Humidity Bounce

M

MobileMe

Active Member
I have two boards setup in my basement. Both have humidity sensors on them. One for the basement to so I can control the dehumidifier. The other is upstairs so I can control the humidifier on the furnace during the winter.

I never noticed it before, but there is about a 3% bounce on them. I didn't notice it because I setup a 15% range to turn the dehumidifier on/off. I recently installed the one upstairs and saw the humidifier turn on/off more than I would like. I gave it a 6% range to see if it helped and it did, but not much.

Is there a way to get the humidity reading to be more stable? I would like to avoid putting in averaging Plc code. Is there a hardware solution?

Thanks,
Mike
 
Hi Mobileme,
 
Humidity sensor is using same ADC logic to read the output from Honeywell analog output.  All current shipping boards have the ADC noise fix already. Thus the humidity sensor reading fluctuate is no longer hardware related, if your board already have ADC fix which is easy to see on the ADC output display.
 
Humidity sensor itself is very sensitive to the hiumidity change. In our lab  testing, if move the sensor above the desk from below the desk, the reading will change.  Since Honeywell claimed they calibrated each sensor, we assume the analog output from it is accurate.
 
If your WebControl board has the ADC fix, which is modified R10 resistor value, then your analog input should steady at zero when nothing attached. You can use analog input to verify that sensor output, just connect a wire from middle pin of humidity sensor to another analog input, you will see the ADC change should reflect the change in humidity reading. The only difference is humidity input also going through a software calculation to convert the analog input into humidity value.
 
The board is a month old with the most recent version of firmware. I connected the middle pin on the humidity sensor to an analog input. For the most part it read zero, but would spike up to 15 when the humidity changed up or down. The other analog inputs are very stable. I must me missing something. I do see the humidity start to vary more when an output is on.
 
Humidity sensor center pin should not read zero, since most time humidity is from 20% to 90% depending where you live. If your humidity sensor most time reading zero, then that sensor need to be checked out.
 
TTL output is not connected to humidity sensor at all. 5V power is regulated, so that TTL output should have no impact on the humidity sensor operation. 
 
Turns out I was using a gauge of wire that was to big. The analog input is bouncing allot when I connect it to the middle pin on the humidity sensor. I would agree that it is the sensor if it was only on one board. Both boards have current firmware. The humidity sensor on one board is directly on the board itself. The analog ports 1 through 3 are very stable which makes this a strange issue. I'm not sure where to go from here.
 
Honeywell's humidity sensor is very sensetive to humidity, as well as to light.  Do you have light shine on it that changing?
Do you have another sensor you can try to see if the problem exist for the other sensor?
 
They are not in any direct light.  I have two humidity sensors on two boards.  I swapped out the sensors and still have the same result.
 
Because using another analog input detected the humidity sensor output line having noise, and you also tested another humidity sensor with same result, the only thing did not replace and test is the wires leading to the humidity sensor.  Is that possible to put sensor next to the board without going through the wire harness you have, so that you can verify the problem is not in WebControl board and not sensor?
 
Actually, before sending the sensor, could you please do one more check.
Using analog input, instead of connecting to the humidity sensor center pin, connect it to the 5V pin, to see if you get steady 5V.
There are two regulators on board, one is 5V, the other is 3.3V. If 5V is half broken, its output may not stable, that could cause the temp sensor and humidity sensor not stable, too, since both of them are powered by 5V supply.  CPU and ADC is powered by 3.3V regulator, so that half broken 5V may not affect ADC measurement.
 
To prevent damage to the 5V regulator, make sure power off the board before doing any work. Check the wiring again before power on.  Most of us do that way when we working on 110VAC, so that we don't get into trouble ourselves. When work on low voltage, changine wires with power on will not harm ourselves, but it could damage the electronics components.  When we handle WC boards and components, we always put on anti-static strips, always power off before doing any wire change, even testing on our lab bench.
 
My 5V supply looks fine on the volt meter (5.1V) the temperature sensors are working great.
But, I noticed a lot of bounce in my RH sensor reading. (up to 5%)
Should check the 5V supply with an Oscope? See if there is any noise on it?
 
Thanks,
Rich
 
Version: v03.02.09
 
 
 
 
 
 
XRinger said:
My 5V supply looks fine on the volt meter (5.1V) the temperature sensors are working great.
But, I noticed a lot of bounce in my RH sensor reading. (up to 5%)
Should check the 5V supply with an Oscope? See if there is any noise on it?
Click to expand...
 
Things I'd do.
Get two resistors, say 1K each although the value isn't critical.
Remove everything else from the board. Put one resistor in the humidity connector between pins 1/2 and the other between pins 2/3.
Run the board back up and observe the "humidity". Probably won't be 50% exactly, but see if it is bouncing around or not.
 
If it isn't, then it sounds like the output from the sensor itself may be noisy. I'd try a 0.1uF cap from ground to sensor output.
 
If it IS then try adding a 100uF cap across the humidity terminals 1-3 (ie, across the supply rail).
 
Some of my older boards had this problem and I found additional filtering on the power supply improved it.
 
Humidity sensor is not really linear.  There is a calibrated logic behind the humidity sensor input, which is not 2.5V for 50% humidity.  From our test, we found out when humidity sensor is about to fail, it will have reading bounce up and down a lot. If you have another sensor to replace it, you can see if new sensor still doing the same.
 
I already had a 0.1 tantalum and a 4.7 uF across the sensor input. I just added a 100 uF to 1 & 3.
Right now, it's cycling up and down 3% and some times 5%. 
Maybe tomorrow I'll try putting the voltage divider on on it. If it goes stable, I'll know. 
I've got a feeling it's the sensor. It wasn't cheap, so I didn't buy a backup.
 
Even with the bounce, it still works well enough for my project. 
Hopefully it will keep working for a while longer.
 
Thanks for the replies,
Rich
 
You must log in or register to reply here.

Similar threads

H
New and Old Somfy Sonesse 30 Motors
Replies
1
Views
144
HomeTechNewbie
H
A
Elk M1G with two additional buildings
Replies
2
Views
121
amacphee
A
M
3-way HAI switch seems to have gone bad - help needed
Replies
5
Views
104
ano
A
M
Can I Automate Force Arm?
Replies
3
Views
1K
Mark S.
M
A
New Inovelli Smartswitch Sounds Interesting
Replies
0
Views
739
ano
A
Back
Top