I used to have a high degree of confidence in the accuracy of the purple air sensors. No longer. Back in March, they broke the system with one or more firmware upgrades, and apparently, they have not found a solution yet. I feel like we’re all victims of bait-and-switch because they USED to be accurate, and now they aren’t.
The temperature and relative humidity effects for the PurpleAir monitors have been known for several years. Two years ago I ran indoor and outdoor PurpleAir monitors alongside Hobo Temperature and RH models with supposedly 0.2 oC resolution for a year or two. The temperature increase (possibly due mostly to the small fan but also perhaps the other electronics) averaged about 8 oF as many of these commenters have been noting. The RH was more interesting. Of course it comes down as the temperature goes up, but the amount it came down was different in the indoor and outdoor monitors. RH indoors was lower than the Hobo by 15% on average, while RH outdoors was lower by 25%. For example, despite there being a reasonable amount of rain in my location (Santa Rosa) the absolute highest RH reported by my outdoor PurpleAir monitor was 70%.
This topic has been discussed before here and I included the details of the findings, but I’m not sure where to find that writeup.
Hi Lance,
The temperature and relative humidity effects for the PurpleAir monitors have been known for several years
Yes, that’s true, but that’s not what this post is about. If you read the original post, the temperature jumped by 15 degrees within one sample period and stayed that way. This isn’t PurpleAir readings being different than more accurate devices, this is PurpleAir readings being different from itself. This was clearly called out as a bug in the firmware by PurpleAir staff who have commented in this thread, and appears to only affect certain versions of the PurpleAir hardware.
The bug is still not fixed, and my outdoor sensor still consistently reads 6-8 deg F warmer than other PurpleAir sensors within 0.5 mi. If you have any insight in how to fix this firmware bug, please let it in this thread.
Hi Garth–
Sorry to hear about your persistent bug. Yes it has nothing to do with the higher T and therefore lower RH readings compared to ambient air. I have no solution for you… Maybe you could offer to trade in your device, which has value in that it’s been so consistent all this time, so the PurpleAir engineers could play with it knowing it’s crazy.
Garth, currently the only fix is to roll the firmware back to 7.02. This will return the temperature readings to their old settings. Be aware that this may cause the VOC issue from version 7.02 to reappear.
If anyone would like their sensor’s firmware rolled back, please email us at contact@purpleair.com with the device ID of your sensor.
What is the VOC issue, please? It would help me to know what might be a problem from the rollback I did.
Thanks, William
Prior to the addition of the BSEC2 drivers, which were added in version 7.03, the VOC value sometimes would get stuck at 25. Thus, this may occur on a sensor that is rolled back to 7.02.
Sorry for being so ignorant, but what does VOC mean and where might it be viewed?
No worries, William. VOC stands for volatile organic compounds. The BME unit (temperature, humidity, and pressure sensor) in the Flex, the Zen, and the Touch includes a VOC sensor. As it is now, the sensor does not differentiate between types of VOCs. It reports the Static IAQ (Index for Air Quality), which is a health metric based on the VOC readings.
This VOC data is currently an experimental feature in PurpleAir sensors. We are working on fully understanding the readings and how to implement them into PurpleAir devices properly. You can read more about this at the following link: BME680 VOC Scale Change on the PurpleAir Map.
This data can be viewed by changing the data layer on the map: Environmental Data Layers.