Cantenna: Part 1

Today I took a soup can and turned it into a cantenna. Once again here is the link to Gregory Rehm's excellent cantenna guide. I tested it by perching it out my office window, taped to a microphone stand. I don't have reference figures yet for the AP with it's original rubber omni-directional antenna, but I measured signal strength using my cell phone (A Oneplus One). According to the app I downloaded for the purpose, I'm getting -64 to -74 dBm across the furthest throw I can measure in the yard. Due to obstructions and where this window is located, I can only test about 300 feet. The can is still uncomfortably close to both the aluminum siding and a first floor roof section.

Additional pictures of this process can be found in the gallery

Cantenna: Part 0

Last night I ordered some parts to make a directional 2.4Ghz antenna.

  • 2x (only need 1) N-type bulkhead connectors
  • 1x N-type to RP-SMA cable, LMR200
  • 2x USB wifi radio
  • 1x 6" USB extension cord

Dosman and I are planning to create a point-to-point mesh connection between our houses. We haven't come up with a purpose for this network connection yet, we're working on the Field of Dreams principle (if you build it they will come). Hedgemage wants to make a shadow internet with lots of point-to-multipoint access. This isn't really feasable at this time for several reasons, but the biggest is simply we don't have access to place multipoint base stations where the public can access them. Another option is to create a redundant internet gateway so that we both have internet access even if one of ours goes down. At the moment we're just doing it to see if it works.

The antenna I'm going to build is a cantenna that is mounted on a disused DISH network reception dish. I'm going to remove the LNBF from the dish and place the cantenna in its location. Dosman has a 2.4 Ghz wire mesh antenna he's going to use on his end.

The radio I'm using is a Raspberry Pi 2 with 2 USB wifi trancievers attached. The raspi is running OpenWRT Chaos Calmer.

I've shamelessly stolen the design of the cantenna from this website.

Bloomington Pumpkin Launch 2016

Heath, Mouse, Daniel, Erin, and I decided to build a trebuchet this year for the Bloomington Pumpkin Launch. We didn't have the time or inclination to design and develop something ambitious, so instead we opted for simply copying someone else's work. Heath found an instructable of someone who'd made a floating-arm type machine in a technical school competition. One of the constraints of this competition was that the machine must fit in an 8 foot cube, not counting the throwing arm.

Obviously building a machine with this limitation in an unlimited class means we never had a chance of winning the distance contest, but it seemed fun anyway. Heath and Mouse did most of the carpentry, I did the welding, Erin and Daniel made the slings from nylon.

We were expecting throw distances of around 200 feet. The day of the contest however, it managed 368 feet. We were last in the distance contest, but set an all-time record in the accuracy portion of the competition.

I'll snap a picture of the trophy and post it here:

Some video of test throws. We were experimenting with different hook angles, weights, and sling lengths. These are not very long throws.

Welding Heath's alternator back together

Heath was removing the alternator in his car to change the brushes or something. Its ear broke off. I don't have before pics on this, it was a quick repair. I grabbed my Tig and zotted it back together. This is the sort of weld where a Miller Dynasty would shine. The concave areas are very challenging to do with conventional 50/50 ac balace and balled tungsten. Anyway, the welds are ugly, but it's back together and working.

Please stop hitting the guardrail!

The guard rail where my driveway meets the road has been under attack. The first occasion after we bought this house in March occured in May. Around midnight a Chevrolet plowed into it at 50 mph. The driver was extremely drunk. We had already gone to bed, and were only made aware of the situation after the first responders arrived (the car was equipped with OnStar). The child occupant was unconcious and thrown into the drainage ditch, the woman was unconcious and unresponsive, the male driver was found hiding from the police along the fence line, bleeding and with torso wounds. He was incoherent and made references to another adult who believed the car would explode. Because of the possibility of a third, and the fact that the police didn't seem to understand how to organize a search, the three of us (houseguest included) staged one. Ultimately there was no third. All three occupants survived. We hope charges were filed.

You can see the damage (and bits of Equinox) clearly in these photos. And then barely a month goes by and this happens:

A kid from the school up the road (and another occupant) drifted off the road at speed, hit the driveway, which lifted the front end and prevented handling or braking and plowed into the same section of guardrail. This time the driver was stone cold sober (and very disappointed in himself). Inattentively drifting off the road and then over-correcting is the most common cause of roll-overs.

Additional images in the gallery.


I have been turned on to a project called Neural Style. I'm terribly familiar with or interested in neural nets. There has been a lot of media exposure lately on the subject and I haven't read any of it. What I have done is stumble upon an amusing imgur post demonstrating what can be done by using a neural net to re-imagine an image (the content image) in the style of another source image (style image). The software wrapper to accomplish this is Neural Style. The software is easy enough to use; if not to install and configure. The software it depends on isn't exactly a prime application of Unix design principles. The various tools and libraries appear to be in a constant state of feature and API flux. But no matter. For quite a while I used only CPU power and system core memory to process these, since the video card I had at first has only 1GB of memory. OpenCL (and CUDA) support are available, but as of now can only address memory in the GPU. AMD has big talk and big plans with regards to their APU work that they will soon have a seamless integration between main and gpu memory for scientific computing purposes. We will be pleased if this bears fruit. Later renders are done with GPU acceleration since I bought an RX480 8GB video card.

Here is a selection of my favorites; they will be featured in the order of content image, result, style.

This is one of the first renders I did. Featured on the left is me sleeping on my cat, Dixie. For the style we have Edward Gorey.

Some stock images of cows and tile floors.

I like how this one turned out. The result basically what I was expected to happen.

This one always makes me giggle.

Steve Buscemi. 'Nuff said.

The old Indiana Theatre, now the Buskirk-Chumley combined with a screen grab from Jet Set Radio.

And the Kremlin recieves the same Jet Set treatment.

World of Warcraft fans will understand a Molten Corgi.

Manezhnaya Square

Rebuilding my A535 transmission

This Dakota has always had a loud gearbox, but lately it seemed to me to be louder than usual. I'd recently removed the carpet from the truck, which frankly was enough to cause the increase. However I was unsure if maybe the noise level was abnormally high due to bearing wear and I was only now aware of it because of the bare floorboards. I've never rebuilt a manual gearbox before; seems like a learning opportunity. It should be noted at this point, that I've always had trouble with downshifts into 2 and 1 on this truck, particularly in cold weather. This will come up later.

Like anyone who is jumping to conclusions without doing any proper diagnostics, I order a rebuild kit first. However, due to reasons involving price tags and sleep deprivation, I only order a bearing kit. Not in fact one that contains new blocking rings (synchros).

Here are a few pictures from the tear-down:

After disassembly and inspection, I conclude that the bearings look outstanding for their age. The blocking rings however are quite worn. In particular the reverse ring (which is plastic) is over 90% worn through. Reverse is not synchronized on this transmission, but neither is it a sliding-idler design. It's continuously meshed like the forward gears. First, second, and third have non-trivial wear on both the conical meshing surface and the tell-tale wear on the dog's teeth of poor engagement.

I also learn from some documentation I found that the wrong lubricant was in the gearbox. When my dad had bought the truck in 2003 it had gear oil (80 weight or so). As it happens, this transmission is designed to run on 10w30 synthetic engine oil. This was the cause of the poor downshifts. The heavier oil was much harder to scrape off the conical section of the gear (where the blocking ring pushes on it during a shift) and match speeds, resulting in forceful shifts being required. Of the various blocking rings, I discovered fifth and second were identical (though had differing amounts of wear). So I swapped them. The others I left where they were.

During the process of changing the bearings, I needed a hydraulic press. So I built one. Neal had been lamenting the lack of one and graciously supplied the steel and bottle jack. I grabbed some old grade-8 bolts and got to welding.

The press:

I changed all the bearings, unfortunately the snap rings included with the kit were either mismatches or of obviously inferior quality so I only ended up using one or two of them. I decided the synchros were good enough for a cheap truck and re-assembled the transmission. The end play on the layshaft didn't change and everything moved smoothly.

Overall the rebuild was a success. It sounds the same, but it shifts much more nicely. Second gear is much more pleasant, fifth a little less so. I would have like to have new synchros, but it's nice to learn just how much the difference in wear between 2 and 5 can affect shift performance. Downshifts have been improved across the range thanks to the change in lubricant.

This big snap ring was a major pain in the mctukhus:

My channel locks were just a hair too wide for the clearance. Much patience was required.

All the pictures I took as re-assembly notes can be found in the gallery.

Fabricating Neal's trailer

Neal asked me to help with the design and fabrication of a flat bed trailer. He'd already acquired plenty of steel for the purpose as well as old tread plates from a car carrier. These are very strong aluminum extrusions and we used them for the deck plating as well as the tailgate. The tailgate disassembles into 2 pieces to form ramps. One end of each ramp has a hook to catch on a lip on the rear of the load floor.

Here's some pictures of the layout:

And here it's mostly finished. The tailgate-and-ramp parts aren't designed or built yet.

The axle, wheels, tongue, toolbox, flower patterned panels, and load floor were all salvage. The lights, wiring, and angle steel were purchased new.

Electrolytic replacement

A large percentage of electronic failures can be traced back to the premature failure of electrolytic capacitors. Electrolytics should last a good long time, but for several reasons they often don't. One is simple cheapness, often circuits are made extra cheap and simply cycle them faster than they were ever intended to go. The device in question is a knockoff and never meant to last more than a few years. The other reason is often explained with an old wives' tale of industrial espionage. It goes something like this: A major Japanese manufacturer had their capacitor recipe stolen by an employee agent, Shenzhen starts making them to the specs, Foxconn and other big manufacturers switch to the cheaper Chinese ones, 2-8 years later all manner of electronics in every consumer market segment start failing left and right. And it's all because the sweet old couple that owned the Japanese cap company never wrote down the stabilizing ingredient formula and only treated the batches at night when nobody was around. Whether the story has any truth or not, undeniably tons of electronics made from about 2004-2012 failed prematurely. Here are a few images from a couple of repairs I did today. They are an LG (stands for Lucky Goldstar, remember those VCRS?) FLATRON L227WTG from 2008 and an Antec power suply from 2005. Sorry for the terrible quality of the pics, I only took them so I had a polarity and position reference to install the new caps.

Here you can see the caps visibly bulging:

They should be cylindrical and have flat (not convex) ends. Bulging and leaks are both sure signs of failure. Absence of these symptoms is by no means an indicator that they're ok. Proper cap testing is beyond the scope of this note though.

Installation of new caps is pretty straight-forward. Just pay attention to the polarity. There are several companies online that offer cap kits for specific models of computer displays, power supplies, dvd players, and other consumer electronics. The prices are in the 10-20 dollar range and are worth every penny. Both of these repairs were successful.

More images in this gallery.