Results 1 to 12 of 12

Thread: Kurt's DIY UV Light

  1. #1
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505

    Kurt's DIY UV Light

    Figured it's time to start my first thread here.

    I like to make stuff; see my signature for what I've been doing over the last... many years. Anyhow, our koi pond which has lasted 22 years is tired out and falling apart. The liner is getting stiff, the deck full of dry rot and termites, the overhead shading collapsed due to the same rot, and the concrete filter container (5' x 10') is developing a rather alarming tilt. Either gopher holes under it are collapsing, roots are raising it up, or both. Anyhow, time to start completely over again.

    The pond is being redone and an all-new filtration system developed. I'm currently building a rotating drum filter with excellent guidance from Zoki51's thread, http://www.koiphen.com/forums/showth...137623-Diy-rdf. My build is similar enough to his that I can't claim any originality; you can read up about my few alternation over on that thread.

    Anyway, as soon as that's done, next will be the bio section with a built-in UV setup, which brings us to this project. Picture will be added as things progress.

    UV Background:
    Years ago I made a 40W UV filter with Home Depot plumbing parts and while it didn't leak, it also didn't really clear the water either. This unit didn't use a quartz tube which is probably why it failed; I recently read that unless a UV bulb is run at around 100 degrees F, its efficiency drops way off. Exposing it directly to 63 degree water like I did meant that it was outputting about 1/3 the UV that it could have been (all while consuming the same power from the wall!).

    It was later replaced by an 80W Emperor Aquatics unit which worked better, but like the previous unit it also failed to clear the water. I never did figure out why; it almost had to be too much water flow yet the measured flow and the UV unit did seem compatible.

    This Project:
    I decided to go with a 120W bulb for not only for the higher exposure but also the longer dwell time due to the longer tube length. Because off-the-shelf 120W units are so expensive, the plan was to fabricate it. The bulb, sleeve, and ballast were about $300 from Pentair (who now owns Emperor Aquatic). I went with them because they're a huge aquaculture supplier and have all the Emperor spare parts, so hopefully they'll continue to carry them for many years.

    The bio section of the filter will be built inside a black Intermediate Bulk Container (IBC), black because I read somewhere that the bacteria does better in the dark, so we'll see. Getting off track briefly, I tried many types of bio media: sponges, matts, lava rock, poly beads, but nothing did better (meaning: didn't eventually plug up) than simi-floating media* in a fluidized bed. A lot of people fixate on surface area and technically they're right, but after a couple years their claimed 50,000 sq. ft. surface area media is probably more like 500 sq. ft. due to internal plugging. The reality is that surface area is useless if it's not accessible.

    Anyway, the IBC is about the right size that instead of building a standalone UV housing, the quartz tube could span the full width of the container. From there things kind of started designing themselves. First was the realization that the media can't be directly exposed to UV else it would always be, well, dead. So that meant shielding the lamp. Then there was how to get the water around the lamp while keeping the media away, as was how to access both the lamp and sleeve from outside the IBC. Currently there are several approaches bouncing around in my head:

    • Have an assembly that resembles a capital "D", with the water leaving half way up the straight side. The UV lamp would be along the straight section and the sleeve extends straight out through the side of the IBC for maintenance. The middle curved part of the "D" would be the inlet side and is screened to keep media out, and the overall assembly shields media from direct UV exposure.
    • Have a 4" tube extend straight through the wall of the IBC via a Uniseal to nearly the other wall, then have it branch off left and right to screened inlets of roughly 24" each in order to minimize media getting stuck to the screen due to suction.
    • Slide the assembly to one corner of the IBC and instead have the pick up end run horizontally, allowing for a full-width inlet screen.
    • Instead of branching off left and right, have a 90 degree bend turn upward to a vertical screened inlet. This is the simplest and cheapest but there may be an issue with media getting stuck to the smaller screen area compared to the other approaches.


    Not sure which approach is best but a key plumbing component known as an "Elbow with heel" was purchased. The quartz sleeve will extend out through the "heel" while the water makes the turn in the large diameter section to flow into the pump. The elbow with heel is this one: https://www.grainger.com/product/MUE...rchQuery=3GUU7

    *So far, the most promising (price vs. performance) semi-floating media seems to be: http://pentairaes.com/sweetwater-swx-bio-media.html. It's within a few percent of K1 Micro in surface area but costs about half as much and isn't as small.
    Last edited by kimini; 12-08-2016 at 06:13 PM.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  2. #2
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    The planned design for the tube and lamp. The reasoning is that if the primary sleeve seal leaks, any water will be stopped by the second seal and drop to the ground instead of causing mischief in the electrics. This setup allows removing the lamp without taking down the system, as well as placing the lamp so that all the UV is hitting actively moving water instead wasting it on the stagnant areas near the heal. A 1/2" sealed electrical connector would work as well except it's too small for the connector to fit through; the wires would have to be removed first from the connector shell and fished through.

    Only one end of the 1"-diameter quartz tube is open, the other end is sealed (think "test tube") so it's just physically supported by a second 4 x 4 x 2 el.

    The idea is to place everything shown outside the IBC, passing the 4" pipe through the wall via a Uniseal with the second 90-degree el against the far wall. Not sure how far up off the floor to place it; probably mid-height so that when it's time to inspect the quartz sleeve, not much water has to be drained, but I have to think about it more.

    Oh, and to keep the setup versatile, all the UV fittings inside the IBC can be assembled and left dry - no glue - so the bits can be shifted around when trying new things in the future.

    So that's the plan, subject to change at any time!
    Attached Images Attached Images  
    Last edited by kimini; 12-09-2016 at 03:33 PM.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  3. #3
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    Being self-critical, one possible safety concern is that the above "stuff" sticks out a ways. Depending where the unit is installed or positioned, clumsy feet (such as mine) could snap off the electrical portion of the assembly. It might be worth putting some sort of cage around it to protect things.
    Last edited by kimini; 12-09-2016 at 03:34 PM.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  4. #4
    montwila's Avatar
    montwila is online now
    montwila
    is almost off extension cords.
    Senior Member ~ Koi Health Care Committee Member
    Join Date
    Aug 2009
    Location
    Auburn, WA.
    Posts
    2,097
    Having broken several quartz sleeves I would never rely on them to be structural. If I am seeing this correctly all the electrical end is mounted on the quartz sleeve.

    The other thing is that the 1" quartz sleeve in a 4" pipe is probably not going to achieve the needed dwell time for algae as a clarifier and certainly not at all for the bacteria or viruses as a sterilizer.
    Last edited by montwila; 12-09-2016 at 06:52 PM.

  5. #5
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    Thanks for the input! The sleeve is only holding the plastic cover which doesn't weigh much. As I said later, as long as it's not forced sideways it should be fine. Regarding the diameter, the thinking - maybe flawed - is that the higher output of this type lamp would penetrate further into the stream, but I guess we'll see! If it's only partly effective then it'll be rebuilt as a 3" diameter set up.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  6. #6
    montwila's Avatar
    montwila is online now
    montwila
    is almost off extension cords.
    Senior Member ~ Koi Health Care Committee Member
    Join Date
    Aug 2009
    Location
    Auburn, WA.
    Posts
    2,097

  7. #7
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    Luck has nothing to do with good - or bad - engineering decisions!
    Build a sports car from scratch! Check out midlana.com and kimini.com

  8. #8
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    I worked through the math and with a 3" ID tube, exposure time is 1 sec/gal with a 3600 GPH flow rate, and 2 seconds with a 4" ID tube. Granted UV transmission will be less, but it gives twice the exposure time. Also, the manufacturer uses these same parts in a 6" OD UV assembly, so that's at least 5" ID. Looks promising for a 4" ID tube.
    Last edited by kimini; 12-16-2016 at 11:56 PM.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  9. #9
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    I'm starting to waver a bit on building it into the IBC. Also, regarding the structural nature of the quartz tube, I realized all the parts associated with the electrics (compression fitting, 1.25" cap, and bulkhead fitting) can be replaced by a simple rubber stopper!
    Build a sports car from scratch! Check out midlana.com and kimini.com

  10. #10
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    I think I'm going to make the UV assembly stand-alone so that there are less-invasive modifications to the IBC. It also makes the various filter subassemblies more stand-alone and easier to modify in the future.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  11. #11
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    Unlike the RDF, a UV clarifier is about a million times easier. Once all the parts are on-hand it's just a matter of "adjusting" several of the PVC pipe parts before assembly.

    One end the housing supports the closed end of the quartz tube so it doesn't need an opening in the housing. A 2"-to-1.25" reducer and a plug are used. The ID of the reducer is increased slightly to 1.00" so that the tube can slide in.

    The other end of the housing requires a sealed connection to the outside. Start with a 1" compression fitting, cutting it to about one third its original length (only half is used, keep the compression seal from the other end as a spare). The OD is too small to be glued directly into the 2"-to-1.25" reducer so a 1.25"-to-1" reducer is also used. Before assembly drill/file/turn the ID of the small reducer to fit the compression fitting, then glue everything together.

    As far as the 4" outer tube goes, either schedule 40 pipe or drain pipe will work because it's not under pressure. Given the choice though, drain pipe more available, available in 10' lengths instead of 20', and is much cheaper.
    Attached Images Attached Images      
    Last edited by kimini; 12-22-2016 at 12:52 PM.
    Build a sports car from scratch! Check out midlana.com and kimini.com

  12. #12
    kimini is online now
    kimini
    has not specified a status.
    Senior Member
    Join Date
    Oct 2007
    Location
    SoCal
    Posts
    505
    Glued in the main PVC tube, added the quartz tube, lamp, and fabricated the electrical end housing. The ballast came with a screw-in part of the OEM assembly but it was a huge pipe thread that I didn't want to buy a tap for. The threads were ground off and it was glued into the second compression fitting. Compared to the first one I made, the closed-end quartz tubes they sell now make the project go together much faster and following a leak-check, will be considered complete.
    Attached Images Attached Images    
    Build a sports car from scratch! Check out midlana.com and kimini.com

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •