Jack Crossfire's blog View Details
Posted by Jack Crossfire | Jan 08, 2009 @ 01:07 PM | 2,741 Views
90 minutes & $10 is what each crash is costing. You're in luck because GPS was in a good mood.

The fog was slightly clearer at 7:50, giving a huge GPS improvement. This detereorated hugely by 8:25 as fog increased & the satellite over Texas disappeared behind the hill. The satellites R in the same place every day, so U can mostly observe the effects of weather.

The slight algorithm changes & the 5Hz upgrade seem to be a success. When the air is still, the aircraft movement is a direct observation of GPS drift.
Posted by Jack Crossfire | Jan 07, 2009 @ 02:52 AM | 2,561 Views
So U think U can avoid the fog by flying earlier eh. Well, the only time it isn't foggy between commutes is when there aren't any good satellites.

During la ninas, Rain Ramon is socked in with fog instead of constant rain from Dec to Mar. In the 1990 la nina, the fog caused huge traffic pileups, killing almost as many people as the 1989 earthquake. La ninas cause very stable, cold airmasses to last for months in the Calif* valleys.

Today's crash was another bug in the landing software which caused her to shoot up 147ft without enough battery power to descend safely.

Today's flight times with no foam baffles:

Good 4Ah: 9m30s
Bad 4Ah: 6m57s
Posted by Jack Crossfire | Jan 06, 2009 @ 01:19 AM | 2,585 Views
So all of last week's styrofoam miracles R busted. The propeller baffles may have increased flight time slightly, but were too noisy. We need stealth in the communist sardine can.

Had fog, drizzle, & no GPS again. Can now guarantee you better keep around some human pilots if U intend to fly low & slow in fog.

One other thing. Our 57600 upstream Comca$t bandwidth may be related to Comca$t'$ new bandwidth throttling software. Those Romanian attackers R using it to turn everyone's $800/year plan into an $800/year brick by flooding everyone's IP address with garbage. Comca$t detects the attack as a bandwidth hog & shuts U down.

Finally, Hale Wayne or Wayne Hale or whatever his name is had a good story about NASA's management training classes this time.

Hale

In the private sector, company training is history, especially company training on mountain meadows in front of lodges. From the gourmet cafeterias to the dance classes, your government sounds like some kind of parallel universe from the 50's. Sheesh, heaven forbid your civil servants have to contribute to a health plan.

The irony is if any taxpayer found out their 70% a year was funding modern dance moves in the mountain meadows, Americans would probably vote to have their taxes raised even more to buy a proper dance floor.
Posted by Jack Crossfire | Jan 05, 2009 @ 04:10 AM | 2,856 Views
The first cam flight was impressive in terms of output, not in terms of flying. During manual takeoff, the takeoff position snapped to horizontal because of ground effect. Once airborn, the autopilot entered spirals until we commanded a turn. There was no vibration in the video. The only blurring was motion. Battery exhaustion occured just 5m35 seconds later.

Automatic takeoff didn't work. The computer couldn't handle the rapid snap to horizontal caused by ground effect. That was the end of the camera mount. The only economical solution seems to be hand launching & hand capturing.

Until your government commands a battery stimulus package, it's going to be mostly pencams. Would U believe they're still selling that crummy Aiptek from last year for the same price?

Quad cam without brightness

quad cam 1 (1 min 3 sec)


Quad cam with brightness

quad cam 1 (1 min 4 sec)
...Continue Reading
Posted by Jack Crossfire | Jan 04, 2009 @ 07:42 PM | 2,379 Views
So put on baffles to try to deflect rotor wash away from the fuselage.

Good 4Ah: 9m16s
Bad 4Ah: 8m0s
3.3Ah: 5m16s

But it was windy & we also changed PID constants. There's a chance with the baffles of having a styrofoam get sucked into a propeller.

The biggest problem is we never did full discharges with the 750mm motor spacing. They were just 1/2 battery flights.

For this week's video, we have timelapse hovers using different collective I constants. Thought I=0.0001 was the best. This video also has a vortex ring state which the computer recovered from.

collective tests (1 min 33 sec)


We have 2 ideas for vortex ring states. Stick with the current collective jamming or program some kind of routine which always ascends & descends briefly.

Comca$t lowered upload speeds to 57600 so it's now 1/2 as fast as the cell phone, so U can actually upload video faster on a cell phone than Comca$t....Continue Reading
Posted by Jack Crossfire | Jan 04, 2009 @ 03:03 AM | 2,507 Views
The theory with this is U can land & take off without being level. The main problem is it's going to give real short flight times. Also don't have a way to stick the styrofoam pieces together or attach it all to the fuselage. May be limited to pencams with the quad rotor.

Moved the motors closer in preparation for the camera mount. Got some current flight times with the closer positioning.

Damaged 4Ah: 5m51s
Good 4Ah: 8m51s
3.3Ah: 4m27s

Reason to believe the fuselage is interfering with the propulsion.

Also have finally converted the A/D values to battery voltages.

15800 = 11.9V
12600 = 11.2V
-1770 = 7.9V
-3830 = 7.5V

We've been quitting at 10.1V....Continue Reading
Posted by Jack Crossfire | Jan 03, 2009 @ 01:49 AM | 2,417 Views
The answer is yes. The uBlox 5 can do 5 Hz. Maybe that's why it's called the uBlox 5. When not locked in, U can force it to update at any rate U want, but it always clamps to 5 when it locks. So does a 5Hz uBlox fly any better than a 4Hz uBlox? Of course not! U think you're going to learn anything valuable from a free blog?

To the Blox's credit, it was windier during today's tests than any previous night test. We'll leave it on 5Hz because it feels good.

4 a good time, go to http://www.sfgate.com/webdb/jobcuts to find out whose head's about 2 B chopped off. You'd be surprised how much doesn't get covered by Time Warner & News Corp.

Some highlights:

E-Loan: 270
Symantec: 269
Boeing: 262
Ebay: 184
Vishay Semi: 97
National Semi: 151
Seagate: 91
Sandisk: 96
Thousands of healthcare layoffs
Thousands of insurance layoffs

But Time Warner said healthcare was an oasis. Fuggeddaboutit....Continue Reading
Posted by Jack Crossfire | Jan 01, 2009 @ 07:04 PM | 2,393 Views
If the Blox is bad in clouds, it's hopeless in drizzle. It couldn't acquire on the ground. Had to fly up manually to acquire, but even then the navigation wasn't there. The EM406 couldn't handle rain either, back in the day. WAAS status crashed on new years eve.

Had our first full batteries of manual flying. The quad rotor is indeed the most stable vehicle we've flown. Manual flying is more efficient than autopilot.

We seem to have done everything possible with the piano mockup. It's not as consistent as we hoped, but that could be mockup inconsistencies. Can't calibrate it without really playing it. The Dec & Jan paychecks went to paying for the HP & the Acer.

Still, our net worth decreased so much in 2008 that we would have been better off buying everything we wanted....Continue Reading
Posted by Jack Crossfire | Dec 31, 2008 @ 05:20 PM | 2,725 Views
will be central planning.

In 2009, you're going to fund $3 trillion of government departments, $1 trillion of economic stimulus packages, & $2 trillion of municipal bond bailouts using a total GDP of $10 trillion. It can't come from China, Japan, heaven, or anywhere else because there isn't enough money. There's only 1 way to spend $6 trillion when U only make $10 trillion: by centrally planning activities.

We don't think your government is going to raise taxes by $6 trillion. Your government is going to do it by central planning. You're not going to be able to do anything U want unless it's part of a central plan. 60% of every day job, movie, meal, vacation, house, SUV, toy, gadget, stock trade, piece of clothing is going to have to be approved.

There is no way to run 60% of an economy through government stimulus if everyone is free to do 100% of what they want.

How do U make people obey the central plan? By selling massive numbers of bonds & making anyone who competes with bonds so poor all their resource consumption has to come from the committee. For example, if U need loans to build UAV's & the central committee sells bonds to fund the auto industry, U best switch from making UAV's to making SUV's.

Building cars, building houses, & lending mortgages will comprise most of the central plan. Perhaps you expected more creative pursuits. Fuggedaboutit. If U have other interests, U won't have them much longer.
Posted by Jack Crossfire | Dec 31, 2008 @ 12:26 PM | 2,778 Views
The bang bang altitude feedback sets the climb rate to no less than 0.3m/s & up to 0.5m/s. If dZ goes below -1m/s, collective gets another +0.1 added. We haven't had any problems with it. It's worked better than the mikrocopter algorithm.

Bang bang feedback was the first algorithm that worked in 2007 with the Corona & it was forgotten because it didn't make sense, but it works.

Landings are faster. Got it descending to the starting altitude at -0.5m/s, then switching to -0.3m/s. The starting altitude is where U flipped the autopilot switch or 10m above where the automatic takeoff began. She magically knows when U want an automatic takeoff.

Got another spiral landing. Suspect it's a gyro bias error from turning. Now just giving up on biasing the yaw gyro since its bias is too sensitive to turns.

We're up to bang bang feedback, ignoring Z gyro bias, & ignoring Z accelerometer. Every Goog search has been thrown out.
Posted by Jack Crossfire | Dec 30, 2008 @ 12:57 PM | 3,362 Views
Go to vicacopter.com & download VicaCopter's brain.

2 months after VicaCopter 2.0, it's VicaCopter 3.0. This code drop has the quad rotor support, UAV programming language, automated takeoffs & landings, curved paths, pirouettes, & turning towards a point.

There R people who spend their entire natural lives programming fixed wing line following algorithms. Beyond a certain speed, VicaCopter also needs a better line algorithm. The best we have is slowing down the fake waypoint as the error grows. It makes 4 some slow flying. U can probably compromise between the amount of slowing & the accuracy of the line.

The piano sensor is down to the actual microchip of the final product & trying everything to make the tone quality reproducable. It's quantizing it to just 3 tone qualities & analysing just the first half of the key press, 30ms of data.

Doubt an external oscillator could propagate far enough & fan enough to be economical so we're stuck with the free 8Mhz oscillator. A trick 4 U PIC fans is to set the overclocking register OSCTUNE to maximum to get 8.8Mhz out of the internal oscillator.




It's time for the biggest story of 2008.

Oobs may be the biggest story for Time Warner, but the skin color story died back in the 60's. U just get excited about it because U do what you're told.

The credit crisis was savings & loan 2.0. We're waiting for the medicare crisis.

It may surprise you, but the biggest story was the space station. Yes indeed. The space station is the most ambitious thing humans R doing. In 2008 the Columbus module, experiment rack, & Japanese module got hoisted up & the solar alpha rotary joint got fixed, all for 1/400th the cost of subsidizing ground based houses.
Posted by Jack Crossfire | Dec 29, 2008 @ 02:43 AM | 2,992 Views
Had a rare moment when the GPS, radios, propellers, batteries, motors, IMU, compass, airframe, computers, weather, & camera all worked simultaneously.

Here U have a pirouetting translation, a nose-in orbit, a translation pointing towards a point, & a simple pirouette before all the batteries were dead. A bug in the arc solver led to a gnarly 2nd landing.

VicaCopter Air Show (8 min 40 sec)


This is probably as stable as it can get without local DGPS, ground based vision, or onboard vision.

Threw another arcane feature you'll never remember into UAVPL. If any lon lat or alt coordinate is set to 0, the 0 coordinate is replaced by the starting coordinate. Set all your altitudes to 0 & U don't have to manually edit Goog Earth waypoints. Variables, relative coordinates, & loops R probably inevitable.

Probably already wrote somewhere that the minimum useful climb rate we ever got with a GPS module was 0.3m/s. Try to move any slower & you're just guessing.

That has caused her to get stuck just short of target altitudes & get stuck during landings. U need to use bang bang feedback for this one.

You'll notice satellites 135 & 138 R geostationary. They're for sending WAAS corrections & those coveted vertical signals for altitude. Unfortunately, they don't seem to do a good job in Calif*. Maybe they're too low. Maybe 2 aren't enough.

#135/Galaxy 15 was built by Orbital Sciences & launched on an Arianne 5 in 2005. It broadcasts backhaul video signals in addition to GPS.

#138/Anik F1R was built by EADS Astrium & launched on a Proton in 2005. It broadcasts HDTV to Canadians in addition to GPS.
Posted by Jack Crossfire | Dec 27, 2008 @ 06:06 PM | 2,749 Views
Well OK!!! We're still tweeking the control laws after all these years. Just like baseball or golf, U can spend an entire lifetime trying different swings.

For today's video, we have pirouettes with translation. This was an automatic takeoff, flight to starting point, translation with pirouette & crash. When she stopped rotating, GPS reported rapid North movement when she was rapidly moving South. There is nothing U can do when GPS malfunctions. For the next flight, she settled on just automatically taking off & hovering.

Translation with pirouette + crash (2 min 5 sec)

Posted by Jack Crossfire | Dec 26, 2008 @ 03:37 AM | 3,486 Views
The GPS satellite map & collective hack seem to work so far. Automatic takeoffs always work & altitude seems to be more stable when the map shows a satellite directly overhead. U don't really need the satellite web page since there is a satellite position in the uBlox output.

If GLONASS was supported, you'd have continuous altitude coverage.

Had the uBlox suddenly drop to 0 satellites & start returning 0, 0, 0 for no reason in flight. It may have been damaged in a crash. Maybe the cell phone interferes with it. The uBlox failure resulted in one of the greatest quad rotor saves ever, but didn't record it because GPS was broken.

Unfortunately, had 1 landing she couldn't finish. She descended for a while, then started oscillating in place until the battery died. In fact, all our batteries have been dying really fast. Also, the cell phone definitely interferes with the 2.4Ghz, so U won't be downloading GPS web pages while flying.
Posted by Jack Crossfire | Dec 25, 2008 @ 08:12 PM | 3,411 Views
Got lucky with the weather & it stopped raining briefly. 6 battery charges without a crash. Yay! Didn't have any sudden losses of altitude. Did have a battery get too low but got lucky.

Reenabled a line of code from the T-Rex 450 which goosed the collective whenever dz got below -1m/s. All roads point to assuming dz below a certain amount always feeds back into itself & have some real high gain below this amount.

For best altitude results, U GPS fans should go to http://www.nstb.tc.faa.gov/Full_WaasSatelliteStatus.htm & time your flying for when a satellite is directly overhead.

Have a video of the piano sensor working. Those of U tuning in without audio are in for a boring video.

Tone quality in an electric piano (1 min 18 sec)

Posted by Jack Crossfire | Dec 25, 2008 @ 03:23 AM | 3,375 Views
The storms don't have gaps anymore. It's a constant rain. Hard to believe anyone took flying for granted.

For years, no matter how hard we tried to avoid it, every invention required op-amps. VicaCopter was the first gadget which didn't require any op-amps. Well, the op-amp holiday is over. The piano sensor would require 1 op-amp for every key for a total of $10. Maybe even a potentiometer for every key: $100

With the op-amps on all the Vref's & photodiodes, the A/D conversion rate can go up to 20khz. The Vref range can be only 0.3V but everything must be op-amped. The photodiodes are so noise free, they don't need lowpass filtering.

So would probably go with 8 * PIC18F4221's in a full implementation. Unfortunately microchip.com has degraded its free sample program even more. Now U get just 2 model types & have to pay a flat $7.50 fee. In another 8 years, they'll be Atmel.

The piano sensor is now just extremely unreliable.
Posted by Jack Crossfire | Dec 24, 2008 @ 01:39 PM | 2,933 Views
Got 1 hour in the field between storms. The altitude doesn't work.

Quad rotors have translational lift during turns. It falls out of the sky when it stops turning. It's quite stable when it isn't moving. Went back to climb rate feedback & need several dry days to live with it.

Rained out for another night, we got the first useful data reduction from the piano sensor. The PIC can sense hand position most of the time by computing just 2 2nd order derivatives.

Unfortunately, the minimum sampling rate to detect anything is 5khz. Still have some options to go through.
Posted by Jack Crossfire | Dec 23, 2008 @ 01:02 PM | 2,999 Views
After all that hustling between commutes, it was another rain out. Your government has us rained out until next week. Replaced the 12 guage wire with 14 guage to try to get more margin. A tiny change in weight, but all the local guy had.

Meanwhile, piano key sensors didn't go so well. Got good data & a plan for compressing it. Put the test article on the floor to eat, then the sensor stopped working. The sensor is freakishly sensitive to alignment.

Calif* is getting flooded by Oaklahoma, Georgia, & Texas refugees. 4 those of U moving or already in Calif*, basically all of U, U need to file your income tax as soon as possible. Calif* is going to run out of money & start paying I.O.U.'s.

Income tax is going up $2500, but it won't keep Calif* from running out of money. U need to have that extra $2500 ready to pay higher income tax of course.
Posted by Jack Crossfire | Dec 22, 2008 @ 04:25 PM | 3,285 Views
Decided to throw U a bone with another flight recording.

Manual XY positioning (1 min 11 sec)


Flight recording of manual XY movement with the quad rotor programmed to always point towards a point. The computer keeps it flying 1m/s in the pilot's intended direction as the nose rotates. The pilot can also command it to point any side of the fuselage at the point. The actual time was 9 minutes.

Get into this mode using

MOVE lon lat alt
TURN_TOWARDS lon lat