The land
What the ground told us
Three holes, a 63.5 kg hammer, and the numbers that decide how this house sits on the earth
7 Aug 2026 · Bharath
There is a stage in building a house where you are spending money on something you will never see, never show anyone, and cannot photograph well. Soil testing is the first of those. It is easy to talk yourself out of. The plot looks solid, the neighbours have built on the same street, the earth is that confident Karnataka red — what exactly is there to find out?
Quite a lot, as it turns out. And for a house made largely of earth, the ground is not only what we build on. It is very nearly what we build from. We wanted to know what we were standing on before we asked it to hold up a house of rammed walls and a dome.
So in July we had three holes put down across the plot.
What actually happens
The test has an honest, almost agricultural simplicity to it. A crew arrives in a van with a bundle of steel rods, some spanners, a hammer, and a heavy steel doughnut. No hydraulics, no computers, nothing you could not carry.
They bore a 150 mm hole with a hand auger — someone turning a T-handle, someone else pulling the auger up every so often to empty the soil off it. When the hole reaches the depth they want a sample from, the auger comes out and a split-barrel sampler goes down on the end of the rods: a 50 mm steel tube that splits lengthways into two halves, like a pea pod.
Then the counting starts. A 63.5 kg hammer is raised and dropped 760 mm onto the top of the rods, over and over, driving the sampler into the undisturbed soil at the bottom of the hole. Someone calls out the blows. The sampler is driven 450 mm in three equal stages of 150 mm, and the blows for each stage are written down separately.
The first 150 mm is thrown away. That stage is only there to get the sampler past the loose, churned-up material the augering itself created. The blows for the second and third stages are added together, and that number is the N-value — the Standard Penetration Test result, done to IS 2131. It is a crude measurement and an extremely useful one: how hard is it to push a known tube into this ground with a known weight falling a known distance.
When the sampler comes back up, they unscrew the cutting shoe, lay the two halves open on the ground, and there is your soil — a clean cylinder of it, from exactly the depth you asked about. That is the photograph worth taking. Ours came up red-brown, gritty, holding its shape, shot through with gravel.
Why three holes
Because one hole tells you about one hole.
Ground varies across a plot — a filled-in pit, an old tank bund, a seam of weaker material — and the only way to find out whether your site is consistent is to look in more than one place. Ours went in at three separate points across the plot: BH-1 out in the middle of it, BH-2 and BH-3 closer to the southern boundary, one to the west and one to the east. If the three had disagreed badly, that would itself have been the finding.
They did not. Which is its own kind of good news.
What came up
Two layers, in the same order in all three holes.
Silty sand from the surface down to about 1.5 m — the red stuff, with enough fine material in it to hold together. Below that, silty gravel: denser, coarser, more than half gravel by weight.
| Hole | Depth | Stratum | Blows per 150 mm | N |
|---|---|---|---|---|
| BH-1 | 1.5 m | Silty sand | 6 · 10 · 14 | 24 |
| BH-1 | 3.0 m | Silty gravel | 13 · 23 · 35 | >50 |
| BH-2 | 1.5 m | Silty sand | 7 · 12 · 15 | 27 |
| BH-2 | 2.0 m | Silty gravel | 18 · 25 · 36 | >50 |
| BH-3 | 1.5 m | Silty sand | 8 · 15 · 17 | 32 |
| BH-3 | 2.0 m | Silty gravel | 17 · 26 · 38 | >50 |
Read the middle column rather than just the total. In every hole the blow count climbs steeply within a single 450 mm drive — 6, then 10, then 14. The ground gets harder the further you push into it. By 2 m the gravel was taking more than fifty blows to move 300 mm, which is where the test stops being meaningful and simply reports refusal.
No water. The holes were left open for twenty-four hours to let any water table find its level, and none appeared within 3 m. That matters more than it might sound, for two reasons. First, this was July — the middle of the monsoon. A dry hole in July is a far stronger statement than a dry hole in March. Second, we are building in earth, and the single thing that destroys an earth wall is persistent damp from below.
What the laboratory added
The samples went off to be classified. The useful numbers:
| At 1.5 m | At 2–3 m | |
|---|---|---|
| Classification | SM — silty sand | GM — silty gravel |
| Gravel | 25–36% | 52–55% |
| Sand | 30–34% | 28–33% |
| Silt and clay | 31–45% | 14–20% |
| Moisture content | 13.2–13.3% | 12.4–12.6% |
| Liquid limit | 29.2–29.6 | — |
| Plasticity index | 6.7–7.0 | — |
| Dry density | 1.51 g/cm³ | 1.53–1.54 g/cm³ |
| Angle of internal friction | 30° | 31–32° |
The line to stop at is the plasticity index: between 6.7 and 7.0.
Plasticity index is the width of the moisture band over which a soil behaves like putty rather than like liquid or like a solid. A high number means a clay that swells when wet and shrinks when dry, and moves a building around with it as the seasons turn. Black cotton soil, which large parts of this state are cursed with, runs several times higher, and people spend a great deal of money getting away from it.
Seven is not that. Seven is a soil with enough fine material to bind and not enough to misbehave.
What it can carry
From the friction angle, the cohesion and the density, the engineers worked out a safe bearing capacity — how much load a footing of a given size at a given depth can put into this ground, with a factor of safety of 2.5 already taken off.
| Hole | 1.5 m, on silty sand | 2–3 m, on silty gravel |
|---|---|---|
| BH-1 | 12.3 t/m² | 23.2 t/m² |
| BH-2 | 12.2 t/m² | 17.5 t/m² |
| BH-3 | 12.2 t/m² | 17.5 t/m² |
Is twelve tonnes per square metre a lot? On its own the number means nothing, so here is the back of an envelope — not a design, just a way of feeling the size of it.
Take a rammed earth wall of the sort we are likely to build — say 450 mm thick and 3 m tall. It weighs somewhere around 2.7 tonnes per metre of its length. Stand that on a footing 900 mm wide and the ground beneath is carrying roughly 3 t/m². Double it for the roof, the floor, the dome and everything else the house puts on that wall, and call it 6.
Against an allowable 12.2. The ground is not the constraint here, and that is a comfortable thing to know before you commit to building heavy.
The settlement check came out the same way: 4.19 mm of immediate settlement calculated against a 25 mm limit.
What this report does not answer
Two things, and this is the part worth being clear-eyed about.
First, it assumes the wrong kind of foundation for what we are building. The report recommends isolated footings — individual square pads, 2 m × 2 m, which is exactly right for a framed building where columns bring the load down at discrete points. A rammed earth house does not work that way. Its walls are continuous, and they deliver their weight as a line, not a set of points. That usually wants a continuous strip footing instead, sized quite differently. The domes add their own problem: an arch or a dome sheds thrust sideways as well as down, and the foundation has to be asked about that explicitly.
The engineers are not wrong. They answered the question that was put to them, for "a building". We now have to go back with the actual structural scheme and have the bearing capacity worked out for a strip footing carrying a line load.
Second — and this is the one people miss — this report is about the soil the house stands on, not the soil the house is made of. Those are different questions asked of the same ground. Nothing in a geotechnical investigation tells you whether a soil will make a good wall. For that you need an entirely separate set of tests:
- a Proctor compaction test, to find the moisture content at which this soil compacts to its greatest density — ram it wetter or drier than that and you get a weaker wall for the same effort
- unconfined compressive strength on stabilised test cylinders, made at several different cement or lime percentages, so we buy the right amount of binder and not a rupee more
- linear shrinkage, to see how much the material moves as it dries
- wet strength and durability — spray and immersion tests, because a wall has to survive weather, not only a laboratory
The in-situ dry density is telling in this respect: 1.51 g/cm³ as the ground lies. A properly rammed wall of the same material should come out far denser than that. The ramming is not a formality. It is the entire difference between soil and a wall.
What we do next
- Go back to the geotechnical engineer with the structural drawings, and ask for safe bearing capacity under a continuous strip footing rather than an isolated pad — and for a view on the dome thrust.
- Decide the founding depth. There is a real choice here: 1.5 m onto the silty sand at about 12 t/m², or another half a metre down onto the silty gravel at 17.5. For a building this heavy, the extra digging may be the cheapest strength we ever buy.
- Commission the material testing — Proctor, cylinders at a range of stabiliser percentages, shrinkage, durability. That is the test that decides the recipe for every wall in the house.
- Find out whether our own soil is the wall soil. The layer at 1.5 m runs 25–36% gravel and 31–45% fines. Whether that is the right gradation, or whether it wants sand adding, is what the cylinders will tell us.
- Design the plinth and the drainage properly. No water at 3 m in July is the best news in this report. It does not excuse us from getting surface water away from the walls.
The honest summary
Three holes, 150 mm wide, 3 m deep. A hammer, a tube, and some arithmetic. Fieldwork in July; the report in our hands in the first week of August.
What we got back: the ground under this plot is consistent across all three holes, gets stronger with depth, has no water table within reach, and is not the expansive clay that would have changed everything. It will hold the house up comfortably.
What we still do not know is whether it will become the house. That is the next set of tests, and the one we are actually nervous about.
