Sunday, April 14, 2019

Day 18, Powerheads be thy bane


It's been a rough few weeks, with a few ups, and a few downs.. more or less up though.  But there have been some developments....

I purchased most of the remaining equipment needed for the tank all at once. I honestly don't even remember everything that was in this order.  It was big.  But I've been slowly working through installing all of it.

First order of business, was tops for the tank. Because no fish were in there yet, I hadn't bothered to make tops for the tank to keep them safe.  I already had some kits on order from BRS, so ideally it was only going to be about 2 days without tops for the little Anthias swarm.  Turns out that was 1 day too long for one of them.  One of them got adventurous on the morning the kit arrived.  Damn.  Either way, I quickly got to work building some screens, and got them on the tank ASAP.

Next order of business, was the Tunze Safety Connector.  This tank is big, and valuable, and I'm really not interested in losing the whole thing due to a brownout or something stupid.  But I want to be really safe, so I went big.

Step 1, Optima marine battery.

Step 2, float charger.

Step 3, Box to hold the battery, to keep it off the floor, and protect it from errant splashing.

Perfect.. sorta...

The problem I have now is, the pump on that side of the tank has been acting up.  I have an original version of the Tunze 6255 on the tank.  Well, I have 2.  The one on the side of the tank that will be getting a battery, well that one seems to be having issues with it's speed controller.  It runs at 100%, or off, nothing else.  And randomly, if you turn it on and off, it won't turn on, and takes a few cycles to get going again.

This is no good for the powerhead that will keep everything alive in a power outage.  I even tested it with the connector, and about 50% of the time it would just shut off when power was removed.  No good.

So I decide, well, lets go get a Tunze Stream 3.  Those things look great.  Not as awesome in power as the 6255, but pretty good, and it will do the job.  Nope.

Tunze shows up, I read the manual, and there is one little gotcha.  The new streams, when set to external control (like 0-10v apex) will shut down completely when the external control is off.  This means in a power outage, the tunze will just shut off. Argh.

So I decide to goof around with the old 6255.  It doesn't have the big fancy box with buttons that the new one does.  It's got this tiny little brainless box with a single knob and some connectors.  Turns out, it runs just fine when power is pulled AND 10v is pulled.  Only problem is, the working one is on the other side of the tank.

So both 6255's have to come out, the right one gets swapped to the left side, the new Stream 3 goes into the right side.  OK, all wired up, seems to work, I'll call this good.

I still have the problem of insufficient flow in the tank though.  A pair of Tunze's does not an 800g tank make.  I've dug and dug and tried to find a way to put an apex controllable powerhead onto 1 1/8" thick acrylic, and nothing exists for that problem.  The WAV claims 1", but that's just shy of working.

So I buy a WAV, and a giant magnet. I mean giant.  Some 4" diameter 1" thick n52 monstrosity off Amazon, that ships with a dozen warning labels.  They both show up, and I try the combo out on some big chunks of thick acrylic that I've shimmed to the right thickness.  The magnet grabs the pump, and then pushes it sideways. Argh.. why?

So what I incorrectly assumed was, aquarium pump magnets were north south through the thickness of the mount.  Instead, the mount contains 2 magnets, one north, one south.  If you try to rotate the magnet, the pump rotates with it.  The giant honking magnet is a single pole, so it grabs one side really strong, and repels the other side.  I guess I need to do something else entirely.

I go to KJ Magnetics, and get a pair of rubber coated 2" x 1" x 1/2" magnets.  About $34 each.  With much difficulty, I manage to orient them north/south like the WAV, and test.  It works!

The WAV however, isn't super happy.  The magnets are kind of at a bad angle, and the WAV whines about POSERR every now and then.  It's a stable hold, but it's probably shaking a bit..

Enter the 3d printer:

OK.  Now What I did here was take the original back magnet from the WAV, and measure the centers of the magnets.  Then, I positioned the big ones so the centers lined up.  I printed this super thick, with tons of infill, so the magnets wouldn't tear it apart.  Does it work?

Why yes, it does!  I've since printed another one, and ordered another WAV for the other side!  Here is the Tunze in it's new home as well:

(The deep yellow color is from the sun being out, and the solar tubes doing their job. It's not that yellow in real life, but the camera does funny things.

What you might also notice in that photo, is the foxfaced rabbit.  He made it out of QT just fine, and has joined the Anthias in the main tank.   The Anthias, thankfully, are doing ok. There are about 23 of the original 26 left, which is roughly what I expected, but it's hard to count, because they don't sit still.  A male is also beginning to emerge, so it's exciting in there. They are all eating well, and looking pretty healthy.  There are a few scrawny little ones that I'm a bit worried about, but they all eat, and there hasn't been any bullying.

Speaking of water flow, the surge bucket made me completely insane this week.  It just stopped working. No obvious reason, it just stopped firing.  Every now and then, it would goof up, and I would have to reset it.   No big deal, shut the feed pump off for 1 minute, let the bucket drain, turn it back on, it fired normally.  This would happen maybe once a week.  But suddenly, it stopped working entirely.  I messed about with it over and over for a week.  Nothing I would do would fix it.  I had literally no idea what was wrong.

So I took the main tube inside it apart.  The zip tie that holds the little RO tube to the side of this was loose.  So I tightened it up, hard.  Then I put it back in, and started fiddling...

Turns out... that zip tie loosened up, and the tube moved downwards.  This caused the pressure equilibrium to come off it's center, and make the whole thing fail.  A few test fires, with me moving the tube up and down a bit, and suddenly my week long nightmare was over.  Thank goodness.  My Anthias need their autofeeder!

However, in the meantime, I figured out some code, to sorta bodge the problem.  I needed to auto-reset the surge whenever it failed.  This way the fish get fed, even if I'm not home to fix it!

This is what I came up with:

Fallback 50
Set 100
If Output Surge_B_Time = ON Then OFF
Min Time 004:00 Then OFF
If Surge CLOSED Then ON
Defer 045:00 Then ON
If Output Cor_Acc_V = OFF Then OFF
Min Time 005:10 Then OFF
If Output Surge_Break = ON Then ON
If Output Surge_Break = OFF Then OFF
Min Time 005:00 Then ON
The "Surge" device is a limit switch at the top of the bucket.  When the bucket fills, right before it hits the apex, the limit switch is activated.  This code, more or less, watches that switch.  If the switch doesn't fire for 45 minutes, it assumes the surge is broken, and turns on Surge_Break.  Once that is set, it sets a 5 minute timer, and turns off the feed pump.  This resets the whole system, and fires off the surge again.   It actually works pretty good!

I also got a pH probe for my kalk reactor.  While I was in there, I decided what I wanted was some code that would run the stirrer in the reactor, for about 20 minutes, immediately following the kalk dose.  That way the newly added RO water would mix in nicely.

If Output Kalk = OFF Then OFF
If Output Kalk = ON Then ON
Min Time 045:00 Then ON
Fallback OFF
OSC 000:00/008:00/352:00 Then ON
Fallback OFF
If Output KalkTimer = ON Then ON
If Output Kalk = ON Then OFF
Perfect!  Now the kalk strirrer turns on for about 30 minutes immediately after the kalk is dosed.  The best part of this is, I don't have to maintain the programming for both.  If I adjust the OSC statement in the Kalk outlet, it just magically works!  Woo!

In other minor news, I upgraded my DI system to a dual reactor, just so I don't have to swap the media out as often:
I got my second QT tank up and running:

And I got some new life!!

Barely visible.  A hermit crab, and 2 emerald crabs.  Also, more trochus snails.

And in the QT tanks:

A royal gramma, and 2 purple firefish.  (Look healthy, hope they do well...)

And what I believe is a male yellow-finned fairy wrasse.  He looks healthy, so here's hoping all of them do well in QT!

Well, it's been a long and difficult day, so I trudge off to bed, after feeding all the fish..

Tuesday, March 19, 2019

Day 17, The arrival of summer

It's March in Arizona, which means, it's basically about to be summer.  This brings with it all the fun concerns of heat and the tank.  But first, FTS!

On March 1st, at pretty much 1am, the Apex woke me up to tell me that the tank was too hot.  77 degrees.  This confused me a bit, because on Feb 28th, it was happy at 76.7.  I run a slightly modified version of the Apex temperature table (the one that varies month to month).  But why is it suddenly a problem, when all winter the tank has been struggling to keep up to 76.5?

Well it turns out, on that table, the coldest month is March.  Jan, Feb, all warmer than March.  This is no good, and also explains some of my pain the last 2 months with the heaters.  So the first thing I did was shift the entire table up by one month, so now the coldest month would be Feb, which is more in line with reality in Arizona.  This should also shut the heaters up a little bit next year.  I have other plans for the heaters, but not today, today is about cooling...

So, knowing that summer was here, it was time to test out the big wall mount fan that is mounted in the room, as it would soon be needed.  Turns out, it was dead.  No point in trying to repair it, I just got a new one.  The old one had a plastic blade, and years ago, I was awoken to a horrific rattle in the middle of the night.  The fan had built up some salt crud on the blade, which increased the weight, and it literally ripped the blades off the fan, which were smacking around inside the cage.  I did replace that blade, but it was a pain.   So goals for the new fan:
  1. Metal blade!
  2. Closer to the wall so I have more room to move around.
Found one on Amazon, swapped it out, all is happy!

A big tank needs a big fan.  This should do the job.  On a setting of 1, it moves a decent quantity of air.

Next step, the humidity.  This fan is going to evaporate some water.  That water will build up in that sealed room, and then pain and sorrow.  On the wall above all the equipment, I have a register in the wall.  It ducts through to a register in the hallway, so air can move out of the room.  At the bottom of that wall, there is an in-wall fan, that blows cold air from the floor level of the hallway into the room.  Problem is, the air just doesn't move around enough.  I used to have 2 120mm computer fans bolted to the register to blow air out, but as with all computer fans, they die in salt, so they were rusted shut.

So I went on Amazon, and found a basement crawlspace fan.  This is a fan specifically designed to reduce humidity in a crawlspace.  I got the AC Infinity AIRTITAN T8.  It has a little digital display, where you can set the temp and humidity level where it turns on and off, and a little remote probe so it senses the room, not the fan itself.

A quick hackjob with the saw, and I expanded the opening to fit, and now we have a fan!

You might notice the fan in that picture says the temp in the room is 84.  84!  It was a bit warm in there.  It's not that hot in the house, maybe 76 degrees, but that room was warm, and the heat rises, plus humidity.  I think the humidity in the room was around 55%.  No good.

Fan immediately spun up to full power, and within an hour, it was telling me the ceiling temperature was 79, so it quickly did it's job!  It also dropped the humidity to about 40%, but it's struggling to keep it there.  My goal is 45, so it runs pretty often.  Once the house AC is on, I think it will be easier, without the AC running, it's just kind of dealing with stagnant airflow.  Overall, for $90, a good buy!  For the curious, the fan in that configuration blows air OUT of the room, not in.  Also, the width of the fan is pretty much the width between 2 studs, so the screw holes line up perfectly to the studs.  Easy as pie to install.

Now those of you who are more observant, might notice that something is different in the FTS shot.  If you said fish, you sir, would be correct.  I kind of did a thing.

Normally I am super conservative when it comes to stocking.  I also have a QT setup all setup and ready to go.  My plan was, go to the LFS, and find a foxface, put it in QT, and in a few weeks, deal with the algae that was starting to grow a little.  Ok, easy plan.

So I get to the LFS, and of course, they have a tank with anthias.  I love anthias.  Why did I want a 800 gallon tank in the first place?  So I could have a giant school of anthias around a big acro head.

Again, normally, a few anthias wouldn't faze me.  But they had all juveniles, so super likely to be female, all same species (I think!) all arrived at once, and there were 26 of them.  Yeah.  26.  In 20 years, I've never seen more than 10 at a store at once.  I couldn't pass it up.  But of course, there is an issue here.  My 40g breeder QT is not equipped to handle 26 anthias at once...

So bad decision dinosaur won the argument, and 26 Dispar anthias went straight into the big tank.  There are no other fish in there, so worse comes to worse, nobody else gets sick.   I could not pass up the opportunity to have the school size that I really wanted, all at once (which IMHO is the best way to stock anthias), and.. well... I had to take the risk.

The foxface however, is currently in the QT tank.  Seems to be doing well.

Interesting thing about the anthias though..  I'm 99% certain they are Dispar..  but when they were bagging them up, I swore for just a second I saw vertical bars on one.  Then later, when the lights went out in the tank, I saw, just for a second, vertical bars on two of them.

I'm not aware of any juvenile mimicry in anthias, but it's possible I have 1-2 individuals of a different species in there, that are mixing together.  They are all acting as a single school right now, but I wonder if there is something else in there.  The only thing that matches the vertical bar pattern I saw was a Sunset Anthias.  It only shows up in the fright or night coloration, and only sometimes, but I suppose it's possible.

Excited about all the new life in the tank, I am barely able to go to sleep...

Wednesday, March 6, 2019

Day 16, Let there be life?

So lots of little things left to do on the tank at this point, as I slowly ramp it up into full operation.  Lights are currently on at 16% (of a max around 65).  So now I want to get a few things running on there, and try to button it all up.

So let's start with the easy stuff.  Since my calcium was low, but my dKH was correct, I decided to just mix up a gallon of CaCl and dose that into the tank over the next few weeks.  Rigged up a simple bottle for that to pump out of, and hooked it into the doser.  That should fix the base calcium levels.

Second, I hooked the Kalk stirrer up.  Went with an Avast K2. I really like Reef Octopus stuff, but I've heard lots of bad things about the motor units on those.  The Avast one, they advertise that you can run that stir motor 24/7, as opposed to periodically, so, ehh..  Went with that one.  Hopefully it was the right choice.

Normally on a big tank like this, the thing to do would be to hook it to the ATO.  As a plan, this is super problematic for me.  First off, the ATO is a Neptune ATK.  This device has an RO feed that goes to a safety float switch.  This way if your ATO turns on, but your water level is high, it does nothing, because the float keeps the water off, despite the PMUP running.   You simply cannot connect a kalk device to this.  If ever the pump turned on, and the float was closed, it would overflow the stirrer.

Second issue is, my tank doesn't have the calcium demand just yet for an ATO dose.  There are two ways to run a kalk reactor when you don't need it's full power:

1) Mix the reactor to full strength, and then attach a dosing pump to the inlet, and dose it as needed.
2) Hook it to an ATO, and adjust the kalk concentration as needed.

The second one there, sounds like a complete nightmare.  Until the coralline kicks off on the new rockwork, I have roughly zero calcium needs right now, and the dKH is rock stable, so, I kind of don't need any kalk dosing yet.  So currently, it's just filled with RO/DI water and nothing else.  But I hooked up the dosing system.

As you might recall, I built a simple doser out of a Kamoer 24v pump connected directly to an Apex 24v outlet.  I connected this doser to the RO/DI reservoir, and then into the Kalk stirrer.  The output of the stirrer goes to one of the sumps.  The doser puts out a known quantity of 83ml/min.  So I have a simple OSC statement in the Apex that runs it once a day right now.  As my needs change, I will up it to a few times per day with math.  This will slowly cause the Neptune ATK to run less often, and at a certain point, I'll have to stop, because I'll hit the evap point.  But during the early stages of the tank, this makes things easier.

However, placing the kalk output into the left sump, raises another problem that has been driving me insane anyhow.  Currently the vinegar doses into that sump, and a few inches away is the pH, ORP, and Salinity probes.  4 times a day, the vinegar dumps in, and the probes all freak out for a few seconds, alarms fire off, and it makes me crazy.  Know what will make that even better?  Dumping 12.0 pH water in there a few times a day.

So the probes all have to move.  3d printer to the rescue.

This little bracket didn't come out perfectly, but it did come out.  Printed from eSun PETG.  It will hold the 4 main neptune probes, and the little clip on the bottom will hold a temp probe from one of those little suction cup thermometers, (I keep on as an emergency sanity check).  Now I can relocate the probes to one of the drain towers on the tank.

And instantly, all my weird accuracy problems have disappeared! No more sump bubbles screwing up the Salinity probe.  No more vinegar screwing up the pH probe.  The only strange effect was, when I moved it up there, the salinity moved down like 5 points, and has remained there.  No idea why.  It's super stable now, just, reading 27.  Oh Neptune.. your silly conductivity probe..

So in the realm of calcium, I returned to an older project.  I have a Reef Octopus CR140D calcium reactor.  I mentioned this awhile back, and that I had lost the O-rings for it, and the pump was dead, so I ordered one from CoralVue.  That showed up.  Time to get the body assembled.

So this is what the pump swap kit looks like.  I have to say, CoralVue hit this one out of the park.  Amazing.  Makes me want to buy more stuff from them.

It comes with the new pump, which basically is the totally wrong physical shape from the one this reactor shipped with.  But they give you this pile of pipe fittings, and detailed instructions.  The fittings kit, has a bit of extra space on the fittings, so you put the new pump into place, and then carefully measure and cut down the supplied tubes, and everything then fits this new pump.  You toss the old input setup entirely.  It came with every imaginable part I would need to do the swap.  All I had to do was carefully cut and measure the bits, and it went together like a dream.



You would seriously never know this wasn't stock!

I also had the O-ring problem to deal with, and as I mentioned, I purchased a bunch off the O-Ring store (amazing..) and had them sitting around.  Turns out, this one fit perfectly:

3.5 x 103mm.  Dead on, super happy.  They even put some saltwater taffy in with the shipping bag.  For what was basically a $10 order.  Amazing service.

Now a quick caveat and warning.  As I plumbed this up, I was trying to make sure that everything lined up perfectly.  This involved gluing the tubes together while the whole assembly was on the reactor.  My mistake, was leaving it to dry like this.  It turns out that PVC cement causes acrylic to craze.  So there is a little bit of ugly now at the bottom of my reactor.  Reader beware.  Luckily, I pressure tested the whole assembly overnight, and there were no issues, it's only minor surface ick.  Oh well.  We learn things.

But all of this brings me to the actual title of my post.  I wanted to add life to the tank.  This was, well...

So I ordered 5 peppermint shrimp, 5 trochus, and 5 ceriths from the Algae Barn.  Went with 3-day because they have the arrive alive guarantee.   The Phoenix post office however, was bound and determined to screw that up.  The box floated around in Phoenix for 3 days before getting to my door.

The trochus were all super sad.  One was unresponsive, the other 4 would barely move.  One shrimp was in 2 pieces, and 2 were dead.  The ceriths were all happy.  Oh well, nothing to be done, it all goes in.

Shrimp went in, immediately disappeared into the rocks never to be seen again.  The trochus were all placed on a rock.  Overnight, 2 of them moved away, and the other three didn't.  Checked in the morning, shells were empty.  My clean up crew found the dead ones, and literally stripped the shells empty overnight.  Effective!

So long story short.  In theory I have 2 tiny peppermint shrimp, somewhere, and 2 trochus, and then 5 ceriths.  My wife of course, said "what were you thinking, you should have bought like 25 of each, do you know the size of this tank?"  The problem is, no. I really don't.  I cannot wrap my brain around the size of this tank still.

Normally, one would buy just a few of these things, and slowly scale up.  You don't want a ton of snails that wipe the algae out and then starve back.  But I was totally off scale here.  1 trochus has managed to barely handle 1 rock in 1 week.  That means I need like..  25 just to pass go.  I probably need more like 50 at this point.  That's an insane number to me.  I would never put 25-50 snails in a tank where I just turned the lights on.

I'm way off here.  I need to hit the store and buy real quantities of life.  The numbers just boggle me, and I can't keep up with it mentally.
If we look at that picture..  there are 30+ rocks there.  There was probably 1.5 - 2 times that much rock already in the tank.  So lets say, 75 rocks, on the low side.

My 120 has, 6 rocks.  My brain says, 800/120 == 6.  So that means I take the numbers from the 120, and scale up by 6.  Right? No.  Scaled up, the rocks would be like 36, not 75.  I need to be operating in the range of 12-15 TIMES what I do/did to the 120.

I am way off.  The 120 would start with like, 2-3 trochus and ramp up.  I should have gotten 25-40.  I should have gotten like 15 peppermint shrimp, maybe 25.  It's insanity to me.

Utterly boggled by numbers, I slip off to bed...

Monday, February 18, 2019

Day 15, The tank that took 7 years to aquascape..

So I finally got a good weekend, with all the ducks in the proper order, and was able to rescape (or finish scaping) the tank.  Now many many moons ago, when I first started all this, I was purchasing monster boulders one at a time direct from Fiji.  The idea at the time was, put one in, let it cycle up, then do another, and keep this up until a mains breaks and floods the house..  err..  I mean it's finished?

So some detours happened along the way, and I wasn't keeping up with the purchases fast enough, and I ended up with half a tank full of rock.  Which you have seen in all the "lovely" full tank shots, where you can see about 5 feet of bare sandbed.  Exiting stuff that.

Of course, because I'm super creative when it comes to fancy rockwork, my plan was "a big wall of rocks".  So to assist with this wall, I built a PVC frame that goes all the way to the bottom of the tank, so the rocks could sit up on it off the bottom a bit.  This gives the fish a permanent cave to hide in, and keeps the deep recesses from turning anoxic.  One of the main returns curves down there a bit, just to give some flow in there (very gentle), just to keep the water moving about.  I figure the sponges will love it too.

So along my new plan of "just get it done, don't spend 6 more years buying 1 rock every 3 months", I went online and got 240lbs of dry rock.

Now the old me would freak out right now and say dry rock is worthless, only Fiji or Pukani for me.  But that stuff is hard to get now.  And I refuse to use Atlantic rock, because I don't want to deal with the hitchikers.  So dead it is.

I got three boxes from Amazon prime, as I mentioned in a previous post.  Then I decided to get 3 more boxes from ArcReef.  I'm super happy with the Amazon Prime stuff.  Cheap, good rock, nice sizes, good stuff that.  The ArcReef stuff..  eeeehh..

So here is my thoughts on the ArcReef rock, and it's mildly whiny.

  • It is 100% the same rock as the stuff that cost 1/3rd as much off Prime.  After I rinsed all the different rock off, I had a really hard time telling it apart.
  • They shipped 51 lbs instead of 45 per box.  So that was nice.
  • I picked ArcReef because they say on the Amazon page that you can request specific sizes.  So of course I say I want 20-45lb boulders.  (Which the page specifically says they will do.
  • I ordered 3 boxes.  The first one arrived, and they required signature for delivery.  They needed an actual signature for a 45lb box of dead rock.  Completely absurd.  So of course I had to sign the little paper and wait for re-delivery.
  • I get the delivery, and it's only 1 box.  There is a note saying the other two were shipped 1 day later.
  • The rocks were kinda small.  Just a bunch of 8" rocks.  The Prime stuff was all 12" or better.  They obviously ignored my note.
  • 1 week later, no rocks.  Fedex says they don't even have the boxes yet.
  • So I send them a note on Amazon.  "Hey, my rocks didn't actually ship.  Also, I wanted boulders, which you said you would provide, please be sure the next 2 boxes are boulders"
  • Next two boxes show up, signature required of course...
  • Just a bunch of 8" rocks again, with a "we're sorry we forgot to ship them, oops"
 Now don't get me wrong.  They are good rocks.  But for the price, I wanted the specific size.  I didn't get that. I have no issues with the rock itself, just that I overpaid for it, because I wanted big rocks.  And the shipping issues were super annoying.

Either way.. Here is 260ish lbs of rock.  Can you tell them apart?  Hint, the big ones were from Prime.

So because as previously stated, my aquascaping theory is "chuck them in", I asked my wife to design an aquascape.  We watched a bunch of videos together, threw some ideas around, and then watched a particularly good set of videos with various shapes of tanks, and ideas.  One of them was a corner tank, and the concept he called "Sleeping dragon".

So we took this idea, because in essence what I really have is two corner tanks connected together (because there is no side-view), and made a huge dragon.  On the left side of the tank, we have the head, and on the right side, a small tail.

Now..  this was no easy project.  It took a day and a half.  Some of those boulders I originally purchased were in the range of 90-100 lbs each.  There were at least 6 that were 80+, and about 10 more that were in the 40 range.  Moving them around was a nightmare.  And oh the sandstorms we caused.  It would get so bad that we couldn't see what we were doing.

In the end, the smaller rocks weren't all that bad, as the big monsters filled in the majority of the space, and then the smaller ones just fit in perfectly to make various shapes, and fill in gaps all over the place.  We also made a small island on one side, just because there is so much space in the tank.  I feel like this structure uses alot more of the available footprint of the sand, and it's now far more stunning to look at.  Overall, I'm super happy with this.

And to put a kick on the end, I have turned on the main lights finally.  I have them set at 15% of full power, and will run like that for probably a week.

In other news, the GFO completely destroyed the phosphate in the tank.  So much so that I'm now shutting it off, because it's caused the chaeto to die back.  I still have a little bit of Nitrate, but that's all easy to deal with.  Hopefully things will balance out once the ugly phase begins....

Turning the lights on has also shown up some dark spots in the tank.  So I need to figure out a few spots or similar to fill those in.  I might just go with some cheap black boxes around the corners to fill in the dead spots, as I don't think I need a ton of light.  I'm also slightly concerned about the amount of white light..  I might swap the COB's for blues.. not sure..  we will see.

Oh, and since you made it to the end.  Here is the timelapse of us doing the aquascape!

Dreaming of a future shipment of peppermint shrimp.. I crawl into bed, exhausted...

Build a 100 watt passively cooled LED cannon for under $200

If you want to build a DIY LED for your tank, there are plenty of resources on the internet to get you started on the usual setup of a driver, and a bucket full of little 3 watt stars to make your lovely strip-style light.  But maybe the strip lights aren't what you want?  Maybe you want something big and stupidly powerful, with a single point of light and deep penetration into the tank.  If so, this build is for you.

This article will guide you step by step through making a large 100 watt pendant LED.  Some of these steps can be modified easily enough, such as swapping in a smaller 50 watt light, or maybe going with a more custom emitter setup.  If you decide to try something a little different, the majority of the steps will still apply to you, you will just be on your own a bit with the adjustments.  None of this however is complex, nor will it require a degree in electrical engineering or great soldering skills.  This is actually about as simple as an LED gets.  1 driver, 2 wires, 1 LED.  Total build time should be about 4-6 hours.

Important note: The heatsink for this is no joke.  The one pictured below is a CoolBay TERA-A, which is 12 pounds.  You need to figure out how to hang this securely over your tank.

This is what we will be building today:

It is basically, a DIY version of this:

So lets start with a list of materials, and I'll try to provide average USD prices for each.

Parts list:

  1. A single COB Led. ($5-$75 depending on type, ebay) (see notes)
  2. A 77mm Lens and collimator (for LED) ($8)
  3. A Mechatronix CoolBay Tera A or Giga A heatsink. ($75-$110)
  4. Meanwell HLG100-36B Driver ($41)
  5. A #10 Stainless steel eyebolt and nut, about 2-4" long ($1)
  6. A #10 Stainless steel nut with nylon insert, and washer to fit the eye bolt ($1)
  7. 4 15mm long M5 screws (stainless steel) ($1)
  8. 4 nylon or stainless spacers, at least 8mm long, 6mm I.D. ($1)
  9. 4 45mm long M3 screws (stainless steel) ($1)
  10. 2 4mm long M3 screws (stainless steel) ($1)
  11. 6 M3 washers (stainless steel) ($1)
  12. A length of 16 gauge 2 conductor wire (long enough to go from your outlet to where you want the device to hang.  I'd just buy a 25'-50' spool for ($5)
  13. An old 3-prong power cord salvaged off any dead appliance ($0, or $1 if you have to buy one)
  14. 14 total Anderson powerpole connectors:
    1. 2 red
    2. 2 black
    3. 2 blue
    4. 2 white
    5. 3 more pairs of colors, doesn't matter what.  I used gray, light blue and green.
  15. 14 Anderson powerpole contacts:
  16. Thermal compound ($8)
  17. A 2.5" Stainless steel disc: ($4)
  18. Urethane Conformal Coating: ($15)
  19. Access to a 3D printer with PETG filament  OR
  20. 5" Stainless steel disc ($4)  
  21. Thermal compound (get some for a CPU)

 Some notes on the bill of materials:

  • For the Anderson powerpoles, you might want extras, in case you mess up.
  • They also sell a nice little package of them that gives you 10 different colors and contacts for $11, maybe just go with that.
  • The powerwerx website sells some really nice 2-conductor wire, which is what I used.
  • Aside from the 45mm M3 screws, I actually just purchased a few variety packages of M3/M5 screws/nuts/washers to get everything I needed, it was pretty cheap and I'll use them for other things.
  • All the screws/washers/spacers are readily available at a local Ace Hardware in the USA, or Amazon.
  • We will talk about the LED itself later.
  • The files for the LED holder ring are here:

Tools you will need:

  1.  Wire crimping tool:
  2. Wire cutters and strippers (above link is a nice combo tool if you don't have one)
  3. Screwdriver or hex key for the various screws
  4. Wrench (adjustable will be easier)
  5. A drill and some drill bits
  6. If you don't have the 3D printer, or a friend with one, you will need a jigsaw with a metal cutting blade, and probably a dremel tool or metal file to smooth out the cut.
  7. A multimeter.

Selecting an LED:

There are a TON of LED's to choose from on ebay.  The easiest solution is to search for "100w COB LED blue".  You want to find one that you can buy a coliminator and lens for.  Look at some of the photos of mine in the article below.  That is the form factor you want to look for.  A decent blue LED will probably cost you about $10.  The lenses/coliminators will probably be in the range of about $8.  I used the search string "77mm lens LED" and found a few that would work.  Make sure to get one with the coliminator on the back, so it connects to your LED.

If you want something more fiddly, and expensive, a seller named AC-RC sells a 60/40 blue/20K chip, which is what I used.

He also has some crazy 5 channel chips which you could use, but would require alot more work on the driver-side to get the voltages right.  You are on your own there.

Wire the driver:

Wiring the driver is surprisingly simple.  There are 3 wires that come out of the driver.  The mains input (green/brown/blue), the dimmer pair (blue/white) and the output (red/black).

Let's start with the mains input.  Usually, I just get an old 3-pronged cord, like a computer cable, and chop it in half.  Then connect the wires of the cord to the 3 wires from the driver.  The driver wire colors are:
  • Green/Yellow - Ground
  • Brown - Hot
  • Blue - Neutral
When you look at the outlet on the wall, the one with the larger opening is the neutral.  You can use this as a guide to figure out which wire is which on your chopped off cord.  Don't plug it in for this:  You can orient the plug so you know which prong is the neutral, and then use a multimeter to probe the wires to find which one is neutral and which is hot.  Typically inside the cord, most neutrals are white, hots are black, but inside the confines of a cord you aren't supposed to cut open, there are no guarantees.  Check.

I like to put some heat shrink tubing on the wires before soldering them together.  I also like to put another larger diameter piece of shrink tubing on the entire cord.  Simply solder the wires together, slide the shrink tubing over the joints, and then shrink with heat.  Once all three are done, slide the larger tubing over the whole set, and shrink that into place.

Alternatively, you could use the Anderson PowerPoles to wire the cord to the driver.  I like to do this, because if the driver were to ever fail, I could swap a new one in easily.

Next, get yourself the 16g wire that will make the run from the driver to the LED.  This is another simple splice and solder operation, just like the last one.  Use multiple heat-shrink tubes to seal it all up.  Just wire the red/black pair to the long 16g wire pair, and call it good.  Again, it might be better to use the PowerPoles here.

Finally, we have the dimmer cord.  The really amazing thing about the HLG model B driver, is that the dimmer operation works 3 ways.  You don't have to flip any switches, or wire it differently, just connect it to the source you want to use, and it works.
  1. 0-10 Volt.  This is what an Apex controller uses.
  2. 0-10v PWM.  (Note: this is not 0-5v, which an Arduino uses)
  3. A resistor/potentiometer.  10K = 10%, 100K = 100%.
  4. Nothing connected (wires left open).  Runs at full power, 102-108%.
I can't tell you how to wire the dimmer into your chosen system, you are on your own there.  But the easiest way to set it up is to simply get a 100K potentiometer and solder the blue/white pair to it, and then use the knob to adjust.  I used a small spade connector strip and some spade connectors to wire mine into the Apex, just for ease of future replacement.

At this point, you want to figure out where to place the driver in your room, and where the cord will run, and eventually, where the LED ends up.  This will help you measure out the cord lengths.  Figure out how long the cord needs to be, and cut to length.  Give yourself a little extra.  It's much easier to coil up extra wire than it is to splice on more if you went too short.

Wiring the LED.

Again, couldn't be simpler.  The LED has a + and - side.  It's a little difficult to figure out which is which sometimes however.  Looking carefully at the LED, the top connector is the + side (there is a tiny + sign there).  If you look at how the traces on the edge are laid out, you can see they wrap around to the left side, where the big red wire is.  You can also see the bottom one, wraps around to the right side to where the big black wire is.  All COB's are different, so pay careful attention.

You can wire either to the side-connectors, or to the tiny little spots on the top and bottom.  Using the ones on the top and bottom involves significantly more soldering skill than the side ones, so choose according to your skill level.  Additionally, if you use the top/bottom pair, you might need to dremel the coliminator a bit to get it to fit afterwards.

Cut about 15 inches of 16g wire, and solder the two ends to the LED.  If you have decided to use the side connectors, after the joints have cooled, take the conformal coating, and cover the entire side connector, top and bottom, with the coating.  This will insulate it from any arcs or other issues.  It smells awful and takes a few hours to dry.  Set the LED aside for now.

Build a hanger for the heatsink:

This is what we will be building:

Take the 2.5" stainless steel disc, and find the center.  Mark for a hole in the center large enough for the eyebolt to slide into. I found a quick 3d printable center finder and printed it to do this step, but there are other ways.  Draw on the disc an X (be precise!) so you can locate where the other holes go.  I like to use a center punch for the center hole, because it makes the next step easier.  If you don't have one, take a sharpened nail, place it carefully in the center of the X, and then lightly hit it with a hammer to make a small divet.

The diameter of the hole spacing for the mounts was 45.5mm on my heatsink.  So I set the calipers to 22.75, and used the little screw on the calipers to hold them into position.  I then placed the tip of the caliper into the divet in the center, and then used the other end to scrape a line across my X, so I knew where the centers of the 4 holes would be. Once you have the 4 locations marked out, it's time to drill.

Here you can see the disc with the marks for the holes.  I used a center punch to make divets for the 4 outer screw holes.  (The brass tool)

I have access to a wide variety of drill bits, so I used ones that were perfect.  If you don't have numbered bits, you can just find ones that fit the screws you will be using.  For the center, you will be using the #10 eyebolt.  For the outer holes, you will be using the M5 screws.  I used a #15 on the center, and a #6 on the outside.  After drilling the main holes, I switched to a much larger bit, and very slightly drilled the holes, just to clean the edges up, and chamfer the holes a bit.  This removes any nasty little metal strings that will cut your fingers later.

Now simply attach the #10 eyebolt to the disc, using the washers, a simple nut on the eye-side of the bolt, and the nylon-insert nut on the other side.  Don't screw the eye bolt in too far, as it still has to fit onto the heatsink.  No more than about 1/2" - 3/4" protrusion on the back side.

Attach the LED:

For this step, you either need to make a disc for your lens out of stainless, or 3d print the holder I designed.  I used white PETG to print mine, and I would strongly advise against PLA, as it might melt and warp under the heat.

If you decide to cut one out of stainless, the process should be simple.  The lens has a small lip on the ouside, and then moves down into the curve.  You want to measure where the curve starts, so you can place your disc onto the lens, and have it hold against the lip.  If you cut the hole the size of the lip, it will just fall through.

Your best bet here is to draw out the circle, and then use a jigsaw to carefully cut around the inside of the circle to make the hole.  Then use a dremel tool, to grind away the inside of the hole until it's the right diameter.  This might take awhile.  Keep the lens nearby to test it a few times as you go.  Finally, you need to drill 4 holes for the M3 mounting screws.  On my sink, they were 113mm diameter centers.  You can use the same X procedure we did for the back mount to make these.  On the face of the heatsink (the shiny aluminum side) these are the second most outer ring of screw holes.

You are now ready to attach the LED.

Center the LED on the sink, and then cover the back with thermal compound.  Place the LED on the sink in the right place, and thread the wires through the convenient hole in the face of the heatsink.

Now place the coliminator and the lens onto the LED. Don't forget to take the protective blue plastic off the LED before you do this!  Because I utilized the inside attachment points on the LED, I had to dremel out the bottom of the coliminator a bit, so it wouldn't interfere with the wires.  If you use the side holes, you don't need to do this step.  Also note, there are 4 little pegs on the coliminator.  These insert into tiny holes on the LED to line it all up.

Now place the steel disc, or 3d-printed part over the lens, and attach it to the sink with the 4 45mm M3 screws.  Tighten them evenly and gently.  You want to press the LED onto the sink so it has a good thermal connection, but you don't want to smash the coliminator or crush the LED.  Just make sure the screws are evenly tightened on all 4 sides, and holding the LED into place.  Don't go nuts here.

At this point, you should have the two wires from the LED sticking out the back of the sink.  Simply crimp on some PowerPole connectors to this pair.

Final touchups:

Now we just need to connect the hanging plate to the back!  Take the 4 15mm M5 screws, use some washers and the M5 spacers, and connect the plate with the eyebolt to the back of the sink.

Now it's time to test fire the device.  I recommend getting a 10K ohm resistor, and connecting it to the dimmer cable for this step.  It keeps you from going blind when firing this thing up.  The output is no joke on these.  At full power, they triggered the auto-darkener on my welding helmet.  Do *not* look directly into the lens on these.

Simply connect the long 16g wire (with powerpoles on the end) to the wire on the LED, and plug the driver in.  If you've made all the connections correctly, your room should be instantly illuminated.

Now all that remains is to hang this beast.  Please be careful here.  This is a heavy heatsink.  You need to make sure to build a strong support to hang it from.  I bolted 2x4's directly to the rafters using massive lag bolts.

And you are done.  You have a huge LED, single point of light, that puts out tons of power, and was cheap to build.  There are a large variety of LED's to choose from, so if you just want pure blues, buy those.  (The solid blue ones are like $5 on ebay)  If you are concerned about the quality of these, then buy cheap ones, and buy a few.  By utilizing the powerpole connectors, it is easy to take the sink down, and swap them out.  Everything we did here makes it easy to replace and edit your fixture as you see fit.  Try out different LED's.  Try out different mounting heights.  Light up your life.

A final note on the 3D printed lens ring.  I used a small cone as a reflector on mine, to tighten the beam up.  This doubled the measured PAR output from the LED.  However, it also gives the edges of the beam a really hard edge.  This might not be to your liking. There is also a file in my design for the ring without the reflector.  You will also have to play with the height of the device over the water, to give you the spread you want.  Mine are pretty high up, which gives an awesome spread.  If you go with a different lens, you could potentially go lower.  With no lens, the spread on the LED's is 120 degrees. Feel free to play around!