June 15, 2015

Yesteryear
One year ago today: June 15, 2014, what’s a Fordson?
Five years ago today: June 15, 2010, electing the unemployed.
(It’s okay by me.)
Six years ago today: June 15, 2009, Flamingo, Florida.

MORNING
           How does this compare to my work a year ago? This is the progress on the joystick from the back view, where you can see the connections. Yes, Nova, it is fully operational and controls all three flight axes. So far it has only been tested with software, but I’ve programmed enough years to know it is behaving exactly as expected. The wiring system as shown here is the most efficient possible, but it would not be used on a real aircraft. I’ll tell you why.

           If you count the wires to the barrier strip, you’ll see there are five. The power and ground leads are common to all three potentiometers. Thus in real life, a failure anywhere in the circuit would result in all loss of control. I doubt such as design would make it past the door, but for demonstration purposes like this apparatus, no harm done.
           The wood surfaces are not finished yet and I intend to run in a ton of smoothing capacitors to cut down on “jitter” that occurs when the servo motors receive power spikes. The three middle (dark blue) wires are the control leads from the potentiometers. The arrangement is very flexible as the leads and potentiometers are all interchangeable.
           Oddly, because this is intended to send output to a microcontroller, this flexibility will likely never get used. Why? Because it is far easier to reprogram the software than to start moving wiring around. Originally I was going to put a shroud around the electronics, but now that seems unnecessary. I want onlookers to see the design simplicity.
           The three output signals are the pitch roll and yaw control. They act independently so it is possible to “fly” this in configurations that would be unsafe or unwise in a real aircraft. There is no feedback to prevent a pilot from simultaneously doing a rudder left and bank right. But if I do block such signals, it would be in the software rather than stopping the pilot from being such a cabbagehead
           How do you like it so far?

NOON

           “Make friends with a man who owns a truck.” --Texas wisdom

          I can tell you what is complicated. Stringing the control wires on an airplane. I learned to avoid mechanical linkages because they must be specially designed before they’ll cooperate. But strings can only pull, so you need them in pairs. This causes you to spend a lot of time staring off into space, since the servo mechanism turns in circles but what you need is back and forth.

          On top of that, the flaps don’t stop turning at the right spots by themselves. So I think I’ve got it, sort of. With the string that pulls the flap up, you have to leave enough slack to let the other string pull it all the way down. The most they can move is the angular distance your servo mechanism can move in one semi-circle. True, I’d rather be playing guitars with Taylor Swift, but playing with fish line was my second choice.
          Let me tell you about the highly touted fish line for robotics. It is finicky. And coated with a lubricant that makes it hard to tie knots that will behave. I confirm the 50-lb line is better for small robotic work as it is the diameter of thick thread. (The 80-pound line has a mind of its own and won't uncoil completely.) And yes, it helps to thread it with a needle. Within moments you will be glad you remembered all your boy scout knots.
           The stuff is not cheap, but since it is not anything fantastic, part of the price could be the hype surrounding the whole business of gouging people who are dumb enough to buy boats in the first place. The smallest spool of this braided fish line is around 130 yards is usually costs at least $20. You always think 130 yards is a lot of line until you make the first ten mistakes.
           My plan for this evening is to bike up to the coffee shop and see if I can’t plan some kind of simple mechanism that lets me adjust the length of strings rather than trying to cut them to the correct length in advance. I’d say things today are about as exciting as they are going to get when I’m in the money-saving mode.

NIGHT
           What I learned new today was about “Darlington arrays”, not the same thing the far humbler Darlington transistor. I have a special place for the transistor, as it is the device that convinced me 99% of the so-called experts on the Internet did not have a clue about the nitty-gritty. For those who missed that lecture, if you look up a Darlington transistor, you’ll get some idiot raving about the immense amount of gain in the circuit. That’s missing the point totally.

           The fact is, the Darlington is a fast switching transistor coupled to a high-power transistor which otherwise are two incompatible designs. The switch would fry with heavy voltage which cannot turn off fast enough. It’s the combination that does the job of fast switching heavy currents.
           And I learned about the array, consisting of a set of seven Darlington pairs which I know I’m going to need when I finally connect up a stepper motor. Here’s a photo of some guy pointing to such an array. I learned about them after throwing out many I should have kept. But again, nobody teaches you what you need to know.
           I decided to rethink the aileron controls. To that end, I biked over to Panera, where the best compliment I’ll give is that at least it is not Starbucks. If there is any accuracy to the rumor that an intellectual crowd hangs out at Starbucks, their aim was way off in Florida. Women in this town have consistently proven they have no idea what an intelligent man looks or acts like. They think he is “playing on the computer all day”.
           It took two refills, but I went over all the alternative arrangements for operating the flaps. I see I am going to need some materials for working with very small pieces in confined areas. Things like tiny screw eyes and eentsy sheet metal screws maybe an eighth of an inch deep. That’s putting it mildly since this involves another quantum learning chore, another leap into the unknown. Why do I say unknown?
           Because everything I thought I had to learn when dealing with robots went beyond the ordinary degree one usually needs to become familiar with a task. It turns out you can’t just “learn some carpentry” or “get good at soldering” and then go build robot parts. But nor am I foolish enough to try to convince anyone about that. Those who think they can just dabble on a topic and adapt that to microcontrollers is woefully mistaken. Trust me, this nation is full of people who tried it and gave up.
           Even something so simple as glue is no longer simple. I keep at least five types of glue on hand. The least used is superglue because it dries brittle. Robots and brittle don’t mix. That’s enough for today.

Last Laugh


+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Return Home
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++