HB HVAC noise supression

[elcook]

I'm setting up an 1-wire system that has HVAC monitoring boards a fair distance (~100ft) from my computer. All is fine with the HVAC units disconnected but when the HVAC inputs (std 24vac sources) are connected, I get numerous exceptions (running with Java software) and frequent sperious readings. Several experiments have definitively located the source of the problems as the HVAC inputs. Tests show that the long cat5 cable is not the problem. A common between the 1-wire system and the HVAC inputs has been verified. Although I don't have the instrumentation (oscilloscope, etc) to quantify it, I suspect the HVAC signals are rather noisy. The air conditioner turning on and off creates the biggest problem.

Any suggestions on a way to clean up the HVAC input noise?

I suppose I could go to the 8 channel I/O board which has optoisolators on the input side. If I do switch to the 8 channel I/O board, anyone have some Java code for it?

Thanks in advance.

 

[wilafau]

I have a similar setup. I probably have 75ft between the computer and HVAC board. I then have about 10ft of cat5 connecting the HVAC board's inputs to the HVAC unit. I was lucky in that the HVAC installers used 8 conductor to run to the remote thermostat. That left a couple of spare wires so I mounted a 18B20 temp. sensor in the thermostat and then used the extra wires to bring the data signal back to the HVAC unit & then to the H-B HVAC monitor. This way I can monitor the temperature right at the thermostat while recording the AC/Heat/Fan signals as well. I used the 4th input on the HB unit to connect before the pan's float cut-off switch - turns out the AC signal from the thermostat runs through this float switch. By reading both sides of this switch I can compare the values and if they are different, then I know my pan is filled with condensate and I have a problem.

I too have noticed erroneous readings however it wasnt with the HVAC inputs, it was with the daisy changed 18B20. I noticed that this only happened if my script was reading the sensor at the very time the HVAC AC kicked on. I suspect Noise. Over several weeks of monitoring 1 sample per minute I only encounter this bad reading about once per week. When my script detects a out-of-wack reading like this it simply uses the last valid one as a substitute.

I am not that comfortable tying the HVAC ground and the 1-wire grounds together, but that is the way this board is designed. I would think that optoisolated inputs would be better and reduce induced noise as well.