Home  5  Books  5  GBEzine  5  News  5  HelpDesk  5  Register  5  GreenBuilding.co.uk
Not signed in (Sign In)

Categories



widget @ surfing-waves.com


Vanilla 1.0.3 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to new Forum Visitors
Join the forum now and benefit from discussions with thousands of other green building fans and discounts on Green Building Press publications: Apply now.




    • CommentAuthornigelm
    • CommentTimeApr 2nd 2017
     
    I am about to fit a new insulated roof to our barn and wanted to monitor the moisture levels inside the roof.

    The roof build up from inside out is:-
    20mm timber board
    Vapour barrier
    300mm timber I beam filled with mineral wool
    20mm wood fibre sarkling board
    Batterns and tiles.

    We have a WUFI model that indicates all should be well but i wanted to fit sensors to compare real world against the model. I have some BME280 combined temperatute, RH and pressure semsors but wanted some advicde as where to install them.

    The stone walls are to be imsulated as wel with 300mm Rockwool batts on the inside and will monitor those as well.

    i
  1.  
    ''The stone walls are to be imsulated as wel with 300mm Rockwool batts on the inside and will monitor those as well.''

    What did WUFI say about that? That sounds like a very cold wall!
    •  
      CommentAuthordjh
    • CommentTimeApr 2nd 2017
     
    Wherever you fit the sensors will be wrong! Also, wherever you fit them, they will go wrong! I'd fit them just above and below the VCL, and just below the sarking boards. In at least two places, with at least one at the ridge (assuming you have a ridge).

    PS Those are interesting looking sensors. I wish they'd been around when I installed mine!
    •  
      CommentAuthorSteamyTea
    • CommentTimeApr 3rd 2017 edited
     
    Posted By: djhPS Those are interesting looking sensors. I wish they'd been around when I installed mine!
    I have just got some (BME280), not got around to wiring any up yet. Got to work out how to get them to sit on the same I2C GPIO as the real time clock

    Does WUFI show any areas of low temperature or high RH?
    If it does, then fit the sensors there.

    Alternatively, place them in random locations, then model from there. Or you could try and work out areas that may pose a problem i.e. junctions, interfaces, high and low airflow areas, north and south facing etc.
    • CommentAuthornigelm
    • CommentTimeApr 3rd 2017
     
    We have modelled the walls as well as the roof and both should be fine. Interestingly we tried both an intelligent membrane and conventional membrane as the VCL and the conventional membrane performed best.
    We are however in France and the weather patterns are different here.

    The sensors are working on an Pi I2C, only two addresses available but the PI zeroW are cheap enough to deploy. The sensors came direct from China at less than a euro each.
    •  
      CommentAuthorSteamyTea
    • CommentTimeApr 3rd 2017 edited
     
    Diverting slightly, I think you can write an address to the 280's and then read them individually. What I am looking at later (after I have done my running around)
    https://stackoverflow.com/questions/17687132/multiple-sensors-on-raspberry-pi-gpio

    http://www.robot-electronics.co.uk/i2c-tutorial
    • CommentAuthornigelm
    • CommentTimeApr 4th 2017
     
    Looking at the data sheet for the BME280 it looks like only two addresses are available selected by setting the address pin to +v or 0v.
    The adafruit I2C multiplexer looks interesting, 8 channels giving potential for 16 sensors.
    • CommentAuthorEd Davies
    • CommentTimeApr 4th 2017
     
    Agreed, one address with SDO high, the other with it low. It might not be obvious how to set the SDO pin, though, depending on the breakout board in use. For the Adafruit board you just ground SDO to get the other address. For the ones Steamy has (which appear to be the same as the ones in the vAir CO₂ monitor if you select the option of having a pressure sensor) it might not be so clear - would need to trace the connections out.

    That multiplexer looks neat but you could avoid the extra hardware by just bit banging other GPIO pins to give more I²C buses. High data rates are hardly critical to monitoring temperatures and humidities in walls.

    Steamy's considering using SPI as it has a chip-select line which you could run off other GPIO pins. It occurred to me today that you could maybe actually use the SDO pin as a chip-select line for I²C. Ground it for the device you want to talk to then address that using the alternative address, leaving all the others on the bus at the default address. Assumes the SDO pin is dynamic - it doesn't just sense its address on startup then stick with it.
    •  
      CommentAuthorSteamyTea
    • CommentTimeApr 4th 2017
     
    After a bit more reading and chatting over at the other place, it seems that SPI allows for longer cabling, I²C seems limited on length. So it may be easier and cheaper to just connect a couple or 4 of them on short cables to a Pi Zero and then get an ethernet adapter and CAT5 back to 'another' computer to do all the interesting stuff on.
    • CommentAuthorgoodevans
    • CommentTimeApr 4th 2017
     
    I suggest ensuring an outside sensor (sheltered) and one or more inside sensors - without that data running in parallel with the 'in fabric' sensors your have no real idea where the moisture is coming form - or how long it takes to get there or what is the best level that could be achieved. Also before installation have all the sensors in one place and calibrate them with each other under various conditions.

    I wager that so long as the vapour barrier is relatively intact that....
    a) the absolute humidity of a sensor just outboard of the vapour barrier will closely track the outside absolute humidity and
    b) the absolute humidity just inboard of the vapour barrier will closely track the internal absolute humidity (provided there is an air gap between the 20mm timber board and the VCL - and the 20mm board is not designed as an air barrier)

    (derive absolute humidity from the RH and Temp - in any units you like (grams/M3 is my favourite it makes for easy calcs for the effect of ventilation)).
    • CommentAuthornigelm
    • CommentTimeApr 5th 2017
     
    I was going to track the inside and outside temperature/humidity as a reference.

    Having checked thesensors there calibration looks good or at least consistent against each other.

    What is the relevence of absolute and relative humidity in relation to the building fabric?
    •  
      CommentAuthordjh
    • CommentTimeApr 5th 2017
     
    Posted By: nigelmWhat is the relevence of absolute and relative humidity in relation to the building fabric?

    Absolute humidity is the density of water molecules in the air (or actually in the space that the nitrogen, oxygen etc also occupies). It has physical significance and has pretty strong conservation properties. In particular, if you take air from outside and bring it inside, it will still have the same absolute humidity, but its relative humidity is likely to be be quite different.

    Relative humidity is to do with the relative rates of evaporation and condensation at a particular temperature and pressure. It's a much more nebulous concept and more complicated to deal with. But it is of great importance because it is largely relative humidity that matters to living things. How comfortable you feel, how well fungus grows (whether mushrooms or rot in wood), how comfortable mites feel, how well bacteria grow and spread all depend on what the relative humidity is. As does the moisture content of wood etc etc.

    https://en.wikipedia.org/wiki/Humidity has some more info.
    • CommentAuthorgoodevans
    • CommentTimeApr 10th 2017
     
    I had a thought today - If accurate temperature/RH within the structure is required/of interest within your fabric make sure your sensor does not significantly heat the measurement zone - isolate the sensor from the CPU unit (have the cpu in the living area), perhaps have the sensor powered off or at least idle for significant periods of time. Test the effects of temperature by embedding a sensor in a block of eps and comparing its temp with one outside the eps. 1 watt of heat output inside a wall with a U value of 0.1W/m2 will change the temperature inside the wall by several degrees over a significant area and render your RH values meaningless (but not your absolute humidity!) - even milliamp sensors will get warm in eps. As RH is of interest, accurate temperature should be also.
    • CommentAuthornigelm
    • CommentTimeApr 10th 2017
     
    The bme280 sensors use 3.6uA at 1hz. I believe the idle current is even lower so hopefully local heating of the sensor will be minimal.

    I have been thinking about read frequency, I am guessing that once every 30 minutes would be sufficient.
    •  
      CommentAuthorSteamyTea
    • CommentTimeApr 10th 2017 edited
     
    It costs nothing to store data. Find out what the local weather stations are collecting at and match them.
Add your comments

    Username Password
  • Format comments as
 
   
The Ecobuilding Buzz
Site Map    |   Home    |   View Cart    |   Pressroom   |   Business   |   Links   
Logout    

© Green Building Press