Right, just looked at the hourly numbers and worked out the rate of change during the day. This is known as the Slope, and it is only the y axis values divided by the x axis values, so nothing special.

0 equals no change and 1 equals infinite change.

Ambient,0.07 °C/hour
No Mass,0.06 °C/hour
Granite,0.07 °C/hour
Water,0.06 °C/hour

So all pretty close to 0 (and below the accuracy of the measuring equipment).

And yes, I am just looking at the UK climate, and I am not suggesting that we all live in tents with 2 metre thick polystyrene walls, though I will claim that as far as temperature stability and energy use, you would not notice the difference.

Chart below shows that the masses track the ambient pretty closely.

Hmmm just looked at your last post ST and something is definitely wrong with your experiment. Adding heat or 'coolth' from the ambient should affect the internal temp the same (just the direction of change waries) but I can tell you that if my house

Posted By: SteamyTeamasses track the ambient pretty closely

I would die (metaphorically speaking). Your graph just does not reflect how high mass houses perform, though sadly I am not informed enough to even speculate why your experiment isn't working.

Posted By: SteamyTeaBy isolating the experiment from solar gain as much as I can and making it airtight, by all having the same initial thermal properties, i.e. U-values, mass, location, shape and size etc, those factors can be discounted.

If you remove solar gain you're removing one of the main factors that thermal mass takes advantage of. A low mass room with significant solar gains (i.e. a S to W facing window) is very likely to get uncomfortably hot in sunny weather. The excess heat can be absorbed by the thermal mass and will significantly reduce maximum temperature.

If there's no heat input and the level of insulation and draught proofing is equally high, I doubt that there will be a significant difference between high and low mass construction.

Our house has elements of both high and low thermal mass. The bottom two floors are solid stone up to 30" thick and mostly with no insulation. The top floor is solid single brick with about 80mm of insulation, which was installed 20 odd years ago. The bottom two floors have very stable temperatures; the top floor suffers from overheating in sunny weather and excessive heat loss in cold weather.

Have you got data for upstairs?
You seem to have a delay of about 3 hours between lowest ambient and lowest kitchen/dining room temperature, but your variation is not that different between days, about 1°C difference (internal), though the ambient temp swings greatly between 8°C between the 29/8 and 30/8 and 19°C between the 2/9 and 3/9, so not so different from my temperature swings between experiments days, big differences in ambient but not so different inside the 'houses'.
Can you send me the raw data as a text file and I shall see if I can do the same tests on it as I have done on mine as there may be an error in the choice of tests.

Just plotted the same time slot as BillT's to see what happens (well to the 7th when I stopped it).
I get a lot of variation in internal temperature and I think this is to do with insulation levels (1 mm of PE has a poor U-Value).
But the main thing, and the point I am making is that adding mass to the 'houses' does not change the internal variation hardly at all. This has made me think about adding insulation to each of the 'houses' and seeing what happens, so I shall go and get some tape and I have some expanded PE and shall wrap the 'houses' up and put then back and see what happens.
Question is, should I insulate the base, sides and top or just the sides and how long should I run the test for, last time was 26 days?

ST, lag, and indirectly decrement, depend on speed of travel of a temp wavefront thro the thickness a thermal mass material. That speed is dependent on the ratio of thermal resistivity to volumetric specific thermal capacity (not dependent at all, BTW, on initial delta-t - in other words greater delta-t doesn't drive the wavefront any faster).

So if you scale the sample down, the wavefront's propagation speed stays the same so the wavefront simply gets thro the sample faster, and the decrement is less, in other words the effect of the thermal mass on the space becomes maybe negligible. Not just scaled down pro rata, so that you'd have to speed time up pro rata - but pretty well extinguished to meaninglessness.

Response to mw116, got ot work out why the initial temperature difference makes no difference to the wave fron tin Toms suggestion before I respond to that.
Not in the centre, it is on the base (call it floor).
I can see that walls could get more solar gain from a window opening, but this is about measuring the affects of just thermal mass in identical containers to take way the other variables.
Using variables to your advantage is a design issue (maybe low mass highly insulated wall to the north and solid walls to the south, solid internal walls that catch the sun, different floor construction between ground floor and successive levels etc.
What I am trying to show is that thermal mass on its own is not what stabilises temperature, it is a combination of lots of things. As a thought experiment (think I have done this one before), imagine a large block of solid rock shaped like a house, pop a thermometer in the middle and measure the temperature, it will probably be around the annual mean ambient air temperature. Create a small void in the middle and measure that air temperature, be somewhere around the same temp, you have decreased the thermal mass, and the U-value slightly, but it makes no difference. Keep doing this and somewhere along the way you will get to a point where the variation in temperature starts to show but only slightly, well within comfort levels, assuming you are comfortable with the mean annual ambient air temperature (so maybe you have to live in Corpus Christi where it is about 22°C). Now from that point on, as you remove mass, you add insulation so that it matches the U-value you are removing, will the temperature stay the same, will it vary wildly, does anyone know, has anyone tried this out?
My initial results show that, within the accuracy of my measuring equipment, adding mass to identical 'houses' makes no difference, so I think it may well be work adding insulation and seeing if that dampens out the variation.

Posted By: SteamyTeaResponse to mw116, got ot work out why the initial temperature difference makes no difference to the wave fron tin Toms suggestion before I respond to that.

Maybe it's a bit like the speed of sound in a medium not varying with the loudness of the sound.

Not sure if you have taken on board the comments about time scale, but if you have you could try using a shorter diurnal cycle (if your houses are small enough and your fridge/freezer is large enough) by alternately cooling and rewarming them every hour or so, and see what happens then.

Posted By: SteamyTeaHaving thought about it a while (takes my mind of what I am really meant to be writing), All I need to do is to have one without any insulation, one with one layer and one with 2 layers, maybe a third with 3 and see what happens. No mass in any of them and insulated all over.
Can always add mass as a third iteration.

That sounds good (or at least interesting) to me.

Right had a few minutes to look at what was happening when I left it in the sun.
The temperatures where higher, but again, within the accuracy of the the measuring equipment there was no significant difference (or statistical).
This was a small, short test so I would not read too much into it.

Results and pictures.
At the Student T-Test results (treatment being varying mass) were >0.05, so no difference there.

Class,Ambient,No Mass,1 kg Granite,1 kg Water
Mean,16.1,16.6,16.9,17.2
Median,15.8,15.0,15.5,16.0
Mode ,15.5,15.0,14.5,15.0
Min,11.0,9.5,10.0,11.0
Max,22.5,36.5,36.0,36.0
Range,11.5,27.0,26.0,25.0
St Dev,2.4,4.8,4.9,4.6