Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[bkenobi]

Good call on the sun "cooking" the sensor.  I just thought about it and the side of the house I was planning on using is a south facing wall.  I'll have to think about whether that location would be a problem or not since it could possibly raise the temperature over the ambient significantly.  I suppose the best thing to do would be to get a 433mhz radio and a thermo-hygrometer and mount it on a different side of the house or just install the Arduino at the other end of the attic (facing north) which would avoid the problem entirely.

[Noam]

For me, it probably would be all that much harder for me to install the outdoor sensor on the north wall. The current location is only temporary (so I can make sure it is working). If I do my modified electrical box idea, it wouldn't be all that noticeable, even. My biggest concern is the brick. I'm worried that it heats up enough, and continues to radiate heat after the sun goes down. I'm worried I'll still get inaccurate readings as a result.

[bkenobi]

Fortunately (for this instance anyway), my house isn't brick so I don't have to worry about latent heat.  My old house had brick for the lower 4 feet in the front and I could definitely feel the temperature difference between the grass a few feet away to the cement walkway next to the bricks after dark (in summer)!  On the other hand, my deck in the back would get extremely warm after I stained it in the summer, but it would drop to ambient pretty quickly once the sun went down.  The thermal mass of wood and stone are obviously very different!   

[bkenobi]

I decided to go with the Arduino approach, so I ordered an Arduino UNO, 4 DHT11 sensors (one extra just in case), and 4 relay modules (2 extras).  These are all out of Hong Kong since the price is ~1/4 of what it would cost anywhere else and maybe less after shipping (US domestic shipping costs more than shipping from China for some reason).

Ironically, I ordered the last of the required components this evening and just now I got a response back from Oregon Scientific on the capabilities of the I300 and I600.  This is what customer support responded with:

Thank you for contacting Oregon Scientific Customer Support.

Yes, I300 and I600 units are capable of having 3 remote sensors.

So, if anyone else is looking to use a weather station for monitoring temperatures/humidity, the I600 is a good choice.  It's available from at least one seller for ~$20 on ebay and includes a temperature/humidity remote sensor.  That's cheaper than the sensor costs on its own!

[Noam]

What are you planning on using the relays for?
I have found that the Dallas DS18S20 and DS18B20 temperature sensors are close to my other thermometers (I have two Vernier "GoTemp!" probes attached to the same machine, sitting side-by-side with two of my Dallas sensors), but they can be off by a few degrees.
I have a 4-pack of the DHT-11's on order (from a Chinese seller on eBay, of course). I want to see how they compare in terms of accuracy, as well as use them to track humidity in the house.

I think you'll find the Arduino pretty simple to use, once you get started. Once I had everything working on a breadboard, I bought a prototyping shield, and soldered it up. That was the hardest part for me. I've been soldering stuff for many years (as the son of an EE, I was taught how to do it early on), but I didn't really have any experience soldering on perfboard (I usually asked my dad to do that part).
In addition to soldering up the shield (during which I made one mistake - I switched the input and output pins, so I had to fix that in the software), I also soldered up the board for the thermostat interface.

My program in the Arduino is pretty basic (since I'm not a programmer, either).
It checks the status of the three thermostat lines (pulled low with a resistor if off, or pulled high by the relay if on). It then turns on or off the appropriate LED for each one (I did the LEDs in the program, mainly so I could see if it was properly detecting the Thermostat cycles - and I liked it enough to keep it that way). Next, it polls the six temperature sensors. It then outputs all the data in CSV format via the serial port. It loops every 10 seconds (or so).

On the PC side, I'm using RealTerm to capture the data to CSV, adding a time stamp at the beginning of the line. I use the DyGraphs Javascript graphing engine to display the data. I'm working on a data logger that will go to a database, but I'm not there yet.

[bkenobi]

My setup will be independent of the PC once programmed.  I'm going to use the Arduino to watch the temperature and humidity in 3 locations.
1) Inside attic section 1
2) Inside attic section 2
3) Outside the house

I have 2 conditions that will be monitored by the Arduino:

• The Arduino will monitor the temperature data to see if either section of attic is above a set point.  If it is AND the outside temperature is xx degrees or more cooler, I will trigger the relay in that section of the attic which will turn on the corresponding attic fan.  I will either have it run for xx minutes and then turn off or I will monitor the attic conditions to determine if it should stay on (haven't decided yet, but it depends on the Arduino's capabilities and the sensor's reliability.
• The Arduino will monitor the humidity data to see if either section of attic is above a set point.  Just like temperature, if it is AND the outside humidity is xx%RH below, I'll trigger the relay.  Again, I'll keep the fan on for xx minutes or until the RH is below the set point or equal to the outside conditions.

My only concern is that the DHT11 is too low accuracy to use in this setup.  It says that it's ±4%RH.  That's means that the RH between two identical sensors sitting next to each other could be as much as 8%RH different, potentially.  If that's true, then I could see large differences in humidity and temperature from inside to outside that really don't exist.  I really don't know how much of a problem this will be, so I'm just ordering the parts in hopes that they will work.  The stuff I selected is all cheap enough that I'm not going broke if I don't use it.  The problem is that it could take a month to find out and another month to get a new set of sensors to try again.   

On the up side, this is all capable of being prototyped, so there's no soldering/desoldering/soldering/desoldering.... to find a workable solution.  It'll simply be plug in and swap the library in the sketch!  I'm pretty confident that I can get a sensor of some kind that will work, but the guaranteed part is far too expensive (SHT75 is $75 on ebay...for ONE).

[Noam]

One thing I've seen done is to use a CD-ROM audio connector (or similar) for the DHT-11, so that you can easily swap it out if you want.  The pin spacing should be correct for the four pins.

As far as the humidity sensitivity goes, you might not be able to be as exact as you want, but a ballpark figure might be good enough. Just set a differential that is far enough apart to account for the error.
I wonder if using two sensors inside and two outside, and then averaging the reading from each pair would give you a more accurate number.

[dhouston]

I wonder if using two sensors inside and two outside, and then averaging the reading from each pair would give you a more accurate number.

Search using ADC Oversampling and ADC Dithering...

• http://www.bing.com/search?q=adc+oversampling+resolution&form=OSDSRC
• http://www.bing.com/search?q=adc+dithering&form=OSDSRC

[Noam]

I wonder if using two sensors inside and two outside, and then averaging the reading from each pair would give you a more accurate number.

Search using ADC Oversampling and ADC Dithering...

• http://www.bing.com/search?q=adc+oversampling+resolution&form=OSDSRC
• http://www.bing.com/search?q=adc+dithering&form=OSDSRC

That would probably only help if the accuracy of the sensors is random. If the sensors ALWAYS seem to read high or ALWAYS seem to read low, then taking an average of a number of readings probably won't help all that much.

[bkenobi]

That's why I'll have to do some testing to see if these will be accurate enough.  If their readings match from one to the next, then I can use the readings as they are.  If there is a delta from one sensor to the next that is inconsistent when they are sitting next to each other, I'll have to re-evaluate if these sensors will work.  DHT22 are another option, but they are close to 4x as expensive.  The next step up is considerably more (~10-20x more than the DHT11 based on what I've found so far).

[systemdm]

Thought you guys would find this tutorial interesting on temperture sensing using the Arduino.


http://www.ladyada.net/learn/sensors/tmp36.html

[bkenobi]

I think I've read that page before.  I'm not too worried about temperature since there are plenty of cheap (or free samples) temperature sensors with great accuracy.  The problem is that there aren't many options for humidity.  And, the available sensors all seem to be digital which limits accuracy to whatever the circuit is designed for.  An analog humidity sensor would potentially provide better accuracy, but they seem to be more expensive by orders of magnitude for some reason.

[bkenobi]

I now have my Arduino and the relays from Hong Kong.  I'm just waiting for the DHT11 modules before I can start playing with code.

FWIW, I ran across a humidity module yesterday that seemed really affordable and is analog, but since it's just a capacitive module, the accuracy is questionable.  I read a project page where someone used it with a 1-wire setup and he seemed to have good results.  The downside was that it was complicated to build the circuit (his words).  He also indicated that the accuracy was questionable since it was not calibrated against anything.  He is using it for the same purpose I am planning though (delta humidity of inside and outside).  As such, as long as it's being used in a duplicate circuit, the results should at least be consistent.

The sensor is a Humirel HS1101 which, according to parallax has an accuracy of 5% but reports an analog value, so it can output values as precise as your A/D converter can handle (10 bit for Arduino).  Here's the info parallax has on this module:

http://www.parallax.com/Portals/0/Downloads/docs/prod/sens/27920-HS1101-v1.0.pdf

They include a simple circuit schematic and some code to get things up and running with a BASIC Stamp controller. 

Here's an implementation and documentation:

http://www.stillhq.com/arduino/000007.html

The modules seem to be ~$5/ea.  I suppose at that price and the fact that there's other stuff involved in getting it up and running (1 circuit/module), it'd probably be quicker, easier, and even cheaper to go with DHT22 for similar accuracy.  Just wanted to post an analog solution.

[Noam]

I got my four DHT11's the other day, and I hooked them all up on the same breadboard, for initial testing.
they seemed to be within 1 or 2 percent of each other, but I didn't have a way at the time to calibrate that against a known humidity value.

I only hooked one up at home, but I also have a separate humidity monitor in that room (one of the wireless ones I bought at RadioShack about 6 years ago). The two of them seem to agree within 1% or so, which is close enough for my application.
Next, I have to figure out how to mount the DHT11 and a DS18B20 (or DS18S20) on the wall in my upstairs hallway. I need the installation to look nice, but still allow airflow to the sensors to be able to get an accurate reading. I was thinking of mounting them inside a surface-mount phone jack, but that would look odd at eye-level in the middle of a wall, with nothing nearby.

[bkenobi]

Hang a picture and install them just above the frame?

I'm planning on playing with the DS18B20 temperature sensors, but what value do you see in installing them in the same location as the DHT11?  The DHT11 reports both temperature and humidity with only the 1 module.  Obviously, the spec sheets indicate that the DS18B20 is more accurate than the DHT11, so perhaps that's your reason.