Grow controller? (DIY)

[Hephaestus]

Should; I'd probably swap em out between grows - but otherwise I wouldn't worry about them; don't get the walmart ones - head over to a pet store, get some half decent ones... Wouldn't bother with anything funky like basswood; in fact I'd probably go with some simple bubble curtains - (basically a pipe with a bunch of holes in it) those would last virtually forever...

 

[medi-useA]

I use those 3-4 inch disk diffusers....and I got a kick-arse micro-fine bubble porcelain areator for b/day as well....not allowed to let them dry out once wet.

Have had bubblers disintergrate on me...they should be replaced every 1-2 grows....and always keep spare bubblers, air and nute pumps as well as a spare lamp or two from when you last upgraded.

It might seem like a pain in the arse, but when things go belly up, you will be so thankful for your spares....and on this I am speaking from experience!

muA

 

[medi-useA]

Hep...I found these...think they might be of use..if only for learning and playing?

this
and this one.


also


this
as well as this...

muA

How's the assembly going on?....or has LIFE interfered?

muA

 

[Hephaestus]

pH probe works LCD works... port expander works... RTC - needs a tweak...

Still waiting on one friggin chip and those UPS sob's to deliver it... Been sitting in winnipeg for 3 days...

Life's somewhat in the way - always is

 

[Hephaestus]

MuA pm'd me this one...
http://www.wireplant.com/
Looks like he's planning similar to my scenario - but right now he's using off the shelf parkfun bits n pieces Hasn't gotten into the port expanders etc that he'll probably figure out PDQ he's going to need...

And yes; I'm 98% sure now; UPS does in fact hate me... Grrrr!

 

[medi-useA]

And yes; I'm 98% sure now; UPS does in fact hate me... Grrrr!

I'm sure all the lovin you'll get when you release the Hephaestus GrowShield onto the growing world will make up for it!...

muA

 

[Hephaestus]

Well if the chips would ever show up... Designs are pretty close to done...

I'm REALLY getting grouchy with not being able to finish this up and shoot the boards off to batchpcb for a run...

Main board - just need a little tweak and test out that last chip's interface... And working on a way to feed 12v up to the Second board...

2nd board is pretty much done (see power feed comment above) . PPM/EC/TDS probe; 4 PC fans (PWM control); 4 inputs for 2wire float switches (using an i2c A/D converter - no more ports used).

Third board - Will be an LED driver board; but its still a long way out; still playing with the theory...

Funds got a bit tight (damn wimmen going overboard for xmas) so probably even if chips do show up - going to have to wait into the new year before I've got some pennies to throw at further development...

 

[Boleman]

I don't know if you are still looking for the right diffuser stones. But the bubble wands that Aerohead and I use are the cat's meow. They don't degrade, and they keep making tiny bubbles, I'm on my second grow with the first set of tubes.

 

[ClamboyBob]

Great thread! I have been wanting to create an automated garden for years and am finally taking the plunge (first time gardener, long time geek). Have thought about this on and off for years but have really been thinking about it recently - so I have a few (many) comments (but of course) Found this thread late last night and did a quick read through but will take a little more time with it today and follow more of the links...

First - if you are not a programmer/hardware geek please at least read (and respond to) the "Your Needs?" section below - it will help me shape my project. Thanks!!!

Hephaestus, I like the direction that you are going in! So here goes some food for thought... And yep, this is going to be long Don't take my comments wrong, I am trying to contribute as well as answer some of the questions swimming around in my head.

Processor - I think that the Arduino Duemilanove would work for a small setup and obviously could be expanded like you are doing with the "port expanders" (and external A/Ds, etc.). However my thought is why add the extra complexity/shields when you can use something like a Mega (or high pin count PIC, etc.) that would give you the additional I/O as part of the processor board package and reduce the need for additional parts. These more powerful boards/processors provide more I/O pins, more RAM, more EEPROM, etc. Having more RAM/EEPROM/pins/etc. is going to serve one better in the end regarding future proofing (room to grow if you will).

Expandability - I think this one has been touched on already and is something I am "fighting with" in my mind. I could build a minimal system that allows me to control the small garden that I am planning or I could build something that is expandable. IMO expandability serves everyone better; I can start with the minimal setup required for my needs and someone else could start with a larger setup that meets their needs - if my needs increase I can easily expand. What I am wrestling with is this:

• (small setup) - do I create one controller for the mother/cloning area and another duplicate one for the flowering area?
• (larger setup) - do I create one setup capable of handling both areas?
• (expandable setup) - do I create something that can easily be expanded to meet my needs (and also more easily meets the needs of others)?

I am leaning more towards the latter two.

External components - Creating a shield, or shields, is certainly one way to go but I feel it lends itself to the smaller setup. Reading through the thread I see that the proposed single shield setup has grown to a multi shield setup - which may answer part of the expandability question. However my thoughts are leaning towards a processor board plus a single I/O board (connected via cable). I am thinking that a single I/O board with lots of capability might be the way to go. When I need another sensor/switched outlet/etc. it is already there. I am also not completely sold on the Arduino; a single large I/O board could also contain the guts of an Arduino (or PIC, etc.).

Connectivity - my solution will most definitely have WiFi for a couple of reasons.

• Don't need a computer sitting there connected via USB, etc.
• Ability to receive alerts on my cell phone/email/etc.
• Configuration via web page, etc., served up by the device

I also love your idea of using Cat5 to connect components. I'll take that a step further and suggest a new standard called "CannaBus" (god I'm clever) which could be used to connect a central processor to multiple satellite processors/devices/sensors. As someone's garden grows and/or new devices become available they can be connected to the central processor via Cat5, WiFi, etc. This would allow a single place to configure/control a setup - and if we were really smart it would be designed in a way that allowed a "slave processor" to take over if the "master processor" failed. Also defining and standardizing CannaBus would allow multiple folks to create their own add-ons that just work with everybody else's devices/processors/etc.

Liability - Was researching logic controlled AC relays yesterday and noticed that SparkFun (one of my favorite places to buy electronic doodads) does not sell prebuilt logic controlled AC relays - but they do sell all the parts (including an empty circuit board) and have a nice tutorial on the subject. Why? Probably because they don't want to be responsible if you burn your house down... Obviously as humans we make mistake - both on the provider side (faulty circuit designs) and on the user side (running a zillion amp light through a 1 amp relay) - I don't want to be responsible for destroying someone's garden or burning down their house.

Programmability - And the best one for last There are all kinds of consumers out there - from folks who cannot program a VCR to folks who whip out the oscope and modify hardware for uses it was never intended for. Given that I ask what programming model fits everyone? How to make it dirt simple and yet powerful? I'll give an example that is floating around in my head. For my design I'd love to have a bunch of analog inputs that are not necessarily "hard-coded" to any particular use - and that can be used as triggers. For instance - Bob is a soil gardener and Jane is a hydro/aero gardener - their sensor needs are going to be different. Bob may want A/D input "A" to measure soil moisture for a plant and when the moisture level gets too low turn on switch "A" to drive a watering pump. Jane may want input "A" to monitor her grow solution reservoir and trigger an alarm when it gets low. There are about a bazillion different combinations (you need an exhaust fan and I don't, I monitor PPM and you don't, you want to vary fan speed with PWM and I don't). How do we make a system that "out of the box" is super configurable to meet as many varied needs as possible yet retains programming simplicity? Is it too much to ask to have the user program the thing (and I'm thinking events and actions here) via some high level scripting language or do you provide some ungodly web/program configuration that allows users to essentially point and click and create associations, etc.? I am really wrestling with this one.

Your needs? - If you have made it this far... This is a call out for what your needs are. I have never grown an indoor garden (yet) so I can think about it in a theoretical way but I don't really know. I'll be building my system alongside of my garden, expanding it as needed. But I do want to keep some of the above principles in mind to keep the best design possible from the get go.

• What size pumps do you use; any idea on what their startup and running amp/watt usage are?
• What size lights do you use; any idea on what their startup and running amp/watt usage are?
• What sensors do you want (and how many)? E.g. temperature, humidity, fluid levels, soil moisture, light meter, PPM, PH, etc.
• What type of alerts/alarms do you want? High/low heat, low water, high/low nutes, lights on (when they should not be), lights off (when they should be), etc.
• Are you interested in having security alarms built into the system? E.g. motion sensor, door alarms, proximity alarms, etc.
• How many controlled outlets do you need? (given, say, a 10 amp max per outlet)
• Are you interested in advanced fan control? This would likely require additional temp sensors, motor controllers, etc.
• Should connectivity be an add-on module? This might allow me to use WiFi, someone else to use USB (to a dedicated computer), someone else to use ethernet, and someone else to use a cellular data connection.
• How do you envision programming/setting up such a beast? I want all of my inputs/outputs to be assignable (as much as possible) and to be able to take different actions based upon different events.

I'm actually getting excited about CannaBus - my system can talk to your system, your device/sensor/module can talk to my controller.

I am thinking soil for the mother and aero/hydro/DWC for the salad. I have mapped out an area of the shop for a 8' tall, 6' wide, 3' deep cabinet. My thought is to divide the space into 2 areas - 8x3x3 for the mother/clones and 8x3x3 for the salad. Don't really need 8' tall but that takes it all the way to the ceiling and will conceal vent ducting which will run up through the ceiling into the attic space above the shop. Will likely go all CFL (at least initially). Just a small garden for my own entertainment.

CB

 

[cyberwax]

Why not just use an old computer and write some kickass software instead? If someone forks up the $$ i could do it. There wont be a problem finding relays and measuring devices that connects easily to the usb and from that its not really hard to make a control interface. Its cool that people make their own control devices but its not really nessecary when computers are invented.

 

[ClamboyBob]

Why not just use an old computer and write some kickass software instead?

Because I can

Really that certainly is an option; I could see a minimal PC running linux sitting in the corner. But then you have a mouse/keyboard/monitor to contend with (or not), software updates, additional PC stuff that needs to be configured, viruses, etc. Its just a bit of overkill using a PC and would not be as stable as a (well written) uc controller (and would be bigger, noisier, consume more power, require Mac, Linux, Windows versions, etc.).

I happen to enjoy working with microprocessors and think this is the perfect application for one. But there is no reason why your computer could not run CannaBus and talk to my sensors

Do you have links to any of the USB measurement devices that you mention? Hacking one of these might be the way to go (rather than reinventing the circuitry).

CB

 

[Hephaestus]

Laptop solves that - but still has issues... And USB bits n pieces aren't exactly cheap either...

The arduino mega - is nice, but it's too hard to get your hands on... Almost guaranteed someone in the closest urban center can provide you a basic arduino kit/assembled - the mega's pretty much 100% an internet order thing... Trying to keep this largely open source / easily sourced parts... I'm not saying it's going to be 100% able to control the biggest rooms - but should handle most cabinets and maybe a basic room

Had thought about the Relays etc - likely final design will encorporate SSRs & fuses to the SSR ratings; better to burn out a fuse than a 50$ piece of kit...

Phidgets (google it) makes a bunch of USB devices for this type of application... But the price is ridiculous if you ask me...

Not sure on the need for a Cannabus type scenario - while it can be done, it increases costs and complexity radically... Yes it'd be nice to just plug in a bunch of sensors/relays/etc into a cable and have each one addressed correctly; the headache comes in that each sensor then becomes very complicated/expensive vs very cheap/simple...

 

[ClamboyBob]

Yeah, laptop would be a nice space saver with peripherals built in - especially if you could keep it running with the lid closed! However, IMO, its overkill and I'd hate to see it freeze/etc. due to some OS issue.

I only ever mail order electronics parts; no place nearby that sells them

Relays definitely! Maybe breakers instead?

Have looked at phidgets before and found them to be overpriced and a bit "closed". Will take another look, thanks!

My thought with CannaBus was not that each sensor implemented it but that each controller (or processor) did. Since I already plan on throwing WiFi onto each controller all it would be is a little work defining some protocol to splat out over UDP (potentially). My thought is redundancy and backup. But it should be pretty obvious that I am in the early stages of planning this out and still trying to gather my thoughts. Never done this before and want to do it well!

 

[cyberwax]

clamboybob: http://www.hannainst.com/usa/prods2.cfm?id=002002&ProdCode=HI 9815

Thats one atleast, tho i know there are more out there and probably cheaper too.

When it comes to what computer to use there are some eee computers along with small atom processor computers that require very little power and are quite reliable. Just make sure to avoid installing anything except a good virus killer(nod32) and a stripped down windows 7 or xp.(or if you want to go hardcore use some sort of linux).

When it comes to relays there are alot of various providers out there, however i have some positive experiences with theese: http://www.controlanything.com/Merchant2/merchant.mvc?Screen=SFNT&Store_Code=NCD

 

[Hephaestus]

Ouch, that hanna pH monitor is 120usd... I like my method better >5$

While the wireless etc is nice to have - it also creates issues... Start having little devices talking back to a primary system - and it wont be long before there are interference issues (phones) or LEO's walking around with a little gadget like those wifi detectors looking for the grow...

Don't like breakers myself... to easy to just flip it and reset - without looking at the problem... Fuse - you've got a trip to the store, and you put it it - blows again... You're going to ask yourself why at least...

 

[cyberwax]

Well there are other alternatives to wifi relays on that site, and the wifi's are actually the most expensive ones, i would rather suggest optical communication for optimal safety.

There are alot of various usb/pci/serial ph probes, sadly tho they are not cheap. Not to be a lazy bastard but what did you actually suggest that can measure ph for 5$? A robotic arm with ph drops?

 

[ClamboyBob]

cyberwax, those parts look nice! But I'm gonna give it a go myself first (got my soil moisture / reservoir water level stuff working and tested today). And why fiber? If you want to go "wired" cat5 would be lots cheaper and much easier to deal with. I did not notice if they had a cat5 product or not.

Hephaestus, I guess if I was out in the middle of the country with nothing around I might be concerned with RF leakage but the reality is that I'm in a very RF saturated area. I really don't think a slight increase in WiFi traffic is going to raise an eyebrow; besides its all encrypted and only numeric data (e.g. no clear strings saying "pot plant 75 needs water")

 

[cyberwax]

Fiber choice because of the inductive load on our ballasts, if you want to use one(or more) of the relays to control the lights. Just an expensive security issue, you could always put a condenser into the fold to diffuse the returning load. The guys making the relays actually has a video howto.

http://www.controlanything.com/Relay/Device/A0009

Also the phidgets ph monitorer is without a probe, so consider that before dismissing the hanna unit..

 

[Hephaestus]

pH probe is just an op-amp circuit... A TL062 is about a buck, couple resistors, and a bnc connector (ok ok... So I left out the pH probe - 20$ or so out of china)

Just saying - wireless with a not so standard protocol (ie xbee) and on an open source project = very findable... You get some LEO who reads a bit, has a friend who's an IT geek - and it's quite possible to search out those signals.

 

[gardenlover]

Hep- looks awesome! You are really committed to this! I checked out your link and its great, however I think you may have the bloom and maybe the veg hours wrong on the website... I believe it said your bloom was 18/6. maybe I am wrong just wanted to tell you.

CLAMBOY- Looks like you have the thinking down. I would not worry about the 1 amp relays powering large problem Just run to another larger mechanical relay that electricians use.