Kontraptioneering!

Hey - Remember the prototype for these? Instagram post by...

Tue, 19 Dec 2017 14:13:35 -0800

Hey - Remember the prototype for these?

Instagram post by Hector Wilkinson • Dec 19, 2017 at 8:02pm UTC

http://www.kontraptionist.com/post/155640938774/cnc-milled-and-programmed-some-stuff-up-for-a

Two Bit Circus‏ @TwoBitCircus May 25“It’s a thin...

Wed, 31 May 2017 10:40:34 -0700

Two Bit Circus‏ @TwoBitCircus

May 25

“It’s a thin line between pair programming and back-seat driving.” #Multitasking Wanna join the carny family? http://twobitcircus.com/careers/ pic.twitter.com/6eqCzx4YQI

Today I did something I should have done long ago - Filmed the...

Wed, 12 Apr 2017 14:19:31 -0700

Today I did something I should have done long ago - Filmed the CNC with the FLIR camera! We have a FLIR E60 thermal camera. I am pretty impressed with how cool it looks, and there’s even a chance that we might learn something here. I’m cutting about 0.006″ off the surface of a piece of white acetal (Delrin ®©™) plastic with an indexable carbide fly cutter spinning at 2500 RPM. and about 25 inches per minute feed rate. I realize that this is probably pretty fast on the spindle and slow on the feed, but it makes for a great video. Sorry for the bad focus. You can really see how a lot of the heat ends up in the chips, and plenty ends up in the workpiece too, but since it’s plastic and a great insulator, the picture is a bit deceiving. Aluminum would quickly dissipate the heat of cutting in to the bulk of the material, but the Delrin lets it sit right at the surface. These are the blanks for winch parts for a simple Drawbot Enjoy!

We made a cloud that rains tequila at Two Bit! This apparently...

Thu, 30 Mar 2017 13:03:00 -0700

We made a cloud that rains tequila at Two Bit! This apparently is how the Mexican government thinks they’re going to convince Germans to vacation in Mexico. It looks good enough that it just might work. It uses ultrasonic atomizers bigger than anything you’ve ever seen - Like, the size of a cheesecake, and a COUPLE of them. The one in your home humidifier is about the size of a dime. We did several iterations of this, and lemme tell you… the shop smelled pretty boozy for a while, triggering mostly bad memories for our staff, and to make matters worse, it’s not like we used the fancy stuff for testing either.

See the cute article on IFLScience! (image above courtesy of)

There are many ways to cut metal. This way is called the Dry Cut...

Mon, 30 Jan 2017 16:03:49 -0800

There are many ways to cut metal. This way is called the Dry Cut Metal Saw which is exactly what it is. A big saw blade with carbide teeth that cuts through steel with no coolant or lubricant. It’s loud. Really loud, and it throws showers of hot burning metal shavings everywhere. It’s the less offensive sibling to abrasive saw cutting, although the metal shavings it makes are much larger, so they carry more heat, so they do a much better job of burning areas of carelessly exposed skin, and are also very effective at getting caught in socks and pockets. What it doesn’t do is fill the air with resin, aluminum oxide and fiberglass dust, which the abrasive saw does quite well. Let’s be clear. I do NOT like this saw. I also don’t like abrasive cutting either, but for really hard materials, sometimes an abrasive blade is the only way to go. For the most part, I’d much rather just use a nice horizontal band saw with coolant. It’s way more humane to the people around it, and they produce better cuts too. BUT when you’re in the middle of a parking lot, sometimes you just have to suck it up.

Today, I’m making blanks that will be machined and tapped to hold shoulder bolts and bronze bushings for a VR experience platform. Photos of that part of the job to follow.

CNC milled and programmed some stuff up for a project for JPL up...

Mon, 09 Jan 2017 13:45:37 -0800

CNC milled and programmed some stuff up for a project for JPL up in Pasadena. It’s a prototype for a moving sign post that points to celestial objects and tracks satellites in real-time. Pretty snazzy. It uses a simple Arduino (for now anyway) to control all the movement and a Teensy that’s eavesdropping on the serial line to cherry-pick out specially formatted comment strings in the movement instructions to change the text and color…. all crammed in to a 4″ pipe AND waterproof. Note that in this version, there’s no pipe covering the guts, and no custom circuit board has been attached, so nothing is weather tight yet. A fun project. I promise more in the near future.

It’s called Cymatics. A fun gadget to demonstrate standing waves...

Mon, 09 Jan 2017 13:35:07 -0800

It’s called Cymatics. A fun gadget to demonstrate standing waves and to drive everybody crazy with in the shop. It’s a speaker enclosure with a thin membrane, topped with some gross powdered drink mix because it was the only convenient powder we had that wasn’t hazardous. This was a prototype for a client job, intended to demonstrate that a 30 foot one of these would be so outrageously annoying that it would be rendered impractical. Ernst Chladni had his plates, I have my Saran Wrap.

Decided I finally needed to get my life together and get a nice...

Sat, 24 Sep 2016 10:27:33 -0700

Decided I finally needed to get my life together and get a nice boombox.

This one features two giant JBL speakers, an 800W Behringer mixer / amp / processor / whatever, all powered by a nice 1KW pure sinewave inverter, juiced by a hefty 24V electric lawnmower battery that I converted from god awful lead-acid batteries to lovely lithium iron phosphate (originally intended for electric car conversion)

It’s loud.

This thing loitering around the shop this afternoon - Makes the...

Fri, 19 Aug 2016 13:49:29 -0700

This thing loitering around the shop this afternoon - Makes the whole place shake!

According to the wikipedia:

The Bell UH-1Y Venom is a twin-engine, medium-sized utility helicopter, built by Bell Helicopter as part of the United States Marine Corps’ H-1 upgrade program. The helicopter is also called Yankee for its variant letter, Y

Fun times with runout! This is on an old 3-jaw chuck on a...

Thu, 18 Aug 2016 12:51:52 -0700

Fun times with runout! This is on an old 3-jaw chuck on a southbend lathe.

Ideally, that needle would be completely still, but there’s crazy run-out. Run-out is a type of mechanical ‘out-of-whack-ness’ that is usually used to describe how much a rotating thing spins and whips around like a cookie dough mixer versus a perfectly straight rotating shaft. In this case, I put an edge finder in the lathe chuck, which by all accounts ought to be the straightest and most perfectly ground thing in the shop. To that, a dial test indicator is placed, and the lathe or whatever you happen to be testing, is turned on to SLOW. The amount of wiggle on the dial is read, in this case, it’s about 0.015″ which is a LOT for something that would ideally have less than 0.001″ of wiggle.

At this point, you take the chuck off, take the jaws out and clean the living crap out of everything, look for burrs and dings on the jaws, clean, re-assemble, and re-test and hope everything gets better. Barring that, you would need to do a whole new operation involving a tool post grinder and a whole bunch of other crap to re-grind the jaws of the chuck so that they’re properly centered.

Now, don’t get me wrong, you can still make nice parts on this lathe, but you have to start with a new part that you remove at least as much material from as you have out-of-whack-ness (runout) and then, after that, you can’t remove the part from the chuck until you’re completely finished. Kind of makes doing operations on existing parts or doing backside operations impossible.

For those, we end up using the 4 jaw chuck and manually dialing it in, which is very time consuming, but really the only way we can achieve good results.

Maybe some day we’ll fix it.

I bought an electric car. I’ve had it a year, so I figure it’s...

Tue, 26 Apr 2016 23:05:25 -0700

I bought an electric car. I’ve had it a year, so I figure it’s high time I write about the experience. I got the Chevy Spark EV, which is the electric version of the regular Chevy Spark. (the gas powered Spark sucks. Two days behind the wheel of a rental was enough for me.)

The electric version, on the other hand, is a savage little sleeper with gobs of torque, a low center of gravity, excellent weight distribution, and VERY impressive acceleration. Like, “How the f**k?!?” acceleration. Merging onto the 110 freeway is actually fun instead of the usual terrifying white knuckle prayer session in most other cars. In short, this car is a blast to drive.

When it comes to electric, GM nailed it. In a tiny little 4 passenger hatchback, they nailed it. Everybody else making EVs isn’t trying hard enough. Look at the Mitsubishi i MiEV. It sucks because Mitsubishi wants it to suck. GM apparently wanted to be good at making an EV, and lo and behold, it worked. I know Tesla is a thing, but since nobody can actually afford them, it’s not worth talking about.

So why did I get the electric? Because I was SICK of working on engines. I was driving a 97 Toyota Corolla for years. I put a new exhaust in it, a catalytic converter, new belts, a new gas tank, a battery, the list goes on, all myself, in my driveway. Filth, impact wrenches, oil, gasoline, grime, stink, all of it. I got it to pass smog, and I vowed never again. I sold it to some chump for the price of my down payment on the lease for the Spark EV. Win. (for me)

So everybody wants to bitch about range. Unless you have an god-awful commute, you will be fine. This is not a car for travelling salespeople, but think about the last time you drove 80+ miles in a stretch. It was NOT fun, and you don’t do it often. The Spark EV goes 80 or so miles, less when you run the heat. The A/C takes off some mileage too, but not nearly as much as you’d expect. I never hesitate to crank it, plus, you can remote start the car, and if you’re plugged in, it will use power from the wall to pre-cool or heat, which is awesome.

The next thing is charging. Let me be crystal clear on this - If you can’t charge at home, you have no business owning an electric car.. Not yet anyhow. When the 200 mile cars come out in late 2016, then maybe you can think about it, but for now, if you street park, forget about it. Maybe you can charge at work, but your car spends most of it’s time at home. Now, at home, you can charge two different ways. One is with a regular old dumb wall outlet, which will give you enough for maybe 50-60 miles of driving for a regular overnight session. Most people never run their batteries flat, and to be sure, this is the slowest possible way to charge, BUT it only takes you 10 seconds to plug it in, so who cares. It’s not like you have to sit there and hold the button down. The next way is to get a 220V charger. You can build your own or buy one for an outrageous amount of money. A 220V (and 110V for that matter) car charger is actually nothing more than a fancy hair dryer plug with a bunch of stuff to keep you from accidentally getting shocked. Nothing more. It’s ridiculous that they cost hundreds of dollars. They don’t change, regulate, or alter the power in any way whatsoever, nor will a bigger unit charge your car faster. The charger is built in to the car, and it is what it is. Deal with it. They will charge your car in about 8 hours from flat dead empty, but nobody runs their car to zero anyway, and again, it’s not like you have to sit there and hold the button down. At the end of the day, the car is cheap to charge. I figure $150 a year on my power bill. I used to spend that much in two months in gas. If I was lucky.

Let’s talk about DC fast charging! Fast chargers are giant external chargers that hook up straight to the battery and dump mind-blowing amounts of power in to your car in a very short period of time. Like, 50,000 watts. That’s enough juice to power your whole block and all their air conditioners in August. Needless to say, it tanks up your car in about 15-20 minutes. The problem is, they are expensive, and while there are a lot of them in California, there are nowhere near enough, and they’re frequently occupied, and often a site only has one or two, and if they go down, you’re screwed. You don’t get range anxiety, you get “Will this fucking Freedom Station be working?!?!” anxiety, or there’s some douche trying to squeeze the last 1% in to his car for 45 minutes. Public charger etiquette is a problem. And again, there aren’t enough of them. There needs to be a station every 10 miles along the I5 from Mexico to Canada. Maybe then there would be enough saturation that people could find open, working chargers.

It’s a small car. It only seats four. 99.5% of the time, this is fine, it’s just me and my daughter going to work and daycare. Baby 2 is on the way, and it’s tight with two car seats, but doable. Don’t expect to haul a lot with a small car. Usually people don’t haul much anyway, except excessive amounts of trash. I DID use a roof rack to move some plywood and drywall, which was HILARIOUS but perfectly fine. Again though, unless you need to haul crap around, you don’t need a big car, electric or otherwise. For big jobs, we have my wife’s car, a Nissan Rogue. Fine car, expensive to drive in comparison, but when we have long drives or vacations, the bigger car makes more sense for all the crap we need to haul. In other words, if you are a two car household, there’s no reason to have both be gas cars… unless you can’t charge at home.

All in all, the experience of driving the Spark EV is more about driving a really speedy and fun SMALL car rather than driving an electric. The electric part is a big deal, for sure, and i do love me some silent rocket ship action, but plugging in at home and work is really the only difference in routine, and never having to stop for gas (which is awesome). I personally will never go back to a gas car. The electric car is perfectly suitable for my everyday use, and even longer trips too, and is just way to easy, smooth, and fun.

Questions?

The Internet of Chickens. Good progress on day 5.

Thu, 04 Feb 2016 19:54:31 -0800

The Internet of Chickens. Good progress on day 5.

Photo

Fri, 29 Jan 2016 17:35:04 -0800

The chickens are dead.A sensor malfunction doomed the four...

Wed, 27 Jan 2016 16:01:14 -0800

The chickens are dead.

A sensor malfunction doomed the four remaining eggs.

Obviously, the temperature didn’t fall to -400 degrees, nor go up to +500, but it sure did spaz out the computer. A simple reset of the Arduino brought it back in to line, only to discover that the internal temperature had reached over 120 degrees. Obviously some fail-safes are in order.

Sadness fills the air.

New milling machine is here! Another Tormach, this time with a...

Wed, 27 Jan 2016 13:36:40 -0800

New milling machine is here!
Another Tormach, this time with a nice beautiful enclosure and beefed up motors and electronics. Looking forward to making copies fly again. It’s been WAY too long.

Well! It looks like the Internet of Chickens is working! Here’s...

Mon, 25 Jan 2016 23:24:42 -0800

Well! It looks like the Internet of Chickens is working! Here’s Egg No. 2 being “candled” by shining a bright LED flashlight through it. It’s a white egg, so visibility is really great. The air sack is clearly visible on the left, along with lots of healthy blood vessels and of course the embryo and her tiny little dark eye (OK, it could just as easily be a boy too) This particular batch of eggs is about 9 days along, with another 12 to go until they hatch. It’s hard to believe that in only 12 days, this thing will completely fill the egg, but I promise, it will. The incubation has been taking place in my internet enabled “Internet of Chickens” built from a wine chiller, utilizing it’s built in thermoelectric cooler (run in reverse to warm instead of cool) and a high precision temperature and humidity sensor. It’s page can be viewed HERE

http://data.openevse.com/emoncms/kontraptionist&id=80

Note the very brief twice daily temperature drops - These are the times that I’m opening the door to turn the eggs and check on them. Future batches will incorporate some kind of automatic egg turner to keep the temperature as steady as possible. Also of note are some really deep temperature drops for prolonged periods earlier in the week, due to a squirreley Arduino and whatever power supply chip and problems it had. This caused the Arduino to crash, sending the temperature dropping. Not fun for these little buggers, Half of the eggs terminated development for one reason or another, and it’s reasonable to believe that this may have been the culprit. Future versions will incorporate watchdog timers to facilitate automatic reset when things go awry, along with better power supply wiring and probably most importantly, not use old Arduinos from the junk drawer with stickers on the bottom labeling them “BAD” like this first run. Live and learn. Or accidentally kill something and learn and try a lot harder next time.

The other thing to note is that these eggs are from the grocery store. No, really, these are grocery store fertile free range eggs that came from chickens of unknown health, and were stored in unknown conditions, aside from full refrigeration for unknown periods of time. The fact that any of them are developing is quite impressive.

Wish us luck!

PID loop inputs and outputs. Here’s a chart from the Internet of...

Tue, 12 Jan 2016 11:53:57 -0800

PID loop inputs and outputs.

Here’s a chart from the Internet of Chickens, via the Arduino serial plotter, which is great (although Serialchart is WAY more powerful)

Light blue is output power, dark blue is setpoint, red is humidity, and yellow is temperature.

As you can see, the temperature is hovering below the setpoint. It seems that the PID loops is allowing this slight error to persist over time. I raised the integral value, which seems to have corrected this problem.

Here’s the best page I’ve found thus far for explaining the effects of different PID tuning parameters.

Behold the Internet of Chickens!!!A link to the Internet of...

Mon, 11 Jan 2016 12:14:20 -0800

Behold the Internet of Chickens!!!
A link to the Internet of Chickens: http://data.openevse.com/emoncms/kontraptionist&id=80

I got it in my head that I was going to incubate a batch of eggs, and I could have bought a commercial incubator and spent a bunch of money, but NO! I have an Arduino!

Here’s what’s going on here - I went down to the local thrift store and bought a broken 8 bottle wine chiller, the kind with a Peltier thermoelectric cooler that takes about 3 weeks to chill a bottle of wine. You know the ones. It ended up in the second hand store because of water damage to the wires and connector that connect the electronics in the door to the power supply in the back. Quick fix.
This thing is brilliant for an incubator however, because if you run a thermoelectric cooler in reverse, it becomes a thermoelectric HEATER. It is after all a solid state heat pump. The box is also very well insulated, is has a clear glass door, and the inside features a fan to keep the air moving, which is important to keep temperature gradients to a minimum. There’s a little LED to light up the inside, and a temperature display on the front (of questionable accuracy)

I kind of liked the temperature display though, so I took the time to test to see if it could report higher temperatures. It did just fine with a space heater pointed at it, so I decided I’d make an effort to keep it running through the conversion, which proved to be easy, since all of the thermometer logic was in the door, along with the LED control via the buttons.

In the back, there’s a jankey 12V power supply and a heatsink and fan for the originally hot (and soon to be cold) side of the thermoelectric cooler.

The plan with this whole thing is to switch the polarity of the Peltier and control it with the arduino, which will be reading a high precision thermometer and hygrometer chip (the SHT11 because I had one laying around) and report the results on a web page so I can monitor and make adjustments if needed.

I originally wanted to use the internal power supply, simply because it was already there, but I discovered that the way that the system worked is that it cycled the entire 12V section of the power supply on and off. This wasn’t going to work for me, since I wanted to keep the fans running at all times, so I put a MOSFET in line with the Peltier so I could control it independently.

FAIL. Cheap switching power supplies do NOT like being run at zero load, and this particular power supply had no safety to shut it down in a zero load condition. It made a weird buzzing sound for a few seconds, then the buzzing stopped. Forever. Whatever, I never liked it much anyway.

I replaced the power supply with a nice 12V 6A power supply that DID have protection features, and this all worked brilliantly. The 12V went through the MOSFET in to the Peltier cooler, and I controlled it with a PWM digital out pin on the Arduino (An Arduino Pro 328P at 16 mhz and 5V) with a Seeed ethernet shield purchased from a defunct radio shack at a steal. The arduino and door thermometer/LED were powered at 5V by a linear voltage regulator from the 12V side of things (a 7805 from another project) and all was well with the world.

The temperature and humidity are read from a chip called an SHT11, which is an expensive, calibrated thing as big as a tic-tac, and is discontinued as far as I can tell, but there are better ones from this company that can be had at better prices. I just had it laying around, and it performs brilliantly. I was not about to trust the temperature to the crappy thermistor or whatever was already built in, plus I needed humidity measurement anyway.

Now here’s the fun part. Chicken eggs need to be held at a pretty constant temperature. As such, a simple “bang bang” control scheme was not going to cut it, as in, “If it’s below 99.5, turn the heater on, and if it’s above, turn it off!”

What I needed is PID control - A proportional–integral–derivative controller (PID controller) is a control loop feedback mechanism (controller) commonly used in industrial control systems. A PID controller continuously calculates an error value as the difference between a measured process variable and a desired setpoint. (thanks Wikipedia!)

The PID controller would be able to set the power of the heater to precisely match the heat loss of the incubator to balance it at exactly the temperature I wanted, plus be able to deal with transients like me opening the door or adding moisture without producing crazy temperature swings or oscillations. Of course you have to tune PID loops, and on a system with such a slow response as mine, this would take FOR EVER. Luckily, some smart person released a library for the Arduino that will do a pretty good job of “auto tuning” to find the ideal values. It does so by varying the power over time to see how the system responds, how long it takes to settle, etc. You more or less tell it what your usual desired temperature is, and say GO. Mine took overnight to figure out settings, and they were several orders of magnitude different than what my first guess was, so…good thing for that. Temperature control is nailed, and works great.

Next, I had to get it on the internet. I ended up taking some code left over from an open souce EV car charger (See the OpenEVSE project) and used that to report my data to an Emoncms server so I could log and graph temperature, humidity and power levels. Emoncms is a pretty awesome open source energy monitoring system that people use for everything from their boilers to solar panel systems to (in my case) electric car chargers. There’s even apps for your phone. It’s also very easy to push data out of small microcontrollers with simple http requests. I can also set parameters, such as temperature from a second, built-in website.

This is proving to be very useful, because it’s helping me ferret out some problems I seem to be having with strange temperature spikes. It has nothing to do with the PID loop malfunctioning, since immediately upon the slightest spike in temperature, the control does its job and drops power, and the temperature increase is what happens first. My hypothesis is that the fan in the chamber is stopping, and then convection happens, which lets hot air settle around the temperature sensor. With quick access to the data, I can spot a transient event and rush out to the garage and see what’s causing the problem. Hopefuly, it’s something I can easily fix with a drop of oil.

For the future, I would like to convert this to a wireless system. My immediate obstacle to this was the fact that all my Wifi stuff is 3.3V, and driving a big MOSFET with 3.3V is a non starter, and anybody that tells you otherwise is a liar. In the upcoming board, a proper “gate driver” is in order along with wire to board connectors and voltage regulators to eliminate all the claptrap, plus, of course, integrated Wifi. Problem there is that I will need to put a plastic “window” on the back to let the WiFi signal out, or get an external antenna. Another item to add is the ability to add moisture and an output to activate an automatic egg turner. More GPIOs in to more more FETs. Stay tuned.

Meanwhile, see the Internet of Chickens! Note the daily fluctuation of output power. The Internet of Chickens is located in my garage, which experiences large daily temperature swings, which can easily be seen in the amount of power required to keep the incubator at a stable temperature. Thanks, PID control!

Guess who’s back. Los Angeles just got a lot more awesome....

Sun, 16 Aug 2015 22:12:16 -0700

Guess who’s back. Los Angeles just got a lot more awesome. CNC power to the rescue. Now, this isn’t some kind of commercial for Laguna. Turns out this machine required a full de-braining and a complete controller retrofit and internal PC board design to meet the stringent K-Lab standards of quality and reliability, and the results are satisfactory for the price. Now if I can figure out a replacement for these horrible hold down clamps it came with, we’ll be in good shape. Bring it on, LA.

We make the most awesome stuff at Two Bit…. This is just...

Tue, 23 Dec 2014 16:00:00 -0800

We make the most awesome stuff at Two Bit…. This is just half.. HALF of an automatic paint cannon system we put together for an ad campaign for Toyota called “Make Your Mark” that was first shown at Art Basel in Miami in early December 2014. People would hash-tag and tweet and so on, and the paintball guns would make an image on a large canvas behind their new Corolla. It was pretty awesome and the tech end went off without a hitch… and it was a SERIOUS rush job too. Major props to our install crew too - Eric, Eliot and Dan.

The ad agency was Brightline.

The paintball system consists of three electric paintball markers (we don’t’ call them guns anymore) in which we completely gutted out the electronics and replaced them with components of our own which are linked to a computer via an beefed up Arduino, which also runs the G-Code interpreter, which handles all of the pan and tilt motors, their accelerations, speeds and so on. The Arduino has access to the bolt and ball sensors on the markers to detect if there are any jams or problems, plus control of gas flow and other safety critical systems. The entire system controls like a CNC milling machine, with some extra special commands of course. Software was no small feat either, but Eric managed to pull off the bulk of it while flying on a plane to New Zeland.

The skunkworks is open for business.