Sunday, October 25, 2015

Automatic Cat Feeder - Electronics

The main electronics

Just a quick note - the green and blue wires you see that are going over and behind the plywood are not going anywhere really - it was just a way to get rid of the 'slack' I had, without cutting the wires to length. I always try to give myself slack when hooking up wires, run cables, etc.

Whenever I start a new electronics project, I break it down into different parts. The reason being is that I do not have a lot of experience in electronics, and therefore do not know much. So, I take the different components that I need to learn about, and proceed to learn about them one at a time.

From my point of view, there's 3 different aspects/components to the electronics for this project:

1. The Arduino parts
2. The motor parts
3. The 'clock' parts

So let's go in a little deeper...

1. The Arduino parts

I actually already knew about the Arduino parts from past projects, so there wasn't much to this part for me this time.

I'm not really going to go into detail about this as you probably already know about Arduino. However, if you are a beginner (like me) and want to know more about how to 'make your own' Arduino, you can check out my Instructable. Basically it shows you how to use your programmed ATMega328 chip on your own circuit. Instead of using up an entire, retail Arduino board ($25-$35) for your projects, you can just use a few cheap electronic components and build it in to your own project, like I did with this automatic cat feeder.

Here's a photo of an 'Arduino' on a breadboard from my Instructable that I just mentioned:

Standalone Arduino on a breadboard

I do want to mention that during the learning/development/testing phase of a new project, I do use an actual, retail Arduino board. I have a couple of older Arduino Duemilanove boards that I always use for prototyping.

2. The motor parts

This ended up being an expensive part of the project. I've never 'driven' a motor from a microcontroller before, and had very little experience with motors in general. Having no idea what the specs needed to rotate the plastic paddle in the food dispenser, I purchased several different kinds and sizes of motors. I tried to keep it cheap, and mostly bought from RS Electronics. None of them ended up being sufficient enough though - they all spun pretty fast which was one problem, and they had very little torque which was the other problem.

What I needed was a motor that spun slow, and had good torque. As I mentioned in the main post for this project, I ended up buying a motor (on eBay) very similar to the one that this guy used. The specs for it:

Rated voltage: 12V
No-load speed: 5RPM
Product diameter: 37mm
Rated torque: 5kg/cm
Stall torque: 30kg/cm
Weight: 190g
Shaft diameter: 6mm
Product dimensions: 37mm (diameter) x 81mm (total length)

Here's a link to an eBay search that may come up with a motor with the same specs.

As I mentioned before, I have no experience with controlling a motor with a microcontroller, so this was all new to me. From what I found, there's a few options - you can buy an Arduino motor shield like this, this, or this. You can build what's called an H-bridge. You can drive it from your Arduino directly maybe (if you want to kill your Arduino). Or, you can go with one of these. It's fairly cheap, and pretty simple to use. It drives 1 or 2 motors (which I needed to run 2), so it worked out great for me.

If you want to use the same motor driver board that I did, check out the Bildr tutorial for it - it's what I used to learn how to hook it up, and run it.

One word of advice though - the last paragraph says that if you set the 2 direction pins HIGH or LOW, the motor will stop. This is not entirely true - you also have to enable standby with:

digitalWrite(STBY, LOW);

Ok, one more word of advice - I think that running a motor at full speed from a stopped position is not a good idea for the motor, so I 'ramp up' the speed in my code. There's a link to the Arduino code from my project's introduction page, but here's a little snippet:

  //Don't spin motor full speed immediately - ramp it up!
  for (i = 1; i < 256; i++) { // loop from 1 to 255
    moveMotorLowLevel(whichMotor, i, 1); //motor A or B, speed from 1 to 255, spin left
    Alarm.delay(10); // give a small delay between each speed change

Regarding dispensing the actual food, I just had to play with how many seconds to run the motor in order to get the desired of food dispensed out - there's no magic formula/equation as far as I know to get it precise the first time.

That about does it for the motor stuff, now on to our last 'component':

3. The 'clock' parts

What good would an automatic cat feeder if it didn't automatically feed the cats? There were 2 ways I could have (that I know of) approached this component:

1. Purchase some kind of automatic switching timer device like this one. There are tons of these on the Internet and brick-and-mortar places like Home Depot, Lowe's, WalMart, Target, etc. Going this route would have simplified the code/software side of things drastically.

2. Add a few components to the circuit to have a built-in clock/timer. This route makes writing the code more involved/complicated, but due to the existence of 'Time'-related libraries, I didn't need to write everything from scratch.

If you haven't picked it up already from reading my other posts, I'll just tell you now - I'm cheap. And, I wanted to learn about building my own clock and implementing 'timers' and the like. Sooo, I went the route of #2.

Turns out that it wasn't that difficult. I actually bought a couple DS1307 Real Time Clock breakout board kits a while back because I wanted to learn about making my own clock, and I knew this day would come where I 'needed' to build one into a project. If you looked at the link, you may have noticed that there are only 6 components including the battery/holder. So for this project, I just integrated those 6 components into my circuit and called it a day!

There was 1 major problem that I ran into doing this however. See Challenge #5 on my previous post, Automatic Cat Feeder - Challenges.

Well that brings this post to a conclusion. Hopefully this information will help someone out there. If you have a question, just leave a comment below.

Sunday, September 20, 2015

Automatic Cat Feeder - Challenges

Below is an overview of the biggest challenges that I had when working on this project. FYI they're not listed in any kind of order.

Challenge #1 - The food dispensing system

My first challenge was what kind of mechanism to make to actually dispense out the food? My father-in-law helped me brainstorm about this and he came up with some pretty good ideas. The one I started pursuing was a system with a 4" diameter polystyrene coupling. I even had a couple of 'discs' laser-cut by Ponoko. They were both shaped like PacMan. One was going to be stationary and the other was going to be rotated by a servo or motor.

If you want to see a video demo of how this was a fail, check it out:

I did come up with the idea of having a small 'vibrating' motor inside of the cylinder, which is to 'agitate'/move the food so that it keeps dispensing instead of getting stuck. I believe this would have worked, but I was having too much trouble with mounting the motor properly centered, and what hardware to use to keep it mounted/stationary without being in the way of the food.

Shortly after I scrapped the idea, I came across this guy's homemade automatic cat feeder system. I loved it, so I based the dispensing system off of his. The actual dispenser is a commercial cereal dispenser. Oh, 1 more important requirement is that it had to be food safe.

Here it is mounted on my 'canvas':

Food dispenser mounted on the 'canvas'

(Note - I'm using the word 'canvas' to refer to the big piece of plywood)
(Note 2 - I use the word 'canvas' very loosely - I am in no way any kind of artist)

Challenge #2 - The food delivery system

This one was probably the 2nd most difficult (the first one was above) part of the project. If you saw the video in my Overview/Introduction post about this project, you may have noticed that I have 3 different food delivery systems in place. The first one was made of regular cardboard. I 'scored' the cardboard and bent/folded wings or rails. I had to play with different angles to make sure that the food doesn't 'bounce' out of the track. It seemed like no matter what I tried, food kept bouncing out. To prevent this, I made a cover much in the same way that I made the base part of the track. Here's what I started with:

Beginning of the first food delivery vehicle - before the cover

I have a generic 'L' bracket attached to the 'canvas' and one of the cardboard 'tracks' is resting on it, and I have a single machine screw with nuts to hold it securely on the bracket. The other track is attached directly to the 'canvas' with a couple of wood screws.

My 2nd method is utilizing a 'core' from a roll of vinyl that I brought home from work (I work at a sign shop so we are always tossing these cardboard tubes).

Vinyl core cardboard tube

As you can probably see, I attached this to my 'canvas' using a couple of carriage bolts (and accompanying hardware of course). On top of the tube is just a cut up plastic Listerine bottle, which just acts as a 'funnel'.

Side view to show a little more of the hardware

The last dispensing 'system' was done out of paper towel tubes, and the chopped-off/top part of a 2-liter soda bottle. The cut-up bottle also acts as a 'funnel' to catch all the food from the dispenser.

Paper towel tubes duct-taped together

I was a little weary about using these cardboard tubes as they're so small and thin, but it actually ended up working quite well. I probably didn't have to use any hardware to 'mount' it to my 'canvas' but I did anyway just to be safe. You see, it fit so tightly behind the 1st chute (the cardboard track), that it doesn't really move. However, I did use a couple of screws anyway:

Can you see the screw through the top hole? (It's too dark to see the bottom one, but there's one in there too)

The way I used screws was just cut a couple of holes out of the tube, drilled a small 'pilot' hole in the cardboard (and canvas of course), and screwed it in. Then I just put duct tape over the holes to cover it to make sure no food comes out:

Holes covered up with duct tape

The only other thing to show as far as the food delivery system is the bottom 'mouth' attachments:

End of the left and center tubes

You probably can't really tell, but for the one on the left, I took a piece of cardboard from a box of Mountain Dew Voltage, rolled it, cut it, duct taped it to the bottom of the vinyl core cardboard tube. For the one on the right in the picture (actually my center one), I just used more paper towel tube, cut a hole in it, duct taped it to the bottom of the chute.

I'm actually not finding any good pictures for the bottom 'mouth' part of the chute on the far right. I just made it out of the same standard cardboard, which I scored, folded, and duct taped.

One of the biggest challenges for all of these was getting it the right angle and the right length. I wanted to make sure that food won't get stuck in there, but didn't want the food to come flying out so fast that it wouldn't stay in the bowl. Additionally, I didn't want it to go too far over the bowl and get in the way of the kitties' heads. So it took some adjusting and testing, gut I finally got it working pretty well.

Challenge #3 - Stationary Bowls

Our cats tend to move their bowls all over the place...sometimes all the way off of the little platform. So my first idea of making the bowls stationary was to 'bolt' them down with a few holes in the bowls, a piece of plywood and some basic hardware:

1st method of keeping bowls stationary

A closeup:

Simple hardware - bracket, screws, washers, wingnuts

I used wingnuts because this is obviously a lot easier to unscrew by hand (as opposed to regular nuts). If nothing else, I knew this would be fairly easy to release the bowls so that we can wash them. Oh, and in case you're wondering why I am using a flathead screw (wrong kind of screw in this application), it's because I didn't have 'panhead' of the right size/length on hand.

Anyway, this seemed to work for a bit, but the screws got loose pretty quick (kind of like my head). I had to come up with something else. I'm not sure where I got the idea, but I thought if they had round (and tapered) bowls, I could make another little platform that would allow them to just 'drop in', so I headed to our local pet store, Petsmart. Found some nice round stainless bowls, so I got to measuring, measuring, then cutting. (Remember - always measure twice, cut once!)

This is what I came up with:

New platform

Side view with it propped up a bit:

Just showing how the bowl sticks down below the main piece of plywood,
but barely just under the entire 'height' of the plaform
I'm much happier with this - better bowls, stays stationary with no hardware, and very, very easily removable (and put-back-able)

Challenge #4 - Button issues

Below are the original buttons that I had put in place. I chose them because they're cheap, and easy to mount:

Original buttons

Well, I had major problems trying to get them to do 'double-click'. You see, I was wanting to have the buttons perform multiple actions - single-click to give the cats a 'snack', double-click to give the cats a bigger snack, and a 'hold' to continuously dispense food until letting go of the button.

I actually had already mounted the buttons before I got to that specific portion of the code. I tried other kinds of buttons and the double-click would work just fine in the code. The Arduino forum came to the rescue again. Someone suggested try using a different 'button handler' (my words, not theirs) - the 'OneButton' library. Well I did some tests with this library and the buttons worked a lot better. However, I had already implemented the 'other' button-handling code, and would have been a huge pain to re-implement the new code. (Remember, I'm a beginner at all of this)

Instead, I bought a couple new buttons and tested them out - they worked a lot better, so I decided to replace the really cheap buttons with not-as-cheap (but still pretty cheap) buttons since they work a lot better:

New/better/current buttons

Challenge #5 - Transferring the DS1307 chip

This was a bit of a strange one. It is purely an electronics problem. The problem is that, during development, I had connected everything up on a breadboard, so it looked something like this:

Circuit on a 'breadboard' (not the circuit for this project...just an example)

The main white rectangular with all the holes is called a breadboard. It allows you to build/prototype circuits, and is a common practice to use them while developing. During the creation of this project, I used a breadboard to build the different parts of the circuit (then finally combining them all).

Ok, so nothing unusual so far, everything worked great on the breadboard. The next step (for me) (I did not want to design an actual circuit board as I was not going to make multiple boards, or sell them.) is to 'move' (or recreate) the circuit on 'protoboard' (a more permanent substrate). An example of protoboard:

One kind of protoboard

Again, no problem, I've done this several times before. But wait a minute...the small chip (the DS1307 - the one that keeps the time) is going to lose power when I move it from the 'breaboard' and onto the 'final' circuit on the protoboard. Uh oh...we have a problem!

How do I move the chip from the breadboard to my actual board:

The eagerly awaiting chip holder on the actual board

Above you can see my mostly complete protoboard without the time-keeping DS1307 chip. Searching on Google gave me nothing. It's probably the search terms I was using. I did have a couple of ideas though. The first one was to build more onto my circuit - the parts that I would need to actually program the microcontroller directly on this board. You see, the way you 'program' the DS1307 chip is through another my case, the Arduino. However, this seems like a waste of time and materials as I'll probably only ever do this 1 time (with this circuit). My 2nd thought was very 'hacky' but would be quick and fairly easy. The idea was to solder 2 wires directly to the chip (sitting on/in the breadboard), hook up a battery (temporarily) to the other end, remove the chip (with the wires and battery in-tow), insert the chip into the 'chip holder' (on the protoboard), desolder the wires, be done.

There was at least 1 possible problem with this approach, so I had to do a test. I was worried that if I had 2 coin cell batteries hooked up at the same time, the voltage would double and fry the chip. However, I was pretty sure that doing that would be fine because they would be hooked up in parallel (as opposed to series), which means that voltage would stay the same, but the run-time would double, which is fine. I did my little test with a couple batteries and my multi-meter, and confirmed my thoughts - it should be safe for the little guy to be transplanted!

Immediately after moving the chip from the breadboard to the proto board:

The chip still has the 2 wires soldered to it after moving the chip to the final board

A closeup:

The 2 wires still soldered to the 2 pins on the little chip

The completed transplant, wires removed:

The 2 wires have been desoldered

As you can see from the above 3 images, I followed the steps I outlined. Then I confirmed that the chip kept the time during this silly process - so we have a successful transplant! Now you can call me Dr. Nerd!

There you have it biggest challenges with this project and how I was able to overcome them. Hopefully this may help someone out there doing something similar.

Next up (and probably lastly for this project) we'll talk about the electronics.

Automatic Cat Feeder - Introduction/Overview

Completed Automatic Cat/Dog Feeder

Above, you will see my latest project - the Automatic Cat Feeder. It is Arduino-based, and built with various off-the-shelf hardware, and some parts that I modified.

How does it work? The Arduino is programmed to dispense food @ 5:15AM and 5:15PM for our cats' daily meals. The Arduino basically turns on the motors (1 at a time) to turn the food-dispensing paddle for a pre-determined amount of time to dispense the right amount of food. The food drops out of the dispenser and down the various chutes you see, and into the 3 bowls.

Before I go any further, I must give credit where it is due. First, I borrowed heavily from this guy for the actual dispensing system. Second, I had a couple of issues on the electronics side of things, but the great Arduino community over on the Arduino forums offered great advice to help get past the hurdles. Lastly, driving the motors was made very simple thanks to Sparkfun's motor driver board

Why I created it
Our cats like to be fed at consistent times of the day, and that's not always easy to do. For various reasons, we can't always feed them at the exact same times daily. Sometimes they may be fed hours later for dinner, and we feel really bad about this.

Now why didn't we just buy an off-the-shelf product? We looked into this and found that they were just too expensive, not very flexible, or were easily defeat-able, allowing the cats to gorge on the food whenever they please. I don't like the idea of an inflexible system that I have little to no control of (yes, I may be a bit of a control-freak). If I build my own, then I have complete control of it, so I can change it, add features, etc. myself.

Before starting this project I had a few requirements in mind:

  • Must be food safe!
  • Must be reliable
  • Must cost less than comparable commercial feeders
  • Must be customizable
  • Must have some kind of audio notification to inform the cats that it's breakfast/dinner time

Is that asking too much? Nah

Ok, ok, I know you want to see a video of it in action, so here you go:

I had in my mind that I would build this with network connectivity to give us the ability to monitor/feed them remotely, but for a couple of reasons I didn't. First, this project already took me a long time to get everything worked out, and just wanted to get it installed for them. Second, an Ethernet Shield or Wi-Fi shield are fairly expensive ($35-$45), especially compared to the new ESP8266 (around $4) that's come out recently. I say recently because I actually started this project a couple of years ago. So why didn't I incorporate the ESP8266? I don't know how to use it just yet.

Ok, I'm sure some of you want to know just how much everything costed me. Price breakdowns are in the Bill of Materials (linked above), but the cost of everything was $190. Yes, that's expensive, and a lot more than I was anticipating. The 2 most expensive items were the motors (about $25/ea) and the cereal dispensers (also about $25/ea). That's $100!. Originally, I was going to come up with my own food dispensing system, which would have saved me over $50. Additionally, I feel that if I would have shopped around and done more research I could've cut the cost of the motors in half. As I'll talk about a bit in future posts, I got impatient and decided to just go the route I did.

Also remember that I'm feeding 3 cats with 2 different kinds of food...your situation may not be as complicated.

In conclusion, I wanted to let you know that there will be more info coming. In the next couple/few posts, I will go over some of the challenges I had with this project, hopefully saving you time in the future. I will also talk about the individual components in more detail (yes there will be videos). I'm not going to do a full tutorial like I did for my Garage Door Controller, but there will be some pretty good detail, hopefully enough for you to replicate if you want to.

Below are some links for you to use if you want to replicate this system, or parts of it.

Arduino code on GitHub
Fritzing file (actual .fzz)
Fritzing file, exported to .png image
Fritzing file, exported to .pdf file
Bill of Materials (Excel)
Bill of Materials (Open Office)

I haven't created a 'digital' schematic yet, but if/when I do, I'll update this post with the link to it.

Well, I hope you've found this post/project useful and/or interesting. Feel free to leave a comment if you have any questions/ideas/suggestions.

Want to know what the challenges I encountered, and how I overcame them? Well, that's exactly what my next article is about

Sunday, May 3, 2015

Source of electronics woes in my garage discovered

It's been around 4 years that I've been having strange problems with 'homemade' electronics devices in my garage.

It started off when I built a (semi) 'automatic cat door'. My little device allowed our cats to push one button (mounted to the outside of the door) when they want to come in, and another button (mounted on the inside of the door) when they want to go outside.

After I built a working prototype that worked flawlessly 100% of the time in my testing, I moved the circuit to a prefboard, installed everything on the door, and thought I had the project completed - (this is a video on my completed project). However, it didn't take long to notice that it was behaving strangely - it would not respond to button presses sometimes, the servo wanted to keep moving even though it hit a limit switch, the servo would go the wrong way, etc. I would have to power cycle it to get it to work properly again. I was never able to figure out what was wrong with it, but the light shone very recently and now I am confident that I know what the problem was...(don't worry, I'll reveal the problem/solution at the end of this post)

A little later in time I purchased this really cool Internet-connected sprinkler system:
But I had a very similar problem with it not behaving properly...see one of my threads about it:!topic/rayshobby/eeiMdxePBUs
Again, myself or the creator of that device was not able to figure out the problem. So I would have to constantly check on the status and make sure that it was working...when it stopped working I had to power cycle it.

Lastly, I developed an automatic Garage Door Controller. It took months and months of development, and I tested it a lot in my 'lab' (read: my very little bedroom in my house), and it worked wonderfully for months before I took it into the garage to test/install. When I did finally install it in my garage, it was so exciting, and it worked great for a couple weeks. Then I started having strange problems where it wouldn't respond to my commands, wouldn't let me connect to it, etc. I would have to power cycle it once or twice a week to get it back working. I took it back into my lab and tested and it would work for weeks with no problem. I couldn't figure out what was different in my garage and my lab. I spent months trying to figure out what it could be, posting on forums, etc. Recently, however, I figured out!

The suspense is killing you, right? Drum roll please:

Ok, so I don't know which one of the above 2 Insteon devices is the culprit. I had them connected together in my garage, and for the test, I pulled both of them out, and haven't tested 1 at a time yet. However, ever since I pulled them, I haven't had any problems with the sprinkler system or the garage door controller (and if I was a betting type of guy I would bet that if I reinstalled the automatic cat door, that would be working great too). It's been about 3 months now, and things are still looking great.

I plan on trying to isolate which device of the 2 is the culprit, then I will try to get in contact with Smarthome/Insteon and see if they know anything or have any useful information about it. I will report back here if there's anything useful I can share.