i noticed that several SDS011 sensors send zeroed values from time to time…
any known solution ?
Hi,
It is no 0 values.
Oxidation or contact issue.
Oxidation or contact issue where ?
i started building these at mid January 
Even with newly built sensors you should expect the occasional measurement failure. The number of SDS011 errors is reported on the status page. You can compare that number to the uptime or number of measurements. Once every few hours is not uncommon.
thank you, so its not a bug, its an option…
Check all the connections between esp and sds011.
I’ve been running one sensor (SDS011/BME280) for more than 5 years now and have been collecting all the statistics from the status page. Below is a graph of the reported number of SDS011 errors as a percentage of the number of measurements since the last reset.
I regularly clean the connections and if needed replace the cables. As you can see there have been SDS011 errors since the start of collecting the data and the error rate is relatively constant. I haven’t investigated possible correlations with other factors but wiring can be safely excluded in my case.
Since I noticed an increase a few months ago I bought a new SDS011 and use that as an indoor sensor (similar board, also a BME280). It shows a similar error rate:
Note the drop the last few days. The only thing that changed there was a change of power supply. So a better power supply might help but there will still be errors.
I do think somehow that there still is some timing issue in the firmware and I seem to remember that Dirk did some research on that a long time ago. For me the failure rate is acceptable.
right now we use cables.
do you think it would be better to use a pcb and plug everything on it ?
Maybe it’s the result of already known issue with sleep mode of SDS011:
I have never seen this behavior, only errors on status.
We use PCB and add 2 type off capacitors for spikes and noise on the power (close to the connector as possible)
We have build more than 140 x SDS011 sensors
And what is the error rate on the menu-status?
My SDS011 sensor has 19 errors from 1300 measurements (1.5%)
well, every day at 00:00:00, i run a script which searches for records with null values the last 24 hours.
then, it imports the results into a table.
and this page, displays the table data…
you can select all or a specific sensor, and last day, last week, last month, from the beginning…
i erased everything from the table and we will check tonight at midnight
Have you checked the voltage (power requirement SDS011)
I have seen low voltage with some USB adapter and Long cable
What is the differences between the high and low error rate sensor?
take a look at this… it has only data from 2 days, but time is on our side…
I have two sensors with an SDS011. One is more than 5 years old and one 5 months old. Both with an acceptable percentage of errors (less than 2%). A few weeks ago however the new one started to show a very high error rate of 20% or more. Cleaning and re-seating/replacing cables and changing power supply didn’t help.
To investigate I changed the debug level from 3 (min. info) to 5. (max. info) and much to my surprise the errors completely disappeared (0%). This is reproducable on that particular sensor, setting debug back to 3 (min. info) and the errors start again, setting it to 5 no more errors.
Note that I run both sensors in Power Save mode. For the ‘faulty’ sensor it also did help a bit to disable Power Save mode but on the other sensor it makes no difference.
Since you’re collecting very nice statistics for your sensors I was wondering what happens if you change the Debug level on some of your sensors to 5. This can be done on a running sensor without rebooting. There is no guarantee that it will help but if it does help it might help finding a possible cause (other than the usual hardware related causes).
i changed my sensor’s debug level to 5.
i collect data every night, once a day, so take a look after 2-3 days..
1 Like
Thanks for testing. Looking at the graph there is no significant difference. Could you try for a while running the sensor in Power Save mode and Debug level 5?