First Adobe Production Run

5 Mar

Two clay and sand sources and mixer.

Based on our first experiment making adobe blocks with the new cement mixer, we did our first production run with the mixer. We prepared for this run by building two ladder forms: one continuous form for four 10”x14”x3” blocks with handles at each end.  On tarps next to the mixer were piles of screened dirt from two different sources: Black Creek Draw (BCD) and South County (SC). (We have so many dirt samples now that we are assigning codes for each sample.)  A five-gallon bucket held the remnants of Portland cement (PC) from an older bag. Another bucket held water, while the 80 oz. bucket was used for measuring the water during mixing.

We delegated jobs between us for efficiency and consistency: Ed would mix, I would wheelbarrow mixed mud to the casting site, clean and pack forms. This allowed Ed to mix batches while I packed mud from the previous batch into forms and cleaned the second form for the next batch.

The previous day, Ed had made test blocks from each of the two dirt sources without cement (unstabilized blocks). He also made a test block from a 50/50 ratio of the two sources. This helped to determine the optimal ratio of the two for the final run. It also later proved to be invaluable for troubleshooting. Like any scientific approach, controls (negative and/or positive) are necessary for accurate comparative analysis and interpretation. Weeks afterward we agreed to institute this as standard procedure for future earthen experiments.

Sedimentation (jar) test of soils.

Jar Tests. In addition to the test blocks, we also did a jar test of each dirt source. We filled each jar to the 400 ml graduation with dirt and added ~400-500 ml of water with a small drop of dish soap. After vigorous shaking for at minimum of 60 seconds, the jars were allowed to stand for at least 4-8 hours before evaluation. In the photograph, the open jar contains BCD dirt; the closed jar, SC dirt. Other than different colors of the sand, both have almost equivalent amounts of clay/silt fraction.  Both jars did not have clear delineations between fractions of silt and clay, but we decided to go ahead and do a 50/50 ratio of the two sources for our final run.

Test blocks of SC & BCD dirt.

Test Blocks.  Test blocks appear in the annotated photos. The two blocks in top row, right to left: cured 100% SC block, cured SC + 2 handfuls of clay + PC. Added cement imparts a lighter cast to earthen blocks. Bottom row: 50/50 SC + BCD block (green) and a half-block made from residual mud from BCD (top ¼) and the 50/50 mix (lower ¼). The single block: 100% BCD (green).

BCD dirt test block.

In mixing the first batch, Ed added all the water at once to the dry components.  We decided to gradually add buckets of water in subsequent batches because both dry and liquids mixed more consistently.

The first two batches of mud were too dry, and pressing mud into the form edges and corners took too much time. Also, gaps in the mud along the edges could be seen when the ladder form was pulled off the first batch of blocks, further confirming the mix was not wet enough. Ed increased the water content (ratio) in subsequent batches, which resulted in a wetter slump that was quicker and easier to press into the forms and no gaps in the formed blocks.

Because some residual mixed mud was always left in the mixer and in the wheelbarrow, Ed reduced the volume of dry components after the first batch. Mud remaining in the wheelbarrow (from ½ to 1 shovelful) was shoveled into the mixer near the end of the mixing time to be incorporated into that batch of mud. The mixing time was also increased from 10 minutes to 15-20 minutes.

Batch mix ratios:


Batch 1 #shovelfuls

Batches 2-11 #shovelfuls

SC 10 9
BCD 10 9
PC 2 ¾ 2
Water 5 5
Mixing time 10 min. 15-20 min.

General procedure:

  • Shovel dry ingredients into drum, alternating 4-5 shovelfuls/dirt source, and finally adding the PC. Let dry components mix for a few minutes.
  • Add water gradually, bucket by bucket. Allow to mix for ~30+ seconds between buckets of water.
  • Mix all components for ~15-20 minutes.
  • Meantime, mud from previous batch is being shoveled into one ladder form with one person pressing mud into bottom, edges and corners making sure that blocks are leveled with top of form. Tops of mud are patted as level as possible. Be sure to lightly wet hands before pressing mud; otherwise, mud sticks to hands and can even lift chunks of mud out of form.
  • While all components are mixing, two people lift form off blocks from pervious run. Person mixing then takes wheelbarrow to mixer to combine residue in drum and place wheelbarrow under the mixer for emptying. Form person cleans the empty form, wiping interior sides and corners with water. Important to have interior of forms moist; otherwise, mud will stick to wood and form will not slide off with ease.
  • Form person wheels the new mud to the prepped form and begins to fill and pack.

Because we had unexpected visitors (locals), we didn’t start our run until 2 pm. Our first 2-3 batches were working out some details on the mix components and the general routine. Once we established the right mix and routine, the entire production run went quicker and smoother. We ran out of cement at batch 11 and at 44 blocks around 5 pm and called it quits there. We were also running out of casting area on the ground. With only enough dirt for another batch, we agreed it was a productive day: 44 blocks (11 mixer batches) in three hours, and a little over a half-hour for clean up.

Casting area and new blocks.

We learned a few lessons from this first production run:

  1. Have three ladder forms available. Pulling the forms from newly cast blocks after only 10-15 minutes was almost too soon with this moist mud. Ideally, we should have given the blocks another 15+ minutes to set before pulling forms.
  2. Place ladder forms on level ground. And don’t try to prevent mud from flowing outside of the form bottoms. I was ‘damming’ some of the open edges and corners inside the forms with loose sand. We discovered half-way through the run that this prevents the mixed mud from being pressed tightly into the corners and leaves large gaps in the blocks after the form is pulled.
  3. Casting area really needs landscape cloth for making adobe blocks. For two reasons: to prevent issues outlined in #2, and to prevent the wet adobe mud from binding with the dirt and sand underneath. We realized this was a problem a few days later when the new blocks were stood on their edges. It became more apparent and detrimental later when we worked these cured blocks with extremely uneven bottoms into the adobe arch of the ramada.
  4. Always, ALWAYS, use fresh Portland cement if you are making stabilized or semi-stabilized blocks. We discovered two weeks later that edges of the cured blocks spalled more than we expected. The edges crumbled with pressure; more than we liked and wanted in our blocks, but not enough to seriously compromise their strength or integrity. Regardless, they did not fully meet our expectations, and we were stumped as to the cause of this spalling.
  5. Always (there it is again) make an unstabilized test block of any candidate soil to be used for adobe blocks. Like a control, these test blocks may prove to be invaluable for later comparisons. Soon after casting in a form, inscribe a code assigned to that soil mix or component into the face of the still-wet block for later identification.

We were quite excited about this first production run until we started using the blocks in the wall of the ramada and discovered that the edges crumbled more than we prefer. My suspicions were that the silt content in either soil was high, or that the bentonite clay content in the BCD dirt was high. For comparison analyses, we went back to the test blocks with and without the Portland cement that we used in our production run.


To our amazement, the unstabilized blocks made from 100% SC dirt, the 100% BCD block, and the 50/50 SC+BCD were as hard, or even harder than cinder blocks! No spalling and the edges were hard enough that they resisted any crumbling and spalling. Then we examined the semi-stabilized SC text block (with cement). The sharp edges of that block crumbled with hard pressure from my fingers and Ed commented that it was much whiter than usual for a semi-stabilized block (which can be seen in the photo of test blocks).

My first suspect was too much bentonite clay in the Black Creek Draw dirt. Cement mixed with dirt high in expansive clays, such as bentonite, is relatively ineffective because the two components tend to oppose each other and result in low bonding strength. When rolling some of the BCD dirt into a ball in my hand, it felt more slimy than the SC dirt. Thus, the BCD may have some bentonite clay in it, but the 100% BCD block (with no PC) had solid and hard edges, resisting finger pressure and crumbling.

Another suspect was silt content, which remains an unknown in both sources. In both sedimentation tests (aka, the jar tests) for the two dirt samples observable strata of silt did not show. Possibly because most of the silt was still mixed in with the sand strata. However, if high silt contents were the cause of spalling, then it would also be seen in the unstabilized test blocks. And it was not.

While discussing these issues, Ed remembered that the bucket of leftover Portland cement he used in that run had clumps near the bottom, indicating that it had absorbed moisture over time. Cement and lime are both hydrophilic, which means they attract and absorb moisture, even from the air.  Both cement and lime require water for their chemical reactivity. Once moistened and reactive, quick (hydrated) lime and cement begin to bind with the mineral components in the mud. As the block cures over time, they become stronger and dense while at the same time, their binding reactivity decreases.

Accordingly, if a bag of lime or cement is exposed to humidity or water in the air, it will absorb moisture and begin to ‘set’, or react with itself. Lime especially is highly reactive even to the carbon dioxide in the air. If not tightly bagged, it can absorb both moisture and CO2 and become ‘stale’. If too stale, the reactivity of lime can be severely impaired for any future use as a building material additive. I have had a bag of Portland cement harden like a rock while still in its bag in my barn near Fort Worth. Several observations here strongly suggest that the culprit was the old ‘stale’ Portland cement. Hence, Lesson #4.

First Adobe Arch

Between previous blocks made with clay and sand from Cedar Springs (CS) and the 44 blocks recently made, we had enough to start constructing the adobe block arch in the ramada’s west wall. That, too, was a learning process as it was our first arch. We are short about six blocks to finish the entire wall and will be doing another adobe block run soon to make up the deficit and have blocks for another project.

By mid-March the west wall of the ramada should be completed. And then we can move on to begin the next major project: the foundation for the earthbag guest house!

Adobe arch of ramada and experimental plaster wall.



2 Responses to “First Adobe Production Run”

  1. Brad April 2, 2013 at 3:14 pm #

    I appreciate the details on your project especially the lessons learned.

    • Macrobe April 12, 2013 at 10:57 am #

      Thank you, Brad. I keep detailed notes like a lab book, probably the scientist in me. 😉

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: