PBP projects for R/C models

[Kenjones1935]

I'm not getting my analog voltage comparator act together.

I want the teacher, at race time, to certify that each car will respond correctly to the bright lights at the beginning and end of the maze.
The photoresister R/C to PIC toggling threshold should be calibrated per car.

The PICkit2 comes with a built in pot attached to pin RA0. Each car while being exposed to the bright light could have its pot turned with a long handled screw driver (avoiding shadows) to the point that the car's control toggles.

This translates to a voltage comparison between the input of RA0 and RA1. The photoresister 10807 reduces its resistance in bright light from 5k down to near 500 ohms If I build a voltage divider with a 5k resistor to Vss = 5V and the 10807 to Gnd and attach it to RA1 will that do the trick? It would give the PIC a swing of about 2V not figuring the input impedance of the PIC.

I've got a some of questions.

First, does the above make sense?
Second, do any of you know some PBP code that does a similar comparison?
Three, I believe I will need to smooth out noise during the transition as the car comes into and leaves the bright light? How would you suggest I do that?

Ken

[mackrackit]

First, does the above make sense?

Yes.

Second, do any of you know some PBP code that does a similar comparison?

I do not have any code but here is how I might try setting it up.
First add a switch on the car and code for the switch at power-up. Also add a LED for calibration indicator.
If switch is HIGH then ADC_SETUP routine. The code will go here once and never loop back to it.
Inside the SET_UP routine have the drive wheels turned off and sub loop reading the ADC from the light and the "calibration" pot.
Have it coded so when the reading from the light and pot are equal with in a certain amount (exact will be difficult) the calibration LED come on.
Then when the calibration switch is turned off the code will go to EXIT_SETUP. This routine will save the ADC reading from the pot to EEPROM then goto the main program.
Now the main program will read the EEPROM to get the value to use for light checking. The reason for EEPROM writing is you will only have to "setup" once and not have to do the setup every time the car is turned on. Maze changes do another SET_UP.

Three, I believe I will need to smooth out noise during the transition as the car comes into and leaves the bright light? How would you suggest I do that?

Turn to the students and say "SSHHH"

Are you meaning light noise? If so this is where the ADC needs to be with in a certain amount so a small range of "bright" light will do the triggering. This partwill take some experimentation for car speed and othe on site variables.

[Kenjones1935]

I took the car to a TV studio today. It was going to be on Access television. I brought the laptop computer and the USB connector. I forgot the battery charger. The car had not been charged for a couple of days. Big mistake.

At the beginning it did just as I thought I had told it, but then as time went on it crashed more, got stuck more. I chose not to do the TV shoot. It sort of worked, but not quite.

I think as the battery discharges, even with the 5volt voltage regulator in the electronic speed control box, everything degrades before it finally refuses to move. I know the DC wheel power motor uses lots of juice

Sound correct to you all?

Ken

[mackrackit]

Sounds correct to me. When the voltage drops below a certain point all kinds of strange things can happen. The PIC ® is probably running "OK" at 8Mhz, but the 5 volt peripherals are probably not happy.

[malc-c]

To be honest I would never use the same power pack to power the electronics. Even when using electric helicopters I would use an RC pack for the receiver, with the main flight pack used just for driving the main motor. 4 x 1.2v Nicads provides a decent supply for the electronics

[Kenjones1935]

I too initially designed with two battery packs. One for the servos and the other for the electronics. My problem is one of packaging. Getting a cheap kit that fits easily into a HPI Racing Touring Car Ready to Race package is way beyond my expertise. Why the HPI car? That is what I have. Also Ten80 Education

( Which I know as http://www.ten80eduction.com)

specializes in using radio control cars for curriculum and lesson plans supporting STEM in our schools. One of that company's sponsors is HPI Racing. I hope to get ten80 support. One battery is much easier to mount than two.

I took my car again for a run. It is using the "there is nothing within four feet of me, I must be under radio control" toggling algorithm. I liked the effect. It is important that the room have lots of room. Clearly a gymnasium fits that criteria. Adding the photoresistor system is more than I can deal with at this time.

Two projects are looming. 1. Get help and advice on packaging (surface mount prototyping). 2. Create a video story board then shoot a video showing what I have built, why I built it, how to modify the code and why I think middle school students will jump to learn.

I discovered a neat trick today. The studio has a small push cart with roller skate sized wheels. I put my car on the cart (back end facing me - it has no sonar sensor) and pushed it around the maze. I could easily see the reaction of the car to each position. It made very clear that the problems are speed, momentum and reaction time.

Ken

[Kenjones1935]

Today after watching my car behave very differently when actually running on the floor as compared with its wheels spinning freely on a pedestal I decided that I need to install the second battery pack and its 5volt regulator. Nuts, more micro-soldering.

Ken

[Kenjones1935]

Please suggest your ideas on competitions that could be implemented with my PIC and/or RC controlled hobby level car. Right now all it does is go through a left turning maze. Programming for both right and left turns is the most obvious extension.

What if I added different sensors and maybe some kind of emitter. I am thinking of a game of tag. Somehow the rest of the cars in the group would need to know which car is "IT".

What fun could be had with a group of these cars?

Ken

[Kenjones1935]

I have installed a second battery pack -- four AA batteries now power the PIC and all the electronics except the radio receiver. The original 7.2 volt NiCad battery power only the servo, the electronic speed control and the radio receiver.

The inconsistent behavior in autonomous control may have been because of noise, but then it could be because of slow reaction time relative to speed travel. Not sure yet.

I need suggestions on how to prototype this package. There appears to be shops that offer short run PC cards. PCB-POOL advertises in the SERVO magazine. You all have any preferences?

Ken

[mackrackit]

I have used these folks.
http://www.expresspcb.com/
Only thing I do not like is you have to use there software. But very happy with the three boards for $51.00. Nice work and quick turn-around.

[Kenjones1935]

Mucho frustrations followed by some successes. The PC board proto challenge is beyond me for now. Created instead a car that toggles itself between radio and PIC control without using a third radio channel and negotiates a maze that includes both right and left turns.

The auto toggle is a function of the car being out in the clear - not in or too near the cardboard maze. The right and left turns were pretty obvious after numerous stupidities.

I hope to post a video next week.

Thanks again for all your support.

Ken

[Kenjones1935]

I think the code is pretty solid. Now the problem is response time, size of maze passageways and RC car speed. A fraction of a second does not seem very long when the car is up on blocks, but it is enough time at speed to crash into a far wall when it should have detected a turning opportunity and done its thing.

I think I need:
1. Techniques using PICBASIC Pro for recording data in memory that can be read back into my PC for analysis. How do I do this? I have not the faintest idea what app to use.

[mackrackit]

This is a tough one.

If the trouble is speed and timing then using more MCU cycles might make it worse.

You can look at writing data to the on-board EEPROM.
http://melabs.com/resources/samples/pbp/eeword.htm

[ardhuru]

I think I need:
1. Techniques using PICBASIC Pro for recording data in memory that can be read back into my PC for analysis. How do I do this? I have not the faintest idea what app to use.

This thread might help:

http://www.picbasic.co.uk/forum/arch...p/t-13317.html

Regards,

Anand

[Kenjones1935]

Thanks for your attention and replies.

I think the PIC is plenty fast enough. I think the SFR05's are responding quickly enough judging from their flashing LED's. My guesses are either SFR inconsistencies and/or weakness in my jury rigged PWM signal.

The servo and the speed control want PWM pulses at 50 per second. I have not been able to do that with my PIC. I get the pulse width correct, but I can not get the rep rate down to once each 20 millisec.

Ken

[Kenjones1935]

The PIC KIT2 programmer has, at the bottom, reference to EEPROM. It has addresses 00 - FF.

I think that if I use the PBP WRITE command at run time I can store data in the inchip EEPROM space. 00 to FF in size. I can then manually read this out with PIC KIT2 programmer.

Sound correct??

Ken

[mackrackit]

Sound correct??

Yup, that will work.

[Kenjones1935]

I need to be able to connect my PICkit 2 to three SRF05 sonar sensors. The easiest way would be to solder J or JR (or whatever - I do not know one connector from another) headers to the board.

I looked carefully at the connectors that came with my radio control car. The pin separation does not match the hole separations in my PICkit 2 forty four pin demo board. Do any of you know the name of the connector header - if it exists - that would solder directly into that row of holes along the sides of my proto card? They are definitely not 1/10 inch centers. I have no means to measure their separation accurately.

Ken

[Byte_Butcher]

They're on 2mm (0.079") centers.

Digikey has a whole variety of them. Here's one flavor with 12 pins, right angle.

http://search.digikey.com/scripts/Dk...me=S5802-12-ND

Steve

[flipper_md]

on the power topic, I've read that most current electric R/C plane systems uses a BEC to run all servos/receiver from the main battery pack(Mostly LiPo these days).
BEC usually have a safety function that will choke power to the motor, but keep servos alive in case that you are running out of juice.

[Kenjones1935]

I'll catch you all up a bit.

I have changed tack.

I am trying to go from a hobby level car with PWM controls to a toy level car most of which have simple DC wheel and servo controls. The purpose is to reduce the cost of kitting for school students. In addition to reducing the cost, DC control of two wheel drive makes it easier to implement a sporty racing-like skid steering.

I purchased (at K-MART) two 1/10 scale toy cars on sale each for $25.95. Their original prices were $50. The 1/10 size gives me more room for adding microchip electronics. The wheels and the steering if driven simply by DC with reversing current are easily controlled using DPDT relays controlled via current drivers by my PIC .

WRONG! The steering servo contains both a DC motor and an angle measuring device which is read by the electronics that comes with the car. A total of six wires go to this servo. Where can I read how this works? Should I just go back to the store and buy 'proper' toy level cars? Your suggestions would be greatly appreciated.

Thank you thank you for the oscilloscope. Without it I would not have the faintest.

Ken

[mackrackit]

A total of six wires go to this servo.

That sounds like a stepper.
A picture of the setup might help...

[Kenjones1935]

Toggling between autonomous control and radio control is much more complicated for the toy level car than for the hobby level car. Why? Toy level drives the steering servo and the wheel motor each with DC in both directions. This requires four wires. There is no common ground. Hobby level cars use pulse width modulation signals for these two jobs. Only two wires are required. The grounds are common.

Toggling between PIC and R/C in toy level requires switching eight wires - four from the radio receiver and four from the PIC. The same operation on PWM wires requires switching only two wires.

Hmmmmm...... That's a pretty good argument for sticking to the expensive machines.

Ken

[Kenjones1935]

I have some Radio Shack solderless prototyping boards. With 22 guage solid wire I can create a fairly robust kit. I need to connect ten signal wires from the PICkit 2 to the solderless board. The Digi-Key S5752-12-ND header might work. The question is whether 22 gauge solid wire will fit and stay in this device. If not #22, then what gauge will work? Or is there a crimp-on male connector that will do the job.

What did the Microchip engineers have in mind when they designed the PICkit 2 printed circuit board? Any ideas?

Ken

[mackrackit]

22 gauge will be snug and if you do your breadboarding like Darrel does it all should stay put.
http://www.picbasic.co.uk/forum/atta...2&d=1278267329

[ScaleRobotics]

What did the Microchip engineers have in mind when they designed the PICkit 2 printed circuit board? Any ideas?

A compact USB device that programs all their MCU's, using 5 to 6 connections?

But what do you mean Ken?

On a tangent, here is what they were thinking for the pickit3:

[mackrackit]

A compact USB device that programs all their MCU's, using 5 to 6 connections?

I think Ken was getting at the 44 pin dev board. It is kind of useless... Mine got the most use helping Ken get started. It is back to collecting dust.

[Kenjones1935]

This describes what I found when I took the "servo" apart. (I got it back together with much struggling so this is from memory.)

It has 2 wires that go to the motor and 4 wires that go to some kind of rotating position sensor. On a small board are 2 etched concentric circles which are in contact with a 4-pronged device mated to the spinning gear. I think 2 of
the 4 prongs touch the outer circle, the other two the inner circle.

Have any of you come across such a device? Any idea its purpose and how I control it. Other cheap cars just go bang bang into the steering motor with a little spring plus the caster effect to bring it back to neutral.

Ken

[mackrackit]

Now I remember seeing one of those.
It is a reversing/limit switch. The cheaper cars are either turn or no turn. Not much in between.

If I remember correctly...

[Kenjones1935]

22 gauge will be snug and if you do your breadboarding like Darrel does it all should stay put.
http://www.picbasic.co.uk/forum/atta...2&d=1278267329

My 16F887 is attached to a Microchip PC board. It is the PICkit #2. My problem is getting ten signal wires off that board an onto a solderless proto board like Darrel used. Are you all suggesting that I do the whole job on a solderless board? If I were to do that, where do I get the solderless header that fits my Microchip provided USB connector? I'm pretty sure my 44 pin square shaped PIC will not fit. I need the long rectangle shape.

Ken

[mackrackit]

I use these with bread boards and the PicKit2
http://www.allelectronics.com/make-a...-ANGLE//1.html

The long rectangular chips are what they call a PDIP package. Much easier to solder when you get to that point.

I would develop the project on a bread board with PDIPs and that might be the way to go with the kit also. It would low the students to experiment...

[Kenjones1935]

Looks like all I need to connect to a PDIP 16F887 on a solderless board are the six wires from the USB device and two .1uf caps between Vcc and Gnd. The rest of the Microchip PICkit 2 schematic I do not use. (Some LED's would be nice.) Don't forget two isolated batteries. One for the DC motors the other for the logic.

This is little considering the six Zettler relays, the SN7407, and a 5v regulator. Hmmmm

Does this seem correct to you?

Ken

[Kenjones1935]

Power, Ground, Trigger and Echo to/from three sonars. Interconnections with the radio receiver and the wires that go to the DC motors and/or the Electronic Speed Control.

If this could all be on one board and the board could fit inside a little car.....

Ken

[mackrackit]

You may also want to look at the 16F886. Same chip as far as programming goes, less I/Os smaller foot print, 28 pins. Same data sheet as the 887.

The same outfit with the headers has a 400 hole bread board that can snap together with others if you need more room.
http://www.allelectronics.com/make-a...dboards/1.html

[Kenjones1935]

Here is a proto board that will carry all my parts and then some. Each board has two rows of 62 pairs of bussed 0.1" centers pin holes.

http://www.circuitspecialists.com/prod.itml/icOid/6888

I need 20 paired pins for the forty pin 16F887, 48 paired pins for the 6 sixteen pin DPDT relays, 7 paired pins for the fourteen pin SN7407, 6 isolated pins for the USB device and three isolated pins for the 5v regulator.

The board is 6 and 1/2 by 3 and 1/8 inches. That fits fine on my 1/10 scale hobby level SPRINT (the belt up the middle driving the front wheels is a problem that can be overcome) and my toy level 1.10 and 1.12 scale toy cars.

Is that a plan, or what?

Ken

[Kenjones1935]

New unexpected problem.

I have not been able to find socket headers that are compatible with the solderless proto board technology. All the headers are designed for PC use. This makes it very difficult to connect my six pin 0.1" centers USB connector device. It also makes it difficult to connect the servo PWM three pin cord, the Electronic Speed Control three pin cord, and the connectors needed for the radio receiver three pin male headers.

I could cut and strip six 1/2" #22 lengths of solid #22 gauge wire and stick then into consecutive holes. The female connector would come in perpendicular to the board.
This is not as reliable as 90 degree square header stock, but what choice do I have?
Remember this is for middle school students.

Suggestions anyone?

ken

[mackrackit]

Are you saying the breadboard you have is not 0.1 centers?
Up in post 311 I gave a link to a place with headers with 0.1 centers.

[Kenjones1935]

The 90 degree headers do not have long enough stems. They do not reach to the springs in the proto board holes. I solved this by bending them straight, but I would rather have the connectors coming in horizontal.

BIG REAL PROBLEM!! My PBP 2.6 version DEFINE pulsin_max
command does nothing. I can see this from the speed at which the sonar LED's blink. I am using a PULSIN command to detect whether the radio transmitter is powered on. If no PWM signal comes in, the power is off and the PIC is on its own to guide the car. This is my new TOGGLE system, but it does not work. When power is off the LED's flash about once per second no matter what I set pulsin_max to be. When power is on they flash many times per second.

Any rumors around campus to which I am not privy?

Ken

[Kenjones1935]

pulsin_max does not cut short the time used by PULSIN? It makes PULSIN ignore everything that happens after pulsin_max duration but the next command in line is not implemented until PULSIN has gone through all its 65,535 ticks.

If this is the case, what can I do to detect within, say, 25 millisec something that happens every 20 millisec or not at all?

Ken

[mackrackit]

I think COUNT might work better in this case...

Those 90 degree headers should be long enough when the plastic that holds them together is slide close to the bend. Works for me that way.