Backyard Aquaponics
A place to discuss aquaponics

Well you've got your valve, you don't have to time the drain part, you can adjust the hight of the water in the GB's by having a flexible pipe from the GB's to the cistern and just changing the cistern's hight. If youv'e got to go electrical flush with a solenoid connected to the ball valve. Easy to maintain and best of all cheap .

Dam'n; Sorry Kuda

Posted these responses on the wrong thread.

np, happens to the best of us

This Arduino thing seems like it would work well to automate and log AP happenings:

http://cachefly.oreilly.com/make/arduinoMAKE07.pdf

http://www.arduino.cc/

Crapload of projects for the Arduino:
http://blog.makezine.com/archive/arduino/

Cool - so who is going to get one and have a play? How much would various sensors cost - eg a temp sensor?

Temp sensors cost virtually nothing just use a thermistor (Temperature sensitive resistor) and make it waterproof. I would love to find a cheap ammonia sensor but they are really upper end.

Yeah cheap. Here a thermistor costs around Rs. 10 or so, which is around US 20 cents or so.

Hi womble and welcome to the forum...stop trying to slink around like that and put up some info on your (potential) BYAP system

And don’t forget LM35 with output voltage linearly proportional to degrees Celsius,
LM 34 calibrated in degrees Fahrenheit

Not widely available but I get LM35 from http://www.modtronix.com.au/product_info.php?currency=AUD&products_id=186

yeh, and like you said, linear................ thermistors are about as linear as a banana

Hi EllKayBee,
Yes just keeping quite at the moment as I am a complete beginner. I am pegging out an area for a small ferrocement tank at the moment to see how it all works, It will only hold guppies and goldfish.

Good point by Steve S the LM35 is a better device.

My thoughts on the issue of a control system for AP:

Warning long ramble ahead but the task is not particularly difficult to do.

I am resonably good with microelectronics and was going to make a microprocessor controlled system for the pump and aeration systems etc but I decided that it would be difficult for non tech people to duplicate (although there seem to be lots of very skilled tech people on this site!)

My thought now is to use an old (free) computer as a controller in a weather proof enclosure and use a control board (lots of cheap types around) to control the system. This would mean that the problem now becomes a software issue rather than a hardware issue which would be easy to solve. The system could also be linked back to your home computer if you have wireless internet so you could manage it from inside.

I have only started to scratch around at the problem this afternoon so you will have to bare with me. Control software should be able to handle any output control board as the output lines would be set through a system configuration file. This would mean that different boards could be selected by non techs in the same sort of way that you would set which printer you want to use with a word processor.

At first I thought a heirarchical structure would be best for the software, for instance one pond has many growbeds. This would allow you to have a heirarchical tree showing the whole system. The problem with this is that there are an infinite number of ways of combining a system. Two ponds could be linked together, growbeds could be flushed at the same time as each other or at different rates depending on the water sensitivity of the plant type in each bed etc etc etc.

Because of this I am now thinking of controlling everything individually. In the example image on this page I am thinking that there would be 2 types of entities.

1) Ponds and growbeds. These would be collection entities that contain data collections such as temperature, water levels etc. They also show the related boolean entities.

2) Boolean (on/off) entities for the rest. Solinoids, water level indication etc. Some things would be data read such as temp and some data outputed such as pumps on.

Rules would be created for each entity. Eg
- Turn Pump 1 on for 20 minutes ----To flood
- Test for max water level and stop pump 1 if true ---- If full
- Turn solinoid 1 on for 10 minutes ---- To drain
- Repeat cycle

This is only rough thinking but if every physical item became a controlled software entity then there would be very little that you couldn't do with the system.

I was going to do exactly this... but I would never use windows or even mac for that matter. There is too greater risk of PC failure.

I want to design a PIC controlled timer which will operate on its own, ie without computer control. But can be connected to a computer to adjust on/off times. Eg. Not enough oxygen, then increase on time, or decrease on time. Temperature dropping/rising. Reduce on time. Etc.

So on its own the system will function fine. Maybe even have a default timer setting hard programmed into the chip. But I still want to monitor and control the system using the parameters from pH, DO, temp and NO3 metres.

I knew this would be the first reply as it was exactly my first thought. My solution to this is the following.

The interface that I showed would just be the control panel. You would use it to set up the system and see graphed data etc, It would also be able to be used on a remote computer. The information would all be written out for a batch processor to interpret meaning that it could run under a much more stable system such as DOS or Linux. (Who ever heard of DOS blue screening with a simple exe?)

A catastropic failure system could also be built in where a failure causes a full reboot. (power kept high to a relay). A relay could also sound a mechanical alarm.

The final system is the one that I would use. Two computers! old computers are free and plentiful and it would not be difficult to get them to poll each other over an ethernet connection or parallel cable and take over in the event of one failing.

I do understand the benefits of a dedicated PLC however as they do have a lot of pluses as well but unless someone produces one commercially then only a few people will be able to do it. The other thing is the output of the program will be a very simple macro file so there is no reason why a microcontroller can't read in that data and do the job of the computer.. I might get around to doing that some time as well.

The new(?) micro's on the market today can do everything you can imagine. They cost almost nothing, use almost no power and are VERY reliable.

I built mine controller with a PIC18 controller and one rs232 device. One other small board with a couple of amps to scales temps and daylight sensors. Bought some solid state relays off ebay for next to nothing.

My times at this point are fixed. They only change at night.

This is a little strange but it was a fast easy approach:
I send messages out the serial port into an old pc and capture them using VBA inside excel. The messages get stored in a table and if they are an alarm, they are passed to outlook which sends me an email to my computer and a txt message on my cell phone.

This is a fairly low tech way of doing this and I will someday try to consolidate everything into outlook and bypass excel or try to write a standalone application. I think I can still log to a file from outlook. Then I'll set up outlook to watch incoming emails so I can and send queries from the house or cell phone....

There are many ways of running these systems. I'm certainly not saying one approach is the best, but if you do automate, it has to be somewhat robust/reliable and where I think the PC is not a bad system for logging and external control, to make your system dependent on their reliability might be asking too much.

For an afternoon of fun, you could try linearizing it in software, (inside a microcontroller) calliberating it with melting ice(0 deg C) and boiling water (preferably distilled) (100 deg C) and noting that the curve of a thermistor is a simple exponential curve.

If you can't find the LM35 that is.