01-19-2003, 10:18 PM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
Proportional Controller EMULATOR
Got a few Bit-G clones for xmas and I gotta say they're nuts!
Took 'em around to a mates place and we raced them all night around a home made track on the living room floor.
During this time all I could hear was this incessant clicking sound from the RC controllers. None of us ever held down the steering buttons for very long since every turn on the race track varied in length / sharpness, etc... So how do you control a Bit-G that doesn't have a proportional control? Click the bloody left-right pads repeatedly depending on the shape/sharpness of the corner, ouch!
This gave me an idea. If I could control the left/right steering as a series of pulses instead of a constant on / off maybe, just maybe I could EMULATE a proportional controller! You could do this for forward/reverse as well, effectively giving you speed control in both directions.
At this point let me say that all I'm trying to achieve is to EMULATE a proportional controller to the best of my ability. There is no way known given the circuitry of the transmitter & receiver to be able to achieve a fully proportional controller since it has been well established that the custom RX/TX IC's use a pulse width encoding scheme that either outputs a 0 (OFF) or 1 (ON) to the motor / steering circuitry on the receiver board.
So, with all of that in mind I came up with a design that sends a pulse to the left/right switch (on the IC side) of the controller. This pulse pretty much emulates the action of clicking on the left/right button instead of the operator of the controller doing this himself. This pulse is controlled by a potentiometer (a variable resistor, found in many computer joysticks) that is used to control left/right functions. This is what you use to steer the Bit-G.
In layman’s terms what does this mean? Well, having the dial in the middle means steering is in its default state, centred with no action. Turning the dial to the left starts a pulse that is very “thin”, this causes the wheels on the Bit-G to turn to the left for a very short amount of time. As the dial is turned to the left even further, this pulse becomes “wider”, therefore the wheels stayed turned for a longer period of time. Turning the dial all the way to the left causes this pulse to stay “on” all the time, therefore keeping the wheels permanently turned to the left. This is the same as when the left button on the controller is pressed.
This is still in heavy testing and development, I’ve been able to get this to work but so far for one direction only. I will need to modify the circuit to add the extra “channel” to be able to turn the car left AND right. The schematic below is for the left. Notice that the circuit requires 6V. I use a 3.3V Zener diode to bring the voltage down the 3V for the controller circuit. I haven’t tried this part yet since charging the Bit-G “might” blow the diode due to excessive currents.
For you tech’s out there the circuit is a basic PWM (Pulse-Width- Modulator).
|
01-19-2003, 11:01 PM
|
|
Registered User
|
|
Join Date: Dec 2002
Posts: 1,237
|
|
i had thoughts to do this but for haven't had the time to really get into it so definetly keep us posted on you progress then with a little more modifications you could put it into a joystic and have a totaly proportional bit
__________________
trust me it's easyer done than said
*** AMC Veteran ***
-- ^-_-^
|
01-20-2003, 11:09 AM
|
TinyRC Newbie
|
|
Join Date: Jan 2003
Posts: 4
|
|
This is a great idea. I had also thought of doing something like this but for the speed instead of the steering. Also for controlling both sides of the steering coils, why don't you duplicate u1d but in an inverting configuration. Then put an RC filter on the steering coils of the car. This might give you a crude analog voltage. But I'm not sure how you can drive only one coil at a time. Also I was wondering what frequency is the PWM running at?
|
01-20-2003, 03:10 PM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
That's a good idea Blue Raven, you can use this for fwd and reverse. The PWM is running at about 10hz, I wanted to duplicate the manual pressing of the button first before I up'ed the frequency on the circuit. Once I get the second channel going for steering both ways I will stuff around with the frequency to get the best result. The values of the caps and resistors aren't final, if anything I've changed them since drawing the first design.
Your suggestion about using an RC filter is an interesting one, I take it you'd be utilising the charge rate combo of the RC network to "smooth" out the square wave pulses going to the steering coils. I "think" (not sure, anyone is free to correct me) that the coils in the steering arm must have some some of alternating pulse (Possibly simulating a sine wave) going into them to be able to generate a magnetic field (since a DC current won't do this). This field alternates depending on what direction you choose. Left charges one coil, Right charges the other. An RC network might smooth out this "alternating wave" so much as to negate any magnetic field being generated and therefore killing any steering action you might have.
The problem before me at the moment is being able to use just the one potentiometer (variable resistor) to vary both left and right signals. When the pot is centred both channels need to be outputting > 1.2 Volts (this is the quivalent to not pressing either left or right buttons). As the voltage approaches zero the channel is activated. I've found that at the input to the Comparator (U1D) the voltage varies between 0.8 and 1.5 volts. Inverting the other channel is worth a try , I will give it a go and find out if it works. Otherwise I was looking at a "dual-gang" pot where one half is used for left and the other half used for right. The advantage of this is that each variable resistor on the pot is electrically isolated for the other AND i can wire one in the opposite direction of the other. So turning one way increases the resistance of one resistor whilst decreasing the resistance of the other.
*sigh* To many nights out drinking during my uni days when I should of been studying!
|
01-20-2003, 05:10 PM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
Hey Smokey, I should be so lucky!
You're probably right, it is a good idea. In testing the circuit I did notice that with the motor is was sputtering along. As the duty cylce of the pulse increases though it becomes a smooth ride.
Also increasing the frequency of the PWM will reduce the jerkiness of the motor, this is something I will try out. With this mod the motor will almost become a stepper motor at the lower end of the duty cycle of the pulse.
Keep in mind that the design is primarily for steering purposes so a clean square wave pulse (with a variable duty cycle) is sufficient for a variable steering control.
Great ideas guys, keep 'em coming.
|
01-21-2003, 01:28 AM
|
|
Poor Student
|
|
Join Date: Dec 2002
Location: Greenville, NC
Posts: 216
|
|
Well here is what i have been planning on using... since my electronic skills are on a somewhat basic level (two electronics classes a couple years ago in high school), i decided to use an already built circuit. Allelectronics.com has some pwm motor controllers for $4. They look to be small and it could be used as a means of pulsing the control buttons (taking out the microswitches and connect to the spot on the motor control board where it would open and close the circuit. I'd have to order and look at the thing to tell what exactly to do with it, but thats my basic idea.
on the same site, there are some "basic stamp" boards that can do all sorts of stuff, including pwm. I have a few from a robotics lab i did in highschool, but forgot a lot about how to use them. they retail for a tall $60
__________________
Fight Crime <--> Shoot Back
|
01-21-2003, 01:30 AM
|
|
Poor Student
|
|
Join Date: Dec 2002
Location: Greenville, NC
Posts: 216
|
|
__________________
Fight Crime <--> Shoot Back
|
01-21-2003, 05:29 AM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
Dude, not a bad suggestion, but the ciruit needs to be built from scratch. I've yet to work out how to tie both left and right steering into the same variable resistor (potentionmeter)
|
01-22-2003, 01:19 AM
|
|
Poor Student
|
|
Join Date: Dec 2002
Location: Greenville, NC
Posts: 216
|
|
f/b and l/r could be run from one basic stamp, and you could use the same pot for input on each. The function would just be inverse. I would recomend picking one up. They are fun to mess with and if you get tired of micro r/c for some reason, you can do many other cool things with them. Like automatic goldfish feeders.
__________________
Fight Crime <--> Shoot Back
|
01-22-2003, 11:32 AM
|
|
Clones Are Good
|
|
Join Date: Jan 2003
Location: St. George, Utah, USA
Posts: 123
|
|
What about robbing the POT's out of an old RC Car "wheel" type Tx? Or, better yet, use the case with the wheel and trigger. That way, you will have the POT's, return springs, plenty of roon to drop in the electronics, and a long telescopic antenna!
|
01-22-2003, 02:49 PM
|
|
Poor Student
|
|
Join Date: Dec 2002
Location: Greenville, NC
Posts: 216
|
|
hey great idea.... I have a really old non operational futaba sport
__________________
Fight Crime <--> Shoot Back
|
01-22-2003, 02:59 PM
|
|
Clones Are Good
|
|
Join Date: Jan 2003
Location: St. George, Utah, USA
Posts: 123
|
|
That is exactly what I was looking at when I got the idea.
I'm excited to hear if it works!
|
01-22-2003, 03:36 PM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
That's a great idea and one I have trying to achieve. the only bugger is that I don't have any old R/C controllers lying around to do this mod in.
Also the other issue is the value (in Ohms) of the variable resistors in these R/C controllers. Unfortunately I can't just use any variable resistor for reasons explained below.
In the circuit I've posted I noticed that the PWM cycles between 0% and 100% duty cycle over a voltage range of approx 0.8v -> 1.5v. For a really smooth control the variable resistor (potentiometer) needs to be able to adjust over this range from it's centre position to rotated all the way to the left (or right).
I'm currently working out the V=IR calculations I need to select the correct values for R6, R7 & R8(pot).
I've been stuffing around with a "daul-gang" pot which is basically two variable resistors on the same rotating shaft. With each reference wired oppposite to the other. This means that when I rotated the pot to the left to output from the comparator (U1d in the schematic diagram) that "drives" the left channel is active, BUT the right channel (with an extra comparator driving the output for the right) is off. Turning the pot to the right causes the left to become inactive and the right active.
Once I've tested and finalised this I will post the updated schematic.
|
01-22-2003, 03:47 PM
|
|
Clones Are Good
|
|
Join Date: Jan 2003
Location: St. George, Utah, USA
Posts: 123
|
|
Right. I think we should be able to use the same set-up, though. Just need to swap the POT's out for the impedence we need?
I know I'm simplifying, but I just wanted the general idea expressed.
|
01-22-2003, 07:07 PM
|
|
Chief Propeller Head
|
|
Join Date: Jan 2003
Location: Melbourne, Australia
Posts: 269
|
|
Hopefullly that will be it, just swapping the POT's to the correct values *SHOULD* do the trick.
I'm currently working on the maths involved to get some proper vaules. I've selected a 5K linear POT to use at the moment because they are widely available. I've pulled apart my old quickshot joystick and found that the X/Y axis POT's were 150K. This sux because there is too much potential (voltage across the resistor) drop which can be balanced by increasing the ohms of the other two resistors each side of the POT. Too high a resistance will create a very small current, possibly not enough to power the Comparator, dunno, not sure............
Hopefully during the next week the design will be finalised, tested and I should have some video to show the darn thing in action, hopefully!
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -5. The time now is 08:51 PM.
|
|