Fault tolerant vacation automatic pet feeder

Finally, I made it. I went on vacation for a full week and my bunny survived. Though I have a pan and tilt cam watching if anything went wrong.

Those on the market are too expensive. And even if I get one I don't trust them. First, they have a single motor. Auger / screw types can jam. If the motor fails, all meals will fail.

I actually built a multiple meal automatic feeder for my dog. It was complete with a hacked digital timer, and low level logic chips like NAND gates. I did that because I always have about a week to design and build it. And it have to work. The problem is, the motors are too expensive, or I couldn't reliably mount them to "open doors". So I only tested it for a weekend or two before he rested in peace.

At the time I hardly know what is a servo. And after I came across it, I just didn't have time to learn the hardware and software of a controller to drive them. And how much 7 servos cost? I used the cheapest motors and they work fine. The problem is that it's hard to glue some latches to the bare motor axle without gears.

This time around it's different. Tiny servos like sg90 cost next to nothing. Arduino kits are cheap and you can do a lot of other things with it. You order it from Amazon and they came quick enough before I setoff for vacation. I hope it's plug-and-play as they claimed. But I saved some money by ordering from other sources. So it's not plug-and-play but I figured it out how to program it.

I'm sort of a hardware and software professional so the rest is easy. The software is supposed to be simple, just wake up everyday at the same time and turn on different motors. That's why I actually built one with discrete logic that could be more reliable, and worked first time. But I don't trust a cheap controller working outdoors. One power outage or interruption the controller can be trapped into non-operational states.

I was prepared to write and test complicated software so if power is stopped for an hour, the operation behaves as if time has been lost for an hour or two, not whole days and not skipping meals.

After more thinking than coding, I realized it's best to drive the Arduino with a digital or mechanical timer. When the Arduino has stable power, it only need to use one location to remember which motor to turn on. The software is real simple. I was preparing to work overnight to make something work reliably. With this simplicity, I just needed trivial modifications to the sample code. There's only a few "critical" lines so I can tested them thoroughly in no time and went to sleep before leaving early the next morning. So to fail the timer has to fail and the board hardware has to fail. They seldom do.

My designs has been for the servos to open something like a trap door under food, and so all the food fall down to a tray. And I have no time to make the custom containers.

What I did was building a see-saw with fat lolipop sticks, with a cup of food on one end and a servo arm on the other end. When the servo turns the cup will drop. Everything is made of fat lolipop sticks that you can find in craft shops, and school glue. They are a surprisingly fast and secure combinations. To guard against earthquick, bigger animals and an angry bunny shaking the pet house, there are little slip guards also made of strips from wooden sticks and the cups are stabilized by cup cake paper liners. The craft isn't too bad but I have to build ad-hoc guards and defences to defend against bigger wild animals and the weather.

Wooden sticks was a surprisingly good choice because you can stack them up with glue to your desired height, or stick them to square dowels, also from craft stores and also hardware stores.

All the wires from the servo, 21 of them, are not soldered, but using the wires from the kit that plug directly into the Arduino board, and also use the tiny development board for the 5V and GND distribution. It doesn't matter if the wires fail to contact. Only one servo will be affected by one circuit failure.

And to ensure a good drop, the first thing was to custom built a big tray and inclined walls for the pet house. I had sewing threads glued to the cups and sticks to ensure they fall as expected but not into the tray to confuse the bunny.

The see-saw isn't the best design but I picked it because I had no time for other things. A long see-saw can balance out a heavy cup if the servo is weak.

As seen in the picture, one of the cup didn't drop. But there's only treats in the cups. There are plenty of other pellet food and hay that the bunny doesn't eat that much unless he has to. The problem was that the little vertical slip guard wasn't vertical enough. So the see-saw stick was trapped. This was a known problem in one or two of the unique handmade mechanisms. I forced the little glued sticks sections apart until it was smoother for the see-saw stick to drop. But one of them revert to the bad position eventually.

I almost built a automatic flush toilet too using the Arduino with servos. But it's too risky first time so the flush was driven by the lawn sprinkler controller for now.

Homemade custom non-slip microfiber car mat / rug / carpet

This driver side mat is customized from a a pure polyester microfiber bath mat. It is non-slip, thick with "deep pile". Less than $10.

Both the top and bottom side are stitched in squares. So it's easy to cut off the excess parts with a scissor, between the squares.

That's it. Completed in 3 minutes. But the cut sides will probably fray (come off) after repeated machine wash. I'll think about that when the time comes. But you can always stitch it by hand. Or I just found out that you can seal the sides with a flame. I'll try it with the excess parts first.

Update: Heat gun! The craft type. It's so easy to control. You can just melt the frayed threads to heal themselves, melt the edge, or melt the whole pile. There's not even smoke unless you heat the mat more than necessary.

It's more tricky on the driver side with all the controls. At the passenger side, I don't even need to cut anything. Because it's totally non-slip on every inch, it hugs the curves on the car floor. It also hugs the side walls all the way up the glove box. There are only little folds that you can hardly see and hardly touches. So, cutting is hardly necessary.

Why? It's cozy. Imagine a thick bath mat. It's highly absorbent for rainy days. And the deep pile can trap a lot of dirt.

Being purely polyester without the need for a non-slip backing, you can wash it often without worrying that the backing disintegrating. My OEM mats have integrated tough rubber backing at the back. You cannot put it in the washing machine. And after 20 years, the rubber mostly disintegrated but still shredding black residue to this date.

Microfiber don't absorb oil. So it's easier to keep clean and I used it in the kitchen. It even doesn't slip on vinyl when I run fast over it at a sharp angle. The only downside is that without a hard backing, it doesn't keep it's shape. It's still very rectangular but the straight lines are not visually perfect. It's a problem if you have to maintain the look of a formal kitchen. But in a car it's hardly a problem. There are curves everywhere. And you can hardly see the edge of the mats.

I don't remember any microfiber giving me troubles, like disintegrating. This will protect the carpet under it unless the mat becomes very thin. It may be less fluffy as time goes on. But it's thick. And it's not a bathroom where you bare foot all the time.

Restart your heater furnance with sand paper and how to upgrade spark igniter

A bit of history. My Trane central heating system from the previous owner must now be over 20 years old. Now Trane isn't in the same sort of furnace business any more. The relay is the 1st to go only a couple of years back. This year the transformer block went South just because one of the plastic socket melted. Part numbers (and replacement part numbers) are no longer applicable but I found one replacement that looked exactly the same on eBay. The wiring and spec are compatible.

Then the furnace is completely dead. I have an electronic (intermittent) spark ignition furnace, the sort that you can hear tick, tick, tick. After avoiding the engine wiring harness diagram for my car for many years, suddenly it made sense to me. After staring at the furnace diagram for so many years, it suddenly make sense to me. I'm sort of professional but I don't know what a pilot relit looks like and what it means. Suddenly I can relate the blocks in the wiring diagram to the boxes in the furnace. The rest is simple.

I checked that the pilot relit has 24V to it when heating is demanded. Basically it's a low to high voltage transformer to create the spark. I immediately ordered one replacement because it is cheap.

Then I thought of lighting it manually until the replacement arrive from the wrong coast. You should NOT attempt this unless someone is in danger of dying from cold and the sort. There is little difference between ignition by a spark and by a BBQ lighter. BUT, the ignition happens inside a room with a door closed. With a lighter you have to open the door and get to within 1 feet of the burner. And also, the furnace cover is open. If anything happens, there's a huge difference. So make sure there's no gas accumulation. Follow manufacturer's procedure to shut down completely, including the gas line, let possible gas leak dissipate, set thermostat to minimum and follow procedure to restart. Set thermostat to above room temperature. Finally put the BBQ lighter tip to where the pilot is and ignite.

After that, I discovered a youtube video showing the same thing. I cannot agree more that you cannot leave the furnace alone after starting. You must shut it down also after desired temperature is reached. Otherwise, the relit will try to spark, the pilot gas will come on, and the whole control will be in some sort of cycle. The gas from the pilot should be adequately dissipate from the vent and that the wast of gas will be minimal compared with when the furnace is started.

After some work, I can see that the sparks aren't completely dead. That led me to wonder if the relit is the only possibility of fault. From the web I vaguely read about that the relit isn't probably earthed, and cleaning the igniter with pin, brush, or sand paper. Now I know that they are talking about, or they don't want to talk about too much just like replacing the spark igniter with a BBQ lighter. Here's what I do.

Of course you check if 24V AC reaches the relit terminals and they are not rusty. Check if the earth terminals have 0 ohm with earth, usually the metal chassis and bare metal gas pipes are earthed. Now the high voltage wire aren't soldered into the relit. You can just pull it out. You may want to cut off a little bit wire and push it back in to ensure better contact.

Inside the ignition hole there are three components in the assembly. The pilot is basically a tough gas tube with a funny shaped end to direct the flames. The flame sensor connects to the gas valve via a bare earth wire like connection. The flame sensor itself is like a thicker stiff earth wire.

The igniter is at the end of the high voltage wire, which is itself a piece of stiff wire. It is usually thin and bent to direct the spark. When it was sparking, I observed that the spark goes to either the pilot or the flame sensor. Indeed both of them are earthed. And my igniter is midway between the flame sensor and the pilot.

So the principle is like lightning find it's way to earth. So I think the cause of not sparking is that the high voltage conductor or the earth conductor are not conducting due to oxidation. If you restore them to shiny metal surface, they should spark if everything else are working properly.

Without unmounting anything, you can see all three tips at an angle from above. You can scratch them to expose a little bare metal using a long pointed object. You can use a long narrow file or sand paper rolled into a tube. If it sparks better, you could take out the assembly and restore the metal surfaces so it will last another 20 years.

To take the assembly out, the minimum is to take out the pilot gas connector. Taking out the whole gas line together with the valve controller saves you a lot of grief.

If scratching a little doesn't work. I thought of mounting an earthed metal tip beside the igniter as a quick fix. It could be a narrow piece of steel sheet metal that can bend toward the igniter tip.

Now for the upgrade. Now manufacturers are all using hot surface igniter, and still some direct spark igniter (that starts the main burners). My intermittent spark igniter (that start a pilot first) are long gone. There's really not much good reason other than that you don't like the tick tick tick any more. A spark igniter last 20 years while a hot surface igniter last a few years at most, though the newer materials last a bit longer.

May be you can't find parts any more. May be your igniter is so bad that you need to replace it. And of course the valve. If you can't find a valve with main and pilot control anymore, you have to replace one with only main valve. Then you need hot surface igniter for the main gas line directly.

It's conceptually simpler if you upgrade to a hot surface. You just need to find a HSI that can be mounted onto the metal sheet chassis and place it close to the pilot tube. The newer rounded type are more likely to fit. A longer one is surer.

Most of them use line voltage at 120V. You probably have that control signal in your system. But to make it simple, you can use the relit input voltage at 24V AC to control a 120V relay, that in turn drives the HSI. The whole thing cost may be $20 that is about the price of a relit.

You need a HSI that is fast acting like a spark igniter. The relay should match the output current of the HSI. One caution is that relays have different types. Solid state relays cannot handle even small motors though their current rating are not exceeded.

Anyway, my relit is arriving next week but I probably do not need that.  My furnace is working now after taking out the igniter assembly and sand paper all over.

After sanding off some oxidation, the relit sparks better but not consistence. I put in my new relit and it works perfectly. This is the 2nd time now. The components was new but in a box that may be 20 years old, from the original parts supplier and original parts number - a collectable. That's how I locate the part on eBay with my 20 year old user manual. Did someone got it from an estate sale of a furnace technician? Did someone retired and clear his stuff? Amazing!

"New" garage organization technique

 First, I didn't reorganize the garage because the stocks kept rising. I hope they stop so I can cash in to get brush nickel finish steel cabinets all around the garage walls.

Then I hope the stocks rise again so I have the money to reorganize the garage properly.

Next I am not sure if I still want the house so it was fruitless to revamp the garage.

Finally, I learned to work with what I have, including the wooden shelves and on-wall planks that the previous owner or his tenants built many years ago. I also have many things brought over the years but never put them on, including two pegboards.

I never use the pegboards much. Say if I took five tools out for the weekend, I wouldn't have remembered where and how to put them back in the same place. I know the usual technique is to draw an outline of the tool on the pegboard but that's too inflexible. I fix sprinklers, do oil change and customize curtains. I need to be able to organize as I go along.

My garage is in a mess because without a plan, I am afraid if I put something out of sight, I would not be able to recall where it is.

In the end, I think I nailed it with the help of computers. Computerization is one big possible solution since the 60's, but the difference is, nowadays it's so easy to take pictures. You have the phone all the time and it's linked to the cloud. The cheapest smart phone will give you enough resolution for documents without scanning, just what spies do.

It's not just building a database. It's like unifying the paradigms, virtually and in real life.

So my pegboard isn't just pegboard, but pegboard with icons! I bet nobody posted those pictures yet. Each icon has a pinched hole for hanging on the pegs. No glue and totally flexible.

For power tools, they are on a horizontal "pegboard". I borrowed the idea elsewhere that is like a long box with partitions. Since I have wall shelves, I immediately got the top and bottom of the box without doing anything. I don't need no partition as they are just for the cords.

I was thinking of many ways to fix the tools in position. Metal brackets are too expensive at over $2 a piece. Can use PVC but I need one PVC plug and coupling for each vertical fat peg. I don't do woodwork since I do not have the space to put up with a table saw. Then it dawned on me that I can simply use long nails! In the picture you can see 4 nails for each tool driving into the wooden shelf board at an angle. The icons have punch holes that hangs on the nails.

It is surprisingly flexible. If I want to change position, I can just pull out the nails. Come to think of it, why we need pegboards in the 1st place? Can we use nails instead of pegs?

Crius cross, Crius square

See what I've done? For PVC structures the Tee is the expensive part and the most restrictive part. Now I can get rid of it in many circumstances.

Say if you want to pack a lot of pipes close together to form a raft like structure. It's not possible with Tee's because of the minimum distance between two pipes. Now you can pack two pipes side by side close to about two PVC pipe wall thickness.

Say if you build a pet cage out of pure PVC pipes. Now you just need a outer frame with rigid Tee's. The "walls" of the cage can be lined with pipes secured by Crius crosses.

Unlike Tee's, now it is not limited to PVC. You can use EMT or PEX for environmental friendliness. And since they are designed to be incompatible with different diameters, so for us there's always something to clip it secure.

Sometimes one clip is not secure enough for the pipe diameter, especially using the Crius cross. You can use concentric Crius clips - the Crius square.

By varying the arc length, it can be very hard to take off the clip unless slipping it sideways. On the other hand, using two clips side by side make it harder to slip sideways.

And if you apply some glue, it's not possible to take them out in many cases.

The problem with professionals writing DIY articles

Looking professional

The article looks good, professional, and the article is written by a different professional writers too. As a result, mostly likely it doesn't accept comments. Any comments would make the article seems unworkable, too difficult for DIYers. So here I am writing this to response to all those who don't accept comments.

Instead of taking time to explain to inexperienced DIYers that have a job to do, these articles have to look professional with a fluent flow of words. Who cares? Under the disguise of plain everyday English, these are concise well written instructions suitable for technical journals and trade magazines. Who cares when I just brought a few dollars of screws and want the job done yesterday?

Too often, the final writer doesn't understand a word what the professional is saying, and sacrifice the content to give the article a professional look - that's how they make money. This is true about all the instruction manuals in the world. The Europeans got rid of words in manuals. The Chinese use their kids to write something that no one will understand anyway, since you ask for it.

Jargons

How many knows what is a door jamb? Do I have to look at the dictionary or Wikipedia or Google every time I came across a term? I know you have a term for everything. But that comes with experience. Since I don't have the money to hire you, can I skip the experience? Is the word jamb necessary? Do I have to learn every part of the door to fix it?

Unnecessary jargons should be avoided. Also a standalone diagram will be great. Some articles seem to be written on the dawn of the internet. They try to save bandwidth by not having diagrams to explain simple things. A pic worth a thousand words. And they may point you to some encyclopedia for the diagrams that's good if you want to get into the trade. I don't.

Two-dimensional thinking

Come on, it's a three-dimensional world. You may have a preferred plane of reference in your trade. But to us mere mortals the world is 3D. What is the center? Of what?

Thinking inside a box

All articles of fixing a sagging door seems to come from the same source. It recommends replacing short screws with 3" ones. It's impossible to screw it in without a pilot hole, I tried. But then you can't find a 3" drill for screws that size. Not in the chain hardware stores. What a waste of time!

That's all for now ...

Tapioca starch miracle: deodorizer , shampoo and body wash

Two to four teaspoons of tapioca starch per cup of water. Pour on dry hair and this alone is a deodorizing shampoo. Massage evenly into hair. Rinse thoroughly after a minute.

I have oily hair. Even if I wash with pure Castile soap and leather twice daily I still smell. If I'm not careful, it's not smell, but horrible smell soon after shampoo. The kind that people accidentally came close within a feet will bounce back and cover their noses. But after I got the smell, a tapioca starch shampoo alone can remove the smell immediately. It's not a cover up. The hair 'smell' better into the day when the very mild starch odour disappears.

I have tried baking soda. A lot of it in a paste enough to pour on hair and massage into it. To my surprise it doesn't work.

My theory is that the root/scalp has oil that repulse water. If you apply soap/detergent, it's like washing a grease filter with soap/detergent. It won't work as detergent works on the surface and cannot handle bulk grease. Baking soda can deal with grease but it doesn't work on scalp. Maybe dry or wet, the coarse particles cannot reach the scalp. All shampoos have the formula to reach the scalp but then they are too strong and strip away oil on the shaft.

Of course tapioca starch (and similar starch) used in cooking is an oil emulsifier. It doesn't dissolve in water nor thicken cold water so water is a good carrier of it. Shake well so the starch is evenly in the water bottle. Just pour and the starch will reach the scalp.

You can use dry tapioca starch but I doubt it will reach the scalp as efficiently as a water mixture.

Now I'm using dilute castile soap and tapioca starch mixture. They don't seem to cancel each other. With soap it's easier to tell if the mixture reached your scalp. It seems the mixture stays on the scalp longer and it is easier to massage into it. The soap also smell nice, with moisturizer and acidic balance.

My hair can take straight liquid Castile soap (or solid bar), which is already mild. Diluted is enough as long as I don't smell. With tapioca I think I don't. I have little helpers to check it. It seems that I can use less tapioca with soap in it.

Wet shampoo isn't a bad thing because that's how to rehydrate skin and scalp. But you need moisturizer to go with it. Castile soap have lots of it. If you don't want soap with tapioca starch, you can add your own ingredient.  But be careful, I tried to thicken the starch water mixture with Guar gum but it's a disaster.

If you search for tapioca, you can find the better home made articles for personal care. It is used in many commercial things like dry shampoo, deodorizer. 

For my soap tapioca mixture, it's a deodorizing shampoo and body wash all in one, how convenient. It's also a 'conditioner' too as you don't need one. It's almost as easy to use as a bottle of concentrated shampoo, but preparing a bottle for each kid is a bit of work. I'm trying to figure that out.