[scrub417]

I am building an egraver that will engrave on cylindrical wooden objects like baseball bats. The end product would take a bmp or jpeg and engrave that on the bat. Initially I would like to be able to do text in various fonts then possibly graphics/logos. I wouldnt consider it a 3D router because the depth would stay fixed. Hardware I have no issues with, I plan on using steppers with an arduino and a dremel. What I would like help with is the software interface. I have pokde around the forums gathering info but I cant seem to nail down the software flow to go from a graphic file to stepper controls. I know gcode is involve but that is all new to me. please feel free to throw some advice my way
thanks

[PaulRowntree]

Depending on your budget, VCarvePro or Aspire (from Vectric) can do wrapped carving. They assume a machine that can move the tool up/down as well as parallel to the rotation axis. You can do the design work in these programs or import the artwork, then output the gcode to your machine. The machine could be running custom driver code or Mach3 or EMC to interpret the Gcode.

You will need a Z degree of freedom to do engraving, since you must come down into the work at some point, lift it to move elsewhere. Plus, you probably want to control the depth of engraving for lettering etc.

[scrub417]

thanks Paul
Initial budget is small, I am little concerned about the resolution that I can get with the Arduino so I want to proof things first then I wouldnt have a big issue with buying a decent front end software. My plan is to take a small lathe and use one stepper to rotate the bat instead of a Y position and the X would be parallel to the rotation axis.
I wouldnt be opposed to adding the Z and a bevel shaped dremel tip(that cut wider as it went deeper) but I wouldnt know how to translate graphic width into depth

[PaulRowntree]

I guess I don't see how you are going to do this with only a rotation axis and movement parallel to this axis. If you manually dial in to make a cut, then you are going to have to dial it out to skip over to a new cut area. Even if you stick to connected script fonts, you have to pick the tool out of the work every so often. And then there are graphics.

I would take a longer view : imagine where you want to be with the project 1-2 years out, and try not to acquire or develop too much that will not be in that configuration. Learning on PicStep boards is one thing, but then that experience is for nothing if/when you decide to use a std interface like a G540 (I know even less about PicSteps than G540's, by the way, but I know my G540 has worked since the day I go it, except for a fuse). If you build a machine that is too limited (eg a 2 axis machine when you need three) then your frustration level will rise before you start to enjoy this process.


And you never want the frustration to be greater than the fun, on anything.

Can you write code? Have you checked into CNC4Free? There are a number of options out there that won't lead you down a blind alley.

[scrub417]

I was planning on just using a solenoid to move it off the bat to relocate, a single output would do that. but I like your philosophy of building for future intent.
I dont know what a G540 is but I'd be interested.
I can write some code, im more hardware friendly (i can build most anything)

[PaulRowntree]

I have never seen a solenoid that would do that, but I am intrigued. It would take quite a bit of force to hold the cutting tool steady, and that may be hard to arrange. Plus, you would need a stroke of at least 6 mm (1/4").
A Gecko540 is an external device that hooks up to a PC via the Parallel port, and can control up to 4 motors plus a bunch of separate I/O lines like estops and home switches. Add the right power supply, motors, and you are away. Many folks here use them for small to intermediate size machines; It is made to handle Nema 23 sized motors as it is limited to 50 V supplies.
I think that most software that drives steppers can go through the parallel port, but not all. It is a good idea to make some software decision before doing too much interfacing.
I design in CorelDraw or VCarvePro, use VCarvePro to translate drawings into toolpaths, then Mach3 to translate the toolpath into stepper pulses. It sounds more cumbersome than it is, really. VCP can also handle wrapped engraving like you want to do; it is worth every cent IMHO, but it isn't free.

[scrub417]

I checked around vectric and they also have Photocarve that is much cheaper(my tiny wallet is rearing it's ugly head). It looks like it is built to do just what I am aiming for. any experience with that?
the 540 seems reasonable in the future but it may be overkill for what I want now. the bats are only 12-14 in long, this would be a small setup, plus I want to build all the drive electronics myself (see earlier wallet comment).
So adjusting my approach, I will look to get Photocarve and I have Mach3.
What specifically would I need to do to the Arduino to 1) talk to Mach3 2) accept the gcode from it?

[RossMosh]

Laser engraver would be the right tool for this job.

[PaulRowntree]

I checked around vectric and they also have Photocarve that is much cheaper(my tiny wallet is rearing it's ugly head). It looks like it is built to do just what I am aiming for. any experience with that?
the 540 seems reasonable in the future but it may be overkill for what I want now. the bats are only 12-14 in long, this would be a small setup, plus I want to build all the drive electronics myself (see earlier wallet comment).
So adjusting my approach, I will look to get Photocarve and I have Mach3.
What specifically would I need to do to the Arduino to 1) talk to Mach3 2) accept the gcode from it?

I have PhotoCarve, and am making progress. Some folks have very nice results. But I don't see this program helping you; It always starts with an image, and tries to find a way of moving the tool in/out of the work to map light and dark areas of the image into high/low regions of the carved material. Unless you converted your engraving text into image format, it seems the wrong way to go (for me).
It is a good question what Mach3 needs to communicate with an external driver box; it has been done, but I assume by working closely with the M3 software coders. EMC is an open source alternative that I have never tried, and I think it is said to be more 'friendly to external hardware. It runs under Linux.
If you have an external embedded processor that understands gcode, then perhaps you don't need M3 or EMC.

[acondit]

If you really want to do engraving, then I would say forget the solenoid. Watch some videos of VCarve cutting a sign. At the corners of the letters, it gradually changes the z-height as it moves the x and y axises to produce a sharp corner on letters. Also certain fonts the body of a letter varies in width, requiring the z-axis to change depth as it cuts the body to create the varying width.

Alan

[Fish4Fun]

scrub417,

There is Arduino CNC firmware that is a three axis g-code interpreter. There is Arduino CNC firmware designed to drive steppers/servos. I can't see a single AVR having the resources to do both simultaneously. Gaining a firm grasp on the finer details of how existing software/firmware/hardware interfaces come together is the first step to DIY. Your job will be MUCH easier if you design each stage to integrate with existing stages. Following is a general outline of the most common path CNC takes from design to product.

1) PC CAD Software is used to define a part's outlines and features in a vector graphic file.

2) PC CAM Software is used to define tool paths based on the individual vectors contained within the file generated in step one; the output is gcode, which is simply a sequential list of machine directives that is NOT machine specific (though typically tailored to a particular type machine, eg Mill, Lathe etc, etc.)

3) Machine Controller, typically something like Mach3, EMC2, Arduino etc, that is setup for a specific machine. That is it "knows" how many steps per inch, "backlash compensation", acceptable accelerations, maximum travel speeds, "home", axis limits, machine "stops" etc, etc. The machine controller then translates the gcode into timed "steps", spindle speeds, etc. It is important to note that this stage DOES NOT directly control the steppers/servos, but rather simply outputs "steps" to any or all of the axis controllers.

4) Stepper/Servo Controller. This stage is typically handled with one or more uControllers dedicated to interfacing with step 3 and ensuring the output is translated into the prescribed motion. Examples of controllers include the G540, Arduino and hundreds of others. The input to these controllers is "Step and DIRECTION", thus, to move an axis 1.001 inches on an axis configured to move 0.0001 Inches Per Step would require 10,010 distinct "step inputs" (and this output would be received from step 3). The Controller is responsible for low level control of the motor, including current control and feedback. For the specific case of a stepper motor driver this low level control can be as simple as dedicated "full step" logic control of a pair of H-Bridges with a resistive current limit, or as complicated as a uController based micro-stepping circuit with PWM based current limiting and encoder based feedback.

By adhering to the above algorithm you ensure that you can use industry standard replacements, either DIY or Production, and ensure compatibility with existing tools elsewhere in the chain. Obviously it would be possible to re-invent/replace any or all of the process for a specific task, but I think you will find the existing process is very mature and quite robust allowing easy migration paths. There are plenty of examples of integrated systems designed to facilitate the process from the user's point-of-view. For example there are MANY CAM applications with CAD capabilities; there are some systems where the CAD/CAM/Controller software is all integrated to interface with a specific machine, a good example is dedicated sign making//engraving packages where the entire system is designed for the sole purpose of rapidly "personalizing" common products. In these "integrated systems" the above algorithm is still intact, it has simply been "streamlined" to facilitate specific tasks.

In most cases it is best to design a machine to be as flexible as possible with respect to the work it can handle. This DOES NOT mean that all machines should "maximize" everything, but most facets should be minimized only after careful consideration (typically budget, lol).

In the specific case of your bat engraving machine, I cannot conceive of ANY reason not to build a 4 axis machine for this task. Even budget cannot remove a mandated axis. You should consider building something like John's (Microcrave) A4 machine and adding a 4th axis.

http://www.cnczone.com/forums/diy-cn...p_machine.html

http://www.cnczone.com/forums/diy-cn...ans_files.html

As far as building the electronics yourself...I hear ya, and there are PLENTY of existing designs floating around the net ready for DIY, but at the end of the day, I don't see a lot of savings over the high-volume, production versions. Buying components and making PCBs in low volume is expensive, time consuming and problematic. The fact that you mentioned an Arduino suggests to me that you have already figured out that the Arduino is in most cases cheaper than building a dedicated PCB for each AVR project, even though the AVRs themselves are <$5 while the Arduinos are ~$30. I don't think you are going to find stepper controllers any different in that respect. @~$250 for a 4 axis controller with all its other features you would be hard pressed to build similar for less actual $$$$ never mind the time. That said, I currently have a thread here on an open source BLDC motor driver that I am working on, but my focus is NOT to build a cheaper controller than is currently available, but to build a CNC oriented driver suitable for cheap BLDC motors available from the hobby industry. I am somewhere North of $500 in my quest to build a $100 controller, LOL. I am NOT saying you shouldn't build your controllers, just suggesting that doing so to save money may not prove a good investment.

I have been playing with the idea of a CNC machine for over 20 years; I have been playing with an ACTUAL CNC machine for a bit over 6 months. If I can give you any advice from my experience it would be: "Give serious consideration to building a proven design using proven components and proven software. DO NOT throw good money at a limited design or components simply because they are within your immediate budget."

I purchased one of Microcarve's unpainted A4s for ~$400. That's the teaser price. The steppers, g540 and Power Supply were another ~$500. The Bosch Colt and requisite mount were ~$200. The Precision Bits Collet kit and a couple of bits were another ~$100. Mach3 is $175, and we won't even talk about the Vectric Software (but it is a GOOD INVESTMENT). While I plan to build a 4th axis, I haven't yet, but I am pretty sure I can build what I would want for <$200. I have spent ~$1k on other additions to my build (dust box & collection system, tooling, lighting, etc, etc), but none of that was strictly necessary (except the dust box, LOL); I would think you could purchase "off the shelf" everything (A4, Steppers/Controller/PS, Spindle/Mount) and use open source software and your 3-Axis machine to build your 4th axis for between $1k and $1.5k total package. If you build something like the A4 from plans, DIY your stepper controllers, find a "deal" on steppers, re-purpose an existing router/dremel as a spindle and DIY your spindle mount, you might end up with between $500 to $1k in it, but you might also end up with a useless pile of parts. I have seen threads here where people do extraordinary things with <$1k, but those are typically seasoned experts with exceptional engineering and scavenging skills who completely discount things like the gas used to find parts, cost of existing tooling (Mills, Lathes etc) and time.

Anyway, I have rambled on much longer than I meant to and given much more advice than asked for, sorry for that, but I think back on all the $$ I have spent thinking that I couldn't afford to "buy" the various bits, only to find out buying what I actually wanted was a lot less $$ in the long run, LOL.

Good Luck with your Build!

Fish

[Darylb]

What if your CNC used cylindrical coordinate like a lathe? You chuck up the bat on the "lathe" and have the rotation of the bat controlled by 1 stepper. Then there is a stepper the moves the router along the axis of the bat and another that moves it towards the centerline of the bat.

I suppose the only challenge here is finding software to accommodate this setup

[CarveOne]

What if your CNC used cylindrical coordinate like a lathe? You chuck up the bat on the "lathe" and have the rotation of the bat controlled by 1 stepper. Then there is a stepper the moves the router along the axis of the bat and another that moves it towards the centerline of the bat.

I suppose the only challenge here is finding software to accommodate this setup

Mach3 has support for lathes. So does EMC2. A bunch of people have shown lathe turnings and wrapped photos on cylindrical wood here on CNC Zone. YouTube should have lots of videos on the subject. Vcarve Pro has wizards for doing some of this work.

CarveOne