Of dehumidifiers, moisture and vapour barriers - experiments

I bought a DHT22 temperature / humidity sensor and wired it up to a Raspberry Pi computer. Thegeekpub.com has a good write up on how to do this. It's basically connected right to the GPIO pins, plus a 10k pull-up resistor. I used an old IDE ribbon cable to connect to the 40-pin header on the Pi, though I had to drill out the missing hole on the IDE connector that used to ensure proper orientation on the drive.

The trickier part was getting the right Python library installed to read it easily.

My goal was to monitor temperature and humidity in the basement, but having a temperature / humidity logging set up, I set out to do some experiments with a dehumidifier. I put one of my dehumidifiers in a basement bedroom that we don't use much, set up the monitoring set up, including a Kill-A-Watt power meter, with a Raspberry pi camera module aimed at it.

I let this run for a few hours with the door closed to gather data.

The above graph shows time in minutes. The green line is power/10. The dehumidifier turns off the compressor from time to time, probably to defrost the heat exchanger, because there is frost on it. I had to enter these values from pictures, so I only fully entered the first hour or so, and only overall level after that, and not at all after the first two hours. The scale on the bottom is in minutes.

The purple line is relative humidity. You can see the relative humidity cycle as the compressor turns on and off. I'm pleased that the sensor is responsive enough to catch those cycles. Humidity changes lag compressor on and off by a minute or two because the compressor first has to cool down the heat exchanger that cools the air before it captures any humidity. After the compressor turns off, it still captures a bit of water on it before the heat exchanger warms up again.

The dark blue line is temperature times 3 (to make it fit on the graph). You can also see the temperature go up slowly whenever the compressor is on. It only goes up by a fraction of a degree. The temperature sensor resolution is 0.1 degrees, and you can see it step by those increments.

For the first half hour the humidity drops quickly, then, every time the compressor turns off for a defrost, it rises up again quickly, until most of the drop is reversed while the dehumidifier compressor is off.

Because it blows air over the heat exchanger to thaw it, I figured it must be re-evaporating most of the water it condensed each time, so it's just sitting there capturing and re-evaporating the same water. So after nearly five hours, I unplugged the dehumidifier just before it hit a defrost cycle so the water wouldn't get re-evaporated (see vertical line on above graph).

But the humidity still rose just as fast as it did during a defrost cycle, and kept on rising, approaching some sort of equilibrium humidity close to what it was before I started the experiment.

Above, a longer term graph. I turned the dehumidifier back on. I also manually entered more of the power levels the next day. After a while, I left the door to the room open, which immediately raised the humidity in the room and increased the compressor on duty cycle.

But for the initial run of five hours with the door closed, I got 710g of water in the bin. The room has a volume of 43 cubic meters. 43 cubic meters at 18°C can only hold about 645g of water at 100% humidity.

Even if I dried the air from 100% humidity to zero, I could not have captured that much water, and the humidity only dropped by 15%. So there are other sources of humidity.

So where could this extra humidity be coming from?

I think a lot of the humidity comes from the concrete. This house was built in 1964, they didn't install vapour barriers under basement concrete slabs back then.

But if humidity is really coming up through the concrete, I should be able to see signs of it.

They do make concrete moisture meters that work on the same principle as contactless wood moisture meters, so I tried my wood moisture meter on the concrete floor in the unfinished part of the basement, and it seems that areas of the floor covered with things like buckets had slightly higher readings, suggesting that moisture is coming up. But the readings were not consistent enough to be conclusive.

I read that the way to measure concrete moisture is to cover it in vapour barrier for 72 hours and measure the humidity under the vapour barrier, so I put some plastic down, with a cheap (and not very accurate) moisture meter under it. After a day, it read 90%. I then took it out from under the plastic and put it on top of the plastic, and, letting it settle for an hour, it read 70%. Actual humidity in the basement was 55%, so the meter is not very accurate. But it proves that the equilibrium moisture of the concrete floor is higher than the air in the basement. Which is to say, humidity IS coming up out of the concrete floor.

But a real "smoking gun" is these fluffy fuzzy crystal formations I see in a few places on the floor.

This is called "efflorescence". Moisture in the floor brings salts and minerals from the concrete to the surface, and as the moisture evaporates, the salts form crystals. There wasn't any wet or even damp spot on the concrete, nevertheless, it brings salts and minerals with it. These fuzzy crystals only form if the concrete isn't actually wet.

Sometimes, with the snow melt, I got water intrusion into my big garage workshop at our last place in the country, and after the concrete surface dried, as it dried, I'd get some of these fuzzy crystals forming there as well.

If moisture is coming up through the concrete slab, presumably it may also come in through concrete walls.

But if you finish a basement, and add a vapour barrier on the inside like what is normally done for houses in Canada, that just traps moisture inside the wall!

Houses built in the last 30 years will have a vapour barrier under the concrete slab, and also some moisture sealant against the outside of the foundation, but that doesn't guarantee that no moisture comes in. And even vapour barrier is slightly moisture permeable, unless its made of metal. So if anything, if you finish a basement wall, the vapour barrier should against the concrete wall, behind the insulation, not in front of it! And reading up on it, vapour barrier is not recommended for basement walls.

Next I moved my monitoring setup to the main unfinished part of the basement and did more experiments. The "GoldStar" dehumidifier that I was testing was the most efficient dehumidifier I brought from the old house, along with my second best one. But a big dehumidifier was left behind in the house, and this one pulls more than 1.5x as much water out of the air per hour as the "GoldStar" one while only using slightly more power. And another GE dehumidifier pulls less than a third as much water out of the air but also uses 400 watts.

So there is a dramatic difference in efficiency. And a 400 watt appliance uses about $1.25 of electricity per day, so buying a newer more efficient dehumidifier can easily pay for itself in electricity costs.

This was another one of those topics where I really dived into a rabbit hole of investigation, and not the first time I investigated some aspect of a dehumidifier

All experiment were with a fan running to mix up the air. I later turned off the fan and moved the sensor up and down. It's fascinating how the air really gets stratified into layers. So having the dehumidifier near the floor is best, and having a dehumidifier that blows air up in a stream also helps to churn up the air. The big dehumidifier unfortunately doesn't do that. But problems from excessive moisture area always near the floor, so having the air churned up to also dry the air at the floor is important.

Even something as mundane as using a dehumidifier can become an interesting scientific exploration!

Two weeks later I bought a BMP280 sensor, which does temperature, humidity and barometric pressure. Overall, this seems to be a much better sensor. The DHT22, I have discovered since, has about a 5% hysteresis inaccuracy on relative humidity. Also the DHT22 only reads temperature to tenths of a degree C, which turns out to be a limiting factor when looking at very subtle changes.

I used the BMP280 sensor to characterize our fridge, to see how much difference dusting out the coils made. It was fun figuring out some of the intricacies of cycles the fridge goes through just from the graphs I plotted.

More home improvement projects on woodgears.ca