Cold Spots

[MikeGrahamT21]

First really cold morning of season has allowed me to go hunting with my thermal imaging camera for cold spots, and as is usual i've found quite a few, but what have others done to sort these type out?

 

• Ceiling joists near the edges of the house, show a temperate difference of about 4C, they were 15C this morning, but the PB without timber behind was 19C. I went up in the loft, and checked the timbers were well covered, and they were, so I can only imagine it must be bridging in from the outside, though even from that side I've run 100mm wool insulation right up the wall and over into the loft space, the timber is still exposed though, as its what the fascias are attached to (rafters), not really sure what else i can do for this aside from internal insulation on the ceiling which isn't going to happen.
• The bottom corners of reveals, both windows and doors - Our triple glazed window in the bedroom is performing well overall, glass was showing 17C (Room temp 17C as I've not had heating on), majority of the frame was 16C, but just on the bottom corners I recorded as low as 13C. I've just recently done all this internally, so I know its correctly foamed around the frame, and there are absolutely no air leaks. In our house (1968) the reveals are closed with medium dense block, so the cold will get drawn through. I'm guessing the answer to this one would be external wall insulation? (its on the cards, but too expensive at present!).

 

Has anyone else had some really awkward cold spots to deal with? I know i'm never going to get rid of them all on a refurb, so its just a case of minimising the ones we do have. This afternoons job is to take some roof tiles off and see if i can see a reason the wall edges (party wall joining external wall) are so cold, I'm guessing i've not run the insulation far enough, so should be easily sorted!

[dogman]

@MikeGrahamT21 I have no experience with thermal imaging but it has me thinking I could try the same with my build before we plasterboard. I am not too worried about the frame but it would be an interesting exercise 

Any one else ran a camera around their frame before first fix 

Did you hire the camera?

[Jeremy Harris]

I ran the camera around our build in a  few places.  No thermal bridging except where expected, the window frames, but even then the temperature difference was small and barely showed up on the image (the frame was less than 1 deg C cooler than the walls, IIRC).  I have a Seek Thermal camera, the Android version, that just plugs into my Sony tablet and seems as good as some of the more expensive kit, although I imported mine from Canada before they were on sale here and it was a lot less than the current price.

 

To get any meaningful result you have to wait until the outside air temperature is low, the colder the batter.  No real point in doing anything until it's at least 15 deg C colder outside than inside, preferably a fair bit more.  Nice, cold, frosty nights are ideal.  Thermal imaging during the day is not a good idea, as even on a cold day the sun will warm up the outside walls that it's shining on.

 

I can try and take some more images later in winter, when the outside temperature has dropped to below 5 deg C or so, as a benchmark for what a pretty thermal bridge free  build looks  like.  The most significant thermal bridges, the door handles, don't show up well though, as their surface emissivity is very low.  This is something else that needs to be taken into account, a thermal imaging camera only measures the radiated heat from surfaces.  If a surface has a low emissivity at the long wavelength IR that the camera bolometer is sensitive to, then you will get a false reading.  This means that the camera will "see" something like a chrome finish door handle as being at a very different temperature to that which it's really sitting at, for example.

 

 

[dogman]

Thanks @JSHarris. I can hire one on the FLIR types locally for a week so will wail till we have another cold spell. 

[MikeGrahamT21]

Yeah I too have the seek Thermal on android, though mine has started reading temperatures higher than they actually are for some reason, but the colours still show the weak spots.

 

Our frames, even the better ones are about 2-3c cooler than the walls, not bad for upvc I guess. Going to get the new 90mm 9 chamber profile which liniar have just launched for the rest, but it's not available in England at the moment, only northern Ireland which is a shame. I'll be patient! Lol

 

It's a good exercise to see the weak spots for sure, you can hire the lower spec flir ones from jewson for not much money

[Jeremy Harris]

The temperature reading depends very strongly  on the emissivity of the surfaces being measured.  IIRC, all these IR imagers are calibrated at an emissivity of around 0.9, using a grey plate emitter with that surface emissivity.  So, if you have something with a higher emissivity, say something that is darker and close to a perfect black body emitter, then the reading will be higher than the true value, and if you have something that has a lower emissivity, something like a shiny bit of chrome handle or hinge, the reading will be lower than the true value.

 

This effect can be quite large.  A perfect black body radiator (emissivity 1.0) will produce an over-reading of a bit over 10%, a chrome door handle, with an emissivity of around 0.1, will under-read by over 90%.  There is a table of emissivity values for different materials here that may be useful: http://www-eng.lbl.gov/~dw/projects/DW4229_LHC_detector_analysis/calculations/emissivity2.pdf

 

One cause of the Seek Thermal (and the FLIR) readings drifting off is if the calibration shutter sticks, or needs cleaning (not an easy job to do!).  Every time the camera clicks, that is the noise of the calibration shutter moving across in front of the bolometer sensor to re-calibrate for changes in the sensor temperature.  In essence, it's a constant emissivity plate that slides across the whole sensor and allows the sensor to self-calibrate for changes.  This self-calibration is never perfect, as there will always be some difference between the sensor temperature and the shutter temperature, and this is especially true of the Seek Thermal, as if the plug-in camera module is held in the hand when in use then the sensor will tend to warm up faster than the calibration shutter.

 

I should add that the hand held IR thermometers are subject to the same errors, but usually they are a bit greater, because they don't have a calibration shutter.

 

 

Edited September 23, 2017 by JSHarris