We had built a DIY RO system awhile ago. It worked OK. But it was still a slow producer to me. So, as time went on we made modifications and purchased high pressure pumps to improve the membrane efficiency. Then, buffering the output with a diaphragm storage tank. And so on.
Tradeoff is that the membranes needs to be flushed more often, keeping the system sanitary took time, lots of manual intervention … but we were producing more product.
Although, with little being automated, waste was high and it just became a real PITA over time.
We ended up purchasing a larger and self-contained RO system at auction. It’s the type of system used by laboratories, photographic development processors, and so on.
We set-aside some space in a basement corner and ran a couple of poly lines leading to taps. This allows us to use the RO for consumption, gardening, cooking, and whatever else.
The RO system has a couple of loops. A generation loop and the consumption loop. The generation loop dumps the output product into a 30 gallon buffer tank. A distribution pump pressurizes the output line/ The pump turns on whenever there is a pressure drop (someone opens a tap). The regeneration loop turns on whenever the RO tank level drops below a predetermined set-point.
The system also has a bunch of automation to include automatic flushing and sanitization cycles. UV-C lamps in several locations within the system minimization/eliminates biological infection. Etc.
With that in place, we don’t really have to do much of anything beyond replacing filters every so often, doing a system sanitization, etc. Workload is reduced to checks and maintenance a couple of times per years or thereabouts.
With a the non-automated system, technically we should be doing membrane flushes regularly (daily) even when the RO is not being used. That’s not happening. Automation helps that.
I forget what it cost but it was around 1K (used). Then add another ~700 for LTL delivery. Plumbing maybe 100-200 dollars in parts. So maybe a couple of thousand invested. It was used and there was a risk there but every thing was working in the end. We had probably spent as much cobbling together a more manual and less capable DIY solution. Such commercial systems still have a costs associated but the expensive membrane elements are much better protected.
The reason I mention this is simply because RO is slow, slow, slow. If you have a system that needs a larger quantity of water, the consumer grade systems can become frustrating. At least it was for me.
Here are some thoughts when figuring out your system:
- Do a paper exercise to estimate demand usage. E.g. 30 gallon demand need. Or, 1 gallon demand need, etc.
- Using a conservative regeneration metric (specifications sheets will be optimistic), determine how long it’ll take to generate the demand. System pressure and temperature is key.
- Consider a buffer tank and distribution pump. And, consider how to keep that sanitary.
- Consider if you need to pressurize the water between the source and the membranes to >80 psi. That would require another pump and plumbing that can handle that. Membranes exist for low pressures but their efficiencies are not great … unless you are willing to give up some RO purity, increase the waste product, and/or increase the time element.
- From the above, figure the level of pain to include time to generate the demand, maintenance needs, sanitization, flushing needs, amount of automation, costs/time.
There is going to be a huge difference in pain depending if there is a demand need for 1 gallon vs 30 gallon. We had made several attempts in the DIY space and eventually could have built an equivalent system but, turns out, taking a risk and investing in a used commercial system has worked out pretty well so far.
Here’s an image of that particular system in the basement. The blue distribution tank in the back also has a UV-C cycle that runs once per day:
I wouldn’t necessarily suggest something like this for most unless there is a regular high demand load. E.g. when you want 20 gallons now, it’s available now.