Go Down

Topic: (Solved) DS18B20 possibly dead? (Read 130 times) previous topic - next topic

intruder82

Feb 05, 2016, 08:02 am Last Edit: Feb 08, 2016, 10:38 am by intruder82
I have a running stable system of 13 DS18B20 sensors that are all connected through one pul up resistor on one wire bus (similar to this http://cdn.instructables.com/F79/RUWE/GTUY9ZXQ/F79RUWEGTUY9ZXQ.MEDIUM.jpg) and reporting to the cloud with visualization service.
Last night all of my sensors started reporting 0 values after few hours one of the sensors changed value to the 85 and stayed like this for the rest of the night..

Can this information indicate that this one sensor with value 85 is the bad one or how can one track the bad sensor in one wire network?

Thanks in advance
Vaidas

Koepel

If all the sensors didn't work very well, then the 1-Wire bus might be bad. I prefer a three-wire bus.

Did you write down the sensor ID's and which one is located where ?

The sensors don't get damaged easily. Is there a high power device nearby with large inductive or voltage peaks ? What cable do you use ?

intruder82

No, I have it on automatic identification. are there some solutions if the ID would be known?

No the one wire bus runs quite far away from the power lines.. I was hoping that only one sensor had a malfunction and this broke down the whole network, so the values of 0 or 85 doesn't mean anything?

Koepel

#3
Feb 05, 2016, 02:06 pm Last Edit: Feb 05, 2016, 02:08 pm by Koepel
No, 0 and 85 doesn't ring a bell. Perhaps if the data was not written, just random data or a previous data was used. A check if the data is actually from the sensor is always needed.

If they are connected one by one, they can be labeled with the ID.
The OneWire has more basic checks than the DallasTemperature, I think the OneWire has an example that prints out all the data and the ID and calculates the checksum. Could you run that sketch ?

Could you measure with a multimeter, perhaps there is a shortcut, or water between the wires, or corrosion.

Nick_Pyner

#4
Feb 05, 2016, 03:42 pm Last Edit: Feb 06, 2016, 02:55 am by Nick_Pyner
I don't think the 0 shouldn't happen at all - in theory.

I have a setup that has worked for some years and suddenly had all sensors reading 0.0. I have no idea what causes this but it fixes itself if the power is turned off. Pressing reset is not enough and, while it is running OK at the moment, the fix may only last a few hours..

I have wondered if it is being caused by one sensor because I have some format glitch appearing too. it seems to return four decimal places, i.e. 23.4500 and the last digit has to go on the next line of the LCD. This only happens with one reading, and a reset is enough to fix it.

edit - I was wrong about 85, el Supremo is right

el_supremo

85 is the default reading of the temperature on power up (read the datasheet). You will see this value after power up if you either read the temperature without starting the conversion or when you start a temperature conversion but don't give the device enough time to finish it. For example, in 12-bit precision, you have to wait about 850 milliseconds between starting the conversion and reading it.
The diagram shows that the power pins are not connected to anything (floating). The datasheet does not mention that this can be done. If you operating the sensors in parasite power mode, the power pins must be grounded.

Pete

intruder82

On the weekend had some time to play around and had luck to find the problem quite fast before I started the tear down.. The problem was in the power supply.. when I connected it to usb port it functioned and when I exchanged the power supply it worked to..

Just don't understand what is wrong with the old power supply.. I connected the voltmeter it shows the same 9,2V.. some kind of strange interference that does not affect arduino but kills the sensors data... (because the arduino functions and sends the data to the cloud just sensors show 0)

P.s. el_supremo you are right it's on parasitic power and the power pin is grounded. took the diagram from internet as example missed the missing line..

intruder82

I don't think the 0 shouldn't happen at all - in theory.

I have a setup that has worked for some years and suddenly had all sensors reading 0.0. I have no idea what causes this but it fixes itself if the power is turned off. Pressing reset is not enough and, while it is running OK at the moment, the fix may only last a few hours..

I have wondered if it is being caused by one sensor because I have some format glitch appearing too. it seems to return four decimal places, i.e. 23.4500 and the last digit has to go on the next line of the LCD. This only happens with one reading, and a reset is enough to fix it.

edit - I was wrong about 85, el Supremo is right
Maybe your problem is the same as my faulty power supply.. only my died completely and yours has not decided yet to die or not :)

Nick_Pyner

I am now beginning to suspect a sensor. It has been OK for a couple of weeks, but only today I found that, when I reset, one sensor reads about ten degrees too high and it slowly comes down to something saner. We are talking about standing temperature in a water pipe here, so I can make a guess at 25C. I have no idea how that would affect the other two. I am going to disconnect everything, replace all sensors and test the originals separately.

Go Up
 


Please enter a valid email to subscribe

Confirm your email address

We need to confirm your email address.
To complete the subscription, please click the link in the email we just sent you.

Thank you for subscribing!

Arduino
via Egeo 16
Torino, 10131
Italy