Indoor temperature 20.4 ; outdoors 4.5

My surface temperatures in the late evening are:-

Frame indoors 16.6 all round; sash 18.5; glass 17; reveal at coldest point 18.5

Frame and sash outside 4.8 ; glass 7.9 varying to 5.6; reveal 4.8

Are the frames insulated? Do you notice a pool of cold air on the floor below the window? In terms of thermal comfort this is the main thing that bothers me with my existing double glazing.

David

Can you get meaningful info from temperature alone? Isn't measuring temperature like just measuring voltage in an electrical circuit (eg you can't work out the resistance or power loss unless you also measure the current).

Do you have any idea how I can 'backwards engineer' to find the u-value of my thermal blinds from the temperature difference across them for a steady state? External temperature 2.5C, internal temp 20C, temp just inside window (single-glazed Victorian sash) 7.5C, therefore temp drop through blind is 12.5C.... any clues? How to assess the surface resistances, if I need to? See technical info on www.thermalblind.co.uk for full report.

Thanks RobL

I'm not quite sure where you are suggesting I take measurements. In December I chose a still, cold day & took measurements over a six hour period on the same window, with the blind up for a period, then down for a period. I took temp measurements in 3 locations, inside the room, just outside the window & between the blind and the window. It reached a steady-state of sorts for each period, which I've graphed. So for one of the 'up' sessions the outdoor temp was 6C, though this was clearly influenced by heatflow from the building, ambient external temp was more like 3C at the time. The room temperature (but again quite close to the window & therefore 'influenced') was 19C & the temp closer to the window was 16C. With the blind down the external temp measured 2-3C (say 2.5C av), the room temp 20C & the 'between the window & the blind' temperature measured av. 7.5C.

You can see a summary of this experiment here - http://bit.ly/tbtechni, showing the thermometer positions and the graph, with a link to a full report which includes photos.

So for your calc, I'm not sure which of these figures I could use. I understand it will not be a completely accurate U-Value but its really useful for me to be able to estimate it without bullshit! (see my previous comments & concerns). I don't want to mislead.

At the moment I'm only able to say "Although it was only 2.5ËšC outside, with the blind down the temperature behind the blind (just inside the glass) was around 12.5ËšC cooler than internal temperature, which was 20ËšC"..... which is a bit erm, long-winded.

Hi Ecohome
it certainly looks like the blinds are affecting things very significantly.

I think you can get a rough U value, but there are a couple of issues:

Using different thermometers for the different measurements can lead to significant inaccuracies, as we're interested in small temperature differences. Most thermometers are +/-2C absolute accuracy, however they generally have resolutions and relative accuracies much higher than this (0.1C common). Imagine you had 2 thermocouples, one was +2C, the other -2C out ! (that would be very unlucky mind)
You could calibrate them all together to largely remove this issue. Just stick them all in a large mug of water that's at ambient temp, and measure them all using your normal way (to take out all inaccuracies). Assume the mean value is correct, and add/subtract values from each one to calibrate to this mean. This won't take out gain differences, but they are usually less important. ***Have just read your report - are the thermometers mechanical ? Best not cal in water ! Still cal them together though ! I've always just used an IR thermometer in fact, as they are usually 0.1C relative accuracy, and cheap & easy to use.

The air circulation will change when the blind is moved, so that the difference in thermometer measurements is down to both improved insulation, and also different air current flow paths. For the "technique of ratios" (tm) to work, we want all thermometers that are used to see the same circumstances for all tests. I don't think this will mess things up much though - I'm mainly worries about the thermometer relative accuracies above.

I'd say that the temp difference from T2 to outside (a long way away outside) will give the best answer, of roughly:

U_blind+window = 5.8*(7.5C-2.5C)/(16C-2.5C) = 2.1W/C.m^2

I'm assuming that the room air temp (room centre) is stable, and that the outside air temp (long way from window) is 2.5C always. I've taken the standard U=5.8 for SG from Pilkington Spectrum.

I think that to improve the accuracy:
--I would use only one thermometer (or calibrate 2 together). You can see from the (7.5C-2.5C) in the calculation how dependant on that small temperature difference you are.
--Measure outside air >1m from the house.
--Then place the same thermometer (or a matched one) onto the inside face of the glass window. A tiny bit of blu-tak / sellotape is fine. Measuring the glass temp instead of air near it means changes in circulating air patterns due to the blind won't make any difference to the accuracy.

The eq will be the same as I did above: 5.8 * ( Tblind_down - Tair )/( Tblind_up - Tair )


Your blinds seem great ! We have many people in old houses in the UK unwilling to change their SG windows (like my parents, grrr), so I wish you all the best with these !

Thanks Rob!

That's really useful, now I think I'll repeat the test with the latest model (better internal fabric) & better thermometers (mine were extremely cheap!), properly calibrated together. I'm still prepping for the NPL hotbox test so hopefully that will give a more accurate, verified & comparative measurement. I'm now making them myself so its exciting to see it all beginning to work well. The first order for one of the Retrofit for the Future projects has been placed so I'll report back when they're fitted (being monitored via Wattbox / EST).

awww, thanks, I guess I should invest in some good ones myself really.....

very interesting!

I think I can wait for the NPL hot-box test I've got lined up (thanks to TSB) for accuracy, rather than cobbling something together myself. But I'm tempted to repeat the test over a more stable (internally) period, also with my new fabric & magnets - I guess I should test at the corner too where there is bound to be some air leakage.

I'm toying with the idea (but this is distracting me from making the blinds!) of constant electronic monitoring, culminating in funky graphing & tweeting - the tweeting blind! This is something I'm concocting with my mate @andysc who has a house that tweets - water consumption, elec use, a power-pie showing what is going where, even a mouse trap that tweets when its caught one ... see http://stanford-clark.com/house_that_twitters.html ... how about constant monitoring of the external & internal temps with eg. http://bit.ly/gZEAdo plus monitoring when the blind is up or down with a reed switch controlled by my stitched-in magnets? A PicAxe or Arduino controller, MQTT sending data online then I have my tweets! mad.

PS. If you're going to Ecobuild, come & see me! I'll be here - http://bit.ly/ecoinnovate

Katy

So far I've avoided datalogging. It's a little bit like my work (electronics design engineer) , and I have so many insulation projects I want to do, that seem to take soooo looong anyway. I took 4 years celotexing the loft :-( Two young kids though :-)

I have a friend at work who logs via I2C - USB to his PC, then he can log in & view data wherever he is. Most impressive. The irony of exactly measuring his planet saving measures with an always on PC...

SteamyTea
So is there a reason why the frame temperature dropped quite a bit when the external temp didn't drop so much?
PS. sorry for hijacking your thread!

RobL
There are much lower power ways of logging than keeping a pc on, favoured seems to be a Viglen mini server thing.... can't recall the name but could find out.

I'm a little confused by the variables in the calculation you suggest. If one should use the 'true' external temp a little further away from the window, then shouldn't you use the 'true' internal temp, so for the given options I get;

window U / external / between / inside / blind U

5.80 / 1.0 / 6.5 / 19.0 / 1.772 W/sq.mK (1C estimate of true ext. temp)
5.80 / 2.5 / 7.5 / 16.0 / 2.148 W/sq.mK (this 16C was temp very close to glass)
5.80 / 1.0 / 7.5 / 20.0 / 1.984 W/sq.mK

I've graphed a section through the materials with the three thermometer positions marked:

Hi ecohome
I'll try & justify the temp points I think you should measure (glass & far external ambient):

The power loss from a surface is proportional to the temperature of the surface compared to the ambient surrounding air. Take ambient to mean so far away, that the window doesn't affect it. It's hard to say what the coefficient of proportionality is in any given instance, as in this case it depends on wind conditions as it's done outside. As you're doing a before & after windowblind test, we can just divide one deltaT by the other, and multiply up by the U *that we already know* is the correct answer in the single glazed case. Thus we neatly divide out the unknown constant of proportionality, and are left with the desired U value.

I can see that various internal temperatures tell you something, and the T2 measurement is going to be close to the glass temperature. However...
The beauty of using the glass temp itself, is that whatever unknown & test-dependant circulating air currents occur inside the room, they won't affect the result of the calculation. Once the glass reaches a certain temperature, that's all you need to know to find the power loss.

The trouble with air currents is that they are so unpredictable. Take double glazing - it's more efficient at 20mm than 100mm. Apparently its something to do with more circulating currents, but who'd have predicted that ?
So that's my suggestion - remove the influence of the different internal air currents caused by the blind being up or down.

Rob

flip....... thanks again!

...but the glass temperature is probably a bit lower than 16C as you get very close..... and considerably lower when the blind is down.

I've looked at SAP 2005 & it reckons a single glazed window including the frame should be 4.8 not 5.8, so should I use that?

It gives me a better value at 1.7 - 1.8!

Table 6e also gives a value for 6mm gap DG (air-filled) at 3.1 & 2.8 for 12mm gap - this means I've surpassed my target! :whoop:

I just had a play with that Pilk software, trying to replicate your setup just for fun.

Glass, 20mm air, Glass has a U of 2.8
Glass, 30mm air, Glass has a U of 2.8
Glass, 50mm air, Glass has a U of 2.8
Glass, 60mm air, Glass has a U of 2.8
Glass, 100mm air, Glass has a U of 2.9

I'm deliberetely not using low e glass, I know your blinds have mylar, but its buried inside, so "hidden" from the air gap and just makes a higher U material. See how I avoid the word multifoil.

Going a bit out on a limb, but I reckon your setup is like the above glass-air-glass, then with extra insulation.
So the insulation of the blind alone is U= 1/( 1/1.8-1/2.8) = 5.04
Thats the same as 5mm of celotex.

Just a bit of fun that, don't take much store in it, the numbers I'm sure are getting lost in the error bars.

All the best !
oo found the smileys