The Planted Tank Forum - Reply to Topic
Thread: Bulkhead/In-line Replenishment Snag... Reply to Thread
Send Trackbacks to (Separate multiple URLs with spaces) :
Post Icons
You may choose an icon for your message from the following list:

Register Now

In order to be able to post messages on the The Planted Tank Forum forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.

User Name:
Please enter a password for your user account. Note that passwords are case-sensitive.


Confirm Password:
Email Address
Please enter a valid email address for yourself.

Email Address:


Human Verification

In order to verify that you are a human and not a spam bot, please enter the answer into the following box below based on the instructions contained in the graphic.

  Additional Options
Miscellaneous Options

  Topic Review (Newest First)
08-04-2007 03:52 AM
scolley Thanks Brian. And overflow and standpipe (with a sump) will do exactly the same thing. I just got jones for solving the "nothing you can see in the tank problem", and wouldn't let it go until I beat it.

And honestly, I didn't actually succeed completely. I do have a tube that protrudes out of the tank floor, with the bit of filter foam on the open end. It's small, and real easy to hide. But technically it is not "nothing" in the tank. But it seems a small compromise.

Thanks for the feedback!
08-03-2007 05:40 PM
bpimm Sweet system,

As I read this thread today I was thinking how to solve your problem, and by the end you had already got there. Great job. the only thing I would have done differently is to have the pressure transducer run the fill valve and reset upon the full state to eliminate the flutter problem. but you got it.

All of my tanks use an overflow and constant inflow to accomplish the same thing, but that requires a standpipe in the tank which is against your goals.

once again, nicely done

08-01-2007 04:00 AM
Conclustion to my problem

When I started this thread I was struggling to find a way to top off/refill a tank without putting visible equipment in the tank. And for me, that inlcuded running anything up the side/back of the tank that could be seen.

I experiemented with lots of pressure switches, but only the one from my last post worked well for me.

Here's a pic of it mounted under my tank.

It's a simple contraption really. I have a hole drilled in the bottom of my tank that has the PVC you see mounted under the bulkhead hardware attached to that hole. Above the bulkhead (in the tank) is a simple slip fitting nozzle with a bit of tubing attached. Over that tubing is a bit of hard filter foam. That keeps snails and such from going in the tubing.

Water pressure comes through that nozzle in the tank, through the hole in the glass, into the PVC you see in this picture. From there - as you can see - the water pressure can be detected by the little sensor shown.

The sensor has two wires that close a circuit when it detects a certain level of water. The level can be adjusted with a little trim screw that is just barely visible in the front of the sensor in this pic. It takes a tiny screwdriver.

And it's a PITA to set, because a tiny fraction of a turn of that screw, and you'll change the level of the tank by an inch or two. But once you get it set at the right level, it's solid as a rock.

Anyway, I've got a fill system that first opens a drain (on a soleniod) for a timed period every night. And then a different solenoid opens that allows water from my household tap into my tank lines. This continues on a timer for a set period of time. All that is controlled by an AquaController III.

When this sensor detects enough water pressure to know that the tank is full, the sensor closes the circuit. The two wires you can see in this pic run to the Aquacontroller, which senses the closed circuit. I've got logic in the Aquacontroller that says... "Even if the time to fill the aquarium is not finished, if you see that the circuit has closed, cut off power to the solenoid that's letting water flow into the tank."

Simple really. And the Aquacontroller has a hysterisis function that takes care of my old problem of people walking across the room, making the sensor go haywire if it was near a full condition - with ripples in the surface making it read "on/off/on/off/on/off... " in rapid succession. With the controllers hysterisis function, I've got it set so that if it detect a full condition (possibly caused by someone walking across the room) and closes the circurit, it stays in that closed condition for a full minute before it will turn on again.

A little complicated. But not too bad. And works like a charm - with two mechanisms working together, providing me redundent assurances that I won't over-fill my tank. First - the fill is on the timer I mentioned before. And just in case the fill time is too long, this little beauty of a sensor shuts the fill off if it gets too full anyway.

Problem solved! Automatic water changes every day, and no equiptment to be seen in the tank! Cool.

Thanks for all the help.
03-29-2007 12:43 PM
scolley OK folks, heads up!

I know I haven't posted my full, functional setup plans. But it WORKS. Works like a charm in fact.

I'm posting now because I just got a notice from Digikey, the only people that I can find that distribute these little water pressure sensors in small volumes. Apparently they are being made obsolete.

So if you want one - get them now. I just bought 5.

Here's a link.

Get 'em before they are gone. At $15 you are never going to see an opportunity like this. I spent a couple of hundred on other sensors before I found this cheap, accurate little jewel. And now apparently they are disappearing.
02-18-2007 01:48 PM
Originally Posted by scolley;374249[COLOR="Red"
essabee - Sorry pal. By definition I'm trying to do something hard here. That's why it has its own thread.
....and I thought that you have nearly solved all except getting back to topped-up position and made a suggestion. Sorry man.
02-15-2007 02:57 AM
scolley hello1 - your friend's approach sounds fine. Thought about something like that myself. But it does not solve my problem. Your friend still had a float he had to hide. Actually, I had to hide something myself - the little open tube at the bottom of the tank. But a little tube at the bottom should be a lot easier to hide than a float on the surface.

PDX-PLT - Welcome to PT! Great first post too, IMO!

I thought about something like this early one, but could not really get my head around it. You have certainly clarified how something like that might work. Thanks. Good thinking!

This potentially eliminates the deadband problem. Cool! But unfortunately the size of the deadband seems also to correspond to the sensitivity of the pressure sensor. I didn't mention this earlier, but the sensors with large deadband also had large variance between the points they considered "full". So I really abandoned those not only because of the dead band size, but because of their lack of accuracy too.

Sorry. I should have mentioned that. Good solution though!

essabee - Sorry pal. By definition I'm trying to do something hard here. That's why it has its own thread.
02-15-2007 01:13 AM
essabee This is too long I could not through it entirely

May I suggest a 'new' approach

First fill a fixed volume from a measured sump then drain to the desired top level.

The caged bear got bored walking to and fore, for a change he started walking fore and to.

I think this should work-no funning- easier to handle.
02-14-2007 07:55 AM
PDX-PLT Greetings Steve!

First post on this forum. I've really enjoyed all your posts. A a techie myself this one particularly is very interesting.

I wonder if you could solve the "deadband" problem with a second solenoid, to "reset" the pressure switch.

In other words, one could have two drainage solenoids, in series. Solenoid #1 goes between the bulkhead and the pressure switch; solenoid #2 goes from the pressure switch to the drain (i.e, the pressure switch is on a "T" between solenoids #1 & #2).

To drain, you open both #1 and #2 simultaneously for the required period of time. When that time is up, close #1 only. This should reset the pressure switch, as it is exposed to the atmosphere, and turn on the fill pump. Then close #2 and open #1, and the pressure switch will shut off when the proper level is reached.
02-14-2007 05:48 AM
congratulations for making the replenishment control

i know its a DIY kinda project but i think u have spent a lot more than me on this, well ur work would be cool cause there will be practically nothing visible in the tank but one of my friend had automated his tank with transparent float switch and the single wire is going out through the bottom and is hidden in the plantation as the float switch goes down the fresh water tap fills the tank the sensor being lifted stops the flow its cool, and it costed him only 38 bucks, waiting for u to post the pics
02-14-2007 01:54 AM
scolley Thanks for the link hello1. And welcome to PT!

But I think you are missing the point. There are hundreds of water level controllers on the market. Maybe thousands. I was looking for a control function that is fairly unique... one that can detect water depth WITHOUT any sensors in the tank, or be otherwise visible, and that can be afforded by mere mortals.

And I succeeded. Now I just need to post pics.
02-13-2007 01:42 PM
I also faced the same prob

Some time back i also faced the same problem i wanted to add 5g of water to my tank as it emptied on an average of 2.5 hrs and was looking for a solution badly but couldn't figure it out myself, i found a solution in this water level controllers, i dropped a mail to this website telling my problem and they replied me with a example movie diagram of installation and working and told me how i can custom arrange it according to my need i think i may save a lot of time and its pretty economical, 150 bucks i think the whole system is worth the money compared to the cost of the switches u installed this page has the installation and working movie at the bottom
01-27-2007 10:58 AM
Originally Posted by Jackfrost View Post

How have you been ?

Long time since I posted to you.

I think you are on travel somewhere this month.

Anyway, I have been reading you threads (I see you have been busy !), with great excitment.

I am in the process of getting my 90 G into setup. I have alll the stuff I need plus some. I will be going for a Discus tank also.

Yours looks great by the way !

I am very interetsed in you auto water changer with that pressure sensor.

Any updates on the design ?

Hey Nick, long time no hear. Good luck with that 90g Discus tank!

I just got back, and really need to document my autowater change. It does work like I originally planned it - water pressure senor through the bottom of the tank, working as a shut-off to the fill process. It's pretty simple really. But I've been though a lot of iterations. Now it finally works great.

I had had an electrical junction box, sporting an AC socket in front, that you plugged the fill components into. And it had a airline tube (filled with water) that plugged into the pressure sensor in that box, with the other end plugged into a bulkhead mount on the underside of the tank. In my tank return lines I have two "T"'s - one to drain water, one to fill. They both are turned on and off by solenoid. And the solenoids are on times.

So, if the timers are set correctly, a couple of times a day the drain solenoid clicks open, and allow water to drain out for and hour or so. The timer snaps that closed, and another timer opens to allow tap water in. That too is on a timer. And it is set so that the refill time is just barely more than necessary - potentially bringing the water a small fraction of an inch higher than it was previously. That's were the sensor kicks in.

The fill solenoid was plugged into the aforementioned junction box with AC socket. And a minute or two before the fill solenoid is shut off by its timer, the pressure sensor can tell it's full, and shuts off the AC power to the socket, killing power to the fill solenoid, shutting the fill line.

This means that if I'm losing water to evaporation, those few extra fill minutes - that usually get cut off because the water is high enough - get accommodated for. The idea is to have the fill timer set so that if I didn't even use a pressure sensor for cut-off, the water level would only creep up a little bit every day. So you get a visual warning, even if the sensor fails.

As it is, the sensor keeps the water within a 1/4" or so of the same place - day in and day out. Between the fill timer and the senor, one of them is going to keep me from flooding the house.

There are several things in the system that complicate this. First water temperature. The water coming in is cold - too cold for discus. So I had to put a second hydor in-line heater on the tank. Before this system I could maintain discus temps with a single 250w hydor. But now I've got an additional 350 watt to keep water changes from chilling the tank.

Second, I have to do chlorine removal. That's easily done with a couple of in-line carbon matrix filters, set between the line from the tap, and the "T" taking tap water into the tank's return lines.

Third, draining is really slow. The diameter of my drain lines, coupled with the tiny 1/8" opening on my solenoid means my drain line has to be pressurized to flow decently. So when my drain timer turns on the drain solenoid, I also have to turn on an Eheim hobby pump (in line with the drain line) to boost drain pressure - otherwise it would take forever. This same narrow solenoid opening, and narrow fill lines is less of a problem on the fill side because the tap water is decently pressurized.

And fourth and last, hysteresis has been my major nemesis. As I understand it, hysteresis is that set of factors that limit/drive when the sensor switches power on and off. The first type of hysteresis problem is called "dead band". That is a pressure range that has to be exceeded for the state of the sensor to change.

In other words, if a pressure sensor had a deadband of 20% of the "set pressure" (a common deadband value BTW) it would mean this - if the pump is set to trigger at 30" (turning a circuit off for instance), then you could lower the water 1" or 2", and the pressure sensor would remain off. It would not turn your power on, even though the water level had dropped. A deadband of 20% setpoint means, in this example, that you have to drop the water AT LEAST 6" (20% of 30") before it would switch on again. So every water change has to be at least 6".

I got around that by finding a sensor that was very sensitive - extremely low deadband. So I've barely got to lower the water at all -1/64" maybe - before it is willing to flip the switch and raising it again.

But with a lower deadband sensor, I ran into a new problem - rapid cycling. So now I can do small water changes, but as the water reaches the full condition, someone walking across the room creates enough vibration that the sensor rapidly cycles through an on/off/on/off condition. And as the fill cycle approached a full point, as long as the fill timer was putting power on the fill process, this would be a problem. No one could walk across the room for the last few minutes of the fill process, without causing the sensor going nuts, trying to figure out if it was full or not. With a bigger deadband you don't have this problem. Once it senses full - even for a split second - it can't be convinced it's not full until you drain out a lot of water again. And that rules out small water changes.

At long last, I've got it all sorted out now. I had to buy an Aquacontroller Jr. though to make it finally work right. The AC jr. replaced my fill and drain timers. That's not such a big deal. But it also let me remove the electrical junction box that the drain solenoid plugged into.

Now I've got my sensor wired directly in to my AC jr. The AC jr times the fill and drain cycles, and turning on/off my fill and drain solenoids as required. And with the sensor wired to the sensor, the AC jr can now tell when the senor senses a full condition, and so the AC jr is programmed to shut off the fill solenoid when that happens. This would still be subject to the rapid on/off cycling problem from floor vibrations but the AC jr has a hysteresis function! That's one of the main reasons why I bought it. So now I can set the AC jr to a 1 minute hysteresis. In this case that means, if it turns off, then stays off for 1 minute before turning on again. In my case, that is enough for someone to have finished walking across the room. And if they are still walking around, that's cool too. It'll just keep turning on, and then off (from the vibrations) once a minute until the people stop walking around, or until the timer on the fill process stops. Now this could leave me a few tiny fractions of an inch from full. But because I have the timer set to fill just a bit more than needed, that tiny gap will be made up within a day.

Wow. Sorry for the long post. Now I need to follow up with pictures.

Works like a charm now - just like I always wanted - no crap in the tank, nothing hanging over the edge or at the top of the water, all the new equipment hidden in the stand, and allowing me to do multiple small water changes through out the day! Too cool!

I'll post pics soon I hope.
01-25-2007 04:17 PM
Jackfrost Steve

How have you been ?

Long time since I posted to you.

I think you are on travel somewhere this month.

Anyway, I have been reading you threads (I see you have been busy !), with great excitment.

I am in the process of getting my 90 G into setup. I have alll the stuff I need plus some. I will be going for a Discus tank also.

Yours looks great by the way !

I am very interetsed in you auto water changer with that pressure sensor.

Any updates on the design ?

10-19-2006 03:14 AM
Don't Need More Help On This One...

Just an update folks. I've got a solution. It's been working for months. But I'm still shaking bugs out. I've already written up the little problem associated with the rapid fire on/off fluttering caused by floor vibrations when the tank hits the full condition, but the timers are still powering the fill solenoid. There are other problems I'm working on too. I'll get it all worked out. But I'm finding that there are a lot of little gremlins out there, waiting to bite you just when you think you've got it.

So I'm holding off on my write up.

I have started to structure my how I hope to document what I've learned. Posted below is how it begins, based on what I've learned so far. It could be months before I'm ready to post the final stuff. So here is the teaser...

Automatic Water Change - A Structured Approach


There is lots and lots of chatter out there at the moment about automatic water changing and top-off. And there are a number of people that are waiting on my long overdue write-up of my DIY solution. But that is a work in progress, and it is clearly going to be so long before it is perfect IMO, so I'm writing up what I've learned to date. I'll document the final system when all the bugs are out.

To make it easier for me to get my heads around the many different aspects of such a system, I've broken it up into a number of easily digestible pieces, organizing the information into sub-systems and critical attributes. If you are thinking about building such a system, breaking it up into these separate pieces might help clarify your thinking about it. It should certainly should help you better understand the impact it will have on your tank, and what risks you might be undertaking.

Sub Systems
Most water change "systems" are a combination of multiple sub-systems that work together, providing a comprehensive solution. I think those sub-systems are typically: drain, fill, water treatment, water storage, and automation. I am not going to try to discuss every option for implementing each of these sub-systems - I'm not qualified to do that. Instead I'll break the description of my solution up into these sub-systems.

Critical Attributes
Any solution, built from multiple pseudo-independent sub-systems will have any number of critical attributes. I believe there is a short list of attributes that must be well understood to have a successful Automatic Water Change system. Those critical attribute are: maintenance requirements, drain/fill capacity, drain/fill schedule, drain/fill duration, water replenishment rate, water temp impact, water pH impact, water hardness impact, fert ppm impact. And of course there is my favorite critical factor: fault tolerance.

To be continued in a few months. I suppose the only reason I'm tossing this out now, is because if you are considering building such a solution, and you do not fully understand the "critical attributes" for your system, you're potentially headed for serious trouble. They are bear understanding before you start. I've learned that the hard way.
10-19-2006 02:58 AM
scolley jdb416 - Man! I hate that you went to all that work! I mean, I DO appreciate it! Thank you. It's just that it doesn't help.

It's a good design. It would work IMO. But early on I rejected any solution that required an external column of water like that. And that solution has some other problems that I've discovered, I just haven't written up yet. Thanks for posting that though!

styxx1 - You are right pal, those solenoids are great. I am absolutely not worried about those failing. My discus took a while getting used to that "snap!" sound. It is loud in a quiet tank. But they do work very, very well.
This thread has more than 15 replies. Click here to review the whole thread.

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

For the best viewing experience please update your browser to Google Chrome