Printable View
If the photocells don't respond fast enough, I wonder if using a photodiode, prebiased with a little bit of voltage, connected to an A/D input, continuously sampled would respond fast.Quote:
Originally Posted by dragons_fire
It would be neat if I could get pictures like that, but I live in a bit of a valley with trees all around. Not much of a horizontal view unless I get up about 100ft...at least I get deer walking thru the yard every morning and evening.
yeah, there not bad. i would like to take better pics, but i like being in the safety of my house.. when we get thunderstorms, we usually get some hail with it.. i would like to go out in my sidekick to somewhere good, but im afraid of the hail going through my soft top.
i think i finally got everything with the 6585 right now.. thanks to everyone for all the help.. im just looking through old threads trying to figure out how to get rid of the "warning[230]" config message...
An upgrade to PBP 2.47 should get rid of it from what I've heard.Quote:
Originally Posted by dragons_fire
I'm still running 2.46 (no need for the upgrade yet and I'm a bit lazy), I just ignore the errors.
i just went in and edited the inc file... added "w = -230"
no more errors..
hopefully i will have a new PIC by wednesday, so i can put this all together and maybe get it moving...
Are you just going to 'fix' the old board or what?Quote:
Originally Posted by dragons_fire
well, i found more stuff wrong with it... so i will probably make a new board... i had the memory card writer header on backwards...
ive been testing the card writer with a 877a, and its pretty neat..
Update:
i havent done too much work on it in the last couple days. the ECM in my truck went bad, so im trying to figure out if its fixable, or if i need to replace it... its a 91 Suzuki Sidekick, adn all the electronics and half the engine parts are made by Mitsubishi.
i have the Rogue Robotics UMMC SD card reader/writer working great now. it just takes serial data and stores it to a card. i can store a "label" and a VAR Byte to it for each one of my variables for this robot.. with a 1gig card, i can fit almost 100,000,000 logs on it and then stick the card into my computer and read it all in notepad. i dont have an actual use for this yet (for this robot, i can think of many uses for other things), but i think its going to be neat to look at everything on my computer..
my new PIC and other parts came today, so i will make the board over the next couple days, and hopefully it will work good this time...
If you feel like showing the new board again before you etch it...
I'm up for another looksy.
<br>
Definetly let us all have a 2nd/3rd/4th look for you (not that you aren't competent or anything, but it couldn't hurt at all). This is too neat of a project to waste any more pieces/parts on. I wanna see something work! :)Quote:
Originally Posted by Darrel Taylor
well here it is... hopefully it works this time.. i have changed a lot around since the last time i posted it...
http://i163.photobucket.com/albums/t...o/scan0037.jpg
thanks fof the help guys, and i will keep everyone updated on how it goes...
Nice looking board...a couple notes...Quote:
Originally Posted by dragons_fire
--Add in vias on any signal/power line you can fit them on, whether you drill them out or not. That way you can drill the via later and plug in extra wiring or make mod's down the road, goes back to cutting traces and making fixes and stuff like that.
--I might get rid of that big ol' ground plane square on the bottom, unless you plan on coating the board with something (a rattle can of acrylic is a good thing I've found out recently and it scrapes off easily with rubbing alcohol and a cotton swab after it's dried). That plane to me looks like a perfect spot to short something out!
--Very few 90 degree corners, very few corners greater than 45 degrees, decent ground run, I might add a few extra power runs on the bottom (and/or thicken up the power traces you've already got) to help the power get around a bit better. The setup you've got is almost a daisy-chain. Optimally, you'd almost want all the power traces to come back to a central point (same for ground, but you're good there).
a very little suggestion. Usually the power rails should be 2-3 X bigger than the signal one.
unless, it looks good, nice job. Don't give up
think you told me all that before, and i just forgot..Quote:
Originally Posted by skimask
i actually have been using "tremclad" clear on the boards after their etched. then i can drill through it, and it works like a solder mask. it seems to take a little more heat, but im mostly soldering headers and sockets, so it shouldnt make too much of a difference..
the ground plane on the top of the board, is right under the SD card reader. so i put it there to kinda block it from anything..
Good call on both points...Quote:
Originally Posted by dragons_fire
Like Mr_E said about the power traces.
NICE JOB!
Just concentrating on the Power&Ground right now.
<img src=http://www.picbasic.co.uk/forum/attachment.php?attachmentid=1578&stc=1&d=117761534 8>
mmm nice catch!
I think the PING's backwards too.
<img src=http://www.picbasic.co.uk/forum/attachment.php?attachmentid=1579&stc=1&d=117761982 8>
<br>
i have never had this many problems with my boards... i have fixed everything that was mentioned... i took out that cap, increased the sizes of power and ground traces (and made them go all around the board), and i flipped the "ping" around too..
i dont know how i got that one backwards, it was fine all the other times i designed this...
thanks guys....
another update... im still working on trying to fix the computer in my car. its made by mitsubishi, adn apparently the caps go bad, and need to be changed.
i made a new board, and it all seems good. i think the old PIC is still good too (im not sure though). i tried to flash an LED on RF0, but its not working.
i can program fine, and then i can read the code from the PIC and its all the same. i think im just forgetting something somewhere.. and remember, im still somewhat new to this...
DEFINE OSC 20
define xt_OSC
CMCON = 7
ADCON1 = 7
start:
high portf.0
Pause 1000
low portf.0
pause 1000
GoTo start
End
yeah i already repaired tons of those beauties awhile back, and yes they had Caps failure... but some has Microcontroller failure as well. Maybe you could have a look at the following which explain it to you better than i could :D
http://www.ecmtogo.com/ECMTips2.htm
And yes, they also have plenty faulty contact/wiring problem too... good old time :D
-------------------------------------
Your config fuse setting is wrong, i suggest you open the .INC file in the PBP folder to see how they set the default one AND to have a look at the following
http://www.picbasic.co.uk/forum/showthread.php?t=543
Make sure you really disable ALL analog stuff with your current settings AND that you disable the LVP mode in your config fuses.
thanks mister e, caps are cheap, so its worth a try..
and the config fuses were wrong.. i pretty much knew they were when i posted.. i have got it t flash leds, and to write t my serial lcd. it is also working t the motor driver, but everythign cuts out when the motors get to a certain speed. i think that maybe my batteries might not be able to handle that much drain..
gonnas work on it a bit more tonight after work, and i will post a video of it running on its own when that happens...
Tape a meter up to the top of it and drive around and watch the meter!Quote:
Originally Posted by dragons_fire
Or maybe even make use of one of your analog or comparator ports, kill the motors when the voltage gets too low, use a seperate voltage reference, blink and LED when the volts are good, no blink when volts drop off, etc.etc...(which is why I said add a load of extra vias even if you don't use them, makes it a piece of cake to add this stuff in there)
Batteries - that's what I'm thinking. If I remember right, your motors are driven off a seperate motor driver chip with a mind of it's own and probably doesn't need a full 5v to operate.
ok... so im not sure whats going on now... i think im getting some kind of interference off the motors.. if i move the board about 3" away from the chassis, everyhting works beautifully. then if i put it back near the motors, it cuts off. i havent added caps across the motor leads and case yet, but i may try that tonight..
im not sure, but when the voltage gets low enough, the PIC should reset right? and it appears from my LCD, that everything is just locking up..
it does kinda move though.. i will keep working on it, and let ya all know when i find something...
mmm, what happen if you try to use the internal OSC (if there's)
Is there a lot of unconnected I/O?
I don't remind your configuration fuses, but you could try to disable the Brow-Out detect or play with it's threshold voltage.
i have played with the fuses, and the threshold... didnt seem to help... im not sure whats causing it to cut out when the motors are close to the chip..
i will try the int osc, and see if that helps...
For S & G 's, maybe throw a couple of thicknesses of aluminum foil between the board and the motors (taped up or coated somehow of course so it doesn't short out) and ground the foil...see what happens... Might help with the EMI if there is any.Quote:
Originally Posted by dragons_fire
And if it doesn't work...at least it'll keep little green aliens from invading your Treadbot with their evil mind control devices! :D
The aluminum foil will work for aliens:) but if the problem is a magnetic one you will need something with a high magnetic permeability. Hold a compass near the motors when they are running and see what happens. Use one that you will not care if it gets reverse polarized.Quote:
Originally Posted by skimask
If this is the problem, all you should need is a thin piece of cheap steel (roof flashing) that is a little larger than the board being that the motors are underneath it.
Ive been playing with it a little bit, and if i move the board up about 3", everything works great... thats what makes me think its an interference problem instead of a "low voltage" thing... i would prefer to keep eveything as close as possible, so its not top heavy..
my sidekick still isnt working, so i have been spending most of my time on it, but when i get it going again, this robot should be driving around my house all the time...
To me that sounds more and more like a magnetic problem. A magnetic field will drop off something like an inverse square, RF does not. (some one will want to correct this):) At least if it was RF, it would most likely carry more than 3"Quote:
Originally Posted by dragons_fire
A little info http://www.lessemf.com/guidelines.pdf
so, i added a sheet of "galvanized flashing" between the board and the motors, and i bent some to make a box that fits perfectly over the PLCC socket, and it seems to help.. its still not perfect, but its a whole lot better than it was before. im going to see what i can do to shield it a little better. i also need to ad caps to the other motor, and im going to see what i can do to make some tubular shields to fit over the motors..
i have also noticed that things get worse if i ground all the shields..
i like the way its laid out right now, but i think it might help if i switch the controller and battery boards, so the PIC would be farther away from the motors. it would also lower the centre of gravity, and make the whole robot a little more stable...
i will try to play with it a little more this evening, and maybe add one more sheet under the one i already have, but idont want to raise the board too much, its already top heavy..
I forgot to mention that grounding the shields will transfer eddy currents.
You might try bending the sheet metal that is between the board and motors down over the tracks (fenders)?
With out a gauss meter the fields will be difficult to map. A compass can sometimes give you an idea.
This is a gadget I use. http://www.forcefieldmagnets.com/cat...roducts_id=100
when i first started building this, my intention was to put the battery between the 2 motors inside the red fibreglass. because i wanted lots of capacity, and a higher voltage, i added them to the top. so now i have lots of room insid the chassis.
i will probably make tubes to go around the motors, and hopefully that will help.
I might also cast a lead "brick" to fill in the space between them to make it a little more stable on angles, and for more grip. the only problem is that i will lose some endurance because the motors will need to haul more wieght.
ive been working for the last ten days, but i get the next 4 off, and its raining like crazy here, so i should get some time this weekend to play with it a bit more.
After i get the "interference" problem figured out, i will have a video up here of it driving around. i have written the code for it to speed up, drive, sense obsactles, slow down and then turn and drive that way..
It was recently (today) suggested in another thread talking about unused pins and running motors and getting interference, etc., to occassionally reaffirm your TRIS assignments for all the pins you're using.Quote:
Originally Posted by dragons_fire
In other words, every once in awhile, or whenever you can spare a cycle or two, redo all your TRIS's, keep your inputs as inputs, and your outputs as outputs.
It seemed to help this other person's problem with motors, magnetics, etc., and it's only a slight software change.
Might help you too...
well, i just noticed that thread as you were replying.. i will give it a try when i get home... the motor controller has a reset line on it that needs to be held high, so i used a pin from the pic, so i can turn it on or off if i need too, and i think hte pic is probably going low on that pin..
do i want to make all the unused (almost all of them right now) inputs or outputs?? they arent tied to ground or anything, because they will eventually be used.
and im not too worried about program space (yet), i have 24K words, and im only using about 300 words right now... i would prefer to get rid of as much magnetism and electrical noise as possible by physical means, and not just block it out with programming..
thanks for the help guys
Good idea...keep pinging that pin in your program to help force it high if need be.Quote:
Originally Posted by dragons_fire
Don't really know on that. Almost anyone of them could be causing a problem. One thing is almost for sure though, if a pin is an output, it generally can't cause an interrupt.Quote:
do i want to make all the unused (almost all of them right now) inputs or outputs?? they arent tied to ground or anything, because they will eventually be used.
But the extra programming doesn't cost anything in hardware! :)Quote:
i would prefer to get rid of as much magnetism and electrical noise as possible by physical means, and not just block it out with programming..
You could also try using a pull-up, and keeping the pin as input. If you want a reset, then in the program change it to an output, low, pause as needed, high, then back to input.Quote:
Originally Posted by dragons_fire
i dont really know whats going on... ive been playing with it for a bit this weekend, and something funny is going on..
when i hold the board about 2" above the motors, everything works fine, if its closer, it will start to speed up, and then just completely stop, and wont reset.
if i plug in my serial LCD and keep the board 2" away from the motors, they run fine, but the LCD will work for a bit, then go blank, and then work a bit more, and go blank and then work more.
with the board beside the motors, the LCd will start, go all black, and then come back and the motors will stop. then the lcd will keep displaying the speed variable changing, so somewhere in there, i think the motor reset is being tripped and its not restarting.
heres a tiny bit of my code...
baudLCD CON 32 'baud rate for lcd
lcd VAR PORTF.0 'serial lcd port
motor VAR PORTD.0 'serial motor port
motorreset VAR PORTE.7 'reset pin For motor controller
baudmotor CON 84 'baud rate for motor
LBAK CON 1 'left backwards
LFWD CON 0 'left forwards
RBAK CON 3 'right backwards
RFWD CON 2 'right forward
speed VAR BYTE
start:
High lcd
High motor
High motorreset
SerOut2 lcd, baudlcd,[254,"G",1,1,"Treadbot"] 'display treadbot on lcd in position 1,1
Pause 1000
For speed = 0 TO 80 'speed up motor to full speed
SerOut2 lcd, baudlcd,[254,"G",1,4,#speed," "] 'display speed var on lcd in position 1,4
SerOut2 motor, baudmotor, [$80, 0, lfwd, speed] 'left motor forward
SerOut2 motor, baudmotor, [$80, 0, rfwd, speed] 'right motor forward
Pause 150
Next
Pause 10000
'drive at full speed for 10 sec
For speed = 80 TO 1 STEP - 1 'slow down motor until stop
SerOut2 lcd, baudlcd,[254,"G",1,4,#speed," "] 'display speed var on LCD in position 1.4
SerOut2 motor, baudmotor, [$80, 0, lfwd, speed] 'left motor forward
SerOut2 motor, baudmotor, [$80, 0, rfwd, speed] 'right motor forward
Pause 150
Next
I built a walking (biped) robot and had trouble with motor noise causing the A/D signals to vary. (I use two A/D inputs to read pots coupled to the legs for position information.) Adding .1uf caps from each motor lead to the motor's case solved the problem.
My motors are the cheaper hobby motors used with the Tamiya double gearbox. If you are using the Pololu serial motor controller they recommend the caps too. I use an analog H bridge and still had noise issues even with seperate motor and logic batteries.
I don't know if it will help you but it is something to try....
By the way, you did an excellent job on the project, I hope it works out.
--John
thanks for the compliment John, i have added the caps across one motor, and just forgot about the other motor. i usually add one from each lead to the case, and another across both leads. its one of those things i picked up from RC cars and planes.
heres a picture of one of my motors ontop of the scroll wheel on my mouse. i cant remeber if i posted it before, but here it is again..
http://i163.photobucket.com/albums/t...o/IMG_7988.jpg
i would like to see a pic of that biped if you have any.. its one thing i want to do eventually. heres a hexapod i made, its kinda the same as the Lynxmotion hexapod.
http://i163.photobucket.com/albums/t...h_MVI_6750.jpg