First off, I want to give full credit to Zoki51 for his generosity in presenting his DIY RDF project which formed the basis of this one. My project here isn’t anything new but it’s my take on fabricating an RDF from raw materials. This thread will be something of a blog/diary so no doubt you'll see a few goofs. The goal is to end up with a kick-butt RDF that’s utterly reliable, built to last decades, and is easy to maintain. That said, I’m not sure how useful the design will be to others because it involves fabrication using various tools and techniques that some may not want to pursue.
Background:
The previous RDF was based on Zoki51's design, but me being me, I strayed from his proven design in a few places. He used a straight-sided drum but not realizing that, I used a more ordinary drum with its typical “fat waist”, where the middle of the drum has a larger OD than the top and bottom. The problem comes when a sheet of filter material is wrapped around it, it causes all sorts of issues with folds or stretching*. Also, the drum wasn’t truly round, wasn’t square, and the wall thickness varied tremendously. There were other things about my implementation that I wanted to improve upon. The IBC container is very flimsy, fine for its intended use, but not so much for an RDF. The steel frame of the IBC is lightly galvanized so will start rusting in its watery environment. It wasn’t easy to ensure a liquid-tight seal to keep dirty water from going around the drum due to the IBC wall flexing. Then there was the sliding seal from the bulkhead to the drum. I “think” it’s working now but it could probably be improved upon – just as soon as I figure out how. The sprayer uses tap water at approximately 60 psi and is very loud, not great in an otherwise-serene pond setting.
RDF v2.0, The Beast
Many store-bought RDFs (in the US at least) tend to be manufactured overseas, are very expensive, and may use proprietary parts. The goal (and challenge) of RDF v2.0 is beat their reliability, flow rate, and cost:
1. High reliability
2. Uses easy-to-source parts
3. Easy maintenance
4. High flow capacity
5. Quiet
5. Competitively priced
Design decisions:
Drum “wrap”:
The design centers around using what’s commonly available. The key elements are the support covering the filter element and the filter material itself. The decision was made to base the design around a standard-size 48” x 96” sheet of thin expanded and flattened 316 stainless, which is cheaper than metal woven cloth of the same size. This size sheet results in a 30.5” diameter by 48” wide drum, which will have roughly double the flow capacity of my current plastic drum unit and giving huge flow capacity. “Cheap” is relative though, at $300, it’s a reminder that this isn’t a toy.
Filter element:
I chose to go with filter cloth again instead of stainless wire mesh. The availability of stainless mesh is a bit of a hodgepodge. That is, there seems to be no standard sizes, sometimes it’s 36” wide, sometimes 48”, and sometimes 54”. Available lengths is even more annoying, where “standard” can be 36”, 48”, or “by the foot” – with a catch. “By the foot” implies you can buy it, you know, by the foot, but they only sell it in strange increments. Want 8 feet of the stuff? You have to buy – and throw away - two feet, and around $70 a foot, it’s very wasteful. Polyester filter cloth on the other hand comes on a roll, so having spare material on-hand is not an issue and you obvious just cut off what you need, makes replacement faster and involves less bleeding when compared to stainless!
[Edit] I changed my mind later, as the polyester filter cloth kept tearing prematurely in the previous build.
Enclosure Material:
What to make the unit from came down to stainless or plastic. While I could have fabricated it from 316 stainless, it’s not fun to work with, makes me bleed, and is really hard on cutters. It also guarantees that practically no one would build this unit, so attention turned to plastics, specifically High Density polyethylene ("HDPE") or polypropylene ("PP"). Which is better depends who is asked though; plastics shops recommended PE, saying it's easier to fabrication, but others said that PP machines better. A variation is something called “HPDE Seaboard”, basically HDPE with additives to make it last in the sun. It comes at a price premium though and due to the additives, may or may not work with regular welding rod…
Yes, welding rod. It turns out that there’s an entire industry that uses plastic welding to fabricate liquid tanks, display cases, and jigs. A plastic welder is sort of like a cross between a hair dryer and a soldering iron and is used very much the same way as when welding metal. Having to buy and use a plastic welder may dissuade some people from pursuing this project, but it’ll at least demonstrate one way of building such a unit, and it’s a chance to learn something new!
Cutting Plastics:
Various methods are used much like when working with wood: a table saw, band saw, CNC router, and water jet. Special cutters are used along with a workflow that minimizes the material melting and sticking to the cutter. The slick way for a home builder to make this RDF is to get everything cut via CNC router or water jet cutter, then assemble it like a kit - unfortunately the problem is cost, but let me back up for a second to explain.
The design needs to address the following conditions: In operation, dirty water flows into the drum, then flows through the filter element and out the bottom. Over time, the filter plugs up and the outgoing water level drops, leaving the drum “holding the bag” so to speak. If the outgoing water level is, for example, 3” lower than the water inside the drum, then the drum is supporting approximately 65 lbs over 19” of its face. That force is directly supported by the filter element, so it has to be well supported to not tear, sag, or pull away from the drum and allow water to bypass the filter material. This is what the sheet of expanded and flattened 316 sheet does, used as a wrap-around collar for the entire drum and fully supporting the filter element. But we aren’t done yet…
When the output-side water level drops, the drum rotates the dirty portion of the filter up to the sprayer and it does its thing. While there is very little actual force against the filter material, it is still being pushed toward the interior of the drum and for the same reasons above it must be supported. Given that we have a clean sheet of paper, I chose to fabricate everything, including the drum, because nothing could be found in that size which was both the right material and perforated (I did find a perforated designer stainless trash can - $800, nope). Fabricating an RDF from scratch is a bit of a double-edged sword. It gives you complete design freedom, but also allows you to do things really wrong since there are no limits.
I decided to fabricate the wall of the drum using plastic ribs. Individually they aren’t very strong, but placing one every inch results in a very stiff yet light assembly (somewhat like an old wooden aircraft fuselage). The downside is cutting costs, quoted at $180-200 per disc. Given that there’d be three slotted bulkheads (top, bottom, and middle) and possibly two smaller supports mid-span, the entire project ended up with a total cutting charge of $1000-1200, not counting material! That was a no-go.
Change of plan
This seems like the perfect time to buy a router, something I’d always wanted but never had an excuse to get. Mounted to a temporary table, fixtures could be made to handle cutting both the slots as well as the large circular sections.
More to come...
*I used polyester filter cloth which is supposedly more resistant to plugging and is WAY cheaper. One issue is that it gets pulled one way when it’s supporting the weight of the water, and gets blown the other way with the sprayer. It’s unknown how long it’ll last getting pushed back and forth. What remains to be seen though, regardless of material used, is if bio growth on the screen will plug it up. It might… and might not. The chlorine in the tap water “may” prevent bio film build up, but we’ll see. Curious what other RDF users have found.