A major update to V5, along with the PCB order can be found at post 63

Original post:

================================================================================


Allrighty,

This has been a long time coming, and I'm proud to finally release the completed, fully functional DIY Automatic Water Changer. It's been through many revisions and testing of various designs, and finally have found what I believe to be an optimal solution.

The design basis I set out for the project was to emulate a standard 10% water change, rather than the peristaltic "continuous change" systems. The reasons were based on efficiency, control, and space considerations. I have used three MJ1200's for each function of drain, fill, and ATO, but any other powerhead will work.

I don't want to bash the continuous change folks, but my reasoning for this design is as follows. A continous change system needs to change out more water to remove the same amount of nitrates from the water column. 10% water changes do not introduce enough system instability to warrant the added operating cost in both water and salt. In addition, they require both a drain and supply reservoir to be connected to the tank at all times, consuming valuable real estate. Also, depending on the pump used, if the supply reservoir empties, the drain may continue to empty the tank without replenishment. Lastly, such a system cannot be accelerated to perform larger than normal water changes if and when necessary to avoid crashes and livestock losses.

The other primary design considerations were inclusion of an ATO, cost, and ease of replication. I wanted this to be easily reproduced by others within the community. The design should also be robust enough to be adaptable to varying tank setups, though the demonstration will be on a tank with a sump.

The current version, V4.2.1, uses only four DPDT 12V relays, and the associated diodes for kickback suppression. I have also included a V 4.2.1 PCB design that can be milled/etched, and is compatible with a standard 0.1" perforated prototyping board. EAGLE files are included below for those wishing to mill a board, rather than use a perforated board.

Circuit:



Board, top view - each square corresponds to a hole/pad on a standard 0.1" perforated board. Green octagons are wire connections, green ovals are the relay pins.



Board, bottom view. Blue lines mark traces, for those unfamiliar with EAGLE schematics. Note this has been mirrored from the EAGLE file, so as to match exactly how it would "look" from the underside.



EAGLE files:

Click to view attachment

EAGLE can be downloaded as freeware from cadsoftusa.com

The board is designed to use a relay such as the sanyou DSY2Y-S-212L or any similar DPDT relay with a DIP 2C pin layout. The circuit, however, will work with any DPDT relay - but you'll have to redesign this board yourself to accomodate it.

I'll try to briefly explain the circuit.

Three float switches are located in the tank. One marks the normal water level, another marks the lowest water level during a 10% change, and the third is an emergency backup, located above the normal water line.

When the AWC is not active, the emergency float switch is closed, and there are no outputs. As evaporation occurs, the water level drops, the high level switch also closes. This activates the ATO pump to add water to the system from the ATO reservoir. This is not any different from a typical floatswitch ATO.

If the high level floatswitch fails, the emergency float switch will be opened by the rising water. This disables the ATO pump and triggers the audible alarm.

To perform a water change, the hose from the drain pump is fed into an empty bucket. A second bucket of freshly mixed saltwater and the refill pump, with a hose fed to the tank, completes the setup.

All that remains is to hold the start button for a few seconds. This activates the drain pump and disables the ATO pump. After enough water is pumped out of the aquarium for the high level switch to close, the Master Control Relay will latch, and the system will complete the cycle automatically. This is a safety feature, so that the start function cannot be inadvertently triggered.

When the bottom float switch closes, it will close and latch the Low Level Latch Relay. This disables the drain pump and activates the fill pump. As the water level rises, the high level switch will eventually open. This unlatches the MCR, which shuts down the fill pump and resets the system. The ATO is now re-enabled.

If the high level float switch fails during a water change, the emergency float switch will also disable the system, as well as trigger the audible alarm.

And there you have it. Clear as mud, I am sure. If you don't understand how relays operate, I'd recommend this as a primer.

Here's a few pics of the demo prototype I tossed together. You will have to forgive the physical crudeness, as this was merely a prototype. I will be incorporating the circuit into the electrical panel of my stand. Please also note that the circuit board is a V4.2, while the board layout above is V4.2.1. Electrically, they are identical, however.

Top:



Bottom:



For diode orientation - The stripe faces up:



The enclosure is a trio of surface-mount electrical boxes bolted together. Two duplex receptacles are each split on the hot side. One receptacle is always on, and powers the 12V AC adapter. The remaining three are controlled by the respective 120V outputs from the AWC circuit, and serve as connections for the unmodified powerheads. The float switches were just twisted together with the correct few wires - again, I would expect better for something other than a temporary demonstration.



There are separate toggle switches for 12V and 120V power to the circuit board, as well as three black momentary push buttons for manual bypass control of the powerheads. These are not shown on the circuit diagram, as they are not part of the board itself. They are entirely optional, and how to add them should be fairly obvious. The green momentary push button is the start button.

Video of the unit in operation can be viewed here


Enjoy!

I'm completely lost.

wat.

NICE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I love your home made pcb with the solder as the tracks!!! Very innovative!!!!

My project just seems to keep snowballing, but ill get some pics up next week, now that my parts have finally arrived.

So whats your next project???????????

This is Art.





Excellent project, even if it does zoom over some heads at 700mph.

know-it-all

ahleluia!! Thats freakin awsome.

That will go good with my automatic saltwater maker and the robot I'm designing to empty my buckets.

Great job.

You sure about that?

Looks cool though. Nice work.

I think he meant to say "others in the electrical engineering community"
Way beyond me man.

I use the same Home Hardware buckets to change my water! (manually though). It is a nifty and smart idea but on a nano I only change like 3 gallons/week. I don't find it too bad to just mix, siphon old, add new. Takes about 15 minutes. But golf clap to you for making it cool. I am just impressed you can wire something like that, the drawings looked like a hamster maze, I was looking for the cheese at the end.

So... Ummm....


WHAT???

nice

Thanks for the comments everyone!

Seriously - if there are any specific questions, or any parts that are particularly confusing - just ask. I'll do my best to try to clarify things a bit.

I do sincerely believe that this is reproducible by anyone who can DIY a floatswitch based ATO. You don't necessarily have to fully understand exactly how the circuit works to solder it together. It might something you've never done before, but soldering really is not a particularly difficult task. All you've needed is a reason to give it a try.

Believe me, compared to the earlier versions - yes, this is "easily" reproducible. I'm sure nobody would want to buy (or build) a $200 PIC programmer, as well as learn how to use the software.

To make it even easier, the EAGLE file can be submitted to a company like custompcb.com, and they'll ship you a nicely milled board for about $20 each. The minimum order at that particular site is two boards, but that's still only $40 - and you could arrange a group buy. Other sites might allow purchase of a single board....

awesome stuff, so incredibly unnecessary, but awesome nonetheless

Thanks for sharing this, It's just what I was looking for.

One question though, if you have submersible pump in your sump to pump out the water change water, won't you instantly create a syphon, and turning off that pump won't stop the water flow? ie, the sump will drain way past the FSLOW until the water level is below that pump and the syphon breaks.

No. The other end of the hose is at a higher point than the low floatswitch. It is also higher than the final water level in the wastewater bucket. Thus, air is introduced and the siphon broken as soon as the pump turns off.




It's definitely unnecessary, but so are many things like Ca reactors and ATOs that make our lives easier.

WILL IT BLEND?

Amazing... I cant wait to see some videos or somethin of this. If all works you should start selling kits.

lol, true


YES

what? enrish prease.

I've toyed with the idea of putting together kits....but I don't really want to incur the cost/hassle of doing so, as well as the hassle of settting up payment and shipping. I'd much rather share the knowledge with the community and let them do with it as they will.

94% of of the people in here have no clue whats going on here.

that's what I was thinkin


It's definitely awesome, but unless it mixes the salt water for you automatically, it's only saving all of like 45 seconds of work that it takes to siphon/pump water out of the sump and dump the fresh SW in...

Still, easier's easier I reckon!

And definitely still some cool factor to it

Its all about cool factor

The next stage of the project will make it more useful, I assure you.

I don't have the correct setup in this apartment, but once I buy a house, the plan is to replace the wastewater bucket with a sewer line, and the refill with a large capacity reservoir. 20g would be sufficient for this tank for a month's worth of water changes, and an even larger reservoir would be even better. Third, connect the ATO reservoir directly to the RO/DI unit, with a separate line for filling the SW reservoir.

So, the system will always replenish it's evaporation without any user input. In addition, one can simply press a button on a weekly basis, or, we can set up a real-time clock circuit to do this for us. SW would need to be mixed on a monthly basis, or even less, depending.


Suddenly more useful, right?

*drool*

+1

What sorcery is this?

Water changes are supposed to be purposefully inconvenient.

way cool!
way beyond my DIY skills. I cant solder to save my life and that is assuming if I can even read the diagrams

Trust me as long as your PCB is silkscreened/solder masked you can do it easily.
I was 13-14 when I was soldering PCBs together for fun,
I "Just"/"Barely" knew(not much,I'm no electrical engineer!)
how it worked and I had tons of success soldering PCBs.
But I'm cheap(I'm putting it off until I do my big reef)
and will do my own water changes until we own a house,
then I'll pour all my funds into a large reef setup.

That's basically the setup I'm trying to put together now. And this module will be a huge help. I still have the problem I mentioned earlier though. I was trying to have a drain line go directly to sewer, but that would create the syphon. It needs some kind of syphon break. I don't have the option of draining above the water line. I don't think a sewer standpipe in the middle of the living room would be tolerated.

Nice work NM.


aaren:

Nice...wanna design one for my 135>?

Folks,

This is *NOT* hard. If the soldering scares you spend the $20 on the PCB fab and get over it. Its an ATO system that is cheaper than a JBJ ATO ($70+) with the ability to handle water changes.

Adding in something to mix the salt water for you would not be too difficult to do as well.

What he is saying is that you hit the red(or green) button which fires up a maxijet that pulls 10% of your tank water volume out and puts it either into a drain or some other form of container.

When a float switch trips it stops. This float switch is supposed to be set for a level in your tank/sump that is equal to 10% of your water volume.

Once that float switch trips it shuts off the first maxijet and turns on a second one. The second one then pumps 10% of your water volume in new salt water from a container back into your tank. It stops when a second float switch which is set to the high water mark for a full tank trips.

There is a third float switch that is setup as an emergency backup in case the second high water mark float switch fails.

No EE degree required.

Nice summary. I don't have an EE degree either.......

I also have it set up so that I use the same pump for mixing the new SW as I do for dispensing it. The hose is simply fed back into a second hole in the bucket lid, and the powerhead plugged into a receptacle near where I store my water.

Setting up a system to actually portion out the salt for the SW reservoir would be something else indeed. A little more complicated, for sure.




I doubt the sewer standpipe in the living room would go well for me either. You have a couple of options, besides creatively hiding a water closet flange below your tank stand.

I thought about it a bit today, and came up with a few solutions. One would be a hole in the pipe at the highest point, with the hole squirting back into the sump. Similar to a siphon break hole on a return line. For more control, a "T" hose barb would be better, and you could add in a ball valve back to the tank that would throttle the flow. It would reduce the flow rate out of the sump, but would also break the siphon pretty quickly.

Another option would be to make use of the siphon, and put the pump inlet just below the float switch trigger level so the siphon breaks naturally. The pump only "starts" the siphon. I don't like this option, I see it being difficult to get "just right".

A third option would be to use a 120V normally closed solenoid valve instead of the pump, and always have the output pipe primed. Instead of turning on a pump, it would open the valve and allow flow through the siphon. The problem here is that there is no guarantee that the siphon will not be lost between water changes (bubbles, etc).

A fourth option would be to use both a pump and a 120V normally closed solenoid valve. The pump ensures that a siphon is not needed, and the valve prevents flow when the pump is off. I'd personally go for this option.

Another, final option, is to put a sewer standpipe in the room behind the tank (assuming it's not outside), and running the outlet hose through the wall.

I'll just have you make one for me when I move home....

AWESOME.

... i think.

Kudos to you on your beautifully simple design. In this day and age, too many turn to programmable circuits to solve their problems, which is overkill for most applications. A correctly designed logic circuit using discrete components is usually more cost effective and very robust.

I have to agree that making the traces from solder at a level that complex takes skills.

Thanks shcudini

You know, that's what I thought about the microprocessors. I was rather enamored with them for a time, and had several designs based on PIC microprocessers. They simplified the design process, but did not reduce the number of components. It actually made more of them. For this project, they were more of a crutch than a solution. Eventually I revisited the relays, and found that I could really bring the component count down to a bare minimum.

The circuit on the board looks like a snaky mess, but with EAGLE, it's not all that hard to design a complex board like that. You draw out your schematic like a normal wiring diagram, and then it generates a board file with the components in a pile. You place the components where you want them, and it indicates what joins to what by the shortest path using little yellow lines. After that, it literally is an exercise of "connect the dots", only you get to cheat and move the dots to wherever it makes it easier for you. I highly recommend the program to anyone working on their own DIY electrical projects.

Actually producing the solder traces wasn't that hard either. I found an easy method for doing it. The first thing I did was solder the relays to the board so I could use them as a point of reference. Then, using a fine tip sharpie, I outlined each trace to match the diagram. After that, just melt enough solder to make the traces. If you get too much on and connect traces where you didn't mean to, just remove the solder with some braid or a desoldering pump. It quite literally moves from a "connect the dots" exercise to a "color between the lines" exercise.



BSD, you get a board etched and bring the components next time you're up, I'll throw one together for you.

Cheers.



Humm.... I would think that a servo or step motor on a worm drive type shaft with a fixed size container on one end would be the simplest route with you pre loading the container.

Or if you want to get more high speed low drag a pressure sensor or scale to measure how much salt is placed into a container then a dump function followed by a water add function based on float sensors coupled with a salinity monitoring probe...

Actually you might even be able to take one of those new RKE controllers and use the Salinity probe module they are releasing soon and have it control salt/water dosing into a container based on readings from said salinity probe and float switches...

<BAM> my head just exploded. I need to stop now else I will be making this my 'next' project for the new tank...

Plus the added complications of programming and making sure it doesn't crash or inf. loop.

I've used PSPICE which does the simulations and layout. I like that EAGLE is free.

I wondered what that black stuff was in between the lines, must be ink from the sharpie. Looks like that board has solder pads around the holes, which probably made the job easier.


I'm going to save your design and build one when I have a larger tank (with as much simple automation as possible).

Well, you said they will take a min order of two... what about splitting the cost of a couple with me?

Yeah, I'd be in for that.

im to lazy to do water changes (once a month if im lucky...sue me)... if you deside to set this up as some sort of unit i could buy let me know... i know i could put it together, but im in grad school and am away from the house literaly from 7am to 10:40pm, so theres no time

any way it anything to make life simpler is way cool in my book.. nice job!


EDIT: i see you may buy those boards... if you set up a group buy on the board let me know... it would be nice to have a weekend project (if i ever get a weekend to myself)

Cool man, get a quote for it and let me know

I'd be in for 3 boards myself. I would estimate that for 100pcs, the board cost should be no more than $5, probably even a lot less than that(maybe $2), but it's been a while since I did an order like this. But you have to find the right PCB house, probably in China, these days....

Here's a good example of pricing:

http://www.pcb123.com/pcb123pricing.php

I could see this working if you a space for a big water change reservoir but as it stands now I still get glares from the GF when I break out my 5 gal buckets. What I wouldn't give for an equipment room.

$5 a board is pretty damn good - but I don't have any sensible reason to plunk down $500 for 100 boards. A group buy would be much smaller than that.

However, since I'm considering buying at least two, it'd cost $97 from PCB123. However, buying ten would cost only $165. It now makes sense to do a small group buy.

Let me think about this one - we just spent about $650 on vet bills, so I don't have the liquidity right now anyways. I think my mind may have been swayed regarding a group buy.




Marteen, I've done a couple more water changes since the initial test, and even though I have to bring in the wastewater and refill containers, it's still a lot easier than without. The 5g buckets your GF hates so much would be out of the living room that much sooner.

Well if you decide to do it, I'm in.

I did a little more research on this and found a few things:

If you only wanted one PCB fabbed for a prototype, and didn't care if it takes a little time, this place seemed pretty good:

http://www.batchpcb.com

I actually created gerber files from Eagle using Neandrathalman's design, and uploaded them to see how the process would work, and it looks like one board would cost $27.50 shipped.

If it was desired to skip the prototype and just order some boards, 100 boards could be ordered from here at $1.50 each plus shipping(from China):

http://www.pcbcart.com/


I'd be in for a few boards at that price.