Category Archives: Rammed earth

Driveway retaining wall

Some really nice weather arrived for the Easter weekend, so instead of sitting out on the deck drinking G ‘n’ Ts we got to finishing off our first little rammed earth wall.

From lessons learned on the first section, we removed the chamfer piece from the top of the form and tried freehand cutting a chamfer at the end instead.

It can’t be seen in this picture, but the walers are bent outwards just enough to ruin the wall’s straightness.

We also discovered the origin of the bowing on the walers. It’s not the ramming of the dirt per se, but rather in the clamping. Our bolt clamps are outside the form, and were meant to provide a simple automagic method of holding the end of the form in place, against the rods. But the walers bend in at the ends ever so slightly as the rods are tightened, which translates to a noticeable bend outwards in the middle.

You can’t avoid having holes in the wall if you’re going to use a through-clamping method of holding the form. So the bolts need to go through the body of the form to prevent the walers from bending. We’ll do that next time.

Getting there

Levelling the top of the wall was a lot easier without the chamfer pieces in the form.

We rammed past the top of the form then used a 1m steel rule to slice the excess off.

We ran out of dirt at the end, and it seemed wasteful to make another batch just for a tiny corner. We left it as it was knowing we’d come back to do the last section the next day.

The bow in the wall is much more apparent with the form off. The slight curve also meant that where the two sections join the second section was slightly wider than the first section and there was a really obvious line on the wall that we don’t want going forward. The formwork has to be perfectly straight to make sure things like that don’t happen.

Chamfering was done with a mini mattock and a mallet, then smoothed with a steel ruler. All freehand. The chamfer is fine for the present purposes but it would be much better to do it against a guide that can keep it uniform along the length of the wall.

The dirt mix was a bit wet, too, and stuck to the form. Bits came away with the form. Again, they’re fine here but we’d want that to not happen going forward. We’ll try a drier mix and reduce the clay content.

The next day it was on to the last section. The PVC pipe we’d put into the second section had moved while ramming, though, and didn’t line up properly with the holes in the walers. We couldn’t get the rods through.

We’d decided that the big rods were too thick anyway, and had some thinner ones at hand. The thinner rods allowed for more play and we were able to get them through the holes. Again, having the through rods inside the edge of the form would make that a non-issue.

As we finished the last bit we realised that we didn’t have enough sifted local dirt. We’d calculated a 2:1 compression ratio from our test cylinders, and it turned out to be more like 2.2:1. Which is not bad considering the test cylinders were only 20cmx10cm.

It was no big deal to get the sifting table out and make some more dirt, though.

We were expecting the last shorter section to only take the morning, but what with the drama with the misaligned holes and running out of dirt it was 3:30PM by the time we had it done and everything tidied up.

All done! Satisfaction achieved.
The end of the wall

We’d bought the cement for the wall some months previously, before knowing how long it would take to sort out the cement mixer. The bags had been sitting on the ground under the house. It’s dry down there, and the bags are plastic-lined inside, but moisture still got in and the outside layer of cement had gone hard. It broke into quite big lumps that the mixer couldn’t break up, so we ended up with lumps of cement in the last bits of wall. The sand was also quite sticky, and contained lumps, so there are intermittent areas of pure sand that aren’t good.

With all the formwork out of the way, we could put the bricks back the next day. We filled the little gap between the bricks and the wall with some concrete. This side of the wall will be under the deck eventually, but for the moment it’s nice to have it neatened off.

The middle section seems to be the best. The mix was drier and it hardened off very quickly. In a way, though, that worked against our favour. Five days later and the section where we’d run out of dirt is starting to crack and lift away. Because the lower portion had dried the top bit didn’t incorporate into it. It’s a relatively thin layer that dried faster than the rest. Leaving something to the next day is obviously another one of those mistakes that we won’t make again.

So, lessons learned:

  1. Make more rigid forms. We’re thinking a shorter form with formply on the outside of the waler as well as the inside, all screwed together.
  2. Through-rods inside the form. They only need to be finger-tight, too. The ramming will tighten them up.
  3. Use a drier mix.
  4. The big lump of wood rammer is good, but could be scraped down to make it more manageable.
  5. Make some kind of guide for cutting the bevel.
  6. Buy the cement only when we’re ready to ram.
  7. Use smaller through-rods.
  8. Always finish off a section for the end of a day’s work. It will be hard by the time you can get back to it the next day.
  9. Don’t work if it’s going to rain.
  10. Have covering ready for the amount of wall you have. It will want to be covered for some time after it’s complete if there’s any threat of rain.
  11. Use sand that doesn’t clump. The beach sand and builder’s sand test cylinders turned out pretty samey-looking, but the beach sand won’t clump.

A bit of rammed earth wall

Now that the cement mixer is all fixed up, and we have sifted and partitioned dirt/sand, we can go ahead and ram some earth.

It was going to be an all day thing and we were due rain in the afternoon, so we got cracking early making the formwork box. Of course, first we had to clear all the leaves away that have accumulated since the footing was done ages ago.

After measuring the desired height using the laser level inside shining out the window, we cut our formply to size. We wanted to have the wall come to the top of the form so that we could scrape the top flat using the form itself.

We want a chamfer on the top edge, because rammed earth edges tend to chip if they’re at 90 degrees. To achieve the chamfer automatically as it were, we nailed a piece of angle to the top of the form.

A 45 degree angle cut on the end of the angle wood so that it mitres with the end piece.

Then it was set up with a temporary clamp, and levelled. The dirt was mostly level already from doing the foundation, but one end needed some shims to get it just right.

Levelling things up. Little squares of masonite in the near corner managed a slight fall in the underlying dirt.

Next it was time to fit the walers, which are supposed to hold the form straight while ramming. We needed to drill holes in them to let the clamping rods through, and we got to use the new drill press for the first time to do that.

All set. Clamping rods are a bit long, also the form end piece did not line up perfectly with the ends of the long pieces. Ramming moved it out so the wall ended up overhanging the foundation by a few mm.

That’s it! We have our box, our dirt, our mixer and our home-made rammer.

Making the box took all morning; we didn’t get to ramming until after lunch.

As the dirt went in we were careful to make sure the sides against the form were well packed. It’s the only part of the wall that will ever be seen, so it’s worth the effort. We went around the edge with a piece of wood and a mini sledge to hammer it down.

Particular attention was given to the corners

Half way through and the rammer gave up the ghost. The concrete lump I’d made wasn’t strong enough and it fell to bits. We smashed the remaining bits off the end of the handle and fixed a piece of wood to it with a screw.

The wood+handle thing actually works pretty well – except that it wasn’t nearly stable enough. The screw hole in the handle quickly wore out and the bit of wood wouldn’t stay attached. We had to come up with something on the spot, so we chopped a spare post and planed off the corners to make it vaguely round on top.

Nature’s ram. Effective, but too heavy for sustained work.

That did the trick, but it’s really heavy and hard on the hands. We’ll modify it and give it a proper handle and then it should be sweet.

This bit of wall has to be done in three parts due to the length of formwork we have. The clamping bolts will need to go through this bit of wall somewhere when doing the next section over, so we put some PVC pipe in there for the bolts. Also made up on the spot.

Nearly done, with the near 50cm or so scraped flat. PVC rod through point visible at the other end.

Almost done, and down came torrential rain. We had to pack away the mixer and cover over the dirt box. We had just enough mixed dirt in the wheelbarrow to complete the job.

The last bit was done with rain pouring down, and the top of the wall got pretty wet. The dirt was not tamped as well as it could have been. We didn’t have a spare tarp, so we covered the wall with the offcut pieces of formply, which was not ideal since the water could still leak through.

We got a tarp over it the next day, but it rained steadily for three days.

Finally, on the fourth day we were able to take the tarp away and removed the formwork.

Well, it’s solid enough! Most of it looks pretty good.

There was definitely some laziness in the tamping towards the top. The piece of angle wood made a nice chamfer, but it was a real pain to work with. It kept getting in the way of the rammer, and it was really hard to get good tamping near it because it wasn’t that strong. It’s probably better to do the chamfer as a separate process at the end, using a piece of something shaped for the purpose. A bit of triangular metal would be perfect.

The walers weren’t really up to the job. They’re 90x35mm hardwood, but the 2.1m span was too much and the wall bowed out in the middle by about 3mm on either side. It doesn’t seem like much, but it’s quite visible when you look down the length of the wall, and it will cause problems when it comes time to put the decking on top. The form could do with being half its length, or having an additional clamp in the middle. We’ll try the second method on the next section of wall.

The clamping rods are too big. You can get away with something much finer, and it would be better to insert bits of electrical conduit in the wall to act as rod-through points rather than those big bits of PVC. We’ll fill the holes later and they shouldn’t be visible, but it’s all just bigger than it needs to be.

Cement mixer fixing

After changing our minds 5 different times with respect to the old cement mixer, we decided to go ahead and fix it up. Getting a new one would have shortened the timeframe to get things done, but it can’t compare to restoring something that would otherwise be rubbish and making it work again. The less rubbish in the world the better.

So after pulling it to bits the bits were stripped and repainted.

The main frame stripped and ready to paint
Main frame and wheels all painted and shiny

There were many learnings along the way. We have an assortment of wire brushes that fit the drill, but the drill heated up quickly and there was concern that it would burn out. It’s not really made for continuous duty. Dumping the drill we got a special paint-removing attachment for the angle grinder, but that wore down faster than lickety split. At $15 a pop, it wasn’t going to cut it. Finally we found rotary wire brushes for the angle grinder which were cheaper and faster than the paint remover, and got rid of the rust at the same time.

The main shaft attached to the bowl was swelled by someone hammering on it, and the bearings just wouldn’t come off. We got a bearing puller from the auto shop but it was useless. Eventually we just cut them off with the angle grinder.

We got this really good enamel paint, but it went on very thinly. It’s good proper tough stuff, but it was going to need 4 coats to give a good finish. So we got some other enamel paint, which went on much quicker, and was nice and shiny, but was a bit too plasticky for comfort. In an attempt to harden it up we put a coat of the original enamel over the top … and the other paint underneath half melted and rippled and now the surface looks like a bodgy job. Oh well, we’re not going to strip it again!

Handle stripped and ready for paint

With new bearings for the pinion and bowl, it runs like butter again.

New pinion bearing

The bowl has so many dents in it that it doesn’t seem worth the effort to strip and paint it. We’ll leave it as is and get a new bowl some time in the future.

For the moment, the edge of the bowl needs neatening up, as it looks to be a pretty nasty scratching hazard as is. The motor shelf and the big wheel at the back are being painted, and we’re off to get a new motor for it today. It’ll be all go again before we know it!

Dirt sifting

With our dirt measuring box ready, we can sift dirt for our rammed earth wall.

We took the seedling table, which has a strong 10mm wire mesh as its main surface, and overlaid 5mm mesh wire. It was positioned below our quarrying pile. A tarp was put down so that if it started to rain we could cover up all our nice dirt and stop it from getting too wet. Then the dirt was thrown off the pile onto the table and pushed through the mesh by hand. The too-big stuff went into the wheelbarrow and down the hill to the orchard level.

The sifting goes at about 0.15 cubic metre per hour, so it was about 5 hours to get 0.7 cubic metres of soil prepared. Then it was off to the hardware to get some sand and like that the soil was ready for mixing and ramming.

Now what to do about that cement mixer?

Prepping for dirt sifting

We’re going to use the end of the driveway as our staging area to prep the dirt going into the rammed earth wall. It needs a bit of attention, though. We want to have the prep area right up against the wall to the right to give space to turn the cars around. So that weedy scree needs to come out and the level squared up.

There was way more dirt in there than expected, but a couple of hours with fork, shovel and wheelbarrow got it neatened up.

Then it was time to make the dirt-measuring box. We need about a cubic metre for our first little wall, so it makes sense to make a 1 cubic metre box. It will have a moveable sliding piece in it that will let us divide the internal area into half-half or one-third-two-thirds or other proportions.

We had some plywood lying around from an aborted furniture project, so we put it to use, routing channels in the right spots so that it slots together and then flat-packs away when we’re not using it.

Great! Now to start sifting.

Cement mixer for rammed earth

Up until now, we’ve been doing all our concrete and dirt-preparation mixing in a plastic tub, using a spade. That’s fairly back-breaking work even for small quantities, and we’re about to take it up a notch so a cement mixer is on the cards.

Cement mixers are expensive, though, and we have a motor in the shed from the old mulcher. The plan is to get an old mixer for next to nothing, attach the motor and Bob’s your uncle.

After not seeing anything on the second hand forums we did a shout-out and one popped up straight away. It’s got no motor or pulley, and the bearings are shot, but it should go with a bit of love. Here it is with shiny new wheels.

The inside bearing on the pinion is the main issue, but we know a guy who’ll send us new ones cheap.

Form comes off

We left the form on the rammed earth wall foundation for nearly a week, mostly as a spot of protection against the rain. Attaching the reinforcing to the rebar uprights was a matter of a few minutes.


And now it’s rammed earth time. We have a job of work ahead of us conditioning the soil to go into this wall. We’re going to sift out the rocks, add sand, cement, plasticure, water and mix. We need to procure the sand, the plasticure and a cement mixer.

One of the issues is that the soil we want to use has been sitting in the rain all winter and is full of water, which will make controlling the water content very difficult. We’d expected that there would have been a good stretch of fine weather by now to dry it out (less than a week until summer!), but it hasn’t eventuated. While the soil is conveniently located (it’s about 3m away to the left of the foundation in the picture above), we really need any soil destined for a wall to be put somewhere dry first. The only dry spots we have are under the house where there’s also no sun to help things out. So it looks like we’ll need to set up a tarp to stage our dirt. At least its ugliness will spur us on to get things sorted quickly!

Wall foundation

The weekend provided some warm, rain-free weather (finally), so we’re able to get on with setting the foundation for our little rammed earth wall.

“Make a box and pour concrete in it”. Sounds so easy.

We decided to use our walers as the box sides: they’re 90mm tall instead of the planned 100mm, but it doesn’t really make much difference and gave us more flexibility for cutting up the formply later. Unfortunately we can’t transport 5.2m pieces of wood around, so we’ve jerry-rigged the box out of 4x 2.4m walers and some pine we had lying around.


The star pickets keep the outside in line, and we used the long steel extrusion to ensure straightness between pieces. We didn’t do too well on one of the star pickets, but a little shim got everything lined up nicely:


We used a spare 30cm-long bit to test the width along the box before filling it in. Everything has to be perfectly straight and level, since the rammed-earth form will used this foundation as its reference. That meant we had to do a bunch more digging to achieve a level bottom so the top could also be level.

Then it was just a matter of mixing up the concrete. By hand. 4 hours and 12 batches later I had a sore arm and a determination to buy a concrete mixer at our earliest convenience. The rammed earth will need about 7 times the volume to be mixed and there’s no way I’m going to do that with a shovel.

But a dead flat dead straight thing amongst the dirt looks just like civilisation and satisfaction levels are high.


First wall prep

We’re happy enough with the rammed earth test cylinders we made last week. After a day drying they went rock hard, and are certainly sturdy enough for the first job we want to used the rammed earth for.

We posted the idea a little while back, and given that we don’t need planning permission for such a small wall, we can just go ahead and do it. Here’s the vision:


The picture is not entirely accurate: the cladded wall at the rear is a little further back, the rammed earth wall is a little closer to the house and not so wide, and it doesn’t extend beyond the side of the house just visible on the right hand side of the picture.

The first thing is to remove the garden bed that’s currently in that spot.


The lavenders haven’t looked any good for ages, and the bulbs are done flowering. They won’t like being transplanted, but they’ll be back next August guaranteed.


We need to bring the base of the wall down to the undisturbed dirt level, which will be under those bricks. So we dig up all the dirt and cart it down the hill. It’s not consistent clay like you’d find deeper down, so we can’t use it for the wall itself.


We also took out all the red gum sleepers. We wanted to save them, but there was only one or two that can be re-used. The rest were half rotten, which makes me a tad nervous about the rest of the retaining around the yard. Some of it is starting to fall to bits, and I’m guessing that we’ll have to get onto these new walls with a bit more speed than what we’ve applied to the project so far.

The garden bed was cleared out in an afternoon, after the morning was spent procuring a post hole digger. The following morning was spent cutting the grass, since rain was expected in the evening and it’s a lot easier to do when it’s dry. Once the maintenance jobs were out of the way we could get back to measuring out the wall.

The deck is going to be 1.6m around this and the adjacent side of the house. This width marries up with the width of the stairs we’ll be putting in to the lower garden, a couple of which are already in place. So we put some stakes in the ground 1.6m away from the house.

We’ll make this wall 30cm thick, which is way too much to carry the load of half a deck and whatever we might put on it, but which is a good balance between ensuring load carrying capacity (a thicker wall) and reducing the amount of material we have to prepare to put into it (a thinner wall). More stakes 1.3m away from the house, requiring us to remove some of the pavers.

We line it up with the end of the house and the nail we put into the house to mark the end of the garage all that time ago. Total length is 5.2m. We set some string down the middle and measure out spots for 5 piles. The piles will go down into the ground a short distance to anchor the wall against the soil and water pressure that will be exerted on it from the driveway. It’s unlikely that they’re actually required – I haven’t done the engineering for this little wall – but it’s all about practicing our techniques at this stage and a bit of over-engineering here is not going to be wasted effort.


Finally all that’s left is to calculate how much dirt we need. We set the laser level up on the floor inside and shone it out the window. That allowed us to calculate a height from the ground to the top of the floor inside, which was 47.5cm. The top of the deck needs to match that. We take off 10cm for the concrete foundation, 7.5cm for joist bits that sit on top of the wall and the thickness of the decking, leaving a 30cm rammed earth wall. The volume of the wall is then 520x30x30cm or 468,000cm3. That’s 468L. From our cylinder experiments we know that compaction around about halves the volume of uncompacted dirt, so we need double the amount in the compacted wall of uncompacted material –  936L. That’s pretty close to 1m3. Our big dirt-carrying tubs are 42L when full, so it’s a little over 22 tubs in total. Or 11 wheelbarrows, since 2 tubs fit in the wheelbarrow. Which is pretty close to a ute-load (our mulch-carting exploits provided the information that the ute holds around 12-13 wheelbarrow loads of mulch).

The 5 piles will be 15cm in diameter, and go down the length of a standard bit of rebar minus 5cm which will extend into the foundation. That depth is 55cm, so each pile will contain ?x7.5×7.5x75cm or 9,719cm3. That’s 48.6L in all 5 piles. The foundation slab will be 520x30x10cm or 156L, not counting that the rebar inside it will reduce the amount of concrete we need to make. One part in 7 of that amount is 22L, which when added to the same proportion in the piles gives us 29L, the amount of cement we need to buy for the foundation. That’s a little bit bigger than a standard bag of potting mix. Interestingly, cement is sold in bags of a specific weight, not volume, despite requirements for cement being pretty much universally volume-driven. Boral helpfully tells us that 108 20kg bags give us 1m3, which means that we need just over 3 bags for our 29L. If we also put a bit into the rammed earth part (2% of 936L), we need 5.2 bags.

Rammed earth test cylinders

After talking about it for possibly decades, now the rammed earth begins.

Step 1 is to test the suitability of your local soil. We did a soil grain size test earlier and found that around here it’s basically 100% clay. We also worked out that adding about 25% sand should give us the consistency we need for a good strong wall. So we’re setting out to create some mixes of local clay and imported sand and maybe even a bit of cement to see which is the best mix to use.

Using the big pile up near the wicking beds, we sieved some dirt first to get a decent grain size that will mix easily with the sand. The dirt is all the same stuff, but some of it is dry, some wet, some broken up and some not. Unless you live on top of a sand quarry you’re never going to be able to just dig dirt out of the ground and expect it to be good to go straight away.

The sieve is a 5mm mesh

We’ll use 100% local soil on the first one to serve as a control/baseline. We also used the first one to figure out how much soil we need to make the cylinder. It was about 17 trowels, which we upped to 20 to make working out the proportions easier. We added two “hoses” of water to each 20-trowel pile of dirt. A “hose” in this case being water that is under pressure in our 30m hose after the tank pump has been turned on and off. Turn the pump on until pressurised, squirt the hose until pressure is lost, repeat.


It’s not really accurate, but the soil doesn’t start out at a consistent moisture level so there’s no point trying to finesse it.

Mix it up and make a little ball. Drop the ball from a consistent (waist) height.


The amount that the ball flattens and the size of the cracks in the side show you whether you need to add more water or not. This one is OK.

The form is 200mm of 100mm diameter PVC stormwater pipe

Our fancy rammer that we made earlier turned out not to be useful for this. It expanded after being let out of the form and now it doesn’t go back in again. It was also really unwieldy for such a small form so we cut the ends off a nice round branch and used that instead.

We started out using the mallet to aid compression, but it unnecessarily complicated the process. The dirt wants to travel up the sides of the form anyway, so it’s never going to be consistent. We ended up just using arm power which is probably better anyway, as that’s what we’ll be using to make the walls.

The form wanted to creep upwards as we rammed, letting dirt squeeze out the bottom. So forms need to be attached to the ground below when doing the real thing.

Gently easing the form off…

The grey bit is some concrete that came off the big rammer while we were seeing if it would go back in the form. I’m thinking that the big rammer will be less useful than hoped.

Well, that clay is very sticky. We don’t want this happening when we take the forms off the walls.

The dirt stuck more at the top of the form. It probably isn’t rammed as well as the rest of it, since it’s hard to really pound the top bits without the dirt flying out everywhere

For the next round, we took 5 trowels of “builders” sand and 15 trowels of local clay. The sand is relatively fine, and my first impression was that something coarser was needed to balance out the very fine particles in the clay.

We can get this stuff by the truckload when the time comes


Working with the sand-impregnated soil was much nicer than with the raw clay. It wasn’t nearly as sticky and compressed a lot easier. It really makes a big difference.

We also put a bit of oil inside the form and it came off without bringing half the dirt with it.

Next up we added 25% white beach sand, which we were expecting to change the colour of the mixed earth but didn’t make a noticeable difference. The grain is a lot coarser and the final mix didn’t feel as robust in the hand as the builder’s sand.

Then we did the builder’s sand again, but added half a trowel of cement which gives around about a 2% cement mixture. Our assistant made signs so we know which is which.



In all the excitement it was dinner time by the time we were done, and no-one had even thought about dinner yet. A quick trip to the wicking beds and we had lettuce for Caesar salad – good times!