OmniStat humidity readings off, two different units

[lupinglade]

Just installed two new OmniStats and the humidity readings seem to be high by about 10%, compared to multiple other readings including readings in the same area by an HAI Extended Temp/Humidity sensor. The hole in the wall is well sealed w/ spray foam. I know for a fact the reading is 35-40% RH but the OmniStats say 50%. I have an older OmniStat that also had this issue initially but seems to have become accurate after the wall hole was sealed w/ spray foam about a year ago. Is this because the thermostats are new and the humidity sensor needs to dry out or some kind of a bug? The new thermostats have been installed about 48hrs ago.

 

[Desert_AIP]

There's a firmware calibration offset for the temperature.
I can't rememebr if there is also one for the humidity, may be worth looking into.
Do you have a standalone Omni temp/humidity sensor you could place next to it to compare (apples to apples)?.

 

[lupinglade]

Unfortunately there isn't one for the humidity. I have placed multiple sensors near the thermostats and they all consistently read 40 and under (37,38,40), the OmniStats read 50-52. That's a big difference when trying to control humidity in an area.

 

[lupinglade]

And yes one of the sensors is an HAI one connected directly to the controller.

 

[Desert_AIP]

I wonder what the rated tolerance for those sensors is.
My guess is a few percent.
Is there a difference in "real humidity" in a 10% increment?

There may be a difference in technique/sensor too.
I think most sensors are capacitive.
Some sensors have a scaled voltage output, others have a sliding frequncy oscillator.
It could be that the frequency conversion is off a bit if a stacking of tolerances in the components occurs (I've seen a DIY sensor using a 555 timer and the capacitor dependent frequency to PW converter).
I would hope there is a calibration of the sensor, but they may have not done it as a cost savings measure judging absolute accuracy to be "unecessary" (guess).

I guess you could recalibrate your program knowing the offset is fairly constant.

 

[ano]

I was going to say that if you are planning to control something like a bath fan with this, you are going to be disappointed. The humidity sensor is very slow to change and reading the humidity is a bit of an inexact science. If you are looking for high precision and exact repeatablity, I don't think this is going to work great for you the way you expect.

Senors for a HAI panel produce a 0-12V signal which the panel reads with 8-bit (0-255) resolution and then it converts this voltage into a humidity number. You just need to find a way to reduce this voltage a bit and your humidity will read lower.

 

[lupinglade]

The problem is that the humidity sensor connected to the panel is actually accurate but the OmniStats are not. I don't see how I can reduce the voltage on the sensor since its soldered on. I also don't mind if the sensor is slower to react (I haven't had this issue though), its the large inaccuracy that is a problem. The strange thing is that the older OmniStat had this issue too but somehow eventually I was able to resolve it or it resolved itself (not sure which as I tried many things). It seems to me like a firmware bug, because all OmniStats seem to have the same issue at some point. I wonder if its an issue with Celsius calculations or something?

 

[ano]

You don't change the sensor voltage at the sensor, you adjust the voltage by adjusting it where it connects to the panel. Depending upon how electric literate you are, it might be as simple as adding a few resistors, one in series and one to pull the voltage down slightly. Or a 1K potentiometer to adjust it slightly. It might take some trial and error.

 

[lupinglade]

That is not possible since its an OmniStat and not a separate temperature sensor. The information is sent over to the HAI panel digitally (serial) via the control connection.

 

[lupinglade]

Would be nice if someone from HAI would help out as I keep running into this with every HAI thermostat. They all read 10% high for humidity...

 

[ano]

Your right. that won't work with an OMNIstat. I wonder if you opened it up you'd find an adjustment inside? When they build these things they have to adjust them.

 

[lupinglade]

Unfortunately no adjustments. I have no idea how these thermostats are going to be able to humidify/dehumidify properly when the amount of correction needed is offset by the 10% error. The space they are controlling is already over-dehumidified...

 

[DIYHO]

Would be nice if someone from HAI would help out as I keep running into this with every HAI thermostat. They all read 10% high for humidity...

I also can verify both the RC2000 and the extended range humidity/temp sensors humidity readings are off, but not too high, too low.

I was trying to determine yesterday why my humidity sensing fan wasn't turning off, so I broke out my Kestrel 4500 to take some humidity readings. What I found was the humidity was actually 15% higher on average than what was being reported by the RC2000 and the other sensors.

I also placed my Kestrel directly on top of the RC2000 with the same result .... approximately 15% higher actual humidity reading than was being reported by the RC2000.

I could understand a +/- 5% variance for these HAI units given they are not intended as high precision instruments, but +/- 15% is entirely unacceptable.

 

[lupinglade]

Exactly. HAI needs to add a calibration offset option for the humidity sensor, otherwise its useless in many cases.

 

[lupinglade]

Just got a replacement unit from HAI. Same problem...10% high again