Spaceclam

well,
i just talked to nathan from fmadirect, and he told me that the receiver is not compatible with the deans antenna. i wonderd about this, becuase the fma antenna is only 18 inches long, but yet is wuite capable of a 100 foot range check. the deans apparently will replace 32-36" antennas. i will try and see if he will tell me what type of wire, what gauge, etc so i can replace it and maybe have it flying well by this weekend.
Cheers,
-Clam

danvel

I still haven't tried the 3 axis accelerometer with my setup but, thinking about that, what I really miss is some kind of absolute SPEED sensor (magnitude & direction) with respect to ground.
TVE what's the range and resolution of your GPS module with SPEED measurement?
As I write this, I look at the optical mouse that lies on my mousepad: It really measures X and Y speed by looking at nearly ANY mousepad surface and I don't think it has a very complex image recognition system inside. Does anybody know how it works? Do you think we could do the same with a very simple, low res B&W camera looking into the ground pattern?

ADI

Ah yes ... this is what I've maintained all along is the most important info to have.
If we're NOT moving, then we MUST be level. On the other hand, if we appear to be level, this is no guarantee that we're not moving.

ADI

bizart

Hi !

TVE, I am sorry for the delay but I have got major office network failure. I have lost many files and almost a week I am trying to restore some office databases. Anyway I am sending you all the videos that have survived. You decide which of them you can store on your space. Currently I am reorganizing my web space so I hope soon to launch my new web site. Thank you in advance.

TVE congratulations too! You have made a lot of progress. Can’t wait to see some new videos!

Danvel you have got pretty stable flying! Really nice machine! And don’t bother using optical mouse sensor. Optical mouse does have some DSP on board but unfortunately the optical part is designed to work exactly at few millimeters from the surface. There are a lot of papers on the net about using camera visual feedback for such kind of systems which theoretically describe what huge amount of work you have to do for some real time, real life visual navigation.

Spaceclam nice to hear for your progress! It is really exiting to know that pitch controlled 4 rotor helicopter is possible to fly. Just think about four methanol engines powered system!!

Currently we have started design of new IMU board and flight control board. I have implemented all of the gyroscopes and accelerometer inertial calculation, radio control, Bluetooth telemetry full duplex communication, PWM control for the motors into one single microcontroller and all these at 150 HZ! I am happy with this result. BTW what are yours main loop cycle times!? All additional sensors will be connected to a secondary microcontroller (these are 2D compass, altimeter, speed sensor, 3D IR thermopiles, RPM hall sensor, ultrasonic sensor, 2 servo camera control, GPS)

Regards to all!

tve

Hi bizart,

Sorry to hear about your loss of data.... I put your video here

http://hanfordsite.com/files/bizart/avionics.avi

I only saw the one so far... I will put any others that show up in my inbox in the same place...

I have been trying several things like I mentioned before. I can get the PID tuned to the point of being very stable except at high frequencies. I'm starting to think that maybe the problem is that I have dedicated so much processor to sampling and integrating the sensors that I might not have enough time dedicated to the PID updates. If that is the problem (I don't think it is), then I will probably need multiple processors as you are doing bizart. If that turns out to not be the problem, I'm thinking it might be some kind of phase cancellation type of thing... and I will have to have multiple (D)erivative terms sampled at different frequencies. Has anyone heard of such a thing for feedback control?

danvel,

I also thought of the optical mouse sensor a while back. AussieMechatronics Guy said that they did not have any success for the same reasons bizart mentioned. I have a link on my site that might interest you. Check the section on motion-estimation
http://www.araa.asn.au/acra/acra2002...rke-Buskey.pdf

RGM3000 gps velocity accuracy
0.1 m/s without SA, 0.05 m/s DGPS

If I were you, I would use the new unit that I posted about recently though.... You could even use it indoors!!


regards,
tve

more later....

Spaceclam

TVE, i will probably finish up your last hub tonight, PM me your address and i will ship it to you, along with one of the experimental blades.
-Clam

Spaceclam

i found the piece of antenna i cut off my fs8, and i put it on with the break free pin that the deans had, and sure enough, second flight with it, it got tangled in my motor, and luckily i used the break free pin rather than just hard soldering it or i would have broken something...
since then, i tried wrapping it around the bottom, which kept it out of the way but i am getting glitches again. i think i need an antenna tube, and to replace my antenna all together.
however, i have found an ATV and expo setting where i can fly it, but it still scoots around a bit. it performs just like it used to in that respect, but it glitched and broke out of the hover. it wont stay in one place like bizart's will, and i can probably hold it within a 12 foot circle. i think unless i find a more efficient gyro, or way to implement it, this is as good as it gets.
-Clam

ADI

Hi All,

I got my 2 way IR comms working great. I'm using a 'learning' IR remote so was able to create my own protocol, so as to save the hassle of deciphering pre-programmed remote control codes.
Now I'll be able to instantly enter new PID parameters and alter joystick sensitivity etc. without a micro re-program.

ADI

tve

Cool....
Once you get the PID right, you could fly it by IR too!

Did you have to follow any kind of standard protocol to get the learining unit to work?

ADI

No ..... I just emulated the approximate baud and sync pulse rate and 38Khz carrier of one of my TV remotes and the learning remote accepted it no problem.

ADI

tve

nice idea... I will have to remember that one

Well,
I just posted a new video http://hanfordsite.com/videos/stable3.asf

Unfortunately it is a large one... ~30Megs
I believe it demonstrates my recent progress pretty well.
Note that the heater/fan in the background is blowing air on the blades the entire video.
I swing the unit back and forth at a high rate to demonstrate acceleration, shake it, let it go, etc.
One thing that I have improved on since this video is the part where it tilts up on one side. It now stays level (or at the setpoint) regardless of how fast you push it up and down and around. I still have not figured out the high-frequency oscillation problem.... should get a chance to work on it this weekend.

Regards!
tve

mdream

I'm back !!!

I don't know if you remember me anymore, but I was with you guys from the beginning (page 2).
After getting my radioamateur lisence, breaking up my previous band, getting divorce from my ex-wife
and moving to my new house, I'm back at business with DF again. I have to say that it's been really enjoyable
pleasure to behold the improvements seen on this forum.

Great fellow ADI has kept me updated about the latest stuff when I've been too busy (to argue with my ex-wife to
do or not to do DF or anything else, hihi) to check out your improvements.
(Big thanks ADI).

And now, from words to acts: Have any of you used PIC processor for DF (ex. PIC16F877).

Cheers,

Janne

kwaiweng

I'm currently using two PIC16F876A to control the DF... The source code is written in C languange and I'm using the PICC Lite compiler

danvel

Nice vid, tve. But if you listen carefully to the audio your DF is saying SET ME FREE! I want to fly away!

ADI

Tve ..... I agree with Danvel. Why don't you let the dog off his chain ? It's cruel to keep em tied down. We need to see your DF flyyyy..yyyy ...yyyyy.
That was a big download, but worth the wait. DF is lookin' really stable.

ADI

Becker

just wondering, what is that 'leash'? is it just power and a serial cable? where is the power comming from?

tve

danvel, ADI
Thanks for the comments guys....

I agree, it does want to fly...

The problem if I let it go in the wild is that it will find that one frequency of oscillation and the results would be very bad

Thanks for downloading that huge video and commenting. Next video will be the thing flying free....


>>
just wondering, what is that 'leash'? is it just power and a serial cable? where is the power comming from?

mikeydude, yes that is the serial cable and power cable. I have a 600watt power supply to use for development. When I actually get to an autonomous flight, it will have a 2600mah 3 cell LiPoly on the bottom of it. I the case of the larger heli, it will be 4 batteries.

Spaceclam

you know your DFs response to control and stabilization will be VERY different with the extra weight of the batteries, right?

tve

Yes,
but I'm getting quite quick at tuning the PID. You can't tell from this video, but it will not be a problem in the slightest...

The work I'm doing right now is not really tied to any particular configuration... it applies more to the general reponse of the controller. I can tune the PID so that it doesn't deviate at all from the setpoint, but then my oscillation problem happens almost imediately if I am not dampening the response by hand. Once I get that figured out (going to work on it right now), it should be very easy to tune the system in a few seconds for different configurations. I'll get some more video to show you what I mean...

Spaceclam

well, i wrecked my right boom today. this thing is very difficult to control. i think i am going to have go buy new gearboxes, and make it with long, square tubing, so as to reduce my control amounts, to allow me to increase my ATV, since small amounts are very difficult to get accurately.

sorry for the off topic question, but do any of you electronics engineers have a small schmatic for a tachometer with a digital readout? i need it to read from 0-2500 rpm, using a magnet or something, not a light sensor. thanks!

tve

I think I have figured out my problem. It is the response time of the motor controllers. Now.... do I go back to my home-brew controllers (I really don't feel like dealing with Microchip pics anymore) or start working on another motor controller design... or try castle controllers and see if they are faster... or I guess pitch control is the other option..
250ms is not a fast enough response for my needs. I probably need something closer to 25ms...

Spaceclam, I'll see if I can find something for you on the schematic a little later...

tve


[edit]

Spaceclam, what are your thoughts on this? Would it produce similar thrust to the Todd's models props?
http://www.hobby-lobby.com/evp-pro.htm

Spaceclam

i suppose it could, but it will not be as efficient. that blade is made so that it will opperate ore or less decently at all blade angles, and produce the same amount of thrust at 5- degrees as +5 degrees. that also requires a hollow shaft. the only difference between mine and theirs is the pitch slider and hte blades i am using. that-s pretty much it. i just need longer booms i guess, so i can increase servo travel.

BTW, i just found this:
http://www.shumatech.com/dro-350/add-ons.htm
and i was wondering if a small circuit could be built to interperate the signal into a digital readout. however, because i would like a bit more precision at lower speeds for things like power tapping, rather than using just one piece of tape, i would probably use like 4 or 6 spaced out or something, so the output frequency would be 6x faster, so that would have to be taken into consideration. while this is not a magnet design, if i were to cover the sides of the IR sensor so that the only light between it and the tape or the spindle is the IR light it shines, that would work fine for me i think. your guys's thoughts?

tve

I think it would work,
here is a counter/display that you could use... look at page 4 for the tach setup. You could use the output from your circuit and scale it to rpms...

http://rocky.digikey.com/WebLib/Red%...0Data/MDMU.pdf


Digi-Key Part Number RLCU00-ND
35.00000 Price

Spaceclam

how would i do this, and how would i hook it up?
also, unfortunatly, the voltage i have to work with is 120v. if i were to tap of a seperate line, and use resistors of the correct impedance for the amps this will draw to reduce the voltage, will that work? this shoud not draw much current...
also, that has a timer and all kinds of functions. all i need is a simple numerical readout that i can attach.

tve

Spaceclam,
35 bucks for a device like that isn't too bad... if you only have 120v, then you will need to step that down and convert to dc, or run the tach using batteries (not a bad idea... especially with lcd) send me a private e-mail. I could help you with this for making those hubs...