DTH11 gets too short start impulse

[robtillaart]

DTH11 gets too short start impulse

[RobTillaart]

Thank you for reporting plus providing a simple solution.

[RobTillaart]

Which processor do you use?
Which IDE version/ develop platform?
Thanks

[RobTillaart]

There is a related issue with the delayMicroseconds(18000).
Some applications using RTOS need to call yield() for context switching, e.g. to handle a Wifi incoming buffer etc.
So the solution will be something more "complicated" than just splitting the call.

[RobTillaart]

@arnowa
Created a Pull Request #73 with a rewrite of the wake-up delay.
Please verify if this fix works for you.

Replaced the delayMicroseconds(_wakeupDelay * 1100UL); with a loop that calls yield() every iteration.
For a DHT22 alike it will be called only once, for the DHT11 alike it will be called 18 times.

  // WAKE UP - add 10% extra for timing inaccuracies in sensor.
  uint32_t startWakeup = micros();
  do
  {
    // HANDLE PENDING IRQ
    yield();
    delayMicroseconds(1100UL);
  }
  while((micros() - startWakeup) < (_wakeupDelay * 1100UL));

Non-RTOS environments (e.g. UNO) implement yield() as an empty function which will be optimized away.

[arnowa]

Sure, will test it today or tomorrow and let you know. Arno

…

On Mon, Jan 31, 2022 at 12:38:18 CET, Rob Tillaart wrote: ***@***.*** Created a Pull Request [2]#73 with a rewrite of the wake-up delay. Please verify if this fix works for you. Replaced the delayMicroseconds(_wakeupDelay * 1100UL); with a loop that calls yield() every iteration. For a DHT22 alike it will be called only once, for the DHT11 alike it will be called 18 times. // WAKE UP - add 10% extra for timing inaccuracies in sensor. uint32_t startWakeup = micros(); do { // HANDLE PENDING IRQ yield(); delayMicroseconds(1100UL); } while((micros() - startWakeup) < (_wakeupDelay * 1100UL)); Non-RTOS environments (e.g. UNO) implement yield() as an empty function which will be optimized away. — Reply to this email directly, [3]view it on GitHub, or [4]unsubscribe. Triage notifications on the go with GitHub Mobile for [5]iOS or [6]Android. You are receiving this because you were mentioned. Message ID: ***@***.***> References 1. https://github.com/arnowa 2. #73 3. #72 (comment) 4. https://github.com/notifications/unsubscribe-auth/AQFEFCPDY4ZFGQ3NAV3KIRTUYZYCVANCNFSM5NENGY6A 5. https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 6. https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign=notification-email&utm_medium=email&utm_source=github

-- Arno Wagner, Dr. sc. techn., Dipl. Inform., Email: ***@***.*** GnuPG: ID: CB5D9718 FP: 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D 9718 ---- A good decision is based on knowledge and not on numbers. -- Plato If it's in the news, don't worry about it. The very definition of "news" is "something that hardly ever happens." -- Bruce Schneier

[arnowa]

Hi Rob, plain cheapduino Uno R3 clone (i.e. ATMega 328) and latest Arduino IDE 1.8.19 on 64 bit Linux. Regards, Arno

…

On Mon, Jan 31, 2022 at 00:51:14 CET, Rob Tillaart wrote: Which processor do you use? Which IDE version/ develop platform? Thanks — Reply to this email directly, [1]view it on GitHub, or [2]unsubscribe. Triage notifications on the go with GitHub Mobile for [3]iOS or [4]Android. You are receiving this because you authored the thread. Message ID: ***@***.***> References 1. #72 (comment) 2. https://github.com/notifications/unsubscribe-auth/AQFEFCMECJNW4FSGJGYFWMDUYXFHFANCNFSM5NENGY6A 3. https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 4. https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign=notification-email&utm_medium=email&utm_source=github

-- Arno Wagner, Dr. sc. techn., Dipl. Inform., Email: ***@***.*** GnuPG: ID: CB5D9718 FP: 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D 9718 ---- A good decision is based on knowledge and not on numbers. -- Plato If it's in the news, don't worry about it. The very definition of "news" is "something that hardly ever happens." -- Bruce Schneier

[arnowa]

Ok, works. Some comments: I thinkt that (micros() - startWakeup) < (_wakeupDelay * 1100UL) will promote all arguments correctly to unsigned long. Hence the calculation will be correct even if micros() wraps as (micros() - startWakeup) will wrap right back in that case. Whether a yield() every 1.1ms is enought when a 19.8ms delay is too long is something I cannot answer.

…

On Mon, Jan 31, 2022 at 12:38:18 CET, Rob Tillaart wrote: ***@***.*** Created a Pull Request [2]#73 with a rewrite of the wake-up delay. Please verify if this fix works for you. Replaced the delayMicroseconds(_wakeupDelay * 1100UL); with a loop that calls yield() every iteration. For a DHT22 alike it will be called only once, for the DHT11 alike it will be called 18 times. // WAKE UP - add 10% extra for timing inaccuracies in sensor. uint32_t startWakeup = micros(); do { // HANDLE PENDING IRQ yield(); delayMicroseconds(1100UL); } while((micros() - startWakeup) < (_wakeupDelay * 1100UL)); Non-RTOS environments (e.g. UNO) implement yield() as an empty function which will be optimized away. — Reply to this email directly, [3]view it on GitHub, or [4]unsubscribe. Triage notifications on the go with GitHub Mobile for [5]iOS or [6]Android. You are receiving this because you were mentioned. Message ID: ***@***.***> References 1. https://github.com/arnowa 2. #73 3. #72 (comment) 4. https://github.com/notifications/unsubscribe-auth/AQFEFCPDY4ZFGQ3NAV3KIRTUYZYCVANCNFSM5NENGY6A 5. https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 6. https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign=notification-email&utm_medium=email&utm_source=github

-- Arno Wagner, Dr. sc. techn., Dipl. Inform., Email: ***@***.*** GnuPG: ID: CB5D9718 FP: 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D 9718 ---- A good decision is based on knowledge and not on numbers. -- Plato If it's in the news, don't worry about it. The very definition of "news" is "something that hardly ever happens." -- Bruce Schneier

[RobTillaart]

Good point, could make it 550.

[RobTillaart]

@arnowa

Which board do you use?
Which IDE?

[arnowa]

plain cheapduino Uno R3 clone (i.e. ATMega 328) and latest Arduino IDE 1.8.19 on 64 bit Linux.

[RobTillaart]

Thanks.
Did some math and every 360 us would match 18 x 1100 = 19.800 exactly. Given the loop has some overhead it would at most 'overflow' 359 us.
3x better than 1100.

[arnowa]

I would go with that and see whether anybody runs into trouble. As this is also better than the original 1.1ms for the DHT22, it should not cause any new problems.

…

On Mon, Jan 31, 2022 at 16:03:36 CET, Rob Tillaart wrote: Thanks. Did some math and every 360 us would match 18 x 1100 = 19.800 exactly. Given the loop has some overhead it would at most 'overflow' 359 us. 3x better than 1100. — Reply to this email directly, [1]view it on GitHub, or [2]unsubscribe. Triage notifications on the go with GitHub Mobile for [3]iOS or [4]Android. You are receiving this because you were mentioned. Message ID: ***@***.***> References 1. #72 (comment) 2. https://github.com/notifications/unsubscribe-auth/AQFEFCMLUPG6KSIJULHVF5LUY2QERANCNFSM5NENGY6A 3. https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 4. https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign=notification-email&utm_medium=email&utm_source=github

-- Arno Wagner, Dr. sc. techn., Dipl. Inform., Email: ***@***.*** GnuPG: ID: CB5D9718 FP: 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D 9718 ---- A good decision is based on knowledge and not on numbers. -- Plato If it's in the news, don't worry about it. The very definition of "news" is "something that hardly ever happens." -- Bruce Schneier

[RobTillaart]

plain cheapduino Uno R3 clone (i.e. ATMega 328) and latest Arduino IDE 1.8.19 on 64 bit Linux.

Ok that allows me to recreate/test a bit

[RobTillaart]

Did some more math/testruns and came to 180.

For DHT11 ==> 19808 us
For DHT22 ==> 1104 us

So quite a perfect (including loop overhead).

  // WAKE UP - add 10% extra for timing inaccuracies in sensor.
  uint32_t startWakeup = micros();
  do
  {
    // HANDLE PENDING IRQ
    yield();
    delayMicroseconds(180UL);
  }
  while((micros() - startWakeup) < (_wakeupDelay * 1100UL));

[RobTillaart]

The problem seems to be calling delayMicroseconds() with a value larger than 16383: https://www.arduino.cc/reference/en/language/functions/time/delaymicroseconds/

The parameter of delayMicros() is unsigned int that implies one could go up to 65535
(it is hand-tuned code, so I stop digging)

---

parameter us is multiplied by 4 here causing overflow...

	// the following loop takes 1/4 of a microsecond (4 cycles)
	// per iteration, so execute it four times for each microsecond of
	// delay requested.
	us <<= 2; // x4 us, = 4 cycles

[RobTillaart]

Solution merged in master