New Products Flash Sale
Jack Crossfire's blog View Details
Posted by Jack Crossfire | Dec 06, 2008 @ 01:14 AM | 2,953 Views
Unlike the lunar lander challenge, the Calif* lander challenge is only 32m horizontal & 10m vertical because because there isn't enough room in Calif*. Also, the Calif* lander challenge doesn't have precise positioning, rocket engines, or prize money because that went to mortgage bailouts.

So got automatic takeoff & landing trimmed out to where it looks pretty slick most of the time. In fact, it's faster to use automatic takeoff than human control. It's a matter of hard coding a lot of parameters to where the least amount of logic depends on GPS, mainly,

1) Getting the ramping speed low enough to not overshoot & high enough to not flop around on the ground.

2) Setting a high enough starting collective to get away with lower ramping speeds.

3) Using the PID controller to pull back neutral collective to what it should be after takeoff instead of 1m/s collective.

4) Trimming the neutral IMU attitude so it doesn't leap sideways during takeoff.

5) Doing 1 manual flight to get the cyclic & rudder reasonably trimmed.

but this has a lot more margin than some of our other robots. If U threw it on the ground & hit the switch, it would probably succeed despite a hairy takeoff.

Also, we're enabling full position hold for takeoff & landing, unlike GA Tech. U need a lot of battery reserve to do automatic landings.

Unfortunately our day job is back to its usual 7 day/week commitment & your government predicts above normal rain in 7 days. Not sure you'll get a daytime video of a Calif* lander challenge.
Posted by Jack Crossfire | Dec 05, 2008 @ 03:21 AM | 4,690 Views
So the propellers from Illinois made it after only 4 days. The propellers from China R still working their way through. Eventually we'll have a collection of 3 blade props as backups.

Got through 3 batteries without a crash. Hooray. The 10x4.5's + a very balanced ship got 9 minute flight times. Motor propeller combination is that important. Unfortunately we're still behind the T-Rex.

Besides flight time, we now have a freakishly quiet ship. Motor adjustment & slower RPM definitely paid off. Holy mother of mufflers it's quiet. Never heard of a powered aircraft so quiet. Even those powered gliders R like semi trucks in comparison.

The piece de resistance was of course, AUTOMATED TAKEOFF & LANDING using GPS ONLY. That's right kids. They say if U can get automated takeoff & landing to work using GPS only, U still can't afford a house but U must be able to afford something. It's much easier on a quad rotor because it's harder to flip over.

Programmed a bunch of rules for takeoff to work around GPS.

1) Minimum collective at power on starts the auto takeoff. Wait until absolute GPS climb rate is below a maximum starting value. Then set the hover position 10m above the stable altitude reading. Neutral cyclic & rudder R set to the starting stick positions.

2) Ramp collective until GPS climb rate is above a minimum ending value. Maintain 0 horizontal velocity & disregard position. U need to ramp fast enough for the attitude...Continue Reading
Posted by Jack Crossfire | Dec 04, 2008 @ 02:02 PM | 2,899 Views
Aviation, photography, & robotics hobbies R a good way to learn about the lives of the super rich.

After reading the Austin Meyer, Ken Rockwell, Philip Greenspun blogs we have 1 nugget of information 4 U. All this stuff about government stimulus packages, treasury oversight, mortgage bailouts, social-change-ism, stock market regulations, it's all garbage to keep the super poor busy.

U need 2 B making & selling something really cheap in massive quantities yourself. That's it.

There R 2 human populations departing each other as we speak. One group is the worker group, encountering very difficult times. The other group is the 1 man business group, becoming super insanely freakishly rich. Cash isn't gaining value. It's losing value catastrophically, making the freakishly rich merely the modern continuation of the norm 30 years ago.

This isn't a small group either. This is a very large fraction of the human race breaking off from the collapsing economy section like a chunk of Antarctica.

When we started Cinelerra in 1997, planned on charging $1 per copy of it & selling millions of copies. Then reusing GPL libraries like libdv looked real attractive for making a better product, so we had to make the whole thing GPL & trade the $1 per copy model for day jobs working on something else for a living.

12 years later, we're looking at the author of X-Plane, who started at roughly the same time, stayed commercial, & now flies full size...Continue Reading
Posted by Jack Crossfire | Dec 02, 2008 @ 10:12 PM | 3,419 Views
While you're still waiting for Chinese air mail, Hale Wayne was lashing out against the SSME again, so we put together some SSME p*rn 4 U. Needless to say, we always get kicked off the NASA blogs, but we don't get kicked off the Chinese, ESA, or ISRO blogs.

This was compiled from our Fl*rida still photos. Click on scroll bar and drag to rotate engine. Didn't bother 3D mapping it. Theoretically there is a tool somewhere for making it into a 3D model.



Engine Porn (0 min 25 sec)


Now today's nugget of information.

Blogs R losing their luster with private individuals & getting replaced by the instant message social network. Notice lots of blogs which used to get a major essay every day now get a few sentences per week. Who wants to do all that content creating when at most 2 people will read it or their old content will just get buried by their mountain of new content?

Blogs R now shifting into community based retail. For selling new products, the interactive blog format works where no forced advertising has. Unlike individuals, retailers want to bury their old content with new content & the readers actually have a connection to the content via their wallets. It's only a matter of time before the gm.coms & the hp.coms become blogs just like these award winning retail blogs, unless they all become government sponsored that is.

thereedspace.com
sparkfun.com
Posted by Jack Crossfire | Dec 02, 2008 @ 02:27 AM | 3,405 Views
Another $85 gone & the quad rotor continues the T-Rex tradition. As expected, the quad rotor eats motors & props like the T-Rex ate landing gear, main shafts, tail booms, main blades, servos, tail shafts, tail blades, batteries, bottom plates, electronics boards, & tail fins.

Fortunately 4 U taxpayers, the local hobby shop is no good 4 quad rotors so there will be no more sudden credit crunches. Everything is a 2 week mail order.

Now some easy reading while U wait for Chinese air mail.

Compare film to digital. The film had a light leak & crummy processing, but we feel it has better color than the digital one. Had to shift everything to the red to get the digital one to come out while the film retained blues. The film was fully automatic & 1 shot. The digital took many attempts & intentional underexposing. Suspect quad VicaCopter's lighting would look much better on film.

Unfortunately, there's no such thing as film. Everything is scanned & viewed on a computer eventually. Film has log color & better dynamic range, but it's still scanned by the same linear photodiodes of every digital sensor.

Even though they're both digital, film scanners do better than digital sensors. Maybe its because they can average data points over a longer time than digital sensors. Maybe projecting artificial light through a negative allows more information to be captured than directly sensing natural light. Maybe it's because Bayer patterns have 2x more green cells than R or B.

Sadly, no more film camera bodies R coming out for EF lenses. U have to get an old one for more than the cost of a digital camera.
Posted by Jack Crossfire | Dec 01, 2008 @ 01:48 AM | 3,075 Views
The answer is yes. Autopilot can crash. The quad is still too underpowered to recover from a freefall & we're still trying bigger collective gains. The APC 9x6's were super strong, but U can't do enough in 3 minute flight times to justify replacing them.

In the mean time, looking at the continuing explosion of autonomous quad rotors in the world reminds U a lot of the explosion of Linux PC software in the 1990's. It's all college student based, just like Linux was.

microdrones.com
www.mikroquad.com
www.mikrokopter.de
uavp.ch


Now some good gains.

Code:
#              P         I          D          PLIMIT  ILIMIT DLIMIT OLIMIT
Z_TO_DZ        -0.1000   0.0000     -0.1000    2.00    0.00   2.00   2.00
DZ_TO_COLLECT  0.0400    0.00020    0.0000     0.30    0.50   0.30   0.50
Now some crash gains.

Code:
#              P         I          D          PLIMIT  ILIMIT DLIMIT OLIMIT
Z_TO_DZ        -0.5000   0.0000     -0.5000    2.00    0.00   2.00   2.00
DZ_TO_COLLECT  0.0400    0.00020    0.0000     0.50    0.50   0.50   0.50

Posted by Jack Crossfire | Nov 30, 2008 @ 01:26 AM | 5,058 Views
2 minute flight times R killing us. It's like flying a rocket. Have enough credit again to upgrade propellers in the next 2 weeks.

Managed to stabilize altitude. It required higher collective limits. Altitude gains have been equal 4 all 3 UAVs. Horizintal velocity gains R identical to the T-Rex. The Corona needed 2.5x more horizontal velocity gain. Suspect it was because the Corona naturally damped horizontal velocities.

This was our first uBlox 5 test with fixed pitch. Fixed pitch rotorcraft give audible cues of collective response & the uBlox 5 is very fast.

Next, U need to align your IMU, magnetometer, & rotors or correct it in software. The result is inability to determine heading & starting lunges in horizontal velocity. Our IMU is off the rotors by a certain amount in pitch....Continue Reading
Posted by Jack Crossfire | Nov 28, 2008 @ 08:02 PM | 5,143 Views
Got the 1st video of the quad rotor flying itself 4 U. Altitude hold is still worthless. May need an MRAC 4 that one. Not enough flight time to do much testing between battery charges. Definitely need more power. This is about as good as the coverage gets when U only have 3 minutes of flight time. Cool the batteries after 3 minutes & U can get another 2 minutes before the quad rotor starts to tilt.

Autonomous quad rotor (1 min 17 sec)
...Continue Reading
Posted by Jack Crossfire | Nov 28, 2008 @ 06:42 AM | 5,053 Views
Now some test flights of 200Hz PWM. 200Hz PWM changed everything.

The stability with 200Hz PWM is freakish. Holy mother of molasses it's stable. It's like flying the Corona again. Full manual mode has returned. With that, the quad rotor was open 4 business.

Sideways lunges went away. Either it was the damping or binding in a motor. The starboard motor seems 2 have a damaged bearing. Got some more life out of it by lubricating it.

The quad did its first reasonable autonomous hover. Take a look at the Goog replay.

Quad rotor's first autonomous hover. (1 min 5 sec)


The hover was 170sec & condensed to 64sec 4 U home gamers. For the first 70sec, we had to command a 2m/s climb rate. Altitude hold doesn't work very well. 170sec was the limit of the battery.

Flight times R still real short. Even if it's balanced, the batteries get flaming hot within 3 minutes. Need better propellers before trying anything longer....Continue Reading
Posted by Jack Crossfire | Nov 27, 2008 @ 06:34 AM | 5,519 Views
Skip Mohatt died on Monday. He died at age 74 from injuries in a car accident. Drive safely.

We were the worst civics student in all history & he was the toughest civics teacher in all history. As far as Calif* civics education was "different" from what the rest of the world was taught, we were a Calif* civics dropout & still are, but Skip had 2 nuggets of information 4 U.

"People become politicians for 1 reason only: power." It's not about change, rescuing the economy, doing good, or making money. It's about power & only power.

The other nugget was "Most of your teachers are now women. In college, most of your professors will be men." That was the first time we ever heard that. Always knew there was something peculiar about gender roles, but this was like the theory of relativity. That was the beginning of the end of our fantasies about marriage.

Would have become obsessed with gender roles eventually, but Skip was really the single point were it began. Every heroine worshipping, heroine robot, heroine galaxy, heroine car, heroine statue, & heroine college, started there.

Now back to the blog.

Got the Suppo ESC's up to 200Hz. That's right kids. 200 Hz. Haven't heard of any castle creations doing that. Maybe a space shuttle main engine can do that. Looks like ESC's, unlike servos, use edge detecting microprocessors at very high clockspeeds. It makes sense because they need a lot more logic than servos to control shutdown modes.

At 200Hz, the quad was very stable on the vertical test stand.

The reason for this is we're getting sporadic sideways lunges in the test flights. Either the radio is dropping packets, the PIC memory is getting corrupted again, or it can't respond to wind. More rate damping is the easiest step.

Good news is the IMU without the Z accelerometer still works. No more IMU glitching in this dumpy apartment. Autopilot is a disaster. The computerized attitude hold doesn't have enough authority to recover from an autopilot malfunction.
Posted by Jack Crossfire | Nov 26, 2008 @ 04:42 AM | 5,029 Views
The answer is yes. The Z axis accelerometer is very problematic. Took a few years, but finally would say the Z axis alone is probably the main point of failure from high vibration. For vehicles which don't fly inverted, U can probably ignore the Z accelerometer & do better.

If the accelerometer shakes enough, it probably goes to maximum in both directions. Lowpass filtered, it probably washes out to 0g.

Unlike normal quad rotors, this quad rotor has so much surface area it actually can glide to a soft landing from a few feet.
Posted by Jack Crossfire | Nov 25, 2008 @ 04:29 AM | 5,255 Views
Well 4 U quad rotor fans, vibration is still as problematic as it was on the unicopters. The quad may vibrate less, but it's so freakishly more unstable than a unicopter, U need higher bandwidth filters. Plus, U need the attitude estimation to work fulltime. There's no manual override if the IMU starts drifting.

Designing a quad rotor from the ground up is like designing a T-Rex from scratch. Holy pile of work. Makes U wonder what those Chinese T-Rex designers consider a mortgage bailout. You're really best off buying finished & proven quad rotor electronics from someone else instead of developing your own.

Now today's discovery.

Suppo ESC's only go up to 2ms on the PWM. Slightly below 2ms, the ESC hits maximum power. If U go higher than 2ms, the ESC limits power to the maximum & shuts down after a certain amount of time, usually in flight. It emits a repeating beep to indicate signal lost & it disarms. Then if U pull back to 1ms it emits a different beep to indicate it's armed again.

The second time it arms, it normalizes idle to 1ms instead of 0.8ms, so U can't balance the quad rotor again until resetting it.

Got the quad rotor up again. Major oscillations resolved. Minor oscillations still there. Manual hover was going smoothly. Seemed 2 B plenty of power, thanks to some software tweeking. Autopilot tests were disasterous. Had some crashes due to the lack of space. For battery #2, had 1 slow heading drift. Then the random nose dives began. Some more crashes & that was it.
Posted by Jack Crossfire | Nov 24, 2008 @ 12:41 AM | 5,102 Views
As another day passes without any flights, it's another delay, another setback. This time, glitches in the IMU readout. The A/D converters were running too fast for the PIC to handle the data + the new rate damping code, so the IMU sampling had to go from 1600Hz to 800Hz. That broke rate damping.

The effect was an A/D interrupt handler getting 2 pins ahead of the A/D buffer handler, accelerometer readouts going into gyro slots, very hard to track down.

Not long ago, dreamed we achieved the wealth & power to own a 9' Steinway, but not the $3 million for the square footage to put it in, so hatched the ingenious strategy of putting it deep in a forest where no-one would hear it.

Sure enough, someone found a piano deep in a forest. Maybe it belonged to someone in the same situation. Maybe a renter who could afford the $3000/month for rent & the $5000 for an accoustic piano, but not the $5 million for a house, faced with the prospect of having his 401k confiscated for the common good, decided now or never & made a forest practice room. Unfortunately, U can't use forests as practice rooms in an age of taxpayers funding $100 million houses, so it was impounded.

http://www.cnn.com/2008/US/11/23/mas...ano/index.html
Posted by Jack Crossfire | Nov 23, 2008 @ 12:06 AM | 5,055 Views
Raytheon Active Kill System (1 min 3 sec)


So this video of a guided missile shooting down another rocket got us thinking, a missile that flies up, turns 90 deg, & flies sideways in a fraction of a second is a super high speed UAV. What else can U do with a super high speed UAV?

Maybe instead of massive vehicles with massive batteries & hour long flight times, U have tiny airborn robots that take off, perform tasks in a fraction of a second & land with just a tiny amount of energy.

Used to wonder if there was a way to launch billions of tiny payloads into space real fast, each payload consisting of maybe a fraction of a gram of raw materials & launched using lasers or some super cheap vehicle. Then these billions of tiny payloads would self assemble in space to form massive structures.

What if U could put a small rocket in a handheld launcher & make it fly out 1m, reverse direction & fly back real fast. Maybe U could make it fly out & back hundreds of times in a second. Every time it flew back to the launcher it would recharge. A perpetual 1m linear orbiting rocket guided by a handheld launcher. U could aim the handheld launcher & the rocket's 1m linear orbit would always point where U aimed it.

Maybe instead of a rocket, it was a mirror with some mode of propulsion & U beamed a high powered laser at it to make a laser saber that was bounded by the rapidly oscillating, flying mirror. Not as clumsy or random as...Continue Reading
Posted by Jack Crossfire | Nov 22, 2008 @ 03:57 AM | 5,224 Views
The quad rotor bailout isn't working. Time for drastic measures. Got to elect a quad rotor despot & implement more regulation on the PIC. We're going to need to nationalize not just all the subprime aircraft, but the prime aircraft too. That's right. All your functioning aircraft now belong to Jack Crossfire, to rescue the broken aircraft. Give em up.

Some photos of Oakland telescopes seem to have missed the Jack Crossfire show, so U still get some content, this time....Continue Reading
Posted by Jack Crossfire | Nov 20, 2008 @ 04:36 AM | 5,293 Views
The quad rotor finally got in the air. Not enough power to control altitude. Cyclic oscillation in the dumpy apartment is the same as in the air. A few seconds in the air with that oscillation & the IMU drifts too much, U start going sideways faster & faster out of control, flip, & engine shutdown. Looks like better propellers R a start.

Higher cyclic gains may reduce oscillations, but these R prone to flipping without more kludges.

Quad tests (0 min 56 sec)

Posted by Jack Crossfire | Nov 19, 2008 @ 03:14 AM | 4,903 Views
Depressed over the quad crisis, it's time to worship The GOOG.

It's time to look over The GOOG's Time Warner photo archive & compare great depression #1 with great depression #2. Mainly the family living in poverty. After photographing our dumpy apartment, in our age of endless propaganda about government safety nets, total regulation & centralization, was surprised to find the typical apartment of the modern 50 hour/week programmer looks exactly like the poverty stricken apartment of the 30's.

U know you're taxed at 70% when U work, under Oooobs consisting of 30% marginal taxes, 20% FICA, 13% sales tax, 1% property tax, assorted infrastructure taxes, & "recycling fees" on everything U buy.

During the unemployed times, & U R going to have a lot of unemployed times, you have no savings. Your purchasing power is provided by unemployment insurance & other government programs. The property, sales, infrastructure, & recycling taxes on those purchases may come out of entitlements, but they easily exceed the 30% of your income U had left during employment, so the total tax you're paying during your lifetime is over 100% of what U earned. U pay over 100% of your income in taxes.

Unlike depression #1, there R no private war chests to start businesses & hire people. That money, over 100% of your income, has been paid into government programs to eliminate the need to start businesses & hire people.

Now what do U do when U can't get yaw damping to work? U bring back the vertical test stand. Most people use lazy susans to test yaw damping, but we only have enough room for a single piece of aluminum which can be reused for a vertical test stand & wind screen. U know the wind screen.

Quad rotor vs. vertical test stand (1 min 31 sec)


Notice heading in the video is fixed except when commanded by the pilot. Cyclic damping is still in crisis.
Posted by Jack Crossfire | Nov 18, 2008 @ 04:30 AM | 5,068 Views
Now that we're not flopping anymore, can see the quad is still capable of nasty vibration. Propeller balancers will be in attendance.

Decided to forget about manual cyclic & begin implementing a full time orientation hold & switchable position hold on top of that. That is a buster to get all 3 modes to coexist simultaneously for quad rotor & single rotor craft.

With enough software functioning between commutes, test flight #2 was still disasterous. Cyclic rate damping is good enough for the computer to hold the attitude, but still not for a human. There is no yaw authority. Yaw damping is ineffective & it takes very large deflections for the computer to try to hold a heading, which it never quite does before 1 of the 2 motor pairs stops.
Posted by Jack Crossfire | Nov 16, 2008 @ 06:18 AM | 5,169 Views
Looks like all quad rotors have to overclock their PWM way above the usual 50Hz. Mikrocopter uses I2C to get still higher update rates. The answer is yes. Suppo/Super Simple ESC's can do 100Hz, but it requires shrinking the resolution. The standard resolution only goes to 60Hz.

Nonlinear feedback curves, lighter propellers, higher gyro bandwidth, & higher frequency PWM has all contributed a bit to stability. It's never going to be as stable as an articulated rotor.

In today's video, U can see the rate is damped enough for a human to hold the angle, but the stick response is prone to violent rotation.

Quad rotor rate damping 2 (1 min 35 sec)


Forget about having enough money for any new parts in November. The Suppos R what U get.
Posted by Jack Crossfire | Nov 14, 2008 @ 02:50 AM | 5,391 Views
Well, Microdrones raised the bar with an 8.5lb quad rotor capable of 50 minute flights on battery power. Dual redundant IMU's & flight computers. Must figure out how Germans got so loaded compared to US.

Anyways, back to bailing out SUV manufacturers.

Decided blogs based on user content R really boring now because no-one's buying anything but T bills, so here R some rate damping efforts. Unfortunately, Goo Tube audio synchronization doesn't convey the relation between movement & throttle changes & sure enough, Goog stock is only $300.

Quad rotor rate damping (2 min 35 sec)


See. No rate damping at all. Disappointed to find the "New 1000mm Quad Copter Design" was using Turnigy ESC's at only double the price, not the 10x price U would expect for the ESC to be the problem. The next step is really full time attitude hold. Everyone uses full time attitude hold. Disable lowpass filtering completely & U start to get some rate damping but it's very erratic.

Also, these props R 72% heavier than successful quad rotor props. Haven't sold the T-Rex yet.