I’m having a hard time locating a low pressure, low cost air pressure sensor to mount into my HVAC system. Without the sensor, I’m not comfortable going fully automatic with WHIA. Most of the sensors I’ve found either cost $800 US or require some pretty bizarre plumbing. I’ve thought about using an anemometer type setup inside the main line, but I’m not interested in flow rate as much as I am pressure. I’ve received some excellent suggestions from some very helpful folks on twitter, but I think I’m going to have to end up home brewing this one. My current theory is that if I cut a hole in the main duct (!) and mount a piece of flexible material over that hole, I should be able to measure the amount of deflection and make an estimate of the internal pressure based on that. I’m going to do more research because this makes my geek sense tingle, but that is at least an option.
WHIA seems to be working like a champ. The vents in the master bedroom for example are pretty much shut all of the time, as is the basement. The main floor is now a fairly consistent temperature and is well within my 1.5 degrees goal. The second floor continues to be a problem. We have an 18 foot vaulted ceiling between the first and second floor that peaks in our office/loft. The office is hot, but I expect it to be. All of that heat from the first floor has to go somewhere and it might as well move on up (yes, just like george and weezy). The problem is that I can’t seem to get the upstairs bedrooms to cool. Those rooms are running full tilt, vents wide open and most of the HVAC air is being forced into there, but I’m still off by 3 degrees. Ceiling fans help the perception of the temperature out a lot, but I’m still trying to work this one out. I’m starting to think that I just may not have enough BTU capacity in those rooms. Not sure how to get around that, but will be a fun problem to figure out!
I figured out what was going on with my RRD graphs. I was just telling rrdtool to graph the wrong PDPs. I fixed that and took the opportunity to recreate the rrd’s to support a longer timeframe at a higher resolution. I finally ran the living room sensor which is located very near the house thermostat and found that the house thermostat is off by about 2 degrees. I trust the DS18S20P sensors far more then that thermostat simply based on my experience with them to date. My outdoor sensor for example is never more then 1 degree off from what the weather underground station a few miles from my house lists.
I’m broke for the next few weeks, so I’m working on designing the PCB for the next version of WHIA. I’m thinking about either using an Arduino MEGA or using multiple Arduinos in an I2C network. Can’t decide yet. The I2C network does have some sex appeal, but the MEGA would be a lot easier to do.
In between tasks I’ve been able to implement a few new WHIA features. I added LED lights to each sensor, and then removed them because they sucked. I’ve started attaching servos to some of the more inaccessible vents in the house which has been…. fun. Some of these things barely have enough space to open and close never mind get cleanly actuated. I’m also having a bit of trouble with my RRD graphs. The daily graph works great, but my weekly only seems to want to show a single days worth of data. I’ve also done some research on radiation shields and it looks like I’m going to have to build the pagoda for my temp sensor.
OK, keeping a sensor from getting false readings due to radiant heat is a serious PIA. I refuse to buy something premade, even if I could find it, but I’m starting to come up with a lot more “doesn’t work” then “almost works” solutions. I’m taking a little break over the next few days to get some honey-do list stuff done (landscape, sheet rock, paint, etc) but more soon.