Primitive water distillation construction notes and lessons learned
- Four different basic ways of distilling water are tested and the lessons
learned are discussed. Gradient steps in complexity of construction are
shown. Any one trying to distill water will most likely be going though
the same learning pains. Hopefully if you read and understand this you
can stand on what I have learned and start at that point to design what
you can build taking into consideration what you have available to use.
1) Can sheet metal or large surfaces be used to distill water evaporated
from an open pan etc? Answer: Technically Yes, but practically No. If
one needs a practical way to produce drinking water before one dies of
thirst this method becomes very problematic. This method will produce a
small energy inefficient amount. I was able to get about one or two
drops every 10 sec by using this method. see open aluminum duct
It is easy to produce way too much steam and overwhelm the cooling area,
which needs to be as large as possible. If one produces too much steam
it keeps the condensation area hot all the time and thus no condensation
takes place. One will not be able to regulate the source of heat in a
primitive environment thus this method could be a problem.
Another big disadvantage is if the relative humidity is too low (say
below 40-50%) then condensation will evaporate before it runs off. This
method then will not then work. The source of heat whether burring
wood, gas, or electrical has a tendency to produce hot dry air above the
fire that will rapidly evaporate the condensed steam. In other words
one needs to transport the steam a bit away from directly above the fire
area before trying to condense it.
Sometimes it helps to slightly cap off the ends to reduce the flow. See
http://home1.gte.net/mikelob/Tea_Pot_ends_closed.JPG . If there are
contaminates in the surrounding air that are settling out (like volcanic
ash) then this open process will allow the distilled water to become
Bottom line: The open-air process is the problem in this method it
introduces too many variables. After a PS in a primitive environment
one needs a controllable process that can be repeated under widely
varying weather conditions.
2) What about using a nearly closed area for condensing steam? This
works better than the above open sheet method but is still not that
practical in a primitive environment due to the need to carefully
control steam flow, and the size of the cooling area needs to be large.
Again too much flow easily overwhelms the surface area and will
produce no or a small amount of distilled water. The heat from the fire
producing hot air needs to be directed away from the condensing area.
For this method one can use 10 foot galvanized steel roof rain down
tubes, air-conditioning duct piping without insulation, or sheet metal
vent pipes. The tests http://home1.gte.net/mikelob/3inch-Ducting.JPG
were done using only one section of aluminum 3” vent flexible pipe 7 ft
long. The open upper end needs to be nearly closed off so that it
limits the steam flow. When properly adjusted, I was able to get 1 cup
of water every 1.5 hour using this 3” by 7-foot method. If one had
enough of these pipes and a way of directing the steam flow into each
then one could produce sufficient water for drinking. One advantage of
is this method is it can be made to limit the contaminants in the air
from entering into the distilled output.
Speculating now, if one had the space to do it, probably outdoors; one
could build a porcupine looking half circle made out of 5-10 rain down
spouts. One would take a bunch of these 8-10 foot rectangular tubes
made out of galvanized steel, and join the open end (down ward
direction) in a hemisphere shape. The upper closed end would have a
small hole that could be adjusted by partly closing it off to maximize
water output. I am thinking the ones that are nearly vertical would
work best with a small air escape hole, and the ones that are more
diagonal would work best with a slightly bigger hole. The water would
collect on the inside as the outside cools off. As the water runs down
the pipe and gets close to the end there could be smaller catching
tubing that could be positioned to allow the water to run off into a
collection point. The fire would be built off to one side and the steam
diverted by use of air ducting into this hemisphere. Heat from the fire
would be directed away from entering the tubes.
Another idea: I bet one could take 2 garbage cans or two 55 gallon tanks
or any 2 large metal containers and make a water distillation unit.
Visualize two garbage cans lying horizontally on something that blocks
it about 4 or 5 ft off the ground. Set the bottoms facing each other and
flush with each other. In the center off to one side (not under the
cans) visualize a fire built with water in a pan above it. Run an air
ducting vent pipe from over the boiling water to divert the steam into a
cut hole in each can. This would be big enough to allow the steam to
enter near the bottom of each can (right over the steam). The cans have
their lids in place and taped shut. A small hole (let the air out) may
be needed in the lid to tune it up to producing maximum distilled output
water. Because of the tapered shape of a garbage can the water would
run toward the lid of the can and out into a small catching container.
All of the above methods use air to take away enough heat to allow the
steam to condense as water this takes a lot of surface area if no
movement of the air is present. It would be much more efficient to use a
fan on the outside of the cooling surface. However, in a primitive
environment this will not be likely due to then need for electrical
power, thus I did not do any testing in this direction.
If one had available a car radiator or car heater unit turned diagonal
and allowed steam into the top in a slow controlled way so that no steam
came out the bottom then one could collect the distilled water as it ran
out the bottom. This would take some cleaning and purging of used
radiators to get all the foreign chemicals out.
3) What about using a bucket of cold water and coils of copper tubing to
condense the steam? This approach is basically building a simplistic
water-cooled heat exchanger. The input reservoir can be a commonly
available pressure cooker of any size. Alternatively the input
reservoir can be made from any able to be closed container. Not much
pressure is involved so theoretically it could even be made from a paint
can in a pinch. I used a 4.2-quart aluminum pressure cooker that took
about 15 min to start producing output once it is turned on from a cold
start. Using the attached test set up
http://home1.gte.net/mikelob/coil_exchanger-detail.JPG I was able to
produce about 6.5 cups/hour of distilled water. The input source was an
electric stove 240 volt x 7.3 amps = 1750 watts. Note that water is
replenished from the upper reservoir at the same rate it is turned to
steam by adjusting the needle valve on the supply bucket and keeping
tack of the level by use of the site gauge (silicon rubber hose) on the
side of the pressure cooker.
I found the water in the heat exchanger bucket gets extremely hot (from
top down) and needs completely changing approximately every 35 min.
This is a messy process in that when the tank is drained the copper
coils get extra hot and the output hose and rubber seals get very hot.
For good tasting water it is not good do to this extra heat on vinyl or
rubber parts and it also tends to introduce leaks in the bottom of the
exchanger tank where the copper tubing makes its seal with the stainless
steel bucket. After a couple of hours of running I didn’t consider this
design to be good enough for day-to-day production unit worthy of a
primitive environment. One could defiantly not walk away from it and
leave it unattended. It would take constant maintenance and attention.
Also the resulting water had a strong vinyl taste to it. As a result, I
don’t recommend using vinyl on any output part of the distillation
system. Along the same line, any hydrocarbons (oils) that are in the
water before distillation will end up in the output having been
evaporated and then condensed back into the output water. Thus taste
will be a problem using this method because of the estimated large
amount of hydrocarbons (oils) found in the after PS available water.
Also the method of measuring the water level in the pressure cooker
needed some redesigning. Silicon rubber fish tank air hose will soften
and swell up and eventually split open or leak at the ends. I initially
tried to use this for a water level indicator and rapidly found it was
not practical. Vinyl is worse it will swell, turn white, stretch and
becomes very soft so it is not very usable on the hot output side.
Glass tubing would work but is not available or practical in a primitive
4) What about using continuous but controlled flow of cold water in a
closed heat exchanger to condense the steam? What about, in order to
improve taste, evaporating the dissolved hydrocarbons before introducing
the water into the boiler? This approach is basically building a
water-cooled heat exchanger by using copper pipes and fittings. See the
linked test set up http://home1.gte.net/mikelob/Exch-Full.JPG. This
approach also has incorporated lessons learned about heating and
evaporating the hydrocarbons in the incoming water to the boiler. Note
by dripping hot waste heat exchanger water into a small copper cup, it
looses it hydrocarbons into the air. An efficient heat exchanger can be
made out of .5” and 1” copper pipe and T-fittings for the end see
http://home1.gte.net/mikelob/Exch-Full-1.JPG. Using the same electric
stove producing 1750 watts and no insulation on the hot spots, after one
hour I measure 2.8 Gal of 140 deg F waist hot water and 6 cups of output
distilled water at about 16 degrees above room temperature or 86 deg F
in my case. This is about 1.5 quarts/hr or 9 Gallons/day. More heat
or more insulation and it would produce more output water. More units
or a bigger pressure cooker and several heat exchangers running off the
same pot in parallel would also produce more output.
General Construction and Adjustment notes:
Be sure to remove the bur off the inside of the ¼” copper pipe that
solders to the open-air-replenish-cup. With the small hole the inside
bur leaves after pipe cutting it will sometimes cause the cup to
overflow (depending on surface tension of the water at the time).
The saddle clamp needle valve was about 2.6 turns open for operation.
The water valve to the bottom of the heat exchanger was open about 1/8
to ¼ turn. I used a typical stop valve that has a rubber washer in it.
This changes dimensions while in operation depending on temperature. The
result is a need to constantly adjust the flow rate. I would recommend
using a ½”gate valve instead. This has no rubber parts and the flow
once adjusted will not change with temperature. See
Rubber hoses works much better than vinyl but one cannot see though it.
5/8-inch car heater hose was used for sealing the ½ inch tubing
(outside diameter) that was slipped into the ½ inch copper pipe (inside
diameter) of the heat exchanger. This same hose was used for the output
wastewater. I don’t recommend using hose on the output of the heat
exchanger for the good distilled water. If one gets steam coming out
(too little water flow in the exchanger) then the taste of the hose will
get into the distilled output water.
The flow of cooling water in the heat exchanger can be adjusted by
measuring how hot outside 1 inch pipe gets along it’s length. If one is
careful one can use ones hand. Start at the cold end and gently move
along the pipe to the point of needing to take ones hand off the unit.
One wants to adjust the water flow so that it never gets hot at the cold
end and is hot to no less than the center of the length. This is all
depending on how much water you want to heat to what temperature.
Generally one can do 130 to 180 degrees F for wastewater out of the heat
exchanger without insulation.
Measuring about 140 degrees at the boiler replenish cup is about
typical. See http://home1.gte.net/mikelob/Replenish_Cup.JPG. If one has
a lot of taste (hydrocarbon) in the result then one wants to get the
replenish cup up to as high a temperature as possible by slowing down
the cooling water in the heat exchanger. This will cause the
hydrocarbons to boil out of the water as it drips though the open air
between the needle valve and the open copper replenish cup. The cup has
a baffled barer in it that blocks the surface of the input from reaching
the output side. This keeps the lighter hydrocarbons (oils) on the
input side until they evaporate. The output .25” tubing is positioned a
bit off the bottom of the cup.
Insolate to keep the heat in the cup on down to about 1/3 of the heat
exchanger and the full length of the exposed ¼ pipe that goes to the
pressure cooker See http://home1.gte.net/mikelob/Replenish-cup-Inso.JPG
and http://home1.gte.net/mikelob/Exch-full-Inso.JPG. Insolate to
keep the heat in the top of the pressure cooker and the pipe leading to
the exchanger. This will make the operation more efficient.
One can always make some charcoal by heating wood in a closed chamber
and burning the gasses that come off to help produce the heat necessary.
The charcoal can then be used to take the hydrocarbons out of the
distilled water by filtering it though a container full of crushed
charcoal. Hopefully this will be un-necessary if one can pre-heat the
water enough to drive off the bad tastes.
Silicon rubber II sealer seems to hold up well in the steam environment
and can be used to seal the pipe going through the top of the pressure
cooker if a rubber grommet is not available. See
http://home1.gte.net/mikelob/Exch-Lid.JPG. A rubber grommet from the
firewall or from the PVC valve from a junk after PS car may be able to
be used. The grommet used for the safety valve in the pressure cooker
may be able to be used. By the way I found the new “step drill bits” to
be quite usefully in drilling the hole in the lid of the pressure cooker
and when needed in the side near the bottom for supply 5 gallon buckets.
A float was used to measure water level inside the pressure cooker. See
http://home1.gte.net/mikelob/Float-Indicator.JPG. It was made out of a
large inline disposable gas filter with the ends cut off and soldered up
to make it air tight. See http://home1.gte.net/mikelob/Float-Hing-1.JPG.
3/32” Brass brazing rod was used for the float hinge mechanism. Any
other type of wire would also have worked including coat hanger wire.
Small copper pipe about 3/16 inch was used for the outer hinge and was
epoxied to the aluminum pot by use of copper epoxy. The copper epoxy is
designed to replace solder and is sold at home improvement stores like
The ½ inch pipe that comes out of the top of the pressure cooker and the
inside pipe of the heat exchanger needs to be just big enough in
diameter so as to not build up a back pressure in the cooker. Bigger
pots and hotter fires may need a bigger pipe or more than one heat
exchanger may be needed..
Occasionally the pressure cooker will need to be shutdown and the
contents empted out to dispose of the salts that build up. The frequency
will depend on the dissolved mineral content of the water.
Water taste: Using the copper coil approach the output had a strong
vinyl hose taste. Using the last approach with the heat exchanger and
the hot water dropped through open air to remove hydrocarbons before
entering the boiler chamber resulted in very little taste.
A bucket feed through for water outlet can be made using a barbed hose
fitting with pipe threads or a needle valve. See
http://home1.gte.net/mikelob/Bucket-feed-2.JPG. Metal washers are used
on both sides of the bucket. A coupling or another pipe fitting is used
on the other side to tighten the washers down on the bucket side wall.
Silicon rubber type II is used to seal the connection. Amazingly enough
standard washer sizes work fine. For example for .25” pipe thread one
uses two .5” metal washers.
Summary: Open air-cooled steam distillation introduces enough variables
that practical day-to-day distillation is not likely. Heat exchangers
that make hot water as a by-product look the most promising for
primitive water distillation.
All of the above methods assume some sort of heat source that can
produce steam. This assumption could be a problem for many areas due to
the near constant rain and lack of firewood, coal, tar or other burnable
resource. Some areas will be able to use hot earths crust to great
Warning: Hot Steam, hot water and hot metal will cause server burns as
it transfers it heat to the skin of the body rapidly.
The amount of heat given off when steam changes back to water is 540
calorie/gram (called heat of vaporization). The amount of heat it takes
to raise one gram of water one degree C is 1 calorie. Thus one can see
a lot of heat is capable of being transferred to the skin for very
little amount of steam.
Using these numbers it is easy to show that one can expect to produce
hot waist water of about 7 (above 190 degree F) to 14 (above 140 degree
F) times more volume of output water than the input water that was
converted to steam in the first place. This assumes a room temperature
of about 70 degree F as the starting temperature for the water.
List of parts for the last approach, number 4:
The unit was built to be broken apart and shipped in a standard home
Depot box no bigger then 18” by 18” by 25”. Recommend using only lead
free solder. Your prices might a bit different I included my current cost.
Quantity - Description:
2 coat hangers (used to make the support for each end of the
exchanger) see http://home1.gte.net/mikelob/Coat_Hangers.JPG
1 29” by .5” copper pipe (about $1.10)
1 23.5” by 1” copper pipe (about $3.00)
2 1” by .5” by .5” copper Ts (about $4.50 each)*
1 14.5” by .5” copper pipe (about $ .60)
1 4” by .5” copper pipe (about $.20)
1 2” by .5” copper pipe (about $.10)
1 .5” copper T (about $2.00)
1 .5” copper 90 degree elbow (about $1.98)
1 10.5” of soft copper tubing .5” OD (about $.80)
1 .5”solder by .75” garden hose water valve ($4.25)
1 20” by .25” copper tubing ($.30/ft)
1 .25” Self – Tapping Saddle Valve or Gate Valve if you can find it
1 1.5” copper end cap ($2.00)
1 6 ft washing machine input rubber hose (3/8” ID) one end cut off ($6.00)
1 6 ft of 5/8” car heater hose ($4.79)
5 hose clamps ($.69/each)
1 Barb 3/8 by 1/4 pipe tread for bucket feed though to hose. ($1.29)
2 ½” washers for bucket feed though (one inside, one outside, silicon
rubber seal in-between) ($.13/each)
1 1/4 inch pipe coupling for bucket feed through (for inside bucket)
1 Purolator (F20030) fuel filter (pep boys) for float in pressure
1 4.2 Qt Aluminum pressure cooker (Mirro brand $17.57 from Wal-Mart)
1 3/32 inch Bronze welding rod. ($.25)
1 3/16 small length section of copper tubing was used for float hinge
1 5 Gallon paint bucket for supply water ($3.00)
1 9/16” id rubber grommet ($1.54/3)
1 ¼” ID rubber grommet ($1.54/6)
1 pipe insulation ($1.19)
Total is about $74 for parts.
* This was the hardest item to find. I needed to go to a pipe supply
place. I in a pinch believe you can use .75” x .5” x .5” pipe and
fittings from local home improvement stores if you can not find this 1”
T. The 1” versions are better and worth looking for see
Assembly instructions after shipping to your destination:
1) With the pressure cooker lid in you lap slide in the output pipe down
from the top.
2) Put the insulation on the top of the lid sliding it around the pipe
if you are using it.
3) Insert the .25” pipe into the lid from top down.
4) Put the lid on the pot.
5) Install coat hanger support on faucet end of exchanger
6) Marry up the exchanger with the .5” pipe on the pot.
7) Tighten all hose clamps.
8) Hook up hoses to the supply bucket.
9) Put a container under the end of the exchanger to catch the distilled
10) Put water in the pressure cooker.
11) Fire up the heat.
12) Adjust the flow of water through the heat exchanger once producing