Printable View
Took me couple of months to make it.
It's controlled by a PIC16F877
3 Axis, 4 Motors
Precision: Logical: 1000'th of an inch, Actual: probably 10'th of an inch :D
The code for the microcontroller I wrote it in MikroC, about 500 lines, and overclocked it to 24MHz :) but still not fast enough :(
The code for the computer I wrote it in VB6, about 600 lines of code.
Here are the pics:
Well done!! I like these kind of projects.
That is brilliant! :eek: It rates up there with Wossy's LCD display.
How did you get the COMM controller working? :confused: I don't have an idea of what goes on with this electronic circuitry stuff. I started trying out the PIC programming, though I didn't get far with it. :(
That's a great CNC. I tried making a CNC just like that once, I only got as far as buying a pen.
Really nice!
No it doesn't. That plotter is about 700 times better than my poxy LCD experiment.Quote:
Originally Posted by BillGeek
Nice work CVM.:thumb::afrog:
I see no difference between the success of this and the LCD, can you write out "Kregg is an Ubuntard" :D Then we can compare.
Seriously, good work CSM
The amount of work CVMichael has done looks massively more involved than my project, I didn't have to program a PIC chip either, that is no mean-feat in itself. I'm filled with awe at that plotter's greatness.
That's awsome CVMichael!
How do you learn to do stuff like that?
Edit: PS. Can we get a video of it in action? :D
shweet!
I will answer more of your questions when I get home (I'm at work now).
I have a video with it, but it's 3:49 minutes (3 min, 49 sec), and 680 MBytes :eek: I have it on my home web-site. The movie is HD quality (30 Fps, 1280x720, Camera: Lumix TZ5), that's why it's so big.Quote:
Originally Posted by Atheist
I can give you a link to it, but I DON'T want to make it public because it will KILL my bandwidth because of it's size.
Send me a PM if you want to see the movie.
PS. The movie is MOV type, how can I compress it ?
Maybe I should compress it first, before I give the link to anyone... *thinking* ... probably I should do that....
Try something like RadTools to compress it, then upload to youtube so your bandwidth isn't hit.
I uploaded on YouTube, but I get "This video is not yet processed."
How long does it take to process ?
Well, there are many things you need to know to do it. And I learned a LOT of things by doing it.Quote:
Originally Posted by Atheist
I started to learn electronics when I was 11 years old. It was my hobby. I was one of those quiet shy guys, that everyone picked on (I don't want to go into details on that). But let's say that I spent a lot of time in house, playing with my electonics, figuring thigs out by trial and error, there was no one to teach me. When I was 16, I came to Canada, where I started right away to learn programming in C and VB as a hobby, and also at school. At school I was basically teaching the teacher, and in college I jumped 2 levels. I did programming classes from semester 4 in second semester, and it was still easy for me.
Mid last year, I realized that I did not do any electronics for like 10 years, and I missed that sooo much. So I set myself a bigger task than I should have: To make a robot.
I have the right combinations of skills to do it: electronics, programming, and some hardware (but that's not that hard to learn).
I started, and not long after I realized that many parts that I needed, have to be custom made. My father works with CNCs at his work, but I could not rely on him (every time I tried it took way to long to get what I wanted). So I dropped the robot (temporarily), and set myself to make my own CNC so I can use it to make the parts for the Robot.
Last year, I did not even know what a Microcontroller was, now I am in love with them :D I bought the full MikroC, and also bought a Programmer (ICD2). It was a very difficult start, but once I learned the basics, everything else is easy.
So... that's my story in short.
Anyways... By the time I finished writing this message, the one YouTube also finished processing.
So here it is !
http://www.youtube.com/watch?v=izudmpqHE74
What do you think ??
Very impressive :)
Can it draw micky mouse?
Not yet, I still have to change the program to draw curves. I guess I could draw curves by drawing small lines.
I still have to do some more programming to make it work the way I want it too...
I guess the next step after that is to make it interpret diferent file formats that I give it, and convert that into my own code.
So, in terms of programming, there is still a lot to do.
[edit]
Do you have a picture of micky mouse that can be easily interpreted to lines & coordinates ?
[edit 2]
This one looks achievable (I think):
http://www.howtodrawguide.com/wp-con...ite-papers.jpg
So it works with the coordinates that you supply it? I was wondering how you were doing it, then i saw the program in the video. Makes sense.
i was thinking something simpler like this:
http://people.mozilla.com/~faaborg/f...onicMickey.png
Here is the actual code to draw the "HELLO"
Code:Private Sub cmdProgramTest2_Click()
CNC.Speed(X_Axis) = 15
CNC.Speed(Y_Axis) = 15
CNC.Speed(Z_Axis) = 11
MakeMove X_Axis, 40000 - CNC.X_Pos
MakeMove Y_Axis, 7000 - CNC.Y_Pos
MakeMove Z_Axis, 30000 - CNC.Z_Pos
CNC.Speed(X_Axis) = 12
CNC.Speed(Y_Axis) = 15
CNC.Speed(Z_Axis) = 10
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove Z_Axis, 8000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, -4000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove X_Axis, 4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, 4000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove Z_Axis, -8000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove X_Axis, 2000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove Z_Axis, 8000
MakeMove X_Axis, 4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, -4000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove X_Axis, -4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, -4000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove X_Axis, 4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove X_Axis, 2000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove Z_Axis, 8000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, -8000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove X_Axis, 4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove X_Axis, 2000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove Z_Axis, 8000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove Z_Axis, -8000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMove X_Axis, 4000
MakeMove Y_Axis, 1000 - CNC.Y_Pos
MakeMove X_Axis, 3000
MakeMove Y_Axis, 0 - CNC.Y_Pos
MakeMoveXZ -1000, 1000, 15, 15
MakeMove Z_Axis, 6000
MakeMoveXZ 1000, 1000, 15, 15
MakeMove X_Axis, 2000
MakeMoveXZ 1000, -1000, 15, 15
MakeMove Z_Axis, -6000
MakeMoveXZ -1000, -1000, 15, 15
MakeMove X_Axis, -2000
MakeMove Y_Axis, 2000 - CNC.Y_Pos
cmdZeroAllAxis_Click
End Sub
It's completely static though, how about dynamic printing?
CVM that is insane!
CNCPrint("Klegg is a ubuntard!");
Yea, I know... There is still a lot of programming to do...Quote:
Originally Posted by MaximilianMayrhofer
Today I will be working to add sensors to know when it reached the "end of line".
For example, If the current position is 1000, and I give the command "MakeMove Z_Axis, -1000000", right now it does not know to stop when it reached 0 (Zero). I am talking about the hardware side. So, when it's executing that command, the motor does not stop even though it reached the end and cannot go anymore.
I already have ONE working sensor on the X axis:
Attachment 70473
But I still have to install 7 more !
CNCPrint("Klegg is an ubuntard!");Quote:
Originally Posted by wossname
When I will make the "CNCPrint" function, I will pass the parameter string through a grammar & spelling check. And when it's wrong it will give an error message.
How about that ?
You could stand to do a bit of OO on that. Come up with a few primitive classes that will draw lines, scalable lines, curves, and scalable curves based on parameters suppied. Then build a series of classes that make use of the primitives to build scalable letters and symbols (font classes), and then you, too, can write "Kregg is an Ubuntard!." with simple calls.
There are 3 ways to draw / move curves (that I can think of)
One is to calculate all the points of the curve, and send the arrays of data to the microcontroller, and handle the points as sets of tiny lines. But this means you need a fast connection between the computer and mictrocontroller and also a large buffer (memory) on the microcontroller.
Second option is to make the code of the microcontroller calculate those values as it's moving, and only pass the parameters of the curve, like center points, radius, start from, end at, etc... This requires a faster microcontroller.
Third option, to pass the formula to the microcontroller along with it's variables (values). This requires even faster microcontroller, but I think the most flexible. (Imagine passing a formula for a fractal :eek: )
And of course a combination of any of the above would be even better.
Unfortunatelly, the microcontroller that I'm using right now does not have a large memory, does not have a fast connection to the computer, and it's already overclocked !
So, I think I will have to buy a faster one (and most likely I will).
How much will a faster microcontroller cost you CVM?
A faster microcontroller may not be the solution. How much faster of a PIC chip can you get? Consider offloading the task to a second chip.
This is the microcontroller that I am using now: PIC16F877, and this is the one I want to get PIC18F4550. There are even faster ones, but this one has the same # of pins, and simmilar pin configuration (input & output).
It's double the speed, and the same price as when I bought the PIC16F877. Since I bought it the PIC16F877, they doubled the price :eek:, before it was the other way around, the PIC18F4550 was more expensive...
When I bought the first PIC16F877, in the local electronics store, it was $20 (I bought one). After a few weeks I found www.microchip.com, and bought 4 of them with $20 !!! The downside is the delivery cost. So I bought other chips, and other microcontrollers also, but unfortunatelly I did not figure out how get them to work yet (I have to admit, I did not try that hard)
"offloading the task to a second chip", well the problem with that is the communication between them, and the connection between them, it takes extra processing for each chip just to "talk" with eachother, and of course extra programming to make it "talk" with eachother, and a separate program on each chip. Synchronizing them would be the most difficult part (I think).
I already thought about that, and when I tried, I found out that in order for them to communicate, I have to use a protocol called I2C. There are 2 problems with that. The protocol is a Master / Slave type of communication. The chips' hardware supports only Master. The Slave can be simulated with software (i.e. program in the chip). MikroC has functions already made for Software I2C, BUT you cannot use Interrupts when using Software I2C. And I have to use interrupts to control the motors.
So that's why I want a faster microcontroller as opposed to linking 2 (or more) together.
And in terms of price, considering what the microcontroller does (i.e. replacing hundreds of other chips through programming), $5 is not a lot at all... Of course, because of the delivery costs, I will buy a few of them, and maybe other chips to make it worth while.
Actually, to draw the curve you would use the same MakeMove(X_Axis) and MakeMove(Y_Axis) statements you are currently using, and simply vary the speed of the motor along the x or y axis at some constant rate. Simple mathematics will allow you to create any kind of curve you like just by varying the speed at the appropriate rate. Of course, I am making the assumption that you can execute an alter speed command while the motor is running.
Otherwise, you would want to try for a dense array of points.
EDIT: Of course, even if you can only execute one command, like a make move command at a time, you can vary the voltage to the motor independently and achieve the same result.
Perhaps its good to look at HPGL or Gerber to get an idea.
PIC programming is a nightmare if you're limited to using chips with a small number of IO pins. It took me 2 weeks to make a 16-pin (total) chip play "Happy Birthday" on a piezo speaker (with 8 associated flashing LEDs). What a nightmare.
Get the biggest, most pin-festooned PIC chip you can afford and then you'll have a ton of resources. Alternatively you could get an FPGA :D
I like Atmel with her AVR family more then PIC. The problem with PIC is also that they change the product line often. But this is all my opinion.
CVM, how do you make the stepper motors work?
Hi wossname,
I cannot find the exact schematic on the net right now, when I get home, I can post the exact one that I'm using.
It's basically this one, with small modifications:
http://graigroup.files.wordpress.com/2008/04/l297.jpg
Here is a thread on an electronics forum about a problem I had with another motor that I was playing with:
http://forum.allaboutcircuits.com/sh...ad.php?t=12814
So the input from that schematic goes directly to the output of the microcontoller. You can actually see this in the pictures on my CNC.
When I started to learn about microcontrollers, I looked up books, and first one I found was programming with PICs, so I invested money to buy books (2 books), and chips using PICs. Only after a while I found out about the other types of microcontrollers. And since I already started that way, I did not want to switch...
I'm still in awe of these machines you've made.
How come the vertical axis is so tall? Is that so you can mount a larger tool on it like a drill or laser or something?
My 2d pen plotter will be a pitifully small feat compared to yours but I'll provide a link to it when it's finished.
Good job.
This CNC is dismantled, I don't use it anymore... take a look at the second one I made.
The vertical axis is tall because I want to be flexible.
Mostly for the reasons you said, and so that I can make (sculpt) bigger object, not just thin and flat object, and...
Another thing I want to do with the CNC is a 3D scanner. For example I put a coffee cup on the CNC, and with a touch sensor (something like this), I can plot each point on the computer, and render the coffee cup on the computer in 3D.
For this I need a large Y axis so that I can put big objects on it...
That would be nice to see !! post it as soon as you have something. Maybe I can (or other ppl) can give you ideas on improvements?Quote:
My 2d pen plotter will be a pitifully small feat compared to yours but I'll provide a link to it when it's finished.
That scanner idea is so cool. Did you make the probe yourself?
I made an attempt to create a touch probe, and it came out pretty good (using a 5mW laser :)), but if I get the chance to buy a real probe (that I can afford), probably from e-bay, then I would rather do that... of course, then I will have to figure out how to connect the probe to my microcontroller...
I'm still actually making the parts for my machine and I'm basically improvising as it goes along. I'm trying to make the tool mount modular as far as possible. So potentially a simiple scanning tool might work.
I was wondering about using a simple electrical contact as a touch sensor. Basically move the Z axis down slowly until there's a signal and then move up again and slide over a bit. Then rinse and repeat until the area is scanned. This sounds much harder than the rest of the project put together actually!
Yea, the programming for that is tricky... (I did not do that yet). Probably a recursive function.
Your idea works only if the object you are scanning is metal.
I'm on my way home (I'll leave work in a few minutes). When I get home I will post some pics (partial pics because I dismantled half of it) of the touch probe I made. If you are interested to see ?
I'm always interested, and we'll always be separated by many hours of global rotation.
My project isn't getting finished in a hurry, so don't worry about it :) No rush.
the screw drive system is slow, you should replace it with a timing belt and a stepper motor drive system.
The timing belt drive will make the system extremely fast, but you need to use a track and a linear bearing for this. The system would be similar to a printer.
For the Y axis you can have another timing belt drive system. So basically you gave one linear motion system mounted on another to get the X and Y motions.
Mechanically its a bit complex making a belt system, the electronic for the system would also get complex as special drivers are required to run the stepper motor, but then you can find many resources on the internet for stepper motor controls.
thanks
a
Hi wossname,
As I said before, I dismantled the probe I made, I only have part of it left (the electronics of it).
I used Google Sketch to design the hardware part (today was the first time to try using Google Sketch, it took me a few hours to make that).
The idea is: A laser points down into a mirror that reflects back the laser into the middle of 4 photodiodes. The mirror is on a spring, and in the center of the spring and mirror there is a thin rod (I used a 1/32 drill bit :) glued to the back of the mirror).
When the rod touches anything (in any direction), the mirror moves, therefore moving the laser pointer towards one (or 2) of the photodiodes. There is a paper on top of the photodiodes to blur the laser pointer.
The photodiodes are connected to the analog of the microcontroller PIC16F88, and also the microcontroller turns on/off the laser.
When the laser is off, it takes the ambient light from the 4 photodiodes, when the laser is on, it can calculate the exact position of the pointer by the intensity of the light on each photodiode.
The thing was soo sensitive that the lightest touch would make a big difference on the angle of the laser.
The touch probe was connected to the microcontroller of the CNC, and it was updating the "touch" about 20 times per second (so the laser was pulsating when it was reading the "touch"). Sadly, I did not get to the part to program the CNC microcontroller to move the motors and do the scan.
OK, since I had the camera in my hand, I took 2 more pics.
One is the current state of my CNC. I removed the top so you can see what's under.
So, the updates (since I made the video on YouTube) are:
I changed the stepper drivers to these: http://www.motiontek.ca/StepperDriver.html, two D42 for X & Y axis, and one D56 for Z axis (the one that moves the whole gantry.
On the right/down you can see a round wheel, that is the manual jog. I am using a small stepper motor (in reverse), when I spin the wheel it generates a signal that is amplified, and goes to the microcontroller (PIC16F88) to detect how many steps were moved and what direction.
Above that you can see a keypad, with 2 displays. One displays the X,Y,Z coordinates in Inch, and the other coordinates in steps (6400 steps/inch). The keypad is used to tell the micro what axis to move (X or Y or Z) when I do the manual jog, and what speed (multiplier). So I could use a 1 to 1, meaning one rotation of the jog it means one rotation of the cnc stepper motor, or I could multiply up to 8, so one rotation of jog means 8 rotations of cnc motor (which also equals to half an inch).
In the other picture you can see what I am working on in the last few days... Yes... that is a ultrasonic range finder. It sends a pulse at 40KHz, waits for the reply, and calculates distance by the time it took for the echo to come back. I am still working on this, as you can see the logic port has many connection on it :), meaning I have problems with it... but I will figure it out soon.
@ Michael,
Did you ever finish the vb.net app (whatever language it was) to control the CNC?
Unfortunately no...
Last year, in October (I think), my C drive just died. I lost a lot of stuff I had on it, including the code for my CNC. I did not have a backup of the .NET code, but I did have an old backup of the C code for the microcontroller. Unfortunately I did not backup everything I had on it...
This is the reason why I did not do anything new for my CNC since then... I will have to re-build the software, and it's not easy.
Since I have to start from scratch, I will do another re-design. I will have it work like a regular (standard) CNC, so that I can use free (or cheap) software that already does this. Then I will modify the electronics side (flip a switch) for when I want to do manual stuff (like the manual jog) or other things I want to add, that the standard software does not have.