Pure Python library for reading DHT sensor on Pycom boards

Jurassic Pork

@robert-hh ²
hello robert,
thanks for your comments. I will change my code with your comments. its seems that your pycom is faster than mine so i will change the size of the buffer.
when it will be ready , i will ask your to test the new code.

Jurassic Pork

@robert-hh
i have made the changes : 20 ms for low pulse and a buffer of 800 bits. It always works for my pycom SiPy and my DTH11 sensor. Can you tell me if it is OK for you. The changes are on my Github projet (file dth.py).

robert-hh

@Jurassic-Pork Just one small change:
The initial high period of 25 ms is too short. That had the effect that the first call failed, and it worked only from the second call on. I had to extend that to 500 ms. You can also set the Pin high in init() and skip the call.
I also simplified __collect_input(), saving quite some memory.

    def __collect_input(self):
        m = bytearray(800)        # needs long sample size to grab all the bits from the DHT
        irqf = disable_irq()
        self.__pin(1)
        for i in range(len(m)):
            m[i] = self.__pin()      ## sample input and store value
        enable_irq(irqf)
        return m

I do not think that the LoPy is faster. I assume that the sensor I issues a different collections of 1s and 0s, either caused by the different coding or the environmental conditions.

Jurassic Pork

@robert-hh
ok , thanks robert .
i have made this changes in the dht.py :
1 - in the init :
def init(self, pin, sensor=0):
self.__pin = pin
self.__dhttype = sensor
self.__pin(1)
time.sleep(0.5)
and also comment this in the read method :
#self.__send_and_sleep(1, 0.025)
2 - in the collect_input method :
m = bytearray(800)

bytearray(800) ---> 857 bytes
[1]*800 --> 6464 bytes

With my SiPy and my dth11 sensor no error in a loop of 100 measures.

livius

@Jurassic-Pork
I am realy do not understand why all do this in this way

for i in range(len(m)):
            m[i] = self.__pin()

i know that time is critical
but why not do this with really shorter buffer
I know that time is critical here but inc should work fast
this is not whole algorithm but:

while  self.__pin():
   m[i]+=1
while not self.__pin():
  m[i]-=1

and result buffer will be e.g.

8, -7, 14, ...

robert-hh

@livius The shortest event to be captured is represented by 3 times calling __pin(). What you suggest is in addition to the pin read and loop maintenance adding the load/inc/store operation, not to mention the bookkeeping for the index I between the loops. ~~It still could be feasible, so it's to be tested~~, and not to mention the additional code required to avoid a lock-up situation.

RobTuDelft

Great! Now working on the Lopy.
Thanks a lot @Jurassic-Pork & @robert-hh!

PYB: soft reboot                                                                                                        
==========Starting main.py==========                                                                                    
                                                                                                                        
<DTHResult object at 3ffe2ed0> <class 'DTHResult'>                                                                      
True                                                                                                                    
Temperature: 21.70                                                                                                      
Humidity: 57.50

Jurassic Pork

hello,
@robert-hh said in Pure Python library for reading DHT sensor on Pycom boards:

@Jurassic-Pork Just one small change:
The initial high period of 25 ms is too short. That had the effect that the first call failed, and it worked only from the second call on. I had to extend that to 500 ms. You can also set the Pin high in init() and skip the call.

why 500ms ? the only specification that i have seen for this is :
When power is supplied to sensor, don't send any instruction to the sensor within one second to pass unstable status.

For this a time.sleep of one second in the constructor of the class (init) ?
and :
Is the time.sleep of one second at the beginning of the read method is still necessary ?

specification for the start signal of the AM2302 sensor : min : 0.8 ms typical : 1 ms max : 20 ms
specification for the start signal of the DTH11 sensor : must be > 18 ms
So i will change the start signal value with a value of 19 ms which does the job for the two sensors.

robert-hh

@Jurassic-Pork the 500 ms were just found by trial. It may be more. Just ensure that the data line is high during that period. If you have a 1s delay in the init, the delay in read is not required.
About the length of the start signal: Just give it a try. I did not check for the maximum length accepted by the AM2302. Only 40 ms was too long, and previous tests with 20ms worked. You could make it dependent on the sensor type, since that is specified at init time.

Jurassic Pork

@robert-hh
i have made a few changes in the dth.py code :

class DTH:
    'DHT sensor (dht11, dht21,dht22) reader class for Pycom'

    #__pin = Pin('P3', mode=Pin.OPEN_DRAIN)
    __dhttype = 0

    def __init__(self, pin, sensor=0):
        self.__pin = pin
        self.__dhttype = sensor
        self.__pin(1)
        time.sleep(1.0)


    def read(self):
        #time.sleep(1)

        # send initial high
        #self.__send_and_sleep(1, 0.025)

        # pull down to low
        self.__send_and_sleep(0, 0.019)

1 - no more default value for the pin
2 - a sleep time of one second in the constructor (init) code.
3 - no more sleep time at the beginning of the read method
4 - a sleep time of 19 ms for the startup signal. This value seems to be OK for the specifications of DTH11 AND AM2302.

no error in a program test of 10000 measures with my SiPy module and a DTH11 sensor;

geda

@RobTuDelft , I can read values from the DHT22 with the lopy...
Did you use exactly the same code that @Jurassic-Pork has commited on github ?

RobTuDelft

@geda Yes works for me too :)

@RobTuDelft said in Pure Python library for reading DHT sensor on Pycom boards:

Great! Now working on the Lopy.
Thanks a lot @Jurassic-Pork & @robert-hh!

PYB: soft reboot                                                                                                        
==========Starting main.py==========                                                                                    
                                                                                                                        
<DTHResult object at 3ffe2ed0> <class 'DTHResult'>                                                                      
True                                                                                                                    
Temperature: 21.70                                                                                                      
Humidity: 57.50

geda

@RobTuDelft Sorry, I cannot read values ... so exactly the same code has work for you ? May be my DHT22 is broken ...

Jurassic Pork

@geda
have you a dht22 module or simply a dht22 component ?
have you a pullup resistor on the signal pin of your dht22 ?
what pin of your lopy are you using for the dht22 ?
hello robert-hh same idea :-) in the same time

robert-hh

@geda Which Pin are you using. Note that P13..P18 are input only and cannot be used for that purpose.
see here: https://docs.pycom.io/pycom_esp32/_downloads/lopy_pinout.pdf

RobTuDelft

@geda Yes, I use the latest library from @Jurassic-Pork.

This example works for me, maybe that helps.

I used a 4k7 resistor between data and +.

AM2302 circuit

geda

@RobTuDelft Thanks ! It works ! I made a mistake on my breadboard and had the wrong GPIO on my extension board.
Thanks for your help, I have really appreciated it !

Temperature: 23.30
Humidity: 43.90

And thanks to Jurassic Pork for porting the library

palmy-nz

I am still getting false. Using DHT11, pins double checked and, module works on arduino.

using @RobTuDelft 's main.py, but changed th = DTH(Pin('P4', mode=Pin.OPEN_DRAIN),0)
as using GHT11 on Pin G11 on expansion board

any ideas?
Thanks guys

Jurassic Pork

hello,
@palmy-nz
what is your power supply for your DHT11 ? 5 v or 3,3 v ?
are you using my library here ?
have you a pull up resistor to 3,3v on the pin signal of your DHT11 ?
have you try another expand board pin ? for example P3

Friendly, J.P

palmy-nz

Hi @Jurassic-Pork, yes I am using your dht library.
and yes I have tried pulling it up to 3.3v and 5v, and have tried other pins.
Not sure what the problem is.
Here is my main py: https://gist.github.com/tairea/3607708a403b7501de142d125b07b7b8
and photo of set up: https://pasteboard.co/1p8ETKipZ.jpg

Explore Pybytes | Official Documentation | Report a Firmware Bug/Issue | GitHub

Pure Python library for reading DHT sensor on Pycom boards

Pycom on Twitter