3

I am new in raspberry pi. In my research project, I would like to monitor ground temperature at 60 points. My team decided to set a monitoring system using raspberry pi.

As a preliminary test, I connected 4 DS18B20 sensors to my pi; it can find sensors and read temperature correctly. Unfortunately, when I connected an additional DS18B20 to the pi (so total 5 DS18B20 were linked to the pi), it stopped monitoring and gave no value. (I tried to reboot the system; however the pi can't read the temperature)

After this weird situation, I disconnected the last sensor (so subsequently total 4 sensors were connected to the pi); then the pi started to read the temperature. I have attached two photos of a breadboard with 4 and 5 sensors.

Breadboard with 4 sensors - work Breadboard with 5 sensors - doesn't work

I really want someone can explain this to me.

Thank you for reading.

UPDATE

I updated config.txt and opend 4, 17, 27 pins. So now I am using 3 pins. As a first trial, I connected 4, 4, and 2 DS18B20s to 4, 17, and 27 pins, respectively. Interestingly, my RPi can find every sensors in this case. Followings are PCB and Terminal screen.

teminal with 4, 4, 2 connection PCB with 4, 4, 2 connection

Then, I connected 5, 4, 1 DS18B20s to 4, 17, and 27 pins, respectively. In this time, my RPI can detect only 5 sensors which are connected to 17 and 27 pins. Followings are PCB and Terminal screen.

teminal with 5, 4, 1 connection PCB with 5, 4, 1 connection

I am guessing that there is a restriction on the number of DS18B20 on a bus. Am I right?

5
  • 2
    What is the electrical connection for these? Are you using powered sensors (3 wires) or parasitic power (2 wires only). Do you have a single pull-up for the one-wire bus (correct) or have you added a pull-up per sensor (incorrect)? Also provide a minimal code sample for inspection that demonstrates how you are reading the values. Commented Jul 2, 2019 at 8:51
  • The datasheet suggests the following for better performance: (1) Use strong pullup 4k7, (2) Supply power at Vdd pin, not using parasite power at Dq. No harm trying. Commented Jul 2, 2019 at 8:53
  • 1
    I am using powered sensors (3 wires) and a single pull-up. Basically I followed the circuit showed in (raspberrypi.org/forums/viewtopic.php?t=167896). I didn't use any code currently as I have a trouble in recognizing sensors. I simply typed "ls" at /sys/bus/w1/devices/. Commented Jul 2, 2019 at 14:54
  • 1
    I have attached photos of my breadboard setting. Commented Jul 2, 2019 at 15:26
  • 1
    Try a stronger pull-up. Commented Jul 2, 2019 at 17:33

2 Answers 2

1

Before going on, please check if you have original DS18B20 sensors. I had some counterfeits with the strangest behaviour. Same vanished after some time. Sometimes after a few minutes, sometimes after a few hours, whereas other counterfeits connected at the same time at the same one-wire bus still worked.

I see that you have the waterproof DS18B20. I've never managed to get original DS18B20 which are waterproof. Even the waterproof DS18B20 sold by Shelly and a large German distributor (Reichelt) were reported as counterfeit.

Currently you can only check with an Arduino. Source code and a lot more information about counterfeit DS18B20 can be found here https://github.com/cpetrich/counterfeit_DS18B20

0

Question

Rpi can handle at most 4 DS18B20s, how come so few? How can greedy me do as many as 60?

Short Answer

Update 2019nov20hkt1014

ds18b20 pullup

Update 2019jul26hkt2030

I have upgraded my Rpi3B+ stretch to Rpi4B buster. I test the same Rpi3B+ program for Rpi4B /buster and found everything work smoothly. The program v71 with updated documentation is listed in Appendix G. I have deleted the old program in Appendix F, because this answer is again getting too long, hitting the forum answer's 32k words limit.

Update 2019jul17hkt1301

I connected 12 DS18B20 to GPIO2, and another 3 to GPIO3, and found measurements stable. I am using the I2C pins for now. I hope to test other pins after I finished testing Rpi4 and buster.

ds18b20 test

*** Begin testGetPrintSensorIdTempTimeList() ********************************************************** Time Testing = 2019-07-17 12:49 tlfong01 Getting sensor temperatures, ............... Begin printSensorIdTempTimeList(), ... ---------------------------------------------------------- Sensor Sensor Temp Time Num ID deg C ---------------------------------------------------------- 1 01161cd97fee 28.12 2019-07-17 12:49:22.365111 2 01161cda20ee 28.12 2019-07-17 12:49:23.325028 3 0316a2795c11 27.81 2019-07-17 12:49:24.284948 4 021313a052aa 28.56 2019-07-17 12:49:25.244955 5 011319f99467 28.38 2019-07-17 12:49:26.204966 6 000001d8e47e 28.44 2019-07-17 12:49:27.164934 7 0316a27926c3 28.25 2019-07-17 12:49:28.124953 8 01131a39c3e9 28.56 2019-07-17 12:49:29.085026 9 0316a279d285 28.25 2019-07-17 12:49:30.044939 10 0316a27907c7 27.69 2019-07-17 12:49:31.004974 11 02131261bcaa 28.38 2019-07-17 12:49:31.965575 12 01131a2edfe9 28.50 2019-07-17 12:49:32.924948 13 0316a27938eb 27.69 2019-07-17 12:49:34.845093 14 01131bb23b09 28.19 2019-07-17 12:49:35.804940 15 000001d8e502 28.44 2019-07-17 12:49:36.764982 ---------------------------------------------------------- End printSensorIdTempTimeList(). Number of sensors tested = 15 Seconds lapsed = 16.69 Seconds per sensor = 1.11 *** End testGetPrintSensorIdTempTimeList() ********************************************************** 

ds18b20 2019jul1402

Long Answer to be shortened or removed

Now I am testing 12 DS18B20 devices, all connected to one GPIO pin 2, configuring as star. Now I am finding that the syatem is not reliable, in that it takes longer to made one measurements, from about 1 second to up ta 1.5 seconds. I need to find if using daisy chain configuration can improve the performance.

Testing 11 sensors OK. All 11 sensors connected in parallel, power 3V3, all output pins connected to GPIO2 (Similarly to shown in the following post . Test program list and sample output shown in Appendix B.

wiring picture

Now I am going to test 9 sensor modules, suing 2 or 3 GPIO pins.

ds18b20 modules

Now I am testing 11 DS18B20 (7 with waterproof probes, two modules, two through hole chips), all on the same bus connected to GPIO2. I found measurements stable. See sample output and program listing in the appendix below.

ds18b20 modules

Long winded Research Notes to shorten or delete later

I guess, DS18B20, using the 1-Wire method, is like the trouble making I2C, cannot have wires too long (preferably less than 30cm), pull up not too strong, and not too weak (total around 4k7) capacitance less than 400pF etc.

For I2C, I had a very hard time and could achieve only placing 8 devices on the bus, but system still very unstable, often getting I/O error (due to impedance overloading, noise etc).

So for this 1-Wire DS18B20, I was humble and my goal was to get at most 8 devices on the 1-Wire bus. I started with onle one DS18B20, and added one after another, until I used up all my 7 devices. Surprisingly I found measurements very stable.

Now let me show off my setup, results, and program.

Setup

setup

I think I confused everybody by using I2C data pin for DS18B20. Actually I can use any other pin. The reason that I choose I2C Data pin (GPIO 2) is that I know I2C CLK and DATA pins have a very strong pull up of 1k8. All other pins (only 90% sure) have pullups of 50k. In other words, I don't need any extra 4k7.

References

  1. DS18B20 Programmable Resolution 1-Wire Digital Thermometer Datasheet - Maxim

  2. Guidelines for Reliable Long Line 1-Wire Networks, Tutorial 148 - Maxim

  3. w1thermsensor DS18B20 Python Module 1.1.2 - pypi

  4. Multiple DS18B20 1-Wire Temperature Sensors - rpi.org forum 2016dec09

  5. DS18B20 Wiring Diagram - raspberry pi forum discussion

  6. DS18B20 MAX DISTANCE -rpi forum geekinthesticks 2013maro6

  7. Joining Multiple DS18B20 Sensors - rpi forum Darranking 2019jan14

  8. 1-Wire - Wikipedia

  9. SunFounder DS18B20 Temperature Sensor Learning Kit

  10. Rpi3B+ Python DS2482 I2C to 1-Wire Bridge Connecting 10+ DS18B20 Temperature Sensors Using 30 Metres Long CAT5e Cables


Appendices

Appendix A - Test Program 2019jul05

Appendix B - DS18B20 Test Program 2019jul11

Appendix C - DS18B20 12 devices testing notes

Appendix D/E - DS18B20 Test Program V50 2019jul14hkt1128

Listings too long hitting the StackExchange answer's 32k word limit, therefore removed.

Appendix F - Program listing of version 62 2019jul17hkt1259

Appendix G - Version 71 Rpi4B Throny python 3.7.3

Update 2019nov20hkt1031

(1) DS18B20 test71 was written for Rpi3B+ stretch 9 and IDLE python 3.5.3. It is still compatible to Rpi4B buster python 3.7.3, without any modification.

(2) The program was developed in an almost pure functional programming and declarative programming paradigm, using lisp/scheme/LUA programming constructs, and OOP style without using OP constructs, only FP. It automatically scales up for any number of DS18B20 sensors. For example, there is no C++ like switch/case, nested if-then-else, numerical for loops. The same program can entertain more or less sensors without modifying any for loop range parameter, ie, "looping for list", not for "range(length ...) etc.

# DS18B20_test71 tlfong01 2019jul26hkt204 # Rpi3B+ Raspbian 9 stretch 2019jul15, IDLE Python 3.5.3 # Rpi4B Raspbian 10 buster 2019jul25, Throny Python 3.7.3 # python3-w1ThermSensor-1.0.5 # *** Setup Notes *** # 1. "$ sudo nano /boot/config.txt" to config w1thermsensor GPIO dtoverlay # 2. GPIO 2/3 (I2C DATA/CLK) with built in 1k8 pull up is used. For other # pins with weaker pullups, a 4k7 resistor is usually recommended. # 3. dtoverlay=w1-gpio,gpiopin=2, dtoverlay=w1-gpio,gpiopin=3 # 4. For two or more GPIO pins, w1thermsensor auto detects connected sensors, # but might hang if GPIO is specified but no senors are connected. # And remember to reboot after making changes to dtoverlay settings. # 6. Terminal mode sensor detection command = "$ ls /sys/bus/w1/devices" # *** Testing Results *** # 1. Program V62 has been tested OK for 15 sensors (12 connected to GPIO2 # and 3 to GPIO3. All connecting cables are less than 1 metre long. # 2. To test later - Rpi4, buster, long CAT5 cables, 5V power and logical # level shifter. # Program Summary and Testing Notes # 1. The three datetime functions in this sections are mainly for demo and testing. They # can be removed after development is completed. # 2. The data structure "sensorIdList" contains one member for each sensor, # Each member consists of the id string of the sensor. # 3. The get/print/getPrint SensorIdList is mainly used for detecting sensors which is # much quicker than the very slow sensor measurement process of around 1 second long. # 4. The data structure "sensorIdTempTimeList", contains one member for each sensor. # Each member consists of the id, temperature read, time read of the sensor. # 5. The getSensorIdTempTime fucntion is very slow, about 1 second to execute. # 6. The main tests can be executed all at once, or comment/uncomment to select lists. from time import sleep from datetime import datetime from datetime import timedelta from w1thermsensor import W1ThermSensor # *** Date Time Functions *** def printDateTimeNow(): timeNowLong = datetime.now() timeNowShort = str(timeNowLong)[0:16] print('\n timeNowLong =', timeNowLong) print(' timeNowShort =', timeNowShort) return def printElapsedTime(): startTime = datetime.now() sleep(2) finishTime = datetime.now() timeElapsed = finishTime - startTime secondsElapsed = timeElapsed.seconds microSecondsElapsed = (secondsElapsed * 1000000) + timeElapsed.microseconds milliSecondsElapsed = microSecondsElapsed / 1000 print('\n Start Time =', startTime) print(' Finish Time =', finishTime) print('\n timeElapsed =', timeElapsed) print(' secondsElapsed =', secondsElapsed) print(' milliSecondsElapsed =', '%d' %milliSecondsElapsed) print(' microSecondsElapsed =', microSecondsElapsed, '\n') def testDateTime(): print('\n*** Begin Test python datetime module, ... *******************************') printDateTimeNow() printElapsedTime() print('*** End Testing python datetime module. **********************************') return # *** DS18B20 Functions *** # *** SensorIdList Functions *** def getSensorIdList(): sensorIdList = [] for sensor in W1ThermSensor.get_available_sensors(): sensorIdList.append(sensor.id) return sensorIdList def printSensorIdList(sensorIdList): print(' ------------------------------------------------------------------------') print(' Sensor ', ' Sensor') print(' Num ', ' ID') print(' ------------------------------------------------------------------------') sensorNum = 1 for sensorId in sensorIdList: if sensorNum < 10: print(' ', sensorNum, ' ', sensorId) else: print(' ', sensorNum, ' ', sensorId) sensorNum = sensorNum + 1 print(' ------------------------------------------------------------------------') def getPrintSensorIdList(): sensorIdList = getSensorIdList() printSensorIdList(sensorIdList) return def testGetPrintSensorIdList(): print('\n*** Begin testGetPrintSensorIdList() ******************************************************************\n') getPrintSensorIdList() print('\n*** End testGetPrintSensorIdList() ******************************************************************\n') return # *** SensorIdTempTimeList Functions *** def getSensorIdTempTimeList(): sensorIdTempTimeList = [] print(' Getting sensor temperatures, ', end = '') for sensor in W1ThermSensor.get_available_sensors(): idTempTimeList = (sensor.id, sensor.get_temperature(), datetime.now()) sensorIdTempTimeList.append(idTempTimeList) print('.', end = '') return sensorIdTempTimeList def printSensorIdTempTimeList(sensorIdTempTimeList): print('\n\n Begin printSensorIdTempTimeList(), ...\n') print(' ----------------------------------------------------------') print(' Sensor', ' Sensor', ' Temp', ' Time') print(' Num', ' ID', ' deg C', ' ') print(' ----------------------------------------------------------') sensorNum = 1 for idTempTimeList in sensorIdTempTimeList: if sensorNum < 10: print(' ', sensorNum, end = '') else: print(' ', sensorNum, end = '') print(' ', idTempTimeList[0], end = ' ') print('', '%.2f' %(idTempTimeList[1]), end = ' ') print('', idTempTimeList[2]) sensorNum = sensorNum + 1 print(' ----------------------------------------------------------') print('\n End printSensorIdTempTimeList().') return def testGetPrintSensorIdTempTimeList(): print('\n*** Begin testGetPrintSensorIdTempTimeList() **********************************************************\n') startTime = datetime.now() print(' Time Testing =', str(startTime)[0:16], 'tlfong01 \n') sensorIdTempTimeList = getSensorIdTempTimeList() printSensorIdTempTimeList(sensorIdTempTimeList) finishTime = datetime.now() timeElapsed = finishTime - startTime secondsElapsed = timeElapsed.seconds microSecondsElapsed = (secondsElapsed * 1000000) + timeElapsed.microseconds sensorNum = len(sensorIdTempTimeList) print('\n Number of sensors tested =', sensorNum) print(' Seconds lapsed =', '%.2f' %(microSecondsElapsed / 1000000)) print(' Seconds per sensor =', '%.2f' %((microSecondsElapsed / 1000000) / sensorNum)) print('\n*** End testGetPrintSensorIdTempTimeList() **********************************************************\n') return # *** Main Test *** testDateTime() testGetPrintSensorIdList() testGetPrintSensorIdTempTimeList() ''' Python 3.7.3 (/usr/bin/python3) >>> %cd /home/pi/Python_Programs/test1321/ds18b20 >>> %Run ds18b20_test70_2019jul2501.py *** Begin Test python datetime module, ... ******************************* timeNowLong = 2019-07-26 20:01:22.402122 timeNowShort = 2019-07-26 20:01 Start Time = 2019-07-26 20:01:22.402724 Finish Time = 2019-07-26 20:01:24.404806 timeElapsed = 0:00:02.002082 secondsElapsed = 2 milliSecondsElapsed = 2002 microSecondsElapsed = 2002082 *** End Testing python datetime module. ********************************** *** Begin testGetPrintSensorIdList() ****************************************************************** ------------------------------------------------------------------------ Sensor Sensor Num ID ------------------------------------------------------------------------ 1 01161cd97fee 2 01161cda20ee 3 0316a2795c11 4 021313a052aa 5 011319f99467 6 000001d8e47e 7 0316a27926c3 8 01131a39c3e9 9 0316a279d285 10 0316a27907c7 11 02131261bcaa 12 01131a2edfe9 13 0316a27938eb 14 01131bb23b09 15 000001d8e502 ------------------------------------------------------------------------ *** End testGetPrintSensorIdList() ****************************************************************** *** Begin testGetPrintSensorIdTempTimeList() ********************************************************** Time Testing = 2019-07-26 20:01 tlfong01 Getting sensor temperatures, ............... Begin printSensorIdTempTimeList(), ... ---------------------------------------------------------- Sensor Sensor Temp Time Num ID deg C ---------------------------------------------------------- 1 01161cd97fee 29.81 2019-07-26 20:01:25.373759 2 01161cda20ee 29.75 2019-07-26 20:01:26.253930 3 0316a2795c11 29.38 2019-07-26 20:01:27.133707 4 021313a052aa 30.00 2019-07-26 20:01:28.014356 5 011319f99467 29.88 2019-07-26 20:01:28.903601 6 000001d8e47e 29.81 2019-07-26 20:01:29.853169 7 0316a27926c3 29.88 2019-07-26 20:01:30.733717 8 01131a39c3e9 29.94 2019-07-26 20:01:31.613679 9 0316a279d285 29.62 2019-07-26 20:01:32.493157 10 0316a27907c7 29.12 2019-07-26 20:01:33.373696 11 02131261bcaa 29.88 2019-07-26 20:01:34.313676 12 01131a2edfe9 29.62 2019-07-26 20:01:35.213715 13 0316a27938eb 29.50 2019-07-26 20:01:36.093166 14 01131bb23b09 29.44 2019-07-26 20:01:36.973689 15 000001d8e502 29.94 2019-07-26 20:01:37.853752 ---------------------------------------------------------- End printSensorIdTempTimeList(). Number of sensors tested = 15 Seconds lapsed = 13.48 Seconds per sensor = 0.90 *** End testGetPrintSensorIdTempTimeList() ********************************************************** >>> ''' # *** End of Sample Output *** # *** End of Program Documentaton *** 
10
  • 2
    Thanks; Do you think that I should rely on I2C to use multiple sensors? Commented Jul 4, 2019 at 6:24
  • 1
    @Sang Inn Woo, Ah I guess you misunderstood my comments on I2C. I said I have bad experience using I2C bus and could not put more than 8 devices on the bus and performance is unstable. So I guessed DS18B20's 1-Wire bus should be as bad. However, I started testing from 1 sensor and adding one by by one. Surprisingly I can add 7 (I have only 7 on hand!) and found everything stable. So I guess there is hope to add to 60! I am using the python3-w1ThermSensor-1.0.5. and I found it easy to setup and test. Commented Jul 4, 2019 at 6:33
  • You might like to read my update, clarifying that I am NOT using I2C. I am still using 1-Wire bus. I can use any GPIO pin. I choose I2C Data pin because it has a built in strong pull up of 1k8. In other words, you can try any other pins. BTW, I don't follow your terminal commands listing 1-Wire devices. Please show me the link for those commands. Thanks. Commented Jul 4, 2019 at 6:44
  • 1
    @ tlfong01 Thank you very much for replying. I am just typing "cd /sys/bus/w1/devices" then "ls" to list DS18B20 sensors that I wired. Is my issue related to powering? In today's trial, when I used 5V 3A USB adapter, I can find four DS18B20 sensors per one bus. However, when I changed USB adapter with 5V 1A, I can find only two sensors. BTW, it would be grealty helpful if you can share details of wiring for DS18B20. Commented Jul 5, 2019 at 13:02
  • 1
    Just as @tlfong01 did, I can now connect 10+ DS18B20 sensors to GPIO 2 or GPIO 3 (which are I2C buses) with stable reading. For GPIO 4, 17, 22, ..., I tried an 1k resister instead of 4.7k for pull-up following the recommendation from @Brick; then I can connect 10+ DS18B20 sensors to those buses. I believe that the problem is now solved with your help. I will try 60 sensors wired to a RPi in early next week and update the post. Commented Jul 17, 2019 at 11:03

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.