HESA Works!

I was working on the Home Environmental Sensor Array at the Milwaukee Makerspace today.  (I took the day off so I was able to make some good progress.)

I had an idea to use an old computer power supply as a project box for the HESA.  It makes sense because I need an enclosure to put the parts in, 120v power, an on/off switch, and a fan.  The power supply box has all of that.  I hope to install all of the parts inside the power supply box and then mount it to my wall.  This includes the PowerSwitch Tail relay, the raspberry pi, and the circuit board.

After taking of the electrical components out of the power supply, I wanted a rectangular piece of wood to attach to the box where the old circuit board was.  This would allow me to attach parts to the wood and not the metal.  I also wanted to use the CNC router to cut the wood just so I could use the router again.  I had to design my own DWG file.  The tool I used to design the wood piece was LibreCAD.  LibreCAD is free and runs on Linux.

The wood piece is a simple rectangle with four holes to attach it to the power supply.  I spent about 2 1/2 hours designing the piece and then cutting it with the CNC router.  It takes a lot of time to measure everything precisely and then design it in the CAD software.  It does not help that I have not used LibreCAD and have to try and retry everything many times.

All that being said, the results were perfect.  The nice thing is that I can re-cut as many of these as I want with the exact same specifications.  In fact, I did cut six more before I left just because I could.

After cutting the wood and mounting it in the power supply, I also installed the PowerSwitch tail in the power supply box.  When I turn the switch on for the power supply, the PST has power as well.  Cool!  Next, I need to figure out how to wire in my raspberry pi to power as well.

The fan that came with the power supply did not work so I replaced it with a different one from the parts rack at the Makerspace.  The fan says it requires 12v input but it will work down to 3v.  My raspberry pi has a 5v output so it does power the fan, albeit a bit slower than it could run.

Now that the enclosure is working a bit I started testing out the new circuit.  First, I wired up the A and B circuits.  Each circuit has two outputs and one input.  One of the outputs sends a signal through a 1k resistor and the other sends a signal through a 100k resistor (not at the same time).  The way it work is like this.  The program on the raspberry pi sends out a signal on circuit A's 1k section.  The input on circuit B looks for a signal.  Then, the pi turns off the first signal and sends another on circuit A's 100k section.  If circuit B does not see that signal, the pi shuts off the signal and tries the same tests sending on circuit B and listening on circuit A.

For whatever reason, this circuit design works perfectly.  The pi always detects a signal when it is sent over the part with the 1k resistor.  I never got any false readings.  I think the problem is solved.

Here is a link to the circuit schematic.  I made it in Digikey's free SchemeIt on-line tool.  By the way, I have tried a few of schematic designers.  Most of them are either not free or too complicated.  I just want to make simple schematics for electrical circuits.  Circuit Lab was nice but is no longer free.  I installed the Electric VLSI Design System on my Linux computer but it is too complex.  SchemeIt was pretty easy to use, and allowed me to easily create my schematic.  I can share the link on the internet and even save it to PDF if I want.  For now, I will stick with SchemeIt.

I also made some videos of the HESA in action that I will upload to YouTube and post a link to.

Another great day at the Makerspace.

Pitcher Holder "Fixed"

The pitcher holder piece that I 3D printed for the automated pet food dispenser does not fit correctly on to the PVC T.  The inside diameter of the pitcher piece is just a bit too small.

I used a drill press and various tools at the Milwaukee Makerspace to rough out a bit of the inside diameter of the part.  I tried different tools (see picture) starting with a little Dremel sander (not pictured), then a 1/4" drill bit, a bigger gear-looking bit, and finally, a wire brush.  The sander and wire brush both worked but are incredibly slow.  The drill bit did not work at all.  The gear thing was too coarse.  It kept gouging out parts of the plastic.

Eventually, I did shave off enough plastic to squeeze the holder onto the PVC.  Unfortunately, I shaved off enough plastic to get to the inlay and crack the piece.  You can see the crack in the picture of the piece.

I will see what Tim thinks but we will probably push on and try to finish assembly with what we have as a beta.  Then, I can re-measure everything and try to re-size and re-print the auger and pitcher holder to fit the PVC dimensions.

I think that will be after I get the HESA project done.

Pet Feeder Assembly Started

Tim and I started to assemble the automated pet feeder parts.

Actually, before we started, I went to Home Depot and bought a 2" circular wire brush that can be inserted into a drill.  Tim and I spent some time using it to smooth out the inside of the PVC T.  After many minutes with the drill, most of the raised part is gone but there were still some higher spots.

We then tried shaving off some of the other auger piece to make the circumference smaller.  That did not turn out so well.  During the process, a bigger part of one edge was cut off.  Then, one of the parts near one end broke off.  Even though it seemed like we cut off quite a bit of the auger, it still sticks a bit in some spots.

We did start assembling the pieces to see how it looks and to feel like we are making progress.  See the pictures.  A pitcher to hold the dog food would go on top of the blue part (that does not fit on the PVC either).  The servo would be attached to the white piece that is attached to the wood support.

I think that I will continue to try smoothing out the inside of the T.  Also, I have an extra auger that was not printed perfectly.  I think I will use that one instead of the broken one.  Hopefully, it will fit in the T if I can get it smoothed down.

Another idea would be to use a plunger instead of an auger.  Maybe that would be more reliable.  I'll have to see if I can find a device to push a cylinder through the PVC.

Finally, now that I've got some experience with this, I would do a few things differently.  First, I would purchase the PVC T and figure out the servo and controller before starting to make anything.  That way, I could measure the PVC and possibly adjust some of the prints and cuts.  Also, knowing what the servo would look like would help to know what kind of end piece would be needed to attach the servo to the wood support.

PVC T for Pet Food Dispenser

Over the weekend, I bought a 2" PVC T and the hardware to assemble the pet food dispenser.  When I got home, I tried inserting the PVC into the wood supports that I cut on Friday.  The hole was just a bit too small.  My theory is that we should have made the cut for the hole right on the line instead of on the inside of the line.

Next, I tried putting the augur in the T.  It went in the first few inches just fine.  Then, it hit the raised part that is made to step the PVC that is inserted into the T.  Doh!  The augur was sized for a 2" diameter but the inside of the PVC is about 1 3/4" with the raised part.  Strike two.

First, I went to Home Depot and bought a different PVC T.  This one still had a raised part inside but it was about half as thick as the other T.  I was hoping it would be easier to make it work.

Then, Tim and I stopped off at the Makerspace to fix the two problems.  We put an abrasive disc Dremel part into a drill press and used that to make the holes on the wood bigger.  It took a few passes to get it right.  Now, the PVC T fits snugly into the wood supports.

Once that was done, we turned our attention to the augur.  At first, we tried getting rid of the raised part on the inside of the T.  We used a hand file and spent a good bit of time with that.  It did make some difference but not enough.  One problem is that the file was flat and the inside of the PVC is round so there was not enough surface area contact.

Finally, we used a band sander to remove a layer off of one of the auger halfs.  That was very ugly but it seemed to work.  I only brought one auger half so we could not finish the job.

After we left, I thought that we should have used a circular wire brush attachment for a drill.  I did not see one at the space.  Maybe I will go to Home Depot and buy one and then use it to smooth out the inside of the T.  I might also have to re-print the augur parts if they will then fit on the inside of the T.

I will chalk this up to a learning experience.

First CNC Router Cut

I spent the evening at the Milwaukee Makerspace.  Tonight I learned how to use the CNC router.  Brant, the President of the 'space spent over an hour training me.  (Thanks Brant!)

I used the CNC router to cut the wood support pieces for the automated pet food dispenser.  The very first thing you need to cut something on a CNC router is a drawing.  The thing you want to make is designed in a 2D software application like SolidWorks, AutoCAD, or Google Sketch.  The application needs to output the results in STL format.  Fortunately, I was able to get the STL file for the wood supports from the Instructables page.

The next step is to load the STL file into a program called CamBam.  CamBam takes CAD drawings and converts them into a script (called G Code) that the CNC software can understand.  Brant and I played around a bit with the settings.  At first, CamBam was confused about how to cut the outside of the piece.  We had to combine all of the individual line segments that made up the outer edge into one long line to make it easier to follow.  Brant spent a lot of time showing me all of the settings in the software.  We got everything working so the machine would cut the hole in the middle, make small indentations for the screw holes, and cut out the entire piece.  The last step was to save all of the commands required to cut the piece as a G Code file.

The G Code text file was loaded into a program called Mach3.  Mach3 is the software that actually controls the CNC router.  It processes the G Code line by line, sends signals to the CNC router and processes the results.  We also had a small amount of setup to do in Mach3.  Basically, we had to center the router in the correct position and height to do the cut completely within the piece of wood.  That only took a few minutes and we were underway.

We programmed the CNC machine to make three passes of .25 inches each.  In hindsight, maybe it should have made more passes.  That might have kept the wood from getting torn up.  The total cut time was only about three minutes per piece.  After the first cut was done, we moved the bed, re-zeroed the router, and cut the second piece.

Here is a link to a short (and somewhat boring) video showing the first cut on the CNC router.  As you will see, we had to hold down the piece when it made the indentations for the small holes.  We fixed that on the second cut by adding more holding tabs and also cut the outline of the entire piece last.

After cutting both pieces, I sanded them and then drilled the small holes for the screws.

I have finished making all of the pieces for the dispenser.  Now, Tim and I need to put it together.

Another great day at the Makerspace.

3D Printing for Dispenser and Getting Back to HESA

Last night, I stopped off at the Milwaukee Makerspace to do a couple of things.

The first thing I did is 3D print the pieces that connect to the wood supports on the automated dog feed dispenser that I am making.  The 3D printer initially had PLA filament in it.  I tried printing with the PLA and it did not work well at first.  The raft printed ok but after that, no plastic was getting laid down.  I cancelled the print, removed the PLA and loaded some ABS filament that I had used before.  I was then able to print the two support pieces.  These are the last two pieces that I needed to 3D print.  Next, I need to use the CNC router to cut the wood.

While I was waiting for the two pieces to print (just over two hours), I got the Home Environmental Sensor Array project off of the shelf and started working on it again.  It took me about thirty minutes to get back up to speed and remember how everything was connected.  Almost everything worked.  The one thing that is not working is the sensor part of the circuit.  As soon as the wire from the sensor is connected to the input GPIO on the raspberry pi, the program I wrote shuts off power as if it detected power coming in to the GPIO port.

I don't understand why this is happening.  Tom G. from the Makerspace suggested that I add a resistor into the circuit.  I did that but it did not help.  Maybe I wired it up wrong.

I also read an article about reading a sensor on the pi.  Its not exactly the same thing I am trying to do but maybe the same circuit would work for me.  It has a resistor and a capacitor.  Maybe waiting for the capacitor to charge is a better way to detect the closed circuit.

I will try to document the circuit I am using now and post it to the Makerspace user forums.  Maybe someone will know what I am doing wrong.

I also need to publish my python source code on the internet and add links to it on my project wiki page.

Even though I did not make much progress on the HESA, it was another great day at the Makerspace.

Printed Auger Part 2

I went to the Makerspace tonight to try printing more of the pet feeder pieces.  At first, I could not get the printer to print.  Nothing would stick to the surface.

I went to get something to eat and ended up at a diner across the street from the 'Space with two other Makers.  One of the guys, Ed, knows quite a bit about the Makerbot 3D printer.  When we got back to the 'Space, he analyzed the problem.  He took the left extruder apart and realized that there was a missing washer on the pin that pushed the filament to a gear.  The missing washer allowed the gear to slip on the filament.  We added one washer and tried another print.  It worked great.

Here is a picture of half of the auger that I printed.  I realized after it started that I already successfully printed this piece and did not need another one.  Oh well.

In other news, I brought a bent metal piece that attaches to my camper to the space.  Dan, the blacksmith, helped me straighten it out.  Cool!

Another great day at the Makerspace.