There is an article here in which greydog describes how to use a Roland TD-3 to extend the voices for the HD-1. You run a MIDI cable from the output of the HD-1 to the input of the TD-3, after which you have access to all of the voices and kits in the TD-3, including reassignment of voices, changing relative loudness of each voice, and stereo voice panning. This works well, but you are still stuck with the HD-1 kick pedal.

The video here, towards the end, shows how to modify the springs in the HD-1 kick pedal to stiffen it up, but that did not seem like it would be enough. I build a wooden contraption to simulate the kick drum, and attached a regular kick pedal to the “drum”:

The kick input to the HD-1 does not accept a piezo signal: it will not trigger with that kind of input.  I discovered, by accident, that it will trigger if you just short the two signal lines together.  So, my original plan was to put a momentary contact push button switch where the beater hits the wooden panel, countersunk so that the beater just barely “clicked” the switch, and trigger the HD-1′s kick input that way.  I came across a free practice pad that had been converted into a DIY drum pad, and mounted that pad where I had intended to mount the switch. The only problem was how to get that piezo signal to trigger the kick drum.  By connecting the new DIY kick drum sensor directly into the kick input of the TD-3, I was able to get the best of both worlds.  I now have all of the HD-1 drums triggering the TD-3 sounds, and also have the new kick pedal triggering the TD-3 kick sound.  As a side benefit, I actually have two kick drums now, one on the HD-1 and one on the TD-3.

It turned out that tying all of those knots in exactly the right places was a frustrating process, so I tried to think of other ways to keep the ropes in place.  While perusing the web, I encountered some designs that used beads rather than knots. Having no beads of that size, we decided to try using washers, instead.  The result is a plant hanger that can be made in under 10 minutes.

To start out, cut four pieces of rope 100 inches each.  Fold them in the middle, to get 8 strands of rope about 50 inches long.  Tie an overhand knot about 3 inches from the top, where the ropes loop back on themselves.  We did not need to, but you could trap a steel ring in the top loop if you wish. This is what the top should look like:

You might think that a simple knot like this could shift, but once you pull each strand tight, it should be solid.

Insert washers from the bottom, trapping pairs of ropes.  The trick is to use two layers of washers, one of which traps pairs of adjacent ropes, with the next layer trapping pairs from adjacent pairs.  It is easier to show you than to explain it:

Note that one of the final washers has to trap one rope from the top and one from the bottom, to turn the whole thing into a cylinder.  This is what it should look like just before you tie a second overhand knot at the bottom:

The next part is tricky.  Put a bowl on the table, place the bottom knot in the bowl, and then place the pot on top of the bowl.  The bowl provided space for the bottom knot so that the pot does not fall over while you are working on getting the ropes arranged around the pot.

Gently lift the top knot to put a little tension on the ropes. You need to start out with all of the washers at the top, and only lower the first layer of washers down until they rest on the top of the rim of the pot.  Make sure that the washers are spread out evenly around the pot.  Then lower the top layer of washers.  These put some side tension on the ropes, pulling them so that the ropes that are holding the pot cannot slide sideways and drop the pot. It takes some fiddling, but after a few minutes, you should end up with something like this:

This design uses ropes in dynamic tension: once you have configured the ropes properly, you cannot let go of the ropes, or you will have to arrange them again.  Be prepared to hang the pot from some temporary hanger if necessary.

Here are some shots of two of the pots after we finished hanging them:

We like the flexibility of being able to chose the rope color and the hanger dimensions, as well as knowing that we’re only 10 minutes away from making another one, should we need one. Note that you can try it with 6 or 10 ropes instead of 8, if you wish.

The original design employs a complex injection molded plastic part that was beyond my capabilities.  I considered attempting a 3D design for the part, and fabricating it at the Vermont FabLab with a 3D printer, but decided to go low tech at first.

I purchased a hole saw and used that to cut disks out of a sheet of 1/4 inch plastic.  I glued magnets to the circumference of the disks and pressed a brass tube through the center of the disks for horizontal stabilization.  I epoxied a brass rod to a ring magnet and dropped a similar ring magnet onto the rod, oriented so that the two magnets would repel.  I then dropped the plastic disks onto the rods, so that the opposing magnets would serve as a magnetic bearing, as discussed in the patent.

This picture shows the component parts of an individual magnetic spin toy.  The brass rod is shown in the lower left.  It is epoxied to the ring magnet at the bottom of the rod; the magnet in the middle of the rod is held suspended by the opposing magnetic force.  The plastic disk is on the right, with the three magnets and the brass tube:

When assembled, they look like this:

I use a steel sheet (actually, a spare music stand) as a base for the magnetic spin toys.  This allows me to reposition them quickly and easily.

You can see videos of all of this as follows:

Assembly of components and demonstration of magnetic bearing and spinning

Two magnetic spin toys interacting

Three magnetic spin toys interacting

Four magnetic spin toys interacting

Four magnetic spin toys interacting in a square array

Four magnetic spin toys interacting in a rhomboidal array

Demonstration of magnetic interaction with distant magnetic spin toys

You might find it interesting to download the videos (I use DownloadHelper in FireFox) and step through the videos one frame at a time (I use KMPlayer) to watch how the disks respond to each other.  I will probably color code the magnets for future videos.

My original thought was to create an array of interacting spinners, one of which could be driven by a hidden microprocessor-controlled motor.  The microprocessor would be set up to drive the motor in varying patterns to ensure that the array never stopped.  I considered linear or curved lines of spinners, perhaps attached to loops.  It turns out that arranging the spinners in tightly packed arrays also works well.  The patent demonstrates that the axes of the spinners do not need to be parallel: you can arrange for the spinners to be at angles to each other, so it would be possible to arrange them on the surface of a hemisphere.  My use of a steel surface for testing could be expanded into using a curved surface for the “final” system.  So.  Lots of things to explore.

I looked around for heated towel racks, but they are either ridiculously expensive, cheaply made, or of the wrong size for my particular room.  I explored heating technologies, and found a product that is intended to be used to heat the bottoms of reptile tanks.  It comes in a 11 inch wide roll of heavy-duty plastic, with a power buss along each edge.  When you attach 110 V A/C (wall voltage) to the two power busses, the heating resistors in the sheet produce 20 watts per linear foot.  You can cut the roll to any length you wish (I used 2 foot lengths) to get as much or as little heat as you require.  The material looks like this:

You can find out more about this product, and purchase it from, www.reptilebasics.com.

I decided to buy some pieces of glass, mount (tape) the plastic material to the glass, and then drape the towels over the glass.  I cut some slots in some wood and dropped the glass into the slots.  The first prototype looked like this:

This shows how the wood was slotted:

This shows the bottom.  I intended to have the two towel warmers side by side, so I cut holes to allow air to circulate between the two towels.  At this point, only one heating plate was installed:

This shows the first prototype with the towel removed:

You can see the two 2-foot long pieces of heating material, taped to the glass with aluminum tape; two other pieces of heating material were used on the other side. The blue masking tape was used to reduce the chance that someone would accidentally brush up against the 110 V A/C connections.

We used this for a week or two. The heat produced was great for drying the towels, but this particular design has a number of problems.  For one thing, the two heating plates would have been too close to easily allow the towels to be draped into the space between the plates, and expanding that space made the unit take up too much floor space.  For another, it consumed more power than I wanted to use continuously (320 watts).  I considered about 5 different alternate designs, and chose one with the second towel directly above the first. This is how it looks, installed:

Notice the space between the towels, to allow access to the wall switches.  This is one reason that I decided to build a custom system. This is what it looks like with the towels removed:

I solved the power consumption problem by purchasing an ELK-960 delay timer, and modifying it so that the maximum time increased from 1 hour to 3 hours.  When you push a button, the timer and heaters come on for 3 hours, and then shut off.   You can see the button and activity LED here:

The bulk of the wiring is hidden behind the bottom panel.  You can see the 12 V power supply for the ELK board on the right, with the ELK board in the middle:

At some point, I may replace the wire nuts with soldered joints, and then enclose the solder in silicone caulk.  Not up to code, but safe enough for careful use by two adults. I attached the top of the rack to the wall with two lengths of chain, to ensure that it does not topple over.

When immature, they have the consistency of a “normal” mushroom, and can be eaten; when ripe, they turn into a regular (albeit huge) puff ball.  The white one weighed 4.25 pounds; the dark one weighted 1.75 pounds.  Both had a leathery skin, and an inside that was more or less homogeneous.

Unfortunately, although the white one was white on the outside, the inside had already begun to ripen, and thus it was not edible.  We placed the mushrooms in likely spots, and hope to have our own next year, and to be able to harvest them in a more timely fashion.

To put this into perspective, the largest one ever found weighed 44 pounds!

A) a mini uninterrupted power supply, consisting of a wall-wart (not in picture) to charge a small NiMh battery.  The power to charge the battery comes in on the black USB cable, and power is supplied to the Arduino via the white cable.  This allows the Arduino to run without interruption during brief power outages (up to a few hours)

B) the Arduino computer/controller, mounted on a clear plastic sheet.  This controls the timing and intensity of the LED lights

C) the motor/power controller.  This converts the control signals from the Arduino so that the LEDs can be controlled properly.

D) the power supply.  I’m using 11 V at the moment, although the LEDs are designed for 12 V.

Here is one of the tanks, illuminated by the LED lights.  Notice, along the top, how there are many bright spots, rather than uniform illumination.

Here is a shot of the lighting system before it was installed, looking at it from “underneath”.  There are 5 1o-watt LEDs, each of which is mounted on a heat sink.  A second system, for a different tank, has 10 LEDs.

Here is a close-up of the lights installed above one of my tanks

The Arduino code counts out 24 hours, during which there is a 1 hour “dawn”, an 11 hour “day”, and a 1 hour “dusk”.  For those of you who are interested, here is the code:

————————
// these are times in seconds
unsigned long dawnLength = 3600; // 60 * 60; // 1 hour dawn
unsigned long daylightLength = 36000; // 10 * 60 * 60; // 10 hour day (was 4294937760
unsigned long duskLength = 3600; // 60 * 60; // 1 hour dusk
unsigned long dayLength = 86400; // 24 * 60 * 60;
unsigned long dawnStart = 0;
unsigned long dayStart = dawnStart + dawnLength;
unsigned long duskStart = dayStart + daylightLength;
unsigned long nightStart = duskStart + duskLength;
// brightness levels
unsigned long moonLight = 1;
unsigned long sunLight = 200; // LEDs are running too hot: try to keep them cool
// time rate
unsigned long oneSec = 1000; // milli seconds: set nightStart) {
      v = moonLight;
      x = 4;
      }
   else if (secs > duskStart) {
      v = map(secs – duskStart, 0, duskLength, sunLight, moonLight);
      x = 3;
      }
   else if (secs > dayStart) {
      v = sunLight;
      x = 2;
      }
   else if (secs > dawnStart) {
      v = map(secs – dawnStart, 0, dawnLength, moonLight, sunLight);
      x = 1;
      }
   else {
      v = moonLight;
      x = 0;
      }
   if (testMode) {
      int xxx = secs % 4; // mod
      xxx = (xxx * 2) + 2;
      analogWrite(pwm, xxx);
      }
   else {
      analogWrite(pwm, v);
      }
   if (useSerial) {
      Serial.print(x, DEC);
      Serial.print(” Secs: “);
      Serial.print(secs, DEC);
      Serial.print(‘ ‘);
      Serial.println(v, DEC);
      }
   }

void loop () {
   currMS = millis();
   if (currMS = oneSec) {
      // one second just expired
      secs = secs + 1;
      prevMS = currMS;
      SetLights();
      }
   // day wrap-around
   if (secs >= dayLength) {
      secs = 0;
      SetLights();
      }
   }

While at BestBuy, I considered upgrading to the latest in the Canon line, the SX 30, but the morons at Canon made the display (in the view finder) about 1/4 the size of that in the SX 10.  How can you take a good product and intentionally make it worse?  I called Canon, and pointed this out, and the support guy basically said “Oh.  Yeah.  I see what you mean.  Have a nice day.”   I decided I was done with that line of cameras.

While the Canon has been great in the super-zoom department (25 mm to over 500 mm effective range), takes OK movies, and can take closeup pictures to within 1 cm of the front of the lens, it does not shine when it comes to low light situations (pun intended).  If you try to push the sensor sensitivity (ISO) past about 100, you usually end up with so much colorful noise in your picture that it is not worth showing anyone.  So, hand-held non-flash pictures are impossible at night with that camera.

A friend of mine, Ray, suggested a number of alternatives, but I’m fairly picky.  I want an articulating LCD screen on the back of the camera (most digital SLRs do not have this feature), did not want to spend $5K or $10K to get a decent “system”, and wanted an Electronic View Finder, a feature that most DSLRs have discarded for reasons that I cannot fathom.

There are a few very small and light high-end cameras out there from Sony and Olympus, but many of them lack the EVF feature. But once I read about the Panasonic Lumix GH1 and GH2, I was intrigued.  Both are part of the Micro Four Thirds camera series that many manufacturers (Panasonic, Olympus, Sony) are supporting.  Micro Four Thirds lenses give you twice the “zoom” of 35mm lenses, so a 100 mm Micro Four Thirds lens behaves the same as a 200 mm lens on a 35 mm camera.

The GH2 is virtually unavailable at this time, and sells at a huge premium (camera body and lens for $1,800 at the moment) over the GH1 (I bought a body and lens for $1,000).  I now have the body and an effective 28 mm – 280 mm lens, a 300 mm – 600 mm lens, and a 90 mm large aperture lens that I mount using an adapter.

The EVF is wonderful: bright and detailed (at over 1 MPixel).  The articulating LCD is perfect (although the GH2 adds touch-sensitivity).  The camera system focuses very rapidly, and can continue to auto focus even while shooting an HD movie.  The two options of HD movie formats each have problems (the AVCHD format is very difficult to edit, and the motion JPG format is limited to 8 minute clips), but the quality is quite good.  The 28-280 lens is quite convenient (although I miss the macro ability that the Canon has), and the 300-600 gives me enough “reach” to take decent nature photographs.  This camera has two shutter buttons, one for stills, one for movies, just as the Canon does, but unlike the Canon, it cannot take a still while taking a movie.

You can see a frame clipped from an HD movie here: http://www.jonbondy.com/P1010109.JPG

And here is a shot of a squirrel: zoom in and look at the detail in the fur:  http://www.jonbondy.com/P1010318.JPG

I’ve written the Arduino software to control the lights, so that they ramp up slowly at “dawn” and back down at “dusk”, but have not added that feature yet.  I should have that done by the end of April.  I also have the LED lighting system set up for installation under the cabinets in my kitchen, but need one full day when I can finish the wiring, since I do not want to get caught with that project half done.

I saw some reviews of the ViewSonic gTablet.  Most of the reviews were fairly negative, but one claimed that the problems had been fixed this spring, and I figured it might be fun to hack the thing, since there were lots of articles about that process on the web.  At $300, it was a $200 savings from the cheapest iPad. All of my friends who have ‘Droid phones are pleased with them, and the gTablet uses the Android OS.  The Internet is always there, and in a moment of weakness, I ordered a gTablet, knowing it would be larger and heavier than the iPad. I bought from Buy.Com (but it was being sold by Tiger Direct); I could have bought from Amazon (but sold by someone else).  Nothing is as it seems.

If you had been sitting with me during the first 30 minutes after I received the tablet, you would have seen me laughing out loud as every application (and I mean every application, including the web browser) crashed.  I was about to return it when I turned it off and on; I could have done a factory reset, too, but didn’t have to.

After the power cycle, it behave properly.  First of all, it asked me the series of questions that it should have asked when I first turned it on (accept EULA, select time zone, etc), so it must have powered up strangely the first time.  Secondly, there were fewer apps visible, as if it had discarded some of the original settings.  But it started to run without crashing.  At this point, I have no desire or plans to load new firmware into the ROMs.

I’ve spent a few hours learning about the Android OS (it is pretty difficult to do some things elegantly when you don’t have a real mouse), learning how to find and install apps, and learning how to acquire eBooks, and convert PDFs to eBooks.  I installed a few free games, but they were either fairly lame/boring, or crashed immediately.

First off, there is no hope of reading a PDF on the tablet.  Adobe’s PDF reader is stunningly brain dead, for a device that is so small.  While you can zoom in, this forces you to manually pan across and down each page, which is tiring and unnecessary.  Score zero for the folks at Adobe, who should have known better.

I also installed a few free eBook reader apps, but they also uniformly were unable to handle the scaling and zooming issues.  Amazing and ridiculous.

In the end, I installed the Kindle reader app (which is fine) and purchased a book from Amazon for $2.50 to get started.  That worked quickly and easily, and reading the book is a pleasure.   Finally, someone who knows how to do something right.

Getting PDFs onto the gTable/Kindle took some more work.  I had to install MobiPocket Creator to convert the PDFs into eBook format, and then move the PRC files over to the Kindle folder on the gTablet, but that was relatively painless.  I now have a series of PDFs available to read, including a free eBook I found in PDF format, and some notes about the impending technical singularity, which I will bring to a discussion this afternoon.  So, it may end up that the tablet is useful after all.  I can also use it to look up info using WiFi during the discussion.

I’m quite pleased with the gTablet, but so far am only using it as an eBook reader and to look stuff up on the web.  Perhaps I will end up using it for other purposes, but the awkward data entry UI makes that seem unlikely.  I’m sure that the iPad is better in a dozen ways, but this was a reasonable way for me to get started with this technology.