DIY controller

[ronc98]

Thanks MotoPsycho, I am not sure if I want to go commerical with this or not. There is plenty of interest in the DIY area. I am hoping once some of these people get their boards stuffed and running we can get some good apps other than what I already wrote. There is not else you can add though. I have pretty much add everything I would ever want, other then auto water changes. From some of the reading i have learned that if I am going to vary seasonal tank temps I should also vary salinty as well. That might be one of my next projects. I am getting to a point where this is closer to the biosphere then a reef tank. The whole controling fever is starting to spill over into Home automation.

R



R

 

[fppf]

Nice looken switch box Ron.

 

[ronc98]

Thanks Fppf I love acad 3d. It makes packaging so much easier.

 

[fppf]

Yeah, I'm drawing my house plains now.

 

[ronc98]

Update.

Progress has slowed a bit. I got a deal of a lifetime on a whole house generator and have been installing that.

As far the controller is concerned today I added email support. There are two modes. One sends an email when the controller reports an error state. Error state would be an over temp, stuck level switch or loss of power. Along with the error I also record and send the sensors and temps for that given day. Second mode sends an email at the end of the day. This email sends the average values for all sensor, tempand weather pattern for that day.

Once all of this is debugged it is time for R&D on the conductivity probe.

R

 

[sharemaster]

Hi Ron

Watching with interest in your project looking very good, I have a few ideas at a not too technical level

The connection to the PC is r232 at present? would it be possible to make a network module maybe a printserver type affair with an IP address to allow remote access from the web to check on the tank and adjust if neccesary?

Maybe even add USB ports to allow a webcam a touch adventurous maybe! but a picture speaks a thousand words!

As a way of allowing other people to do the same what about a DIY kit? with PCB, components, connectors, box, cd with the latest software etc so we can build one ourselves!

tagging along.

 

[ronc98]

right now you can either connect it to the pc via USB or RS232. I am working on a ethernet web server solution. This will take the place of the PC program completely.

Webcam would be cool however I am not heading that direction at the moment.

As far as a kit You can get it either way built or unbuilt. Interest has been 50/50 When I started this I got enough stuff to build 20 of these I am almost ready to buy more parts.

 

[55chadman]

so whats the cost gonna be for another diy are you gonna make a kit of this or go full out and build and sell this unit to anyone

looks like its got way more potential then the aq and a little cheaper any chance of beta testing this unit on another set up

thanks

chad

 

[ronc98]

Update:

Got the control boxes back from my miller today and put one of these together. Much tighter and much better looking then my attempt to make holes in it. Now I have to decide if I am going to get the labels milled into the plastic or just use a label maker.

 

[natual]

I have to commend you on your effort and to all of you that understand even alittle of what is being said in this thread.
This thing is amazing.
Truly congrats and... DOM Dom dom... when can santa, easter bunny, Uncle Sam(july 4th)pick one up for me.
But seriously I thought of trying to take on a controller for a long time and came up with a wish list maybe it could give you some ideas. You have done alot of them already.

I am building a list of things that hopefully I could eventually put togather to basically leave my tank for a month without ever doing anything (except feeding).
The stipulation is that if you don't have the money to put it all togather up front then add-ons would be as simple as buying a new probe (or whatever) and plugging it in (maybe alittle calibration).

I have looked into alot of threads, webpages, etc.

Philosophy:
Its construction and software, firmware have to be free and available and easily updateable;customizable(the way
linux started), PCB board cookie cutter design and skematic so that everything can be done by people completely
inept (like me).
Parts need to be available, universal, and not easily outdated (i.e. simple and old technology).



Lighting
on/off
each light controlled (assuming that 1 ballast per light)
shut off if tank temp rises too high
lunar cycles for a night light
lights based on Ph needs(option)

Heating and Cooling

Have the ability to have hood fans work either separate or with this as in conjunction have a separate
temp measurement and adjustment for Hood Temp
Override due to over temp
control of a peltier device
control of a chiller
sensing water temp.
sensing Canopy and stand temp.

Electricity

Know what watts and amps each pump, ballast, chiller, etc is using.
Calculation of power usage per month.

Wave making


* Controlling PH's for feeding, random cylces etc

tank parameters


Highwater Flood Alarm/controllers
Low Water Alarm/controllers
makeup water reseviors Level sensor/controllers
shut down circuits
power back-up systems
Sump level sensors
Nilsen reactor controller
O2 probe
main pump output at head and at tank.
Saltwater pump on/off mixing and salinity after 1hr-24hr.
CO2 usage/ Pressure (monitor)
Additional ion exchange measurements for specific elements/compounds
Current Transformer to detect if a device is powered.
Battery backup power check and system check.
RO/DI pressure and auto shutoff/controllers.
programable pumping cycles
Wave timer cycles
pH measurements for tank and Ca reactor.
ORP (and for OZONe after skimmer).
Temperatures
PAR sensor(s) on main tank(s)/refugium(s), out of tank(to know how much light is reflected out of tank),
independent waterproof sensor for monitoring PAR a particular light makes at a given distance and from air
to water, etc. during sunrise/sunset, noonday.
Infrared sensor in water(easy with led type sensor)
UV sensor in water(easy with led type sensor)
TDS
Salinity in main tank(s), QT tank, brine, phytoplankton, rotifer, and other tanks.
dimmable on-demand red led nightlights
dimmable led moonlights according to lunar schedule
monitors 4 float switches and/or sensor type level monitor.
Detect, Turn on/of humidity of room and dehumidifier.
Auto fishfeeder and hopper control
Tunze-type variable speed control
seaswirl monitoring.




Data Logging

Logging of watts, KWt/hr., and amps of each electrical tool is using/second,/day,/month,/year,/since start
Day versus night watts,KWt/hr., and amps using.
Heating times, (everytime heater(s) kick on/off make a log).

Software
Calculators for KWt/hr. used, plugin for current $/KWt/hr.
Calculators for Salt used in wtr. chgs.
Calculators for time spent on maintaince (show you how much you have saved by automating)
PC page and internet function when parameters have strayed it'll call you.
Web broadcast of tank.
Heating increase after differnet lighting systems or equipment is turned on
Program flexible for many different configurations of board and probes.
Ph at night with kalkwasser dosing
Feeding mode.
suspend mode.
Delay program (user define param.s for set period).
Time, day, month, year. Stamp on everything
timer
countdown timer
CO2 Low tank Pressure warning


System design

All modularized for easy add-ons.

user friendly text boxes for changing programming settings.
Power indicated LED
audible alarms (tied into pager and network) with certain ticks
to tell you what is wrong
Double or Triple microprocessors for control of different things
2 CO2 solenoids to controll ph during photo period with Co2
additions.
i can call home modem and check my tank with my work laptop
If things gets careless...It logs time off and restart, turns on
backup power ( 12v )
Automatic and manual control of backup systems.
Able to monitor up to 20 Tanks(80 sensors, 200 pumps,lights,
etc).
Software able to monitor up to 4 units.
110v, 220v (AC/DC) for overseas, and battery backup
(12v,9v,etc).
X10 modifications and acesss to x10 and other interfaces.
Interface
RS232 (serial) PC interface.
USB, ethernet, parallel ports(USB to Parallel Port Adapter).
LCD w/ backlight and buttons for unplugged standalone.
joystick controller(posablity instead of a button).
Expandable memmory (flash, EEPHROM(sp.),other) slots
atleast for a month of standalone logging.
Portability

 

[NY_Englishman]

Hi Guys,

This is a great thread. It's kind of spooky that I just stumbled across this thread as I was in contact with Joe at mcuresearch today about the apparent lack of a fail-safe mindset in controller design. I w rote the following:

"I am thinking of developing (as a future DIY project) a fail-safe aqua controller. By fail-safe I donââ"šÂ¬Ã¢"žÂ¢t mean just employing redundancy but having a fail-safe self-monitoring circuit that checks for simultaneous switching of duplicated critical circuits, as well as checking itself ââ"šÂ¬Ã¢â‚¬Å“ therefore single component failures will be detected and the system will shutdown (safely) and alert the customer. At present, on most systems out there, it looks like if the temperature sensor fails, or the relay, or its drive circuit inclusing uC, controlling the switched 115VAC, the heater output stays on - there is no fault detection (other than a email/SMS message) and the tank ââ"šÂ¬Ã…"œboils.ââ"šÂ¬Ã‚ This bothers me and just adding a second relay or even a whole second parallel channel wouldnââ"šÂ¬Ã¢"žÂ¢t do it unless the circuit is able to detect an individual component failure. I guess whilst there is much technology employed within aquarium electronics and commercial level sensing etc it seems no-one (I could be wrong) has approached these problem with a fail-safe mind-set (ie single component failure detection). I guess if you loose a $50K home aquarium itââ"šÂ¬Ã¢"žÂ¢s not as big a deal as losing a limb on a commercial guillotine.
"

Anyway, one of you guys mentioned fail-safe in this thread somewhere and I was just wondering about what you've done in this regard. Does your current circuit design detect any semiconductor failure, or relay contact failure and immediately abort to a safe condition? Or do you "just" have some redundancy that is itself not checked. Or do you check redundant circuits with a non fail-safe monitoring circuit. Or is it fail-safe with respect to single component fault conditions?

Anyways, respect to you guys and keep up the great work. And now I know of a local S/W guy!

Cheers,
Englishman

 

[fppf]

I went as far as I could with keeping cost down.
The sensors are triple redundent. Each senor in the set is on its own set of IO pins. For the outputs if the signal is lost they all turn off. For the heater and chiller outputs they are SSR's backed with a mech relay. If the system senses that the heater or chiller is stuck it will open the mech relay. The only other layer that I missed is having triple micro's. The micro has external watch dogs that will shut everything down if it fails.

I work in the aerospace industry, we take "Fail Safe" very seriously.

 

[NY_Englishman]

Thanks for your explanation and I'm pleased to see you've thought about your FMET and DFMEA; I'm glad you added redundancy (this is good) but you have no automated fault detection (this is bad, and, admittedly, a difficult black art to grapple with).

For example, if the N/O contact (s) on your "mech" relay fails to open you don't detect that failure (that is not fail-safe, even though there is no catastrophic event yet). A subsequent failure to your SSR (however unlikely) would lead to the heater being permanently on. Not good. The non-detection of a single fault which is then followed by a subsequent fault is considered still to be a single fault condition (in my world). The non-detection of a single fault condition implies the system is not fail-safe - it just means it is fault tolerant which is a different kettle of fish (no pun intended).

In a real world situation, it is unlikely that you'll get multiple failures and I would much rather have a system like yours that has redundancy, but wasn't fail-safe, than one system relying on say a single uC port driving a single transistor/smartFET/SPI bus relay driver which in turn drives the heater relay.

Having worked extensively (H/W design team leader) in the automotive and, less so, aerospace industries (amongst others) I agree that the level of redundancy is often impressive (the number of redundant channels of RR Trent EMS was staggering) but, alas, not fail-safe - just catastrophic-failure-unlikely because of all the redundancy. The most stringent standards I've designed to are Machine Guarding Systems (eg BS6491) - these systems are used, in all the usual hazardous environments as well as, in Nuclear Power Stations (eg Sizewell B in the UK).

I only mention all of this because people often go beyond two channel (ie one level of redundancy) instead of making the design fail-safe. It is more cost effective, and results in a more elegant design, if you do the latter rather than the former. It's just more difficult to do.

Anyway, that's my 2 cents - I'm not trying to start an argument or have a go at you or anything, I just wanted to know more about your design! I'm sure the feature list is excellent and the software well written. Thank you again for documenting this.

 

[ronc98]

Update:

About a week ago I was at a point where I was happy with both firmware and host pc software. All the wish list of things I wanted to get done have been finished and debugged so I get to start the fun stuff. I have been playing around with ASP code to give this a true webbased interface. That is cool and all but you still need to be at a PC to monitor it. So with a little more research and playing around I have decided to go along the routes of mobile internet technology. Bascially I am going to write the asp in MIT so you can access this with a webbased cell phone or PDA. I am going to link the asp into the PC code so all the work on the PC end is not wasted.

I have also been doing some research on ion specific probes. From what I have read so far their outputs are very similar in nature to a PH probe and will connect right into the circcuits I already have designed. A little code work and some testing and I will have a means to measure Ca, ALK, N3 and a few others.

This weekend I start to breadboard my conductivity circuit conductivity is rather easy to measure all the work is in the probe itself and I am going to buy it instead of build that.

A word about measuing Probes (ph,ORP ECT) in electrically isolated tanks. This has not been as easy as I would have thought. The outputs for these probes are super super small and if the tanks are not connected via water there is really no easy way to get them electrically connected. The potential for ground loops is always there even with grounding probes. The only way I have been able to get good stable readings is to use isolation circuits to seperate the probes. This gets expensive! I am curious how the commercial controllers claim to do this. There are two ways run the probe in a differential setup which is impossible with the probes we use or isolate the inputs. I doubt they do either and just live with the offset.

Here is a graph of the two PH probes I have running my tank. The bottom is my Calc reactor the top is the tank PH. The spikes you see are from the KALK top off water dripping into the same area as the Tank PH sensor.

 

[BeanAnimal]

Ron from my understanding the ION specific probes are very short lived. You will get a few weeks or a month or so out of each probe.

 

[ronc98]

I will have to do some more research. From what I have read so far the life is more like a 6 months to a year.

 

[BeanAnimal]

the verneir probes are pretty cheap and they tech support folks are fairly informative. You may want to give them a call.

Even at 6mos $75 a probe is kinda prohibitive. I don't have a PH probe yet, but was told by most of the manufactureres not to expect more than 12-18 months out of them, and that 24 is the absolute max. I am talking about constant emersions... of course storing the probes in distilled water and using them only when needed will surely result in a much longer lifespan.

 

[ronc98]

I think life span with anything is more of a question of the enviroment they are in. I have run PH probes for two plus years in my tank without problem. You do have to recal but still I do not see a significant change in the call coefficients over time. I am going to get a few Ion specific probes just to play around with and just see. I guess this is what R&D is about.

 

[BeanAnimal]

Please do keep is informed, as this is an area of great interest. When I get around to my controller I would love to log as much as possible.

 

[natual]

<a href=showthread.php?s=&postid=6608736#post6608736 target=_blank>Originally posted</a> by BeanAnimal
the verneir probes are pretty cheap and they tech support folks are fairly informative. You may want to give them a call.

Even at 6mos $75 a probe is kinda prohibitive. I don't have a PH probe yet, but was told by most of the manufactureres not to expect more than 12-18 months out of them, and that 24 is the absolute max. I am talking about constant emersions... of course storing the probes in distilled water and using them only when needed will surely result in a much longer lifespan.

Bean,
You really don't want to store pH probes in distilled water, they need to be in a buffered sys. I don't mean to hyjack but Ion specific and even general probes are very sensitive and quality of probes is a big issue, specifically pH probes that are all glass usually have a better durability and larger surface area. Denver instruments makes a good probe and so does Orion but some would disagree. Plastic probes do not seem to hold up as well. Most can be stored in .30M KCl or Friskolyt or equivalent.
Conductivity and calibration I can help w/ too.

If you want more info pm me.