Sensor Variation

10 posts / 0 new
Last post
Walter
Walter's picture
Sensor Variation

I'm currently testing a new scale. It's sitting on a table in my office while I complete calibrations and work out a few bugs. Up until yesterday, only one DHT22 was working. Last night I corrected the problem and now both are working. They're sitting just a few inches apart so my expectation is that they should both display the same information. This is almost true for temp, only a 0.2F difference. With humidity I'm getting roughly a 10% difference. One sensor shows rh=53 while the other shows rh=58. I'm not sure what the tolerances are on these sensors or if this represents a problem? Perhaps it's a common issue and I'm worried about nothing. :)

Nate
Nate's picture
Humidity accuracy

Out of the box, Mine are quite far off as well. I'd say the humidity realistically +/- 20%.

A calibration is probably what you would need to do to get them in line... You can currently do this with HiveTool by adjusting the zero and gain- You cannot do it with HiveControl

Walter
Walter's picture
Is there any way to know what

Is there any way to know what value to set without a hygrometer?

Dennis
Erratic Humidity readings

My in-hive temp/humidity sensor was looking pretty good for the first few days I had it running. But now the humidity sensor is looking a bit whacky with continuous stream of erratic spikes. Anyone else seen this before? I can't tell if this is something to do with the bees jostling the sensor of if it means I've a loose connection or something.
McQ

Image: 
Paul
Paul's picture
Which driver?

Which DHT22 driver are you using? The DHT22 (wiringPi libraries) or the DHTxx (PIGPIO libraries)? WARNING the GPIO pin numbers change from RasPi to Broadcom when you change drivers. We are really trying to clean up the noise from bad readings. Image 0.7.4 adds the PIGPIO based drivers with more checks on data values. Using the DHT22 driver there were some weird timing interactions between the two sensors which have gone away with the DHTxx driver.

Dennis
Erratic Humidity readings

Paul --
I'm using the DHTxx on GPIO 27.

Paul
Paul's picture
What about the temperature data?

How does the temperature data look? Is the inside temp as erratic as the humidity?

Also, when you swap the two sensors, does the problem move with the sensor?

Dennis
Humidity Wierdness Fixed

Just to follow up on my earlier question about erratic humidity readings: We finally had a sunny day in the high 50's yesterday so I went into the hive for a quick look around. It turns out the bees had managed to shift the position of the DHT22. It ended up in a heavy traffic region. I suspect they were bumping into it or something. In any case, I shifted it up to a location above the brood, and the output seems to have settled down.

Image: 
Paul
Paul's picture
Outside/inside air mixing?

Very interesting. I wouldn't have expected this but it looks like it really was "good" data. A couple of things about the DHT22. It transmits it's data as 5 bytes: 2 bytes temp, 2 bytes RH and 1 byte checksum. If all five bytes add to zero, the data is supposed to be good, although I have seen bad data get through (like temperature readings of hundreds of degrees). Also, RH is driven from temp, so if temp is all over the place, RH will be too.

Here, temp is stable and reasonable. The RH readings are within a reasonable range too. You said it had moved to "a heavy traffic region". Notice from the first graph that shows inside and outside humidity - the peaks on the inside RH are about equal to the outside RH. So maybe what you are seeing are are currents from outside mixing with the internal RH of about 40to 50%.

So, the lesson here is to keep the internal probes away from drafts? Remember the amount of variance you see now. As they propolize it, the response will slow and it will look more like a straight line. Keeping porpolis from blocking sensors could be a whole area of research. (Wrapping in window screen is probably a bad idea.)

Paul
Paul's picture
Outside/inside air mixing?

Very interesting. I wouldn't have expected this but it looks like it really was "good" data. A couple of things about the DHT22. It transmits it's data as 5 bytes: 2 bytes temp, 2 bytes RH and 1 byte checksum. If all five bytes add to zero, the data is supposed to be good, although I have seen bad data get through (like temperature readings of hundreds of degrees). Also, RH is driven from temp, so if temp is all over the place, RH will be too.

Here, temp is stable and reasonable. The RH readings are within a reasonable range too. You said it had moved to "a heavy traffic region". Notice from the first graph that shows inside and outside humidity - the peaks on the inside RH are about equal to the outside RH. So maybe what you are seeing are are currents from outside mixing with the internal RH of about 40to 50%.

So, the lesson here is to keep the internal probes away from drafts? Remember the amount of variance you see now. As they propolize it, the response will slow and it will look more like a straight line. Keeping porpolis from blocking sensors could be a whole area of research. (Wrapping in window screen is probably a bad idea.)

Log in to post comments