2

I have an Arduino with a servo, INA219 (current and voltage sensor) and 2 DS18B20 connected to it.

DS18B20 are connected in full-power mode with 5k resistor connected between DATA line and +5V.

I use DallasTemperature library: https://github.com/milesburton/Arduino-Temperature-Control-Library

I get the temperature readings, but the problem is that my servo is jittering slightly almost every time temperature reading is done. If I disconnect DATA line from pin 2 (where it's normally connected) then servo is not moving anymore and everything is fine.

Any ideas what might be causing it? I would expect servo interfering with OneWire readings, not the other way around :)

Here is the full code: https://pastebin.com/xcES4PVT

Here is the minimal example with which I can reproduce the problem: 

    #include <DallasTemperature.h>
#include <Servo.h>

#define ONE_WIRE_BUS 2
OneWire oneWire(ONE_WIRE_BUS);

// Pass our oneWire reference to Dallas Temperature. 
DallasTemperature sensors(&oneWire);

Servo myservo;

long previousMillis = 0;

long tempRequestedMillis = 0;

bool tempRequested = false;

void setup(void)
{
  // start serial port
  Serial.begin(115200);
  Serial.println("Dallas Temperature IC Control Library Demo");

  myservo.attach(9);

  // Start up the library
  sensors.begin();
  sensors.setWaitForConversion(false);
}

void loop(void)
{ 

  unsigned long currentMillis = millis();
  if (!tempRequested && currentMillis - previousMillis > 2000) {
    previousMillis = currentMillis;

    sensors.requestTemperatures(); // Send the command to get temperatures

    tempRequestedMillis = currentMillis;
    tempRequested = true;
  } else if (tempRequested && currentMillis - tempRequestedMillis > 2000) {
    Serial.println(sensors.getTempCByIndex(0));
    tempRequested = false;
  }
}
RubberDuck
  • 332
  • 1
  • 9
Leonti
  • 123
  • 3
  • 1
    maybe something is taking up too much time in the loop ... start by commenting out all the `serial.print` statements – jsotola Feb 03 '18 at 02:18
  • I have the same problem that Leonti. BUt in my case, i have a duemilanove and the lcd shield, that which occupies pins 9 and 10 that I read are the necessary ones in the duemilanove to work with this library. Is there any possibility of using other pins? Is there another library that uses different pins? I have free: A1,A2,A3,11,12 and 13... – Juan Vivo Sep 23 '18 at 09:31

2 Answers2

6

Arduino's Servo library on AVR uses interrupts; if interrupts are disabled for any amount of time, the servos will twitch.

1-Wire protocol (used by DS18B20) needs to disable interrupts for the protocol to work. Therefore, servos glitch.

There is a great explanation of the problem here: https://learn.adafruit.com/neopixels-and-servos/overview , just mentally replace "Neopixel library" with "OneWire library". Their TiCo_Servo library sounds just like what you want.

theamk
  • 176
  • 1
  • That was it! After using TiCoServo library there are no issues! Thanks a lot! – Leonti Feb 03 '18 at 11:53
  • The OneWire library was poorly implemented IMO. It globally disables *all* interrupts when it only really needs to disable interrupts *on the data line*. If other interrupts are well behaved and return quickly, it won’t disrupt the One wire timing. I don’t know how to turn my avr code into an Arduino library, but there’s an alternative implementation in my digital thermometer project. https://github.com/rubberduck203/digital-thermometer – RubberDuck Feb 03 '18 at 14:05
1

Reading a DS18B20 the naive way takes up to 1500ms because you have to trigger a temperature conversion, then wait for it to complete.

If you have them full-powered, you better trigger the conversion, continue with your program and read the result 2 seconds later, when it's done. You can also trigger the conversion on all connected DS18B20 simultaneously by using the Skip ROM command before sending the Convert T command.

Please see the documentation of Maxim/Dallas' library how to do that.

  • I tried to use `setWaitForConversion` which shouldn't wait for conversion and then read the results 2 seconds later - still the same problem. I've attached the minimal code which has the problem, would you mind taking a look at it? –  Feb 03 '18 at 02:52
  • Looks okay, are you sure you aren't accidentally running the old code? –  Feb 03 '18 at 03:04