Update Aug 2016 –
Watermaker still working, but not without some issues. The gauges seem to rust – the flow meter balls don’t seem to like the pickeling agen used over the summer, so in the future, you might find different gauges. Otherwise the hp pump is ok, but the water puppy has needed a new impeller.
pixs seem to have dissapeared – need to fix.
Update May 21, 2014
I’ll be updating his project, creating a one year status report, reporting on additions we’ve made, etc.
Part I Feb 7, 2013
Water Maker Project
This article is to describe the watermaker we made for s/v cetacean our Tayana 37 cutter. The project had a few goals, be as cost effective as possible, have it fit in the available space ( and since when is there any available space in a 37 foot cruising boat, with two people aboard? Be as efficient as possible or at least not waste energy. I was thinking that running the engine for re-charging and desalinating water at the same time constituted a good use of time and resources, so that was the goal. Lastly, use readily available parts so that if repairs or replacement parts/filters/membranes was necessary, we would have multiple sources. With the above in mind, I’d like to start with a description of osmosis since I found it helped when needing to make decisions on parts or layout. After the basic osmosis discussion we’ll get into planning, i.e energy source, layout, then parts and finally construction, operation and maintenance.
Reverse Osmosis Desalinators are the usual and practical way for small, remote communities or yachts with access to sea water to make fresh water. We are not really making water, but watermaker is the popular name, so it will be used in this article. It is a relatively energy intensive process, needing to overcome osmotic pressure and frictions, so how to power it is a major consideration. Osmosis is the movement of a solvent, like water, across a semipermeable membrane from an area of low concentration of dissolved solids to an area of higher concentration of dissolved solids. Osmosis will effectively try to equalize the concentrations by diluting the more concentrated partition.
Figure 1: Osmosis
In figure 1: osmosis , left side, the vessel has different concentrations on each side of the semipermeable(permeable to water but not the dissolved solids, like salt, so its semi-permeable) membrane (green dotted line). On the right, osmotic pressure has pushed water (but not the dissolved solids) through the (green dotted line) semipermeable membrane. The difference in heights of the spaces represents the osmotic pressure. Osmotic pressure is surprisingly large. Reversing that process , takes energy, and pushes the water with dissolved solids through a semipermeable membrane, leaving the dissolved solids. Because the membrane only lets the solution through, the salt (dissolved solids) stays on the outside.
Figure 2: reverse osmosis
In a watermaker, a high pressure pump is used to push the salt water solution backwards thru the membrane. About 40% of the source water is pushed through the membrane (single membrane). is now fresh drinking water. 60% of the water remains as concentrate and exits the system. To make 20 gallons of fresh water you need to start with 50 gallons of sea water in our simple system. If the amount of recovered water is important then there are ways using multiple osmosis membranes to raise the efficiency. But with yacht systems, there is a fairly large body of water to draw from, so I wasn’t worried about amount of water used. Osmosis is why people can’t drink sea water to quench a thrust. When we drink sea water the kidneys try to filter the salt from the blood. But they can’t, they need more water, so water is osmotically drawn from the cells to help make urine, and the cell dehydrates, not a good thing. Osmotic pressure is variable and based on the concentration of dissolved solids, but ordinary sea water at 75 degrees F exerts an osmotic pressure of about 310 lbs/sq in. In order to make water we need to exert enough force to overcome osmotic pressure plus overcome any frictions in the system. Just a little aside. I was thinking that 310 lb/sq in is a lot of pressure, so wouldn’t it be possible to make use of that force to generate energy. A quick Google search resulted in a number of experimental projects using osmotic pressure as a power generating source. Pretty cool. Reverse osmosis removes, won’t let pass, anything larger than about 0.001 micro. Large molecules, viruses, bacteria are all larger. As you can see from Figure 3: What does reverse osmosis remove (figure is from esp water products). The membrane essentially filters anything larger that small molecules from the water. That includes most trace elements, virus, bacteria, etc. That leaves, water, H2O. It also shows the need to prefilter the system. Since the membrane is such a good catcher of almost everything, it is easily is messed up.
Figure 3: What does reverse osmosis remove
Osmosis will remove just about everything that we don’t want to drink, like viruses, bacteria, heavy metals and salt, but there is always a chance that some part misbehaves or the water is contaminated by another means, so some people put short wave UV sterilizers in line with the product water before it enters the tank. We don’t have or currently use a sterilizer, but I’ll touch on the subject later.
This project is not really complicated or hard. It does take a long time. But it does require a basic understanding of plumbing and electrical techniques. So if you’ve never heard of Teflon tape or don’t know what NPT stands for than this project is a bit much. But…. anyone with a moderate knowledge of basic plumbing, mechanics, electricity should be successful.
We’ll go one subsection at a time:
- Through hull and pre-filtering
- Primary or pre-pump (before the high pressure pump).
- High pressure pump and pressure vessel and hoses
- Control panel and outputs
Where should the sea water come into the boat? Some watermaker manufacturers mandate a through hull be dedicated to the unit. I suppose the reasoning is: if other devices share the water from the same thruhull then the watermaker might not get its proper share and run dry. I also suppose that some boat owners don’t know their systems well enough to know what “other” systems connect to any given through-hull further some of these systems may be automatic, for example generator cooling water, so may be used during seemingly random times. Since I and anyone else that builds their own ro system understand how the through hulls are used, we can safely share a through. Just don’t let the watermaker run dry! It’s important to pick a through hull as deep as possible. Not only will this prevent air when the boat is heeled, but it also eliminates sucking surface scum and oils. Don’t use through hulls that share with an engine, washing machine, dish washer or other high volume, high frequency device (not that we have most of that stuff). The through hull I decided on is normally about two feet below the waterline and with multiple uses:
- Anchor wash-down intake
- Galley sink sea water foot pump. On Cetacean we have two foot pumps – one for fresh the other for salt water. BWM (Before Watermaker) we used the salt water pump for pre –rinse of galley dishes, cooking spaghetti, etc. Now I’m not sure we will continue those practices. But it’s still connected. None of these systems uses water unless specifically operated. So even if the through hull is “on” the other systems don’t use any water.
It’s necessary to have a series of filters prior to the membrane. The coarsest is a sea strainer, immediately after the through hull. We picked one with a clear bowl – easier to see dirt inside. It usually pays to put only high quality parts adjacent to a through hull. I read once, that ABYC suggest that through hulls and adjacent equipment pass a step on test. That is, you need to be able to stomp or otherwise bang on any through hull related connections so that physical impacts won’t damage the through-hull. To quote: “A seacock shall be securely mounted so that the system will withstand 500 pounds of static force applied for 30 seconds to the inboard end of its connecting fitting, at any point in its most vulnerable direction, without the system failing to perform as intended.” I think that’s good advice, and for the stuff downstream of the seacock as well. After the sea strainer, there’s a set of filters (micro-filters) to keep stuff from clogging the membrane. A filter of 5µ or less is highly recommended. To keep the 5filter from clogging there is a courser filter preceding it, I use a 20 µ.. I used a standard cartridge filter housings from Home Depot. These filter housings come in a few standard sizes, the most common is the 10” length. They are quite inexpensive (compared to most marine products), and readily available. Note that they come, with different pipe size connections. However, even though I bought the housings at H.D. , the filter elements available there were not suitable for saltwater. Polypropylene filters are used in salt water and had to be purchased from other sources. In Cetacean, there wasn’t enough room to put all these filters near the through hull, so they are mounted in the galley, about 20 feet away. Insure good access to these two filters because they are changed relatively often and the wrench to open the canister requires some swing room. I didn’t get clear canisters but that would not be a bad idea. The size of the water tubing used to connect parts is significant and should be planned for the system as designed and expansion. I choose ¾ inch ID for the tubing for everything up to the high pressure pump input. There were a number of reasons:
- The length of all the above mentioned tubing ended up at about 30 feet, so ¾” was selected to limit frictional losses.
- It allows for expansion – increase the capacity later.
I now believe that ¾ is overkill, even for a 30’ run and that ½ ID would be more than adequate. The canister filters come with any pipe size you want. I choose 3/4 inch to fit the tubing size but they are available as ½ and I believe 3/8 as well. I used http://flexpvc.com/WaterFlowBasedOnPipeSize.shtml as a reference but this is for PVC pipe so it’s only a rough guess for flexible water tubing. I figure four gallons of brine for every one gallon of product meaning a 20 gal/hr system will need 100 gallons per hour of input flow. So, assume 4X the end product volume of water to brine when making your calculations.
Depending on where the high pressure pump is located you may or may not need an electric pre-pump. The high pressure pumps require water to be “fed” to them. If the high pressure pump is mounted sufficiently below the through hull, then gravity may provide enough pressure and you won’t need this pump. In our installation the high pressure pump is higher than the through hull so we needed this pump. The pump sucks the water out of the through hull and pressurizers it between 10-30psi, more than sufficient to feed the system. I used a Jabcso water puppy for the electric primary/feed pump. I found them on sale for around $60. You just want something that will lift water up from the through hull and apply enough water pressure to the high pressure pump. I wired the pump to an unused circuit breaker on the main panel. I connected a pressure gauge to the output of this pump and plumbed it over to the control panel to monitor the input. This gauge can be used to monitor proper operation and health of the pre-filters . Since I’m kind of anal, I added a air release to the tubing near the gage – to bleed air out of the system. Depending on where you have the gauge there could be a relative sucking so purchasing a 30-0-30 type gauge is probably a good idea.
High Pressure Pump
I used a Hypro 2345B for its price, availability, footprint and capacity at 4.8 gpm @ 2500 psi running at 1725 RPM max. Since my system is design for 40 gal/hr (but I only installed one membrane so it’s currently a 20 gal/hr system) that translates to less than 4 gal/min so this pump is adequate for the current 20 gal/hr system and future expansion. The Hpro Company also provides CAD drawings to help in making brackets and making models to find room for the pump. Since the head of the Hypro is brass it has to be kept free of salt water during storage. If that seems to be a problem for you, most of these pumps are available in stainless too, but at a premium price. I had a local welder fabracte a bracket to hold the pump in position. The best way to mount the pump, if it’s engine driven, it to mount the pump directly on the engine and not to the boat. That eliminated engine viration from effecting the system. That’s not what I did. But I did a fairly exastive study of how the engine moved in operation and convinced myself that it was ok.
There really wasn’t room for a electric clutch in our install but it would have been nice. My bracket is pretty simple, but I need attache the belt by hand every time we make water – also there’s a wing nut and spring to tighten belt tension. It’s not hard but kind of a pain. I calculated from the manufactures specifications what my oprerating RPM needed to be and bought “v” pullies to match. My range is fromengine idle to a engine at moderate RPM to allow water to be made at anchore or maybe moving slowly. The pump will not accomidate full cruising RPM and pump operation at the same time. In my installation, the difficult part was fabricating a bracket to hold the pump. I ended up making a foam model of the pump, fastening the model into the proper position and then taking measurements for a bracket that would hold the pump into that exact position.
I lag bolted the bracket to the engine compartment. I elected the engine driven method, but an electric motor either 12VDC or 110VA can be put into place instead. If you want to use an electric motor, then you need to do the homework as to the size of the motor. The calculations are straight forward. I used the engine because I didn’t have the space for the electric motor and I eliminated the motor losses. In retrospect, if I had the time, I think I’d use a DC motor and put my money into more solar, wind or maybe a generator.
Add a pressure release to the pump output in case too much pressure is ever generated. That can happen reasonably easily if the needle valve is closed on start-up. I used an adjustable unit. I could have paid to have it shipped already calibrated but did not. Unless you feel confident you can calibrate the release then purchase an already calibrated unit. I set the unloader valve for 900 psi. To calibrate I connected the high pressure pump to the release valve with a gauge attached. Nothing else. I then turned the pump – first time I turned the pump by hand – and adjusted the release valve at my limit which was 900 psi. After a few trials the release valve reliably vented at 900 psi. I believe that these valves are also known as unloader valves. I ran the overflow water outlet to the bilge through a hose.
High Pressure Hose
I spent a lot of time picking a hose – but in the end I used what was available. If I could do it again I’d opt for an armored hose. My pressure vessel is located about six feet from the high pressure pump and I needed a high pressure hose usable for pure water. Lots of hoses are available for hydraulic use, like tractors, etc, but I didn’t trust the hose to not contaminate our water. I ended up with an Eaton product in the Synflex line, made for high pressure pure water applications. The hose is usable past 2000psi. But I have to tell you – the company I ordered the hose from also swaged the fittings. I ended up going back to the company three times after the swages failed at about 700 psi, three times. Once it released in my face. After the 3rd time I insisted they go back to the manufacturer and they found they had the wrong swage specifications – it’s has worked perfectly since then. I had a second piece of the same hose made to connect the brine side of the pressure vessel to the control panel – needle valve .
Pressure Vessel and Membrane
I purchased a pressure vessel made by Codeline , the M2500 with aluminum clamp-on end caps. I was told this type of endcap was easier to assemble so opted that option. The end caps are threaded for ¼ NPT fittings, for feed, brine and product (feed, concentrate, permeate). The pressure vessel has one input and two outputs. The input side is from the high pressure pump. The two outputs are for brine (concentrate) and product water(permeate). These holes were not marked on my unit, so I ended up taking great pains to make sure I got it right.
I purchased high pressure brass fittings for all these connections and used Teflon tape at all the fittings. The membrane is a Dow Filmteck SW30 – 2540. This membrane is optimized for sea water so is a good choice for boat based systems.
The membrane housing is mounted to the boat with four plywood brackets (shown in fig 2) I used shock cord screwed to the plywood on one side and a small loop created with seizing to fit over a screw on the other side of the bracket. The housing is easily removed for maintenance by unhooking the shock cord.
The panel is a good placeto mount all the gauges and flowmeters. Essentially you are adjusting (restricting) the flow of output brine to build up the input side past the osmotic pressure of 800psi. It probably makes sense to talk a bit about terminology. In reverse osmosis parlance the product water is called permeate, the brine or waste output is called concentrate, the input called feed.
Its worth using and remembering these terms so when you read the reference literature mentioned at the end of this article you understand what they are talking about. I put the needle valve, test/tank three way valve, flow meters and pressure gauges on the panel. If one were to use a permanently mounted TDS (salinity meter) it might go here too. My panel is a piece of stainless steel sheet over a piece of plywood with a teak frame. It took some time cutting the holes but I think it came out ok. When you order the flow meters, needle valve, etc you will need to consider the mounting.
I’d recommend you check the pressure rating of all the brass or stainless fittings. Many of the hardware store brass fittings are low pressure only. Also, I used the large size flow meters, if I had to do it again I’d use the smaller ones.
Flush and Pickle
I make enough extra water each run to back flush the entire system and store the watermaker with pure product water between runs. The flush system I added a three way valve at the sea strainer. Attached to the valve is a length of water hose used to suck product water into the system. When I’ve made enough water I direct the product water to the “test” hose to fill up the bucket. I’m pretty careful to keep containments out of that bucket. The bucket then provides the flush water – the hose is inserted, the three way valve is flipped and the though hull is closed until the bucket is empty. To pickle- repeat the above but add the pickling solution to the bucket before flushing. I bought Sodium Metabisulfite from a Beer and Winemaking store for a lot less money than the watermaker companies want. I mix two tablespoons of the powder for every gallon of water. Costs I probably spent around $1500 on parts and brackets. But I probably have 500 hours of labor in planning and then in the install. The pump bracket was probably the hardest single portion and took the most time. I found rounding up the parts to be the 2nd hardest part of this project, but once done – it goes together pretty easily. Parts I have a parts and
- check oil in pump
- open thruhull
- put flush/sea water valve to sea
- check brine faucet in head (it needs to aim into the sink)
- get bucket for flush
- put test hose in sink
- put test/product valve in test position
- put belt on pump adjust belt tension
- check needle valve for position
- check valve on flow meters turn on primer pump
- check for leaks turn on engine
- check primary side pressure record created/hours
- check water level in tank make water put flush/salt water valve to flush –
- when done flush with fresh water
I’m probably going to add a UV sterilizer to the system.
Membranes can be compromised, harbors are polluted.
A friend has added UV sterilizer both before the primary pump and after the membrane (permeate) to his system. So, not needed in a perfect world, it seems like a good idea. He used the SC2.5 on the input and the sc1 on the permeate.
There’s a lot of information on the DOW site.