIO ser6(4, 6); //tx/rx (whatever your board is wired to)
void setup() {
  delay(5000);
  Serial.begin();
  ser6.begin(115200);
  delay(100); // Make sure everyone is alive and running on the 8266 to respond
  ser6.write("AT+GMR\r\n");
}

void loop() {
while (ser6.available()) {
    Serial.printf("%c", ser6.read());
  }
}

[vslinuxdotnet]

yes that's one problem the slow response, and there's more that matter:

1º 8266 with AT Firmware takes 2000ms to boot after reset.
2º 8266 is slow to answer, and read all data in serial port takes 3 or 4 cpu loops to read 5 lines.
3º must not be delays in loop or some data will be missed.
4º 8266 with AT Firmware don't accept commands < 2 seconds, because takes some time do process one and ignore all the others.

I solved like that:

unsigned long currentMillis = millis();

 if (currentMillis - heartbeatMillis >= 3000)
{
    heartbeatMillis = currentMillis;
    ser6.write("AT+GMR\r\n");
}

  while (ser6.available()) {
    Serial.printf("%c", ser6.read());
  }

@earlephilhower thanks for you time.
Regards VS

[earlephilhower]

If you can rewire the ESP to a SPI port, look at https://github.com/earlephilhower/arduino-pico/tree/master/libraries/lwIP_ESPHost which lets you use all the built-in networking here at much higher speed than an AT-based serial port.

-edit- OOps, I see you're rocking an ESP8266. ESPHost is only on the current ESP32 crop, not ESP8266. Sorry!