A multi-ducted fan design is for all intents the same as the multi-rotors already discussed, although they will have slightly greater efficiency due to less blade tip related losses. If you want to build a proven successful design, then by all means, you'll want to stick with the multi-prop design. I personally can't help but to go off the beaten path for many projects, although this of course is the reason many of them have been unsuccessful.
All this technology stuff is cool, but we need to get back to Ray's original post. Let's build one here from scratch instead of trying to reverse engineer others.
Keep it simple for everyone and just start out with a stick built Quad or Hex craft. You can grow from there.
I bet we can come up will some pretty cool stuff.
Just look at what Old Man Earl came up with on his own last year:
Just do a forum search for posts by "Ole Man Earl" for the last 12 months in the Propeller forum and it will turn up a lot of stuff that he did here for a year that pertains to this thread. Alas we have lost Earl to DIY Drones, but he seems happy over there.
I live minutes from Hobby Lobby and have seen these models in person. They are impressive but they employ 2 servos per thrust duct. Given the small output of each, that would require at least 4 - 6 thrusters. 8-12 servos may be a lot to control. And I don't think there is enough thrust for vertical takeoff. Well, maybe with 6. They work well because of the good wing lift for horizontal flight.
Jim
markaeric said...
A multi-ducted fan design is for all intents the same as the multi-rotors already discussed, although they will have slightly greater efficiency due to less blade tip related losses. If you want to build a proven successful design, then by all means, you'll want to stick with the multi-prop design. I personally can't help but to go off the beaten path for many projects, although this of course is the reason many of them have been unsuccessful.
I am not sure of the stability from a beginners perspective. Usually they seem to like to get the props spread out quite a bit.
For a futuristic idea, I think it would be nice to build a UFO Saucer housing and mount the motors/props in a housing formed with the shell. I know this will require stronger motors as there is the resistance of the shell. Perhaps an EPP foamie. Anyway, that's for later.
hover1: Thanks for pointing out that half of the motor/props run counterclockwise. I certainly was aware but it hadn't been mentioned.
Hanno: I have been thinking re 1 or 2 props (chips)...
My reasoning for 2 props was that the sensor inputs could come from a seperate pcb and the 6DOF or 9DOF would also work as a replacement.
Prop chip "Sensor" ...
1 cog for the 'supervision'
1 cog Serial (for output to the "controller" prop board)
2 cogs -·SPI driver &·I2C driver (could be combined)
Gyro 3 axis
Accelerometer 3 axis
Digital Compass 3 axis
Pressure & Temperature
Ping))) for ground detection
1 cog serial driver
GPS input
Of course, it would be easier for the Gyro, Accelerometer, etc to each be on their own cog as the driver could be more specific and may already be in OBEX. This would make·5 cogs, for a total of 8 for the whole "Sensor" module. This would be a seperate pcb and would have other applications such as robots.
Prop chip "Controller" driving the motors (via ESC's)...
1 cog for the 'brain' to do the kallman?? filtering, etc
1 cog (minimum) to read the RC info (6 channels)
1 cog to get the serial input from the sensor board
2 cogs for 4 motor/escs (uses 4 counters in PWM)
3 cogs if 6 motor/escs
Future options ???
1 cog (at least) to write to microSD (data logging) - later?
1 cog to transmit (zigbee?) data to earth - later?
1 cog to control camera - later?
To drive ESC replacements, I am not sure if 1 cog per motor may be required. I would expect it to be likely, or at most 2 motors per cog.
As you can see, I have thought about camera control later. However, I think it would be better that the "Controller" prop just output the info in a serial stream to another prop which could then be used for the transmission, including overlaying the data on screen. I would like to keep things in a modular form.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
Below i post a link from china or TW? I use to order cheap things from there. There you can find BLDC motors ESC's etc. Those might be not the best quality parts but cheap and might be good enough for prototyping http://www.dealextreme.com/products.dx/category.801
Cluso,
" 1 cog (minimum) to read the RC info (6 channels) " There are some ready made spin codes in the object exchange for reading RC receivers. Those are using 1 cog/2 channels. We can achieve higher speeds with pasm ... (I newer made a calculation to see if speed is enough for more channels but they say max 2ch/cog in spin)
Regarding gyros:
we could use the ADIS16080 but price is a little high (around 50 USD)
or ADXL345 3 axes accelerometer (aprox 15 USD)
I will spend some more time to find some cheaper gyros if any
Thanks for the information. I liked the Super Simple ESC. The reasons to build your own ESC are, 1. faster response time 2. weight, 3. size. Most hobby ESC don't list their update rate and when listed it is typically 8 to 10 kHz.
To keep things simple I would start with an existing ESC. If this proves to be inadequate, then consider building and programming an ESC. Building and programming an ESC is a project in itself.
Cluso99
I think your organizational structure is probably the way to go. One other thought would be to mix CPUs. I am concerned that the amount of Propeller memory may be inadequate for processing. The Propeller could be used to receive sensor data and possibly a second to output data to ESC, camera servos, etc. One CPU could be dedicated to the 'math' i.e. Kalman filter. Really depends on how much information you want to process(Hovering, Obstacle avoidance, Navigation, data logging, image processing, flight stability). The Brain CPU lets call it would need to be fast, need to address enough memory, FPU, etc. Of course this makes programming more complex.
Hover1, just try getting your hovercraft a few feet in the air with no skirt and then we can talk [noparse]:)[/noparse]
BuildThings, I was not in anyway trying to question the awesomeness of ducted fans however awesomeness alone will not make the design work and that is a shrouded rotor not a ducted fan.
markaeric, go for it, I'm a fan (lol) of difficult: www.indoor.flyer.co.uk/buzzard.htm Still doesn't fly properly but I haven't touched it for some time.
My job is trying to design/build flapping winged micro air vehicles that can fly like insects and hover etc, yep I know difficult well [noparse]:)[/noparse]
Szabi: I found the LY530ALH & LPR530AL from Digikey for $8.90ea. They are analog so ano $2+ for ADC. I have been looking for gyros with I2C or SPI but nothing at a reasonable price, and still 2 chips (roll/pitch & yaw). The ADIS16080 is a Yaw only I2C so too expensive.
The ADXL345 $6.16 or MMA7455 $2.91 for 3-axis accelerometer and HMC5843 $19.20 for 3-axis digital compass. Need to see if the response is adequate - does anyone know??
I prefer pasm so I think we can probably read the whole 6 RC channels in 1 cog. If I understand correctly, they are sequential anyway.
BTW: Everyone is making a big thing about prop memory. It has 32KB Ram in the hub. Take a look at other micros - which ones have 32KB Flash or Ram??? Not many!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I found the ITG-3200 which is a new 3 axes I2C gyro chip for a price below 3USD for mass production and retail price is in range of 10 -15 $. (Just google for it and you will find lot of news about it. ex.: http://www.i-micronews.com/news/WORLD’S-FIRST-SINGLE-CHIP-3-AXIS-DIGITAL-GYROSCOPE-AVAILABLE,3684.html )
The chip it's in a 4 x 4 x 0.9 mm QFN package and available for selected customers only...
Not sure from where but let's see if we can get some
Graham, flapping wing flight seems extremely cool, yet endlessly complex. Just give us a heads-up when you release your armada of killer robo-bees [noparse];)[/noparse]
The ITG-3200 & MPX-200? are not yet available and they are aiming for millions of units for the games arena. We have no hope of getting any quantities, so I owuld not design a pcb for this yet. There other products are analog :-( and only 2-axis like the others. InvenSense is the company.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
LOL! Sorry, its hard for me to get my head wrapped around 32k and 80 MHz again. I have been learning over the past month how capable the Propeller and similar processors are. Many many years ago I did assembly language on 6502 and Z80 CPUs with less memory and 1 to 4 MHz clock speed. I have gotten use to thinking in terms of GHz and GBytes. The Propeller is really changing the way I look at computing and its refreshing.
Great discussion! I had a lot of trouble finding an inexpensive gyro for my DanceBot a couple years ago- I ended up very happy with a $30 piezo RC gyro. You send it a 1ms pulse, and right away it returns a pulse whose length is determined by the sensors rotation. Worked very well. The "tilt" object (described in the Parallax book, and available on the ftp site- I'll have to post it to obex) uses 1 cog to read the accelerometer and perform the floating point kalman filter to update the tilt reading. The gyro is read by the main spin loop. For this we'd need 2 of each...
I was "eating my own dog food" and came up with this program in 12Blocks:
Think it'll fly?
Hanno
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Co-author of the official Propeller Guide- available at Amazon
Developer of ViewPort, the premier visual debugger for the Propeller (read the review here, thread here), 12Blocks, the block-based programming environment (thread here)
and PropScope, the multi-function USB oscilloscope/function generator/logic analyzer
Dynamically selectable full scale: ±250/±500/±2000 dps
SPI, I²C digital interfaces
16-bit data output
Fully configurable low-pass and high-pass embedded digital filters
Programmable interrupt generator
There are lots of new chips becoming available. Warning... don't use them as they will be in short supply because the high volume game manufacturers will be taking them !!! You may be the lucky one to get one but the rest of us will likely miss out. I want this to be a shard project with readily available chips and cheap. No use laying a pcb out for chips you may not be able to buy.
The Epsom chip is analog and not available. The SG is not available and analog. Seems like we will have to use a pair of analog gyro chips for 3-axis
Hanno: Ah, if only it was that easy · Your 12 Blocks is brilliant
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I also found the MG1101 from Gyration. This is a dual axis (supposedly low cost) gyro. Google shows only broken links on google to gyration website :-( I could not find it @ gyration. If they are using this gyro in their units, we just need to buy some mouses etc for about 25 USD
Anyway... while searching for gyros for days would be easier to use an analog chip attached ADC.
We also could make a separate PCB with analog gyros and using a (for ex.) PIC16Fxxx with analog inputs, so analog to digital conversion will be solved separately in one chip.
The MG1101 looks expensive and it's about 14-15mm cube, 2-axis. We could mount 2 on the pcb to get 3-axis and it's digital I2C. I couldn't find a source and it's quite old in this arena.
Think I am going to settle on the following for the sensor pcb. Basically it's similar to the 9DOF from Sparkfun with a prop instead of the Atmega, plus pressure, temp, and maybe ultrasonic for ground proximity.
Before anyone comments why don't I use the Atmega - it's a prop project!!!
Propeller & EEPROM etc
LPR530AL (or LPR550AL)
LY530ALH (or LY550ALH or LPY versions)
ADC 4 channel or micro TBD
ADXL345
HMC5843
BMP085
PMB-248 or PMB-648 or #28146 GPS Module (GPS from Parallax)
Any suggestions for ground sensing?·(maybe need to detect from 5m) Perhaps the Parallax·Ping))) sensor could be used???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
It seems Ole Man Earl got the Ping to work well with close ground measurements last year, but I'm having problems finding it! (Parallax Jim, please hurry on the Forum software upgrade!)
I did find some info on his project over at DIY Drones:
Cluso99 said...
The MG1101 looks expensive and it's about 14-15mm cube, 2-axis. We could mount 2 on the pcb to get 3-axis and it's digital I2C. I couldn't find a source and it's quite old in this arena.
Think I am going to settle on the following for the sensor pcb. Basically it's similar to the 9DOF from Sparkfun with a prop instead of the Atmega, plus pressure, temp, and maybe ultrasonic for ground proximity.
Before anyone comments why don't I use the Atmega - it's a prop project!!!
<UL>
* Propeller & EEPROM etc
* LPR530AL (or LPR550AL)
* LY530ALH (or LY550ALH or LPY versions)
* ADC 4 channel or micro TBD
* ADXL345
* HMC5843
* BMP085
* PMB-248 or PMB-648 or #28146 GPS Module (GPS from Parallax)
</UL>
Any suggestions for ground sensing? (maybe need to detect from 5m) Perhaps the Parallax Ping))) sensor could be used???
There are several Ping sensors with different beam projections but the ultrasonic sensor would not work if there is moderate pitch due to angle of reflection. The copter would have to be almost flat I would think and relatively close to a reflective surface.
One could suspend a sensor platform from underneath the copter that used servos to tilt the platform in order to keep parallel with the ground. But then who knows if the ground will be flat and what its reflectance characteristic will be.
The same would how true for infrared distance sensors I would think.
There is always stereo vision processing if the ground has features. This of course requires a lot of memory and computation. OpenCV frameworks has stereo vision routines.
Valid point. I would think that the utilization of the proposed GPS along with the barometric sensor would bring the bird back very close to home, (a designated flat landing area and probably the launch point), and the Ping sensor would be used to guide the landing in the last few feet. And of course the 6DOF would keep the craft flat for the landing.
Jim
Robert T said...
There are several Ping sensors with different beam projections but the ultrasonic sensor would not work if there is moderate pitch due to angle of reflection. The copter would have to be almost flat I would think and relatively close to a reflective surface.
We can buy a 9DOF pcb from Sparkfun with the gyros, accelerometer and compass with an Atmega328 assembled & tested $125. We can do a 2 wire serial interface to our prop controller board.
If I do a pcb for this section then parts will cost arount $55-60 + shipping to me usually costs me 50% (yes OUCH) unless I time it with a friend which saves a bit. Then the pcb and a screen because the assembly will require a paste screen because of the·QFPN packages and a small volume so likely $15 per board. Certainly we would have a prop on the pcb for coding and the extra parts to be added (GPS, Pressure, temp). Doesn't sound too viable does it???
The "controller" pcb is fine. A prop + interface to 6-8 RC inputs and 6-8 PWM outputs (motor esc's etc) and the serial to the 9DOF board. We can add the interface here for the GPS, pressure/temp, ultrasonic. Even a Prop ProtoBoard could be used here. We can do the complex code on this pcb for the prop.
Your thoughts???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I'm not sure what the need is for so many sensors/degrees....
For leveling you should only need 2 gyro directions (pitch and roll), and for facing a single axis (yaw) compass should be fine. Those coupled with a tri-axis accelerometer should be enough for what we want. That's what Jason Dorie had in his "Propeller with Propellers" project, and it looked pretty darn stable in that last video. Why have 3 gyro and 3 compass directions on top of a the tri-axis accelerometer? I guess I am missing something?
I'm currently building my cuircuit using all Parallax parts: Prop Protoboard USB, H48C tri-axis accelerometer module, HMC6352 compass module, and 2x LISY300 Gyro modules. I currently have the modules setup on my PPDB breadboard along with the stuff to connect to the reciever. I've got Jason Dorie's code to read 6 RC channels from the reciever in one cog and source files for working with the modules from the parallax site. I've got raw sensor/reciever data for everything dumping out to PST so far... Now I need to mix in some Kalman filtering and PID (again Jason Dorie's code provides a good starting point for me) and then plug the results into PWM32v7 to drive the 4 ESCs. Maybe, I'm being overly optimistic, but I hope to attempt some flights this weekend.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
I like that idea a lot. The 9DOF board by Sparkfun would simply output the info to the propeller. The Propeller could then integrate that information with any other sensors(GPS, Barometer, etc). The only code on the Atmega would only collect the data and respond to requests from the propeller. The whole 9DOF board would simply be treated as a complicated sensor.
Roy
I agree the 9DOF is over kill for the functions mentioned but not all sensors need to be incorporated. What would be the cost of your present setup if done the way Cluso described. If there is a big price difference then your way may be better. Do you think the compass is adequate to control the Yaw if you are using the two opposite motors to control this? Or would you treat Yaw like you would Pitch and Roll(accelerometer and gyroscope)? Some of the videos I have seen of quads do show Yaw drift and I have wondered how this is being controlled.
What is a three axis magnetometer? I read the data sheet and my guess is that the magnetic field is read along the X, Y and Z axis and combined into a single direction. Does this compensate for tilt(vector math)?
Roy, I think you're totally on the right path. At minimum, you need 2 gyro+2 accelerometer to level the craft and make it flyable. My code (above) uses 2 kalman filters to process those 4 sensors into 2 usable measurements. Thanks for the pointer to Dorie's code for reading 6 rc channels. I think to fly the craft you just need 4 channels- up/down, rotation, fwd/bkw, and left/right- again see my code above. Instead of using a compass to tell direction I would use what rc helicopter people have been doing for decades- use a gyro. I haven't had much luck with compass sensors- they're not very accurate, give bad readings- especially near magnetic fields like 4 motors, and are sensitive to tilt. Good luck!
Hanno
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Co-author of the official Propeller Guide- available at Amazon
Developer of ViewPort, the premier visual debugger for the Propeller (read the review here, thread here), 12Blocks, the block-based programming environment (thread here)
and PropScope, the multi-function USB oscilloscope/function generator/logic analyzer
Robert T:
My way is using modules made by parallax that contain ADCs with serial interfaces and support circuitry, so overall it's more expensive. ~$168 for the Prop Protoboard USB, 2x LISY300, H48C, and HMC6352. Then add a bit for connectors, passives, and such. I wanted to build something within my electrical/circuit capabilities (beginner), and using parallax modules fits that well with lots of documentation and easy interfacing to the Prop.
I think my compass will be enough for yaw (it says it's a 2-axis compass, but the sample code just reads out a single heading value). I won't really know until I try it. I do have the H48C in conjunction with the compass and gyros, so that may help.
My understanding is that a magnetometer is another name for a digital compass.
Hanno:
If the compass doesn't work out, then I'll get another LISY300 module to use instead.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
Post Edited (Roy Eltham) : 3/19/2010 4:48:53 AM GMT
Comments
Upon looking for a good source of model ducted fans, I came across a thrust vector kit (http://www.hobby-lobby.com/360_degree_thrust_vector_upgrade_set_rc_airplane_spare_parts_390761_prd1.htm) for a R/C SU-34 (http://www.hobby-lobby.com/su34_twin_360_degree_vectored_thrust_jet.htm), and now my imagination is running wild!
Keep it simple for everyone and just start out with a stick built Quad or Hex craft. You can grow from there.
I bet we can come up will some pretty cool stuff.
Just look at what Old Man Earl came up with on his own last year:
http://forums.parallax.com/forums/default.aspx?f=25&m=350669&g=350724#m350724
Just do a forum search for posts by "Ole Man Earl" for the last 12 months in the Propeller forum and it will turn up a lot of stuff that he did here for a year that pertains to this thread. Alas we have lost Earl to DIY Drones, but he seems happy over there.
Jim
Post Edited (hover1) : 3/17/2010 2:37:41 AM GMT
Jim
http://www.societyofrobots.com/robottheory/Energy-Efficient_Autonomous_Four-Rotor_Flying_Robot_Controlled_at_1_Khz.pdf
I am not sure of the stability from a beginners perspective. Usually they seem to like to get the props spread out quite a bit.
For a futuristic idea, I think it would be nice to build a UFO Saucer housing and mount the motors/props in a housing formed with the shell. I know this will require stronger motors as there is the resistance of the shell. Perhaps an EPP foamie. Anyway, that's for later.
hover1: Thanks for pointing out that half of the motor/props run counterclockwise. I certainly was aware but it hadn't been mentioned.
Hanno: I have been thinking re 1 or 2 props (chips)...
My reasoning for 2 props was that the sensor inputs could come from a seperate pcb and the 6DOF or 9DOF would also work as a replacement.
Of course, it would be easier for the Gyro, Accelerometer, etc to each be on their own cog as the driver could be more specific and may already be in OBEX. This would make·5 cogs, for a total of 8 for the whole "Sensor" module. This would be a seperate pcb and would have other applications such as robots.
To drive ESC replacements, I am not sure if 1 cog per motor may be required. I would expect it to be likely, or at most 2 motors per cog.
As you can see, I have thought about camera control later. However, I think it would be better that the "Controller" prop just output the info in a serial stream to another prop which could then be used for the transmission, including overlaying the data on screen. I would like to keep things in a modular form.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
some more details about Microcopter or Microcopter like I2C ESC's inculding sch, and source you can find here:
http://freenet-homepage.de/alex_konze/bko/blc4428/blc4428.htm#bild01
http://plischka.at/Regler.html
http://www.rcgroups.com/forums/showthread.php?t=1039478
i know some links are all in German but, if someone needs something to understand I can help with translation.
Below i post a link from china or TW?
http://www.dealextreme.com/products.dx/category.801
Cluso,
" 1 cog (minimum) to read the RC info (6 channels) " There are some ready made spin codes in the object exchange for reading RC receivers. Those are using 1 cog/2 channels. We can achieve higher speeds with pasm ... (I newer made a calculation to see if speed is enough for more channels but they say max 2ch/cog in spin)
Regarding gyros:
we could use the ADIS16080 but price is a little high (around 50 USD)
or ADXL345 3 axes accelerometer (aprox 15 USD)
I will spend some more time to find some cheaper gyros if any
Br.
Sz
Thanks for the information. I liked the Super Simple ESC. The reasons to build your own ESC are, 1. faster response time 2. weight, 3. size. Most hobby ESC don't list their update rate and when listed it is typically 8 to 10 kHz.
To keep things simple I would start with an existing ESC. If this proves to be inadequate, then consider building and programming an ESC. Building and programming an ESC is a project in itself.
Cluso99
I think your organizational structure is probably the way to go. One other thought would be to mix CPUs. I am concerned that the amount of Propeller memory may be inadequate for processing. The Propeller could be used to receive sensor data and possibly a second to output data to ESC, camera servos, etc. One CPU could be dedicated to the 'math' i.e. Kalman filter. Really depends on how much information you want to process(Hovering, Obstacle avoidance, Navigation, data logging, image processing, flight stability). The Brain CPU lets call it would need to be fast, need to address enough memory, FPU, etc. Of course this makes programming more complex.
BuildThings, I was not in anyway trying to question the awesomeness of ducted fans however awesomeness alone will not make the design work and that is a shrouded rotor not a ducted fan.
markaeric, go for it, I'm a fan (lol) of difficult: www.indoor.flyer.co.uk/buzzard.htm Still doesn't fly properly but I haven't touched it for some time.
My job is trying to design/build flapping winged micro air vehicles that can fly like insects and hover etc, yep I know difficult well [noparse]:)[/noparse]
Graham
The ADXL345 $6.16 or MMA7455 $2.91 for 3-axis accelerometer and HMC5843 $19.20 for 3-axis digital compass. Need to see if the response is adequate - does anyone know??
I prefer pasm so I think we can probably read the whole 6 RC channels in 1 cog. If I understand correctly, they are sequential anyway.
BTW: Everyone is making a big thing about prop memory. It has 32KB Ram in the hub. Take a look at other micros - which ones have 32KB Flash or Ram??? Not many!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
The chip it's in a 4 x 4 x 0.9 mm QFN package and available for selected customers only...
Not sure from where but let's see if we can get some
Graham, flapping wing flight seems extremely cool, yet endlessly complex. Just give us a heads-up when you release your armada of killer robo-bees [noparse];)[/noparse]
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
·http://www.st.com/stonline/stappl/cms/press/news/year2009/p2440.htm
but Digikey and Mouser do not even list it yet, but Arrow does have a backorder option.
Analog output, but it does cut out one chip.
Jim
·http://www.epsontoyocom.co.jp/english/info/2010/0223.html
No pricing available yet.
Jim
I was "eating my own dog food" and came up with this program in 12Blocks:
Think it'll fly?
Hanno
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Co-author of the official Propeller Guide- available at Amazon
Developer of ViewPort, the premier visual debugger for the Propeller (read the review here, thread here),
12Blocks, the block-based programming environment (thread here)
and PropScope, the multi-function USB oscilloscope/function generator/logic analyzer
http://www.st.com/stonline/products/families/sensors/l3g4200d.htm
Dynamically selectable full scale: ±250/±500/±2000 dps
SPI, I²C digital interfaces
16-bit data output
Fully configurable low-pass and high-pass embedded digital filters
Programmable interrupt generator
Jim
Post Edited (hover1) : 3/18/2010 12:11:13 PM GMT
The Epsom chip is analog and not available. The SG is not available and analog. Seems like we will have to use a pair of analog gyro chips for 3-axis
Hanno: Ah, if only it was that easy
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
Anyway... while searching for gyros for days would be easier to use an analog chip attached ADC.
We also could make a separate PCB with analog gyros and using a (for ex.) PIC16Fxxx with analog inputs, so analog to digital conversion will be solved separately in one chip.
This,·The "tilt" object (described in the Parallax book, and available on the ftp site- I'll have to post it to obex)
sounds very interesting, where can i find it
regards
Rob.
Bottom right of this product page:
http://www.parallax.com/Store/Books/Propeller/tabid/171/CategoryID/45/List/0/Level/a/ProductID/637/Default.aspx?SortField=ProductName,ProductName
Jim
Think I am going to settle on the following for the sensor pcb. Basically it's similar to the 9DOF from Sparkfun with a prop instead of the Atmega, plus pressure, temp, and maybe ultrasonic for ground proximity.
Before anyone comments why don't I use the Atmega - it's a prop project!!!
Any suggestions for ground sensing?·(maybe need to detect from 5m) Perhaps the Parallax·Ping))) sensor could be used???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
I did find some info on his project over at DIY Drones:
http://diydrones.com/profiles/blogs/the-autoprop8-has-a-home
Will search some more.
Jim
One could suspend a sensor platform from underneath the copter that used servos to tilt the platform in order to keep parallel with the ground. But then who knows if the ground will be flat and what its reflectance characteristic will be.
The same would how true for infrared distance sensors I would think.
There is always stereo vision processing if the ground has features. This of course requires a lot of memory and computation. OpenCV frameworks has stereo vision routines.
I realize that stereo vision is beyond this project at least for now but check out http://www.surveyor.com/stereo/stereo_info.html
Jim
We can buy a 9DOF pcb from Sparkfun with the gyros, accelerometer and compass with an Atmega328 assembled & tested $125. We can do a 2 wire serial interface to our prop controller board.
If I do a pcb for this section then parts will cost arount $55-60 + shipping to me usually costs me 50% (yes OUCH) unless I time it with a friend which saves a bit. Then the pcb and a screen because the assembly will require a paste screen because of the·QFPN packages and a small volume so likely $15 per board. Certainly we would have a prop on the pcb for coding and the extra parts to be added (GPS, Pressure, temp). Doesn't sound too viable does it???
The "controller" pcb is fine. A prop + interface to 6-8 RC inputs and 6-8 PWM outputs (motor esc's etc) and the serial to the 9DOF board. We can add the interface here for the GPS, pressure/temp, ultrasonic. Even a Prop ProtoBoard could be used here. We can do the complex code on this pcb for the prop.
Your thoughts???
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
For leveling you should only need 2 gyro directions (pitch and roll), and for facing a single axis (yaw) compass should be fine. Those coupled with a tri-axis accelerometer should be enough for what we want. That's what Jason Dorie had in his "Propeller with Propellers" project, and it looked pretty darn stable in that last video. Why have 3 gyro and 3 compass directions on top of a the tri-axis accelerometer? I guess I am missing something?
I'm currently building my cuircuit using all Parallax parts: Prop Protoboard USB, H48C tri-axis accelerometer module, HMC6352 compass module, and 2x LISY300 Gyro modules. I currently have the modules setup on my PPDB breadboard along with the stuff to connect to the reciever. I've got Jason Dorie's code to read 6 RC channels from the reciever in one cog and source files for working with the modules from the parallax site. I've got raw sensor/reciever data for everything dumping out to PST so far... Now I need to mix in some Kalman filtering and PID (again Jason Dorie's code provides a good starting point for me) and then plug the results into PWM32v7 to drive the 4 ESCs. Maybe, I'm being overly optimistic, but I hope to attempt some flights this weekend.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
I like that idea a lot. The 9DOF board by Sparkfun would simply output the info to the propeller. The Propeller could then integrate that information with any other sensors(GPS, Barometer, etc). The only code on the Atmega would only collect the data and respond to requests from the propeller. The whole 9DOF board would simply be treated as a complicated sensor.
Roy
I agree the 9DOF is over kill for the functions mentioned but not all sensors need to be incorporated. What would be the cost of your present setup if done the way Cluso described. If there is a big price difference then your way may be better. Do you think the compass is adequate to control the Yaw if you are using the two opposite motors to control this? Or would you treat Yaw like you would Pitch and Roll(accelerometer and gyroscope)? Some of the videos I have seen of quads do show Yaw drift and I have wondered how this is being controlled.
What is a three axis magnetometer? I read the data sheet and my guess is that the magnetic field is read along the X, Y and Z axis and combined into a single direction. Does this compensate for tilt(vector math)?
Hanno
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Co-author of the official Propeller Guide- available at Amazon
Developer of ViewPort, the premier visual debugger for the Propeller (read the review here, thread here),
12Blocks, the block-based programming environment (thread here)
and PropScope, the multi-function USB oscilloscope/function generator/logic analyzer
My way is using modules made by parallax that contain ADCs with serial interfaces and support circuitry, so overall it's more expensive. ~$168 for the Prop Protoboard USB, 2x LISY300, H48C, and HMC6352. Then add a bit for connectors, passives, and such. I wanted to build something within my electrical/circuit capabilities (beginner), and using parallax modules fits that well with lots of documentation and easy interfacing to the Prop.
I think my compass will be enough for yaw (it says it's a 2-axis compass, but the sample code just reads out a single heading value). I won't really know until I try it. I do have the H48C in conjunction with the compass and gyros, so that may help.
My understanding is that a magnetometer is another name for a digital compass.
Hanno:
If the compass doesn't work out, then I'll get another LISY300 module to use instead.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
Post Edited (Roy Eltham) : 3/19/2010 4:48:53 AM GMT