cnc

Ive made a few cnc, and updated an industrial one with the same gear.
Hit ebay my friend. Dont go ardy as its not quick enough imo, gives shitty results. Get a usb breakout board off ebay with usbcnc for $30-80aud, controllers for $7aud each. Easy cheap build if you already have steppers, or buy them too

Forgot to mention, dont forget your board will need seperate stepper controllers for any stepper larger than a cockroach, and a spindle controller.

Anyway your board was on ebay recently for $30.

Usb isnt always supported by software, so find what you want to use first.

Thx for a response!

>>1026779
"your board will need seperate stepper controllers for any stepper"
controllers=drivers?
I'm confused about this: if I buy usb board, I don't need anything else electronicwise except stepper drivers, power supply and pc?

>>1026779
>Usb isnt always supported by software, so find what you want to use first.
I think I would automatically disregard any software that didn't have a serial-over-USB option--especially if you had to pay for it.
The thought of buying software in 2016 that still requires a fucking parallel port is mind-boggling.

Then again, I find a number of aspects of typical DIY-CNC matters amusing.
Paying for special software to make special shapes (cams, gears, 4+ axis control),,, requiring a workflow of 3-4 separate programs just to actually design and mill a part....
All this horse-shit dates from olden moldy times, long before gigahertz PCs and $2 arduinos.

I ain't built a CNC mill or lathe yet, so you all can still say I don't know shit.
I would definitely use a arduino/usb/serial connection, definitely try writing my own CAD/CAM program, and definitely do it on Windows.

File: mdrivers_65n6kj6n.jpg (143KB, 1554x784px) Image search: [Google]

143KB, 1554x784px

>>1026952
>controllers=drivers?
>I'm confused about this: if I buy usb board, I don't need anything else electronicwise except stepper drivers, power supply and pc?
(not that guy but anyway)

to control motors with a microprocessor or breakout board, you need a motor driver board. (it is not like a "computer driver file" at all...)

you can get motor driver boards that will drive several motors, but many people just like to get single driver boards for each motor. this makes troubleshooting easier if one board fails.
Also when you get into using larger motors, that's the only way to do it,,,, each motor ends up with its own driver board

pic related:
both pictures are sized so they are about actual-size, relative to each other. the motor driver boards in both are circled.....

on the left is an electronics kit for a 3D printer, they usually support 5 motors (usually nema17 size). the motors are small and don't need much power, so the controllers are very small.

on the right is a (basic) 4-axis CNC router kit. the motors are larger (motor size is nema23) and they are under a greater physical load, so the driver boards are larger. (these are TB6560 driver boards, they run on 12 to 24 volts, they are good for any nema17 and smaller nema23 motors but may overload on big nema23's).

Also note: the CNC router kit on the right is not really "complete", some smaller bits you would need (like limit switches) aren't included. That kit just has the power supply, a controller board, 4 motor drivers and 4 motors.

>>1026956
First guy here.
Usb is often overlooked as the bandwidth wasnt there. Unless your spending big $, last i checked the available gear sub $100 was garbage, and resolution / cutting speed will suck.

As the other guy said, motor drivers basically take the signals from your breakout board, and add the power your steppers need to drive. If your running a bunch of tiny tiny steppers, it may be possible to run them direct, but usually its asking for trouble.
Imho, as with my last several and current builds, go with a mach3 compliant serial based breakout board, or mach3 usb board if your up for the cSh. Get motor drivers that put out more than your steppers need, or you will replace em later anyway.

Save cash by using an old atx power supply, google how to run them with a resistor off fleabay for $1. They make a good source for these projects.

Also, ensure your steppers or servos are compatable with your boards. Eg : # of steps per rotation, wire count and as mentioned power needs.

Good luck anon :). Ill check in tomorrow

File: portal_73sk.jpg (304KB, 1396x1038px) Image search: [Google]

304KB, 1396x1038px

>>1027084
>Usb is often overlooked as the bandwidth wasnt there. Unless your spending big $, last i checked the available gear sub $100 was garbage, and resolution / cutting speed will suck.
This is part of what I meant tho, when I said I find a lot of DIY CNC matters amusing.... these ancient parallel-port systems use a dumb breakout (without its own logic) and so the PC must send every-single-pulse for every axis of motor. It's part of why I want to write my own CAD/CAM software.

With an arduino mega on a USB-1.0 line and set the serial speed to 115200 (the highest speed that the Arduino IDE monitor will read) then you can send about 800 serial message per second, total, at 128 characters each.

A 3-axis CNC mill can really only cut straight lines. To make curves, it cuts a whole bunch of really short straight-line sections that look like a curve.

For a 3-axis mill, you could send a statement that designates a straight-line segment, like this:
[6.500][x=-21800][y=38750][z=5910]
(that's only 34 characters)

And what this would mean is:
"In a total time span of 4.5 seconds, move the X axis -21800 steps, move the Y axis +38750 steps, and move the Z axis +5910 steps"
and the Arduino mega could figure out the step intervals for all three motors itself, and do all the real stepping itself, in immediate-mode.
In only 34 characters, that line tells it how to correctly make 66,460 steps--and that is nowhere near the limit of the 128-character message length.

When the arduino is done cutting that straight path, it signals back [OK], or it says which limit switch it hit [LIMIT=#3]....

So then, you would not be limited to 800 total motor steps per second: you would be limited to 400 straight-line segments per second, and the length of any single line segment wouldn't be an issue.

Plus in that serial message--you still have room for sending a spindle motor speed, a coolant flow signal, and a bunch more axis data too.

pic unrelated

>>1027142
>And what this would mean is:
>"In a total time span of 4.5 seconds,
Well nuts
that should read "In a total time span of 6.5 seconds..."

I bumped up the step counts, as I was writing, to demonstrate the effectiveness of the notation,,, but didn't change that 4.5 to a 6.5.

-------

Another nice thin you could do with this system is you could tell the arduino what direction to pull away from, to avoid marking the workpiece where the tool stops at the end of the segment (while it is waiting for another line instruction from the PC).
Something like this:
[6.500][x=-21800][y=38750][z=5910][px=-10][py=+10][pz=-10]

which would mean, at the end of this line segment, pull the x-axis -10 steps, the y-axis +10 steps, and the z-axis -10 steps.

When the arduino receives the next line segment, before it starts cutting it just reverses the last three "pull-away" commands so it is in contact with the workpiece again....

This way, you don't need a PC running in immediate mode. The arduino will only cut a straight-line segment when it has all the data to do that, and at the end it will automatically pull the cutter away a bit and wait for the next line segment instruction.

>>1027142
Seems legit. Hence why others have made usb controllers, and ardu ones with standalone tft screens & memory card readers/ usb drive connection for source files. Difference here is price. You want cheap crap, deal with it.

As for your argument with usb vs serial, as mentioned its price vs software. If you write up the whole deal more power to you anon. That said its already been done, and is mostly shit by fools that are against using naughty copies of software.

Seems you want a nice hi spec build, but dont know shit about how the electronics work, as per your second post, and expect pro tier industrial gear for cheap, with the "ill fix the shortfall in coding".

Protip. Ardu results will be poor & slower in comparison to the cheap serial units from ebay. Whats more, your software will be shit, given you will have to keep rewriting it everytime the free crap you rely on is updated, eg extra linefeeds or decimals added into code generated by inkskape or whatever your running to create. Just google the shit, your walking into a minefield here

OP here

Whoa, thanks a lot to all of you, especially to first guy, really helpful info! :)
I think I now have a far better understanding on making CNC machines comparing to couple of days earlier, aaaand I think I'll go with mach3 usb board.

>>1027183
I realised that ardu won't do me any good, especially I'm not that good with editing code, I need something as close as possible to plug&play (I know it's impossible in diy cheap cnc attempt, but anyway - arduino is a no no)

>>1027183
>>and expect pro tier industrial gear for cheap, with the "ill fix the shortfall in coding".

Actually coding is not my stronger side, the thing I'm relaying is the hardware (perfect steel frame/ feed rods which I manufacture myself, high quality ball screws etc)

My advice again (first guy, and ps theres only two guys responding in the thread), go with cheap serial first. Mach3 is a bitchy thing, and most sub $100 boards that have usb are mis-described. The usb is a power source only, not a connector.

Alternately, there are serial boards with onboard motor controllers, around $50aud - should be ok, if your not driving anything too hardcore.

My latest rig in production is actually a serial breakout, 5axis. 3 x stepper controllers, 1x turret head 1 x laser head, just for shits n giggles. Laser is a "15 watt uv" special i picked up a while ago. Frame is from alum extrusions i had lying around (60% complete, off to hardware when wife not around lol), servos, drives, switches etc etc from old xerox copier.

Was thinking of adding the stapler from the xerox too, why not lol.

Remember power supply - atx with resistor.

Remember a bloody emergency stop switch that triggers the board, and preferably cuts power to at least the spindle if not all drives. Cant tell you how many bits ive broken over the years with shitty cnc mills and no decent stop!

Remote motion is a bonus. Ebay that shit.

File: photo_1024x1024.jpg (266KB, 1534x510px) Image search: [Google]

266KB, 1534x510px

I am looking into some entry level cnc machines myself right now, I came across some Chinese variants that have good frames but those archaic parallel port interfaces. I am thinking of marrying a tinyG for its amazing motion control with some beefier controllers. the ones onboard the tinyG would fail to drive even modest nema23 motors. My only gripe is that the only good software for the tinyg is web based.

>>1027534
>I am looking into some entry level cnc machines myself right now, I came across some Chinese variants that have good frames but those archaic parallel port interfaces
(I am the guy who wants to [eventually] build his own USB-based setup...)

The problem with "being stuck with parallel port" is that a lot of CAM software doesn't support anything else. The software doesn't have an in-program option to use any kind of USB-port controller,,,,, That's why the other guy said "decide on what software you want to run first, and then look for a controller that will work with it".

If you want something that is easy to get running right away, don't do anything yourself. Just use what most other people use (however annoying it may be).

Also a warning with CNC controller boards from China: sometimes the hardware (and the price!) looks really nice, but the lack of documentation/lack of product support/hardware and software bugs kill the concept.

>>1027183
>Protip. Ardu results will be poor & slower in comparison to the cheap serial units from ebay. Whats more, your software will be shit, given you will have to keep rewriting it everytime the free crap you rely on is updated, eg extra linefeeds or decimals added into code generated by inkskape or whatever your running to create.
There won't be no Inkscape....
I am planning on writing one program to do it all.
The program will be what I use to generate CAD files, the CAD file will already contain general data describing a toolpath, and that data will be interpreted by the PC in real-time and sent to the CNC machine to actually cut.
It won't be standard CAD files, but I don't care about that.

,,,,,,,

It would probably be better if this board had the colored ID tags for posts... like /biz/ has. Makes it somewhat-less anonymous, but it's a lot easier to keep long conversations straight.

>>1027597
Yeah, I am looking at one of these things from the far east, but I don't feel like dealing with parallel port software (it would likely be mach3 or linuxcnc) I was planning on gutting the electronics for the shit in my earlier pic >>1027534
as well as upgrade the motors to ones with 400 steps/rev for higher resolution (my main use will be pcb milling)

TinyG (the controller board) has a good community behind it and it is in another league than tard-uino based controllers. The controller next to it even supports better microstepping and higher current capability than the onboard controllers, as well as active breaking and the like. Almost all the stepper driver boards are based off a toshiba chip and all operate similarly.

File: 505451410_646.jpg (85KB, 800x600px) Image search: [Google]

85KB, 800x600px

>>1027615
forgot pic

>>1027597
Same guy from before. Put simply, ive been playing with this shit for around 20 years, since sperry days, both sw and hw dev.

If as you infer, you are capable of development of software solutions to run this, and hardware dev to implement your sw solution, your amazing. Seriously. Lets consider the hundreds of cumulative man hours of dev in just the gear cutting solution of the gear cutting sources for inkscapes gcode tools, and then the parsing and manipulation of codes itself.
I recently spent around 100 hours coding a solution for converting gcode from generic codes to suit a custom controllers codes from around 15 years ago, maxed out at win xp & vesa connectors. Then when it was finished, client decided to take initial recommendation and update hw to modern serial comms from custom vesa shit.

As i said in the second or whatever post, listen fucker. Accept your abilities, based in your price point. Unless your spending $$, which you arent based on your posts, your stuck with serial and probably pirate sw. That said, there are reasons, and limitations as ive said despite your constant ramblings about arduo and usb bandwidths, that make serial a more suitable port for this based on cost.

Put simply, either :
A) spend $$ and stop shitposting
B) spend fa, go serial, stop shitposting
C) keep shitposting and take it to /b/

Lets be honest pal. Your like the twat that wants to make a arduo dev to pick locks, to open his mothers car that the keys were lost to and his arthritic dog is locked in. Not happening /b/ro.

>>1027142
What happen if you use servomotors instead of steppers?
Can you control the speed of servos to create a true curve?

>>1028738
>What happen if you use servomotors instead of steppers?
In the industrial jargon--a "servo" doesn't specify any particular type of motor, it just means *some* type of motor or actuator with a positional feedback device on it that allows you to monitor its actual position. A servo can use a BLDC motor, or a stepper motor, or a brushed-DC motor, or a hydraulic motor, or a linear actuator, or any other kind of motor.

RC servos are not real servos, since they don't really give closed-loop control.

>Can you control the speed of servos to create a true curve?
The axis motors isn't the reason.
The only way you can cut a "true" curve is if the work piece rotates around a longitudinal axis, and you use a stationary cutter--such as the way you normally cut metals on a lathe.

You can use a vertical 3-axis CNC mill to cut out round pieces and holes in stuff, and it can be done so fine that it looks and feels perfect--but it you measured it with the right tools, you would see that it is much less consistent than the same kind of cut done on a lathe.

OP here. After reading this whole thread and researching every cnc related forum/blog there is, I'm starting to think that maybe cncing is a bit too difficult for me (no coding skills, no electronics skills, just a raw will to reduce my manual labour). Maybe a "spend a lot of $$" is the best solution for me lol

>>1029045
i ask because DC servos and brushless servos ofthen allows to control the speed.
Controlling the speeds of two servos will allow to create a vector that is tangent to the curve that you are cutting so it will be more "true".
But i'm not sure how accurate the servo itselft is in follow the designated speed.
Of course the friction of the machine will also introduce and error in that curve.