DIY controller

[ronc98]

Here is my final DIY controller, I have been working on this for about 8 months and I have gotten to a point to where I am ready to share this with everyone. Those of you that have been wanting to make your own controller to tweak it for your tank this is it.

The plan for this is to be an open source/ schematic project so I can get some more programmers to hammer out some code. At this point it is 99 percent done for my tank however there are some wish list things that would be cool to add to it. Anyone interested playing around with this let me know.

Here is a description along with schematics.

The unit is 68hc11F1 based with a Real Time clock and 32K of NV ram and 64 of PROM. PC communication is standard RS-422 multidrop(small converter hangs of the PC RS-232 port)

Firmware is written in C.

There are four connectors that supply 7 open collector outputs for use with relays or outlet strips. I have also made an outlet strip that is optocoupled with high current triacs for outlet switching. You can really design your own strip depending upon what you need. These outlet strips are hardwires however then can be remote to about 20 feet as long as you can run a wire similiar to CAT-5.

There is one connector that supplies four level switches(contact closures) as well as 4 open collector outputs.

Two RTD temp probes.

two 12-bit DA(0-5V) outputs.

Four BNC probe inputs that can be configured for PH/ORP/ect. All these circuits are similar and can use the same types of circuits with different components. All of these connect to the HC11 via a 12 bit ADC.

Current firmware

There is calibration and temp compensation for three PH probes and a direct reading of ORP. There is not much need for ORP calibration since it is a direct MV reading. I have not played with any other probes however I doube it will be too hard. Each of these probes can be software routed to control an AC output for controlling purposes

There is Calibration routines for both Temp probes and PH/OPR Probes.

All outputs are programmed via Time on and Time off. These are also linked to one of the Temp probes so you can turn on heaters, fans and then start turning off lights when certain temp thresholds are reached. Again this is really user defined and the user can modify the firmware himself to meet these needs.

There is an input level switch and outlet used to does and mix a Kalk reactor. Mixing interval and duration can be programmed. Will not dose to the tank unless mixing has been off for 15 minutes.

Eight random\period AC outputs used for wave making and pump control.

All of this can run standalone however I also wrote a Host program that connects to the controller and monitors all the probes/temps/outputs and inputs. These can all be saved to a ASCII file and graphed. All of the time and setpoints can be viewed and modified via the PC. Since the controller is RS-422 it can communicate with multiple controllers with the same program. The cool thing about this is you can VPN to your computer to monitor while away. This saved me a few times.

In the works:

phase moon controller.
Intensity control for lighting.
Dosing pump control.

[/IMG] Main controller schematic

[/IMG] Main controller PCB

[/IMG] Outlet schematic

[/IMG] Outlet PCB

[/IMG] Bottom view of finished controller

[/IMG] Rear View of finished controller

 

[cmhollis]

Ron,

How much are you looking to get for an assembled unit? Does this have the capability to connect to my home network so that I can check on the system without having to connect a computer to it?

Thanks,

Mike

 

[csammis]

I don't care so much about the hardware, but I'm simply mad about firmware-level C

Got a link to the code?

 

[AcroSteve]

Ron, this looks hardcore ... and sweet!

Brings back memories from college.

 

[fppf]

Why did you use an 11?
Thats a really old chip and wont be around much longer.
Even a lot of the HC 12 family just got changed to "Not Recommend for new design".

 

[ronc98]

FppF

I built this on the 11 because I have the kernel done. Granted it is old but I do not see all the doom and gloom. I do not think getting the hc11 will be that hard.

 

[fppf]

Its not that hard to port up to the 12.
Should get some SMT on there

 

[ronc98]

Not to be an rude or anything. but dont you think if I designed this I would know what it takes to port to a different processor? If I went anywhere I would go with a microchip device for the flash ability.

 

[fppf]

Ok, no problem, HC 12 has Flash on it too.
My next design will be on the 32 bit chip. Don't know why I went with a 16 bit, guess it was lazzyness. You can get a lot of bang for the $5-$10 now. The new Blackfin chip from AD is only in the $10 range and it clocks up in the 600 MHz range with 40 bit accums. You can even get it with a dual core. Not that you need that much for what we are doing but hey, why not?

 

[ronc98]

fair enough but I wanted to keep it simple for the coders out there. IMO using a 32 bit processor is way over kill.

 

[csammis]

edit: nm, I accept that I have no experience with the chips in question

 

[zapata41]

where's the digi display.

 

[BeanAnimal]

Nice job roruting the board. How much was manual and how much was autorouted?

Moon phase and Intensity going to be PWM or digital->analog dimmer?

 

[hms]

Hello everybody ,

May I ask a stupid question regarding to the " How the probe (PH meter , ORP ... etc )work " ? Are they current sensitive or voltage sensitive parts and what the vendor and vendor part number for the OP amplifer (for example U41 , U40 )?

Thanks for help !

 

[BeanAnimal]

Most read on the mV scale. I would imagine any opamp would work, but am not sure what part is intended here.

LMC6001 Seems to be the most popular for DIY probe amps...

Maxim/Dallas MAX406 is also mentioned in an app note at the maxim site.

I have seen a few LM324 designs as well.

 

[ronc98]

The board was completely manually rounted. Autorouters are worthless IMO unless you have a really expensive one($50K).

Both ORP and PH are standard circuit that require an opamp that has a very high input impedance. Anything with a jfet type input works well. I used op07 in my circuit.

By the way the display 4x20 backlight LCD is on the other side of the box along with the 4 key keypad.

 

[BeanAnimal]

I have played with protel for hours on end. Not being trained in it's use, and not being formely trained in PCB layout, I found it rather frustrating to say the least. The manual routing function that highlights the NETs is pretty cool.

I don't see a lot of vias or jumpers (compared to my poor layouts). You must do quite a bit of layout and routing.

*drool* and you can afford solder mask (and I bet silkscreen on the other side). I have been using the poor mans method... etch em myself and then go nuts trying to drill centered holes.

 

[ronc98]

This board is to dense to etch them myself. I have a PCB miller that I used for the protype. Not many vias but there are enough. If you get a board made you have to get a solder mask and silk. It does not cost much more plus it helps.

I do this stuff everyday so not a big deal. Took about a day to layout.

the last 6 months have been writting and debugging code.

 

[fppf]

LOL
I was the other way, more time in layout and less time in software.

You can get proto boards made without mask and screens. Saved me about $50 on my boards. When its all done and ready to go I will order boards with masks. Its much easier to prototype without masks, you can solder on to traces anywhere.

 

[ronc98]

All of my prototype boards are done with a circuit miller. It basically mills the traces out from a copper covered board.

In this project it is 90 percent software. Hardware is easy it is all the features that you always end up adding in that tends to drag software development out. Since this is standalone and can connect to a Pc to setup and record data you double the programming required.