Doghouse automation design advice

The idea is to provide as much water as the dog needs and have it be as free of pollutants as it can be (do you know how much bacteria your "watering bowl" accumulates in just one hour?).

I'm sure much less than the dog's own poop...

I (partially) agree with @x2dz based on a recent experience. We were taking our dog to a trainer. The place had several filled water bowls scattered about the place. The bowls were filled each day with fresh tap water. As expected, our dog got thirsty from the activity and drank some of the water. The net result was a near death experience and a $500 vet bill. Even the freshest water put into a dirty bowl can KILL A DOG.

For an outside dog, water should ALWAYS be available from a CLEAN BOWL.

One thing I would seriously keep in mind is that you're giving Arduino quite an important task, it will more or less be playing God for the dogs.

Searching for a PH-meter-design I stumbled on a thread in an aquarium-forum of a guy wanting to automate his. The thread/project was looking nice until he came home one day while his controller kept looping in the feeder-routine, the fish had gotten food for a month in one day. Since the heat was turned on by the thermostat-function, which wasn't called for any more, the aquarium more or less turned into fish-soup.

It sure is fun to automate a doghouse, but trusting it to work perfectly could be far less fun...

That sounds more like an incident of one dog drinking after another dog without being in close contact with each other; cross-contamination, so to speak. I'm not saying dogs or any other animal should drink from a dirty bowl (even when my parents kept chicken, we would clean out the old sink we used for watering them every once in a while). Should you be overly concerned about the bacteria content over the course of a day or two, though, if its just your dog(s) at your own house (no other stranger dogs, or other houses)? Probably not. Our dog has never had a problem drinking out of our fountain in the 10+ years we've had her, and I don't recall my parents being anywhere near as concerned about the same for any of the dogs we owned either (outside or inside). None of them ever died prematurely from water intake.

All of this is really off-topic, though - so I'm not going to comment more on it; do with your dogs as you wish, I guess...

:slight_smile:

I read that they would still server their purpose even though they're not connected directly across the coil - is that correct?

It may be, but I've never seen it done like that. I can't see how that will clamp a positive spike.


Rob

Simpson_Jr:
One thing I would seriously keep in mind is that you're giving Arduino quite an important task, it will more or less be playing God for the dogs.

Searching for a PH-meter-design I stumbled on a thread in an aquarium-forum of a guy wanting to automate his. The thread/project was looking nice until he came home one day while his controller kept looping in the feeder-routine, the fish had gotten food for a month in one day. Since the heat was turned on by the thermostat-function, which wasn't called for any more, the aquarium more or less turned into fish-soup.

It sure is fun to automate a doghouse, but trusting it to work perfectly could be far less fun...

This sure got offtopic quick! :smiley: I'll say this - I've looked after dogs for most of my life and I would NEVER dream of leaving them unattended in the "Arduino hand of God" :smiley: for more than a couple of hours. This automation will be done solely for the fact that it would be more comfortable and healthy for everyone involved :wink:

@cr0sh - I do believe that people should be registered to have people, too - why do you think there are so many twisted people out there and the jails seem to always be overcrowded? I think that in order for you to look after any living thing you should have the living standard (i.e. funds) AND the necessary psychological mindset. If something happened to an animal I was looking after, God forbid, my conscience would be clear - because I would have done everything the way it should be done. If you were in that situation - would you be able to say the same? I know dog-owners like you and I'll tell you what I tell them - you're on a slippery slope. Be careful because the simplest lazyness or ignorance on your part could spell life or death for your pet. Enough said.

Does anyone have any comments on the actual design? Any criticisms or ideas?

Well, there's the post right above your last one...

@Graynomad commented about diodes in your schematic. Did you not respond because you agree and feel no further discussion is necessary?

Oops... I was thinking about it and then I forgot to write it in the post... My bad.

Would someone who has done inductive load control via transistors jump in and say what's what? Would this configuration work? Because I'd prefer the diodes be placed on the PCB for safety reasons (the motors and valves will be exposed to the elements and I want the part count to be as minimal as possible)...

Thanks again to everyone, who took to time to read this thread!

x2dz
While I disagree with your over regulation of your dog, and some of your interesting thoughts on bacteria in standing water, based on my observations of our cat, the sheep and cattle in our paddock.
This is besides the fact that in the reply two up they weren't going to be in the hands of the Arduion god for more than 2 hours.

I've had lots of pets over my life, and haven't managed to kill one yet, but I blown up the odd electronics.
They usually make noise too, (sometimes it sounds like a relieved sound).

Transistors and FEts are different beasts.
Your drawing shows Fets with the internal clamping diodes. These provide both 'flyback' and overvoltage limiting.
The idea is that when an inductive load (relay) is de-energised the collapsing magnetic field induces a voltage in the coil windings.
This voltage is opposite to the original (antiphase/reverse...) and depending on the inductive load can be as high as the energising voltage. It wants to go somewhere and since it came from the relay, it ain't going back there.
The power supply capacitors act as a short circuit to this voltage (remember its the wrong way around, and a short duration spike) and low and behold the little old transistor/Fet is open and waiting for a reverse voltage, that usually exceeds it rating (for reverse voltage).

Usually the voltage spike is very short duration, and creates havoc on the supply line, upsetting the electronics. Often the bypass capacitors can't cope. (see below about the lack of them). So stopping it before its created is always good.

The diode across the relay coil will short out the induced voltage before its created.
The clamping diode across the transistor protects the transistor/fet, and the voltage (current as a result of the voltage) gets absorbed back into the coil, after it goes around through the capacitors and through the flyback.
Either one does the job.

second point
the power supply section seems to be lacking any sort of filtering (much like that horrible bucket smoothing out the water)
You need both larger values of capacitor and the 0.1uF ones as well. The rule of thumb is 1000uF for every 1Amp the circuit will supply, so add up those relays and make sure you have enough filtering.
Unlike the dog, electronics will only use what is needed, so you can go overboard, but doen't forget the 0.1uF as well.

general point
I tend to use an optocoupler on the output of the uP. Over the many years in the electronics game, I have noted that whenever something goes wrong, the transistor always dies. if you are lucky its between collector and emitter, but since murphy always appears, the base also gets clobered.
If the base gets clobbered the poor old uP (or other delicate electronic bit) usually receives a smack as well.
So I try to avoid this.

If you were concerned with running wires everywhere, the otion of a remote unit using an optocoupler and transistor/fet/triac on a seperate board is also avalable, and you only need to feed the 5/3.3v signal out.

Now that we are back into health care for animals....what exactly are your qualifications in regard to animal husbandry/vet/animal care (not sure if the same term is used overseas).
And btw from my life experiences, its often the people with money that don't look after their pets as well as those without money...come to think of it their cars and kids could almost be in that same boat.

i also recall reading somehwere that the old cowboys (we never had them here, they were all young) usued to eat poorly and drink water that was less that ideal, and then when they reached town, drunk lots of spirits, which neutralised the bateria.
Perhaps there is something in the mystery of why my cat prefers to drink from a manky bowl outside that could probably grow penicilin, rather than his fresh cleaned (yes we do actually wash it, but not in the dishwasher) bowl in the kitchen.
Perhaps hes topping up before cleaning out my booze cabinet.

Mark

Well, that's one hell of a reply :smiley:

OK, let's get started...

  1. The schematic does in fact show FETs with internal clamping diodes. The reason I included an external one is two-fold
  • It could be left in as an optional component - i.e. if I weren't able to find protected FETs at a reasonable price (protected = back EMF, overcurrent and thermal protection) I would always be able to put in some regular old FETs and any run-of-the-mill clamping diode. I like having options.
  • The second point is that even if I were to put in protected FETs I read a couple articles claiming that the internal diodes are often sub-par to an external one - be it current capacity, recovery time and so on. Some of the articles, in fact, went so far as to say that you should always include an external diode, regardless of it being engineered in the FET.
  1. From what I understand you're saying it's OK to have the diodes being across the FETs rather than the load? Is that right? :fearful:
  2. As per the power supply - I was thinking of using a ready-made supply with enough filtering to make additional filters unnecessary (i.e. a computer power supply). But in any case would you be as kind as to tell me the parts needed for, say 10A? (if I understood you correctly its 10 (1000uF caps) + 10 (0.1uF caps) /I'm guessing for HP and LP filters/?) How would they be wired?
  3. Ah... galvanic isolation... My sweet, sweet friend... :smiley:
    I also tend to like having galvanic isolation when having more than one voltage level in a circuit, but...
    The reason I didn't use optocouplers is again two-fold:
  • I decided that for a 10 channel output profile as the one I'll be building, the cost of an Arduino blowing (atmega328p, $6.59) was low enough vs the price of the 10 optocouplers I'd need to get.
  • This one is a little trickier :slight_smile: Again per some documents I found out that I'd have to select not just any plain old optocoupler... It said that most optos would be able to supply enough current to signal the output, but not enough to fully saturate the FET. This would require an opto with a supply connected to the receiving side as well. It all started to feel a little redundant and I didn't like the idea of having to scavenge some exotic opto out of the thousands available, order it and have it delivered two weeks from today... Call me lazy :smiley:
  1. My expertise is mostly in the realm of dog's psychological and behavioral aspects, but that doesn't mean I haven't read a good deal about dog physiology and health-care. Most of the points I have stressed throughout this discussion are based on scientific FACT, not personal experiences, as is the case with the other participants and them using the "if I can't SEE the harm in it, then it must be the right thing to do" dogma. That's just not me.

p.s.
The people with money vs those without point... Here's the thing -

    • People WITH money COULD look after their pets IF they wanted to (i.e. the people themselves wouldn't go hungry because of them supplying food of decent quantity and quality to their pets /in my scenario I'm talking about to dogs with a combined weight of over 100KG/).*
    • People WITHOUT money would potentially go hungry (I should note that I'm not talking about rich vs poor people... I'm talking about some sub-par standard, necessary to look after a pet properly) even if they wanted to look after their pets in a decent way.*

Again thanks everyone for the input!

  • I decided that for a 10 channel output profile as the one I'll be building, the cost of an Arduino blowing (atmega328p, $6.59) was low enough vs the price of the 10 optocouplers I'd need to get.

What else will be on the Arduino side? A computer?

What do you mean? Would you please elaborate?

An Arduino is fully capable of running without a host computer. Despite this fact, do you plan to include a host computer?

x2dz

  1. From what I understand you're saying it's OK to have the diodes being across the FETs rather than the load? Is that right?

. I thought I said that, but you have read some other stuf as well, by the sounds of it.

I was thinking of using a ready-made supply

. Yes but then you stuck a 5v regulator in between.
If you are using a computer based supply, the 12v side is nowhere near as well filtered as the 5v.
A good practice...is to always put one as close to the regulator as possible.
If you use a 470uF either side you should be fine, given the intention of using a supply with caps in it.

  • 10 (0.1uF caps)

Not sure where the 10x0.1uF came from.
Good practice is to put one either side and as close to the regulator as possible, PLUS one across the supply pins of each IC.
In your case there's only the 328, so thats one extra one. (total 3).

Ah... galvanic isolation

Either you read alot and pick up terms, or have some additional electronic knowledge than you lead us to believe, given its an Arduino forum, and not an electronics engineers forum (no offense to the other electronics experts)
If its for yourself, then no sweat at $6.59 you understand the economics.

If its a commercial design, then the ability to have it serviced at your local TV repair shop, and the thing with legs replaced is good practice.

I'm tired of the new version of explorer making the bottom disappera, so thats it from me.

@CodingBadly: No, I don't... If you were mislead by me mentioning a computer PSU, I just happen to have some spares that I KNOW output a good stable, filtered supply. That's it.

I may decide to include some sort of LAN interface to the Arduino later on...

@markB:

  • What I read about the clamping diodes concurs with your opinion. I was just making sure I understood what you said. :slight_smile:

  • About the supply filters - I really did forget to add the caps next to the 7805. I'll correct that - thanks! I guess it slipped my mind, since a couple weeks ago I did another design and I just checked - it appears I put them in there :smiley:
    Would you just clarify if you meant - a total of two 470uF (on either side of the 7805)
    and a total of 3 0.1uF (either side of 7805 and across the 328 supply pins) - just for certainty's sake?

  • The galvanic isolation bit - I'm currently studying 'Automation and Information technology' at Uni and I guess that would explain some of the phrases I used. But apart from that up to this point we've had very little actual practical application and in that respect I'm pretty much self-taught.

Thanks again to everyone, who responded! :slight_smile:

x2dz

Would you just clarify if you meant - a total of two 470uF (on either side of the 7805)

. Yes at the very least. if you have physical room for a larger one on the 5v side, it doesn't hurt.

and a total of 3 0.1uF (either side of 7805 and across the 328 supply pins) - just for certainty's sake?

Yes
again this is the least.

I'm currently studying 'Automation and Information technology' at Uni

Ah yes theoretically uni is a great place to learn.
personally I spent 30 odd years learning it in the real world, with the occasional courses after I understood the basics.

Since this is about automating your dog ..sorry the dogs environment, how come your tutors at uni haven't been able to review your design.

mark

Since this is about automating your dog ..sorry the dogs environment

  • Haha... Still on that, are we? :smiley:

how come your tutors at uni haven't been able to review your design.

I have asked on or two of them if they could take a look at it, but I thought I'd try here in the mean time and iron out anything stupid I'd done :smiley:

p.s. Besides some of them are wonderfully helpful if you ask them stuff about the lecture they just presented, but if you take them a little out of their comfort zone (for example hitting a problem they're not certain about), they tend to get short with you if you know what I mean...

Here's the latest PCB mock-up...

Thinking of going into production soon :slight_smile:

Hi guys,

I'm currently selecting the MOSFETs for the design and would like to hear your opinions on whether the IRL1004 is a good match :slight_smile:

The largest load will roughly be around 2.5-3Amps@12VDC - this will be the motor, driving the awning for the houses. I don't think it will run for longer than 15-30 seconds at a time.

This Power MOSFET is in a TO-220 package, has a R DS (on) = 0.009 Ohm with a V GS = 4.5V (it is a logic level MOSFET) and according to my calculations should have no problems whatsoever handling the task it will be performing (and I'll more or less be wasting its potential on the sub Amp loads - valves, etc.). I used P = I (of the load)^2 * R DS (on) .

The only thing that puzzles me is that there are two Power Dissipation figures mentioned in the sheet - one is approx. 50W (for the TO-220 package, in the beginning of the datasheet) and the other is 200W as a Maximum rating for the FET. Shouldn't these be closer to each other?

I'm attaching the component's datasheet and would be greatful to hear your opinions. I will, of course, be running practical tests prior to ordering in quantities, but nevertheless - better safe than sorry :slight_smile:

Thanks again in advance!

irl1004.pdf (94.6 KB)

p.s. Would I be able to get away with using IRL2203N or IRL3103 or even a IRL540 FET since I don't really need the power of the IRL1004 (plus they're cheaper)?

I'm attaching their sheets :slight_smile:

10x

irl2203n.pdf (225 KB)

irl3103_2.pdf (218 KB)

IRL540.pdf (996 KB)