Rain water harvesting and slow sand water filters

The water clarity of the output from filter 4 has been getting worse and worse up until 2 days ago. An air vent on the top of the barrel was added after I noticed the output pulsing while the pump was in operation. I realized at that time that the pump was actually drawing water from the filter’s output because the water had reached the top of the barrel beyond the overflow drain just below the top and was forming a seal so as to allow the pump drawing water to create a vacuum strong enough to briefly cause water to be forcefully drawn through the filter much faster than it could handle. This was (apparently) causing the increased turbidity and (possibly) the biological contamination of the output water. So. . . know that:

The storage container, especially when used with a pump, must be vented above the water level in the container or the pumping action from the small recirculation motor  (or any pumping action) will actually forcefully draw water from the filter at a destructive rate.

When the filter was first set up the flow rate was too fast, a regulator valve was added.

.25mm sand on the top 4 inches;  .45 mm sand bottom; pea gravel bottom 4 inches;  carbon granule filter on output;  3 weeks running after extended freeze. As of today March 7:  24 degrees F.

March 6 2012:  flow volume:  37.5 litres per hour ; flow rate: .140 meters per hour

Compared to the other filters I have put together, this one is slightly different. The majority of the sand is .45 mm effective size. This filter appears to be more subject to variations in the input water quality. Heavy rain was followed by the water output clouding up slightly at the beginning of the filter’s initial set up. Also, filter 4 appears to be more readily effected by freezing weather. Freezing weather was followed by the output water becoming highly turbid for 10 days after the thaw. The filters with .25mm  sand in the bottom layer did not do this. It is important to note, however, that this filter was set up within weeks of the freezing weather and the biolayer was not fully developed when the filter froze. The other filters (filter 1,2,3 and the pond filters) had 6 months to develop a biolayer before winter. More later . . .   .

Feb. 28 6:30 AM PST: The output water directly from the filter is now perfectly clear based on visual inspection. The filter did not freeze last night, although the temperature was 30 degrees F.

Feb. 27 5 PM PST : The water directly from the filter output is now “clear” like it was before the cold weather set in and froze the filter. From this perhaps we can conclude that the freezing weather caused the filter to produce more turbid water temporarily?

As of Feb. 26, the output water was running clear. Now, as of Feb 27 AM, the outside temperature has dropped to 26 degrees. The filter is frozen, and there is no output.

As of Feb. 25 there has been more improvement, but very slight. I almost cannot see any cloudiness in the water at this point in time. The uv filter would probably work now, but not at 100 percent efficiency. Check the link below for an explanation of how a uv filter works. This is a good learning example. These filters will be effected by freezing weather and changes in water input turbidity. On the positive side note that the output could still be free of coliform even though it is cloudy, but we do not know for sure until the water is tested. Also know that even if the water “looks” perfectly clear, it still may be contaminated with harmful microbes, or toxic heavy metals such as mercury, or lead.

To keep out contamination, the system must be sealed, tested regularly, and the roof must not contain any non-biological contaminants; such as lead, copper, mercury, zinc or asbestos. All these are chemicals that can be found in toxic amounts in runoff from roofing material and / or gutters in a worst-case scenario.

As Feb. 24 2012 there is a very slight improvement – not enough, yet –  but it does “look” better.  Perhaps  the carbon filter has helped. . .   ?

Feb. 22, 2012: Filter 4 is putting out cloudy water. (If you can “see” cloudyness in water, it won’t allow a uv filter to work.) Specifically, the turbidity of the output has increased substantially – last week the water “looked” perfectly clear based on a visual observation – now the water “looks” slightly cloudy. I do not know why for sure, but my best guess is that the effective grain size of the coarse sand is too large – .45 mm is likely just too large – but that is just a guess. This has happened before, to a much lesser degree, on all the filters I have put together – after freezing weather. Another possible explanation for the increase in turbidity  is that the outside temperature has increased substantially immediately after the freezing temperatures, and there has been a substantial amount of rainfall in the past 3 days. I have added a carbon granule filter, but so far it has not made much difference. I’ll update as soon as there is new information.

If I were using a UV filter to assure absolute purification, the high turbidity would prevent ultra-violet light from adequately sterilizing the water. (the preceding link will open a new page, on that page click on the “GET A COPY”  link at the top of the dialogue window) A turbidity level of 5ntu is the maximum level at which a uv light will work – any cloudier and the uv system will fail. Five (5) ntu is just barely visible to the average person. The water may still “look” clear.

The water is not clear enough for a uv filter to operate at 100 percent efficiency (because I can “see” the cloudiness in the water – and that is too much – remember err on the side of caution) , and there would be a very real possibility of contamination in the output following the uv filter.  A week ago, the water was very clear; and a uv filter would have worked. This is why it is extremely important to note that a DIY slow sand filter is not foolproof; and furthermore, THERE ARE NO GUARANTEES THAT a filter built exactly like this one will work all the time. Again, I must state that I DO NOT RECOMMEND using water from any filter described on any of my websites for drinking water. The owner/operator of any slow sand filter is TOTALLY responsible for the water quality, and type of use, irregardless of what they read or hear from me; and contamination can come from anywhere at any time.

Here are some (large) photos of the sand sizes used in filter 4. The .45 mm is the lower level of sand , and the .25 mm

is the upper level.

NEW

I spoke with the testing lab this afternoon regarding filter 4. The filter is working. “Working”, means that the filter is removing coliform bacteria completely from the input water; however a word of caution here: this is only for this filter at this location and is only one test. It must be kept in mind that contamination can come from anywhere and each owner is solely responsible for the operation of the filter they build and/or own. It is totally possible to build a filter identical to this one and have bacteria show up in the output at any time. Also note that the “log reduction, (usually called “log credit”) on this filter has only been shown to be 3.08 with 1200 cfu/100ml input. This means that unless the water running through this filter is equally or less contaminated than 1200 cfu/100ml the output may be toxic.

For example, let’s say the water coming off the roof suddenly has 2000 cfu/100ml (this can easily happen anytime – contamination levels in roof water will vary SIGNIFICANTLY) ; then this filter may not work to remove all the pathogens. Or,  perhaps someone else builds an “exact copy” of this filter and starts running their roof water through it. There is absolutely no guarantee that the filter will work – it “might” work. There is no way to tell. It might produce horrifically toxic water. A test must be done on each filter, and furthermore, the water should be tested and monitored regularly with adequate safeguards for possible contamination. Even if a UV filter is used, if the turbidity goes above the limit for the particular type of light, then it will not inactivate pathogens, so the turbidity level must be monitored closely also. This filter has not had a turbidity test done yet.

The results are:

Post-filter: no detectable coliform, this means no e coli or fecal coliform are in the output water.

Pre-filter: 1200 cfu/100 ml.

Storage: 3 cfu/100 ml. no e coli or fecal coliform.

The results can be viewed here: http://www.roofwaterharvesting.org/test_results.html

Specifically, this means that water directly out of the filter has no detectable coliform bacteria in it. Water that has been going into the filter on an ongoing basis has a high coliform count, and water from storage is slightly contaminated with 3 coliform per 100ml but contains no e coli. This does show that contamination can come from anwhere, as I had meticulously and extra carefully cleaned out the storage container with disinfectant and made sure the container was closed tightly. Still, from somewhere some coliform has managed to get into the system after the filter output. However, inside the filter at the output and inside the output pipe there is only purified water. The actual paperwork will be coming back to me in several weeks at which time I will post it online.

This is a quick video I made to demonstrate the operation of filter 4. Refer to the previous post of November 13 2011 “Slow sand filter construction” on this blog for more information about the parts of the filter system shown in this video. The recirculation procedure is the same as on filter 1. Filter 1 and the recirculation system has been working for over 3 years now. The dry season here leaves us with no rainwater flowing through the filters so we must have some way to make sure water gets circulated through the filter to keep the oxygen and food available for the living biological layer.  This works well by occasionally adding 5 gallons of water from the stored unfiltered water in the large 55 gallon barrel used as the diverter. Only adding a small amount of the stagnant water every week or so is enough to keep the filter alive and working during the summer months.

January 31, 2012:
Filter 4 (posted on November 13, 2011) has been frozen intermittently since the end of December 2011. We had a massive snow/ice storm here in January, and I am still cleaning up the damage (nothing serious, but some very close calls with large trees), and this will be weeks of work. Because of these situations, the output has not been tested yet. The output water is “clear”, and has a “fresh water smell” similar to the fresh air after a rain shower, and the output water does not form bubbles on the water surface as it flows into the container (indicating reasonable purity) but still the turbidity is not satisfactory. Visual inspection with a low power red laser indicates that the turbidity has not improved significantly since the water started “looking” clear. As of now (Jan 31, 2012) there has been a 5 day period of continuous running time since the last freeze. If the current non freezing weather holds, the tests will go in, hopefully, at the end of this week on Feb 3 2012. More later, as I get time.

Input and output water for Filter 4. This is output and input on an ongoing basis.

After a very informative conversation with Jim, the author of several of the comments on this blog, it has become obvious that some additions to the websites and this blog are in order. First, where and how to get the sand needed to put together a small slow sand filter.

Links to sand suppliers:

Under “A variety of applications”, this website says they have sand that will work for “water filtration and purification”. I have not checked this one out closely yet. . . . . :
the Cemex company

Standard Sand and Silica company in Florida

Lane mountain sand company in Washington state

Red Flint sand company in Wisconsin

Sri Supreme in Marysville, CA

target sand company in Canada and USA
A subsidiary of the QUIKRETE® Companies

Unimin (this link is to the company’s locations page) A local supplier here in Seattle is Salmon Bay Sand and Gravel.

Manufacturers Mineral Company in Renton, WA (no website found) they have a .25 mm effective size sand called “TB Sand”.

1215 Monster Rd SW
Renton, WA 98057
(425) 228-2120

These suppliers may not sell directly to individuals, so it will be necessary to find a local sand and gravel company or hardware store to order the sand for you. The large hardware stores may be able to order it for you if you have the supplier’s name and the product description. Alternatively, one could dig sand from a riverbed, (taking extreme care to avoid destroying ecosystems), wash it, sterilize it, dry it and then run it through sieve to size it appropriately.

If you live in a place where there is lots of sand readily available, this may be the best option. . . however it will be extremely time consuming, one would have to purchase the proper sieve devices which would likely be super expensive. I have already tried using un sized sand directly from the ground. It works but not very well. In an emergency situation, sifting sand through a makeshift sieve would probably work and would be able to provide safe water, but it would not be ideal.

Can “play sand” be used for a slow sand filter? Yes, but it is difficult to wash and contains lots of silt and furthermore the sizes of the grains vary widely. As a result the filter will plug up very quickly. Can mason’s sand be used? Maybe, depending on what is in it. Again, the effective size and uniformity coefficient may not be suitable for filtration purposes. What about sand from the beach? Probably not, unless the sand is washed and sterilized which would likely cost WAY more than just buying filter sand. If the “beach” is next to salt water, then the sand would be nearly impossible to clean. If the “beach” is contaminated then the sand would also be contaminated. If the “beach” is on fresh water and contamination is minimum, then the sand may be usable, but be aware that there may be horrific contamination from anywhere. The sand would need to be washed, and then heated to very high temps to be safe.

Much more info later as I get time. . . . .