Electronics Rundown

So Adam, you’re building a robotic cutting machine… How do you plan on getting the cutting tool to draw the pretty pictures on your workpiece? Hopefully this post will answer that question at least from the hardware perspective.

30,000 ft View of Components

The machine has a gantry that moves across the x axis of the workspace. The cutting tool is mounted to a carriage that traverses the gantry in the y direction. I’m using x and y fairly arbitrarily as I’ve seen different machines oriented differently. This carriage also raises and lowers the tool in the z axis.

This motion is actuated by stepper motors through rack and pinion gears. More on steppers later, but for now just know that they’re a type of electric motor able to rotate a specified number of degrees and stop.  Some folks use servo motors here, but they’re beyond my requirements and budget, so I’m not going to get into that discussion at this point.

Stepper motors are connected to controllers that tell the motors when and how far to turn. Each controller is connected to the power supply and the break out board. The break out board interfaces with a PC via parallel port. The break out board may also have inputs for limit switches, emergency stops and torch height control. Torch height control measures the voltage of a plasma cutter’s arc to determine its distance from the workpiece. (Similar concept to why I kept blowing the breaker when I was stick welding and let the electrode get too far from the weld…) This is handy when you’re cutting material that isn’t perfectly flat.

Street View

There are several options when it comes to acquiring and assembling the electronics portion of a CNC table. One can purchase components separately, as part of a package deal that include various combinations of parts, or as a turn key kit including motors, controllers, power supply, torch height control, enclosure and wiring.

Ad Hoc Partial Kit Turn Key Kit
Motors  $296  incl.  incl.
Controllers / BOB  $260   incl.   incl.
Power Supply  $120   incl.   incl.
Enclosure  $100  $100   incl.
Torch Height Control  $268  $268   incl.
Misc. Cables, Switches, Cooling  $250?  $200?   incl.
Time  30 hrs?  20 hrs?   incl.
Kit  $0 $668  $1575
Support None None Included + 2 yr Warranty
Total  $1294 + time  $1256 + time  $1575

I put together the above table to try to compare the options for putting together the electronics for my table. The “Ad Hoc” column is basically sourcing different components off ebay. The “Partial Kit” column is using the supplier recommended in the Bill of Materials that came with my gantry kit.  This kit uses slightly less powerful motors than those quoted in the other two columns.  The “Turn Key Kit” column is the Dragon Cut 620-4 kit from candcnc.   The THC quoted in the first two columns is a really basic unit. I’ve seen it demoed on youtube, and it seems to work all right. The candcnc THC seems more robust though. It also has custom integration with Mach 3 software and is made and supported in the USA.

I think the table pretty much speaks for itself. There’s no way the time and (most likely) frustration of putting together one of the DIY options is worth the $300 difference.  This is especially true if you take into account the warranty and support going forward.

On that note, I wrote Tom at candcnc an email asking for clarification on which of his systems would be best suited for my project. His site advises using the 620 oz. in. motor kit for gantry weights over 50 lbs.  I figure that’s almost exactly what my gantry will weigh. I wasn’t sure if that advice was assuming one or two motors to drive the gantry. He responded within an hour with a very informative email.  He informed me that they assumed two motors driving the gantry, so the 620’s are the way to go.

Water Table Design and Checking Squareness

Hello! I got a message this morning that the rest of the steel I need for the frame is waiting for me to pick it up next time I’m in Seattle.  I’m trying to think of things I can work ahead on. Finally got around to adding the water table and gantry (with my own digital versions of the shopdroid brackets) to my design in Blender. I was happy to see the gantry (shown in white) fit up nicely.

Latest Blender Render

Latest Blender Render. Purple in honor of my Kansas State Wildcats!

I really haven’t decided if I’m going to do the water table right away or not. I did this mostly to make sure it could be retrofitted if I decided not to put it in initially.  As I understand it, the argument for is that the water traps the vaporized metal and greatly reduces dust in the shop and in your lungs.  Since I’m in borrowed shop space, I’m leaning toward installing water from the start so I don’t coat my benefactor’s airplanes with metal dust. In either case, it’ll work. I’ll just have to alter the design of the slats (a grid of vertical strips of metal that supports the work piece during cutting) a little based on which way I decide to go.

If I go with the water table, it’d be 50″ x 98″ and could sit on top of the frame, unattached. On my list of things to do is looking into sheet metal sizes to figure out if I could make this out of one piece.  I actually haven’t even determined what material and thickness would be best for this.  I’ve heard of some folks using spray on truck bed liner for their water tables. Any thoughts?

Hip to be Square

I went to the hangar this morning to check the squareness of the one end of the frame I’ve fabricated. Pythagorean Theorem said it should be 59.1 inches diagonally across this section. Both diagonals were just over 59 inches. That should  mean it’s  pretty square. I’m planning on using this section as a jig for the other side and keeping them clamped together till the new one cools. Thanks to John (aka zappafan1 on cnczone.com) for the tips!




Stay tuned!

MIG Welding the First Table End

Had an interesting day welding up the first end of my CNC table with the Millermatic 250 MIG welder the other day. I had this section all tacked together, so I just had to run beads along the seams in such a way as not to warp the heck out of everything. I got to the shop and clamped my work down flat on the table. I checked the settings on the welding machine and got to it. I ran about an inch of bead before I realized I’d forgotten to listen for the gas… AGAIN. This time instead of turning on the wrong bottle, I’d forgotten to turn the gas on at all. Shame. I need a checklist for this apparently.

I did my penance of grinding out the porous weld, turned on the gas and sat down to try again. This time I didn’t make it an inch.  As soon as I struck an arc, I flashed myself in the eyes. Another item on the checklist needs to be to switch the auto-darkening welding mask setting from “grind” back to “weld.”

After rubbing my eyes and blinking for a minute, I finally ran my first bead of the day. I think it turned out more or less decently. The only issue was I had my mask set too dark this time and I couldn’t see where the end of the joint was. As a result I kept going till the arc went out because there was no material underneath it. This left an arc mark as I finally pulled it away.


I welded this originally in the flat position. Note the arc mark to the right resulting from my inability to see I was at the end of the joint. I turned down the darkness setting on my mask to improve visibility.

After the early hiccups, I finally settled down into a bit of a groove. I had the work piece clamped down flat at first, welding flat and moving around to different areas of the piece to keep too much heat from building up in one area and to give plenty of time for cooling. This should help prevent warping. Then I flipped it and did the other sides before turning it on its various ends for the fillet welds.


Got a little skinny in the middle of this one and forgot to leave the gas on the end of the weld for a moment at the end. Think that’s the cause of the dimple.


Not sure how I ended up with this porosity. Had to grind that out. remembered to switch back to the “weld” mask setting this time…

At one point I was feeling cocky and tried to do a vertical weld. I should have waited and reviewed the technique on that as I ended up with a pretty ugly weld with a big droop at the bottom.


Less than stellar results of my first attempt at vertical welding. After this I reoriented my work so I could weld flat until I do some more research and practice with the vertical position. Also note nice cut quality from the Rage II cold chop saw.







I’m not sure what the brown precipitate is on some of these welds. I think I probably need to clean the joints better before welding. I’ll try that next time. Seems to come off easily enough but it leaves indentations in the weld.

So to recap, today I learned:

1. (Again) Double check the gas bottle is turned on before welding.

2. Double check welding mask reset from “grind” to “weld” after grinding.

3. Don’t forget to hover over the end of the bead with the gas for a moment to keep the puddle shielded as it cools.

4. Don’t set the mask so dark that you can’t see where the end of the weld should be.

5. Going to try cleaning the joints with a wire cup brush on the angle grinder next time to see if that gets rid of the brown precipitate.

6. I wasn’t satisfied with how the corners blended together. Next time I’m going to try to weld around the corners and mate the beads up on the flat sides.

While there are lots of ways I hope to improve, I’m pretty happy with how this is turning out so far. It’s cool to be learning and practicing welding on something that’s useful but not critical for life and limb. I’m going to pop in for a minute tomorrow after it’s all cooled off and measure the piece on both diagonals to see how I did at keeping it square. I’ll try to post some pics of that.  I’ve got a call in to my steel supplier in Seattle and hopefully next week I’ll get the rest of the tubing I need.

Thanks for reading and don’t forget you can subscribe and get email updates as they happen.


Let the Build Begin!

I finally got my Chop Saw! So far I’m reasonably happy with it. There were a few chips in the paint when I pulled it out of the box. I was annoyed by this, but not enough to delay my project a month by sending it back. At about $280 plus shipping, it was a little more than half the cost of the comparable DeWalt I could buy locally. If it lasts through this project, it will have paid for itself. As far as cutting goes, it’s like butter so far. I’m glad I went with a cold cut saw versus the slow, hot, abrasive wheels.

Bar stock cut with the Rage saw.

Bar stock cut with the Rage saw.

So having the welding more or less figured out and the ability to cut steel, the only thing left was to get some steel.  In Juneau, steel is about 75% more expensive than in Seattle because it has to be shipped up on a barge.  I have access to discounted shipping so it makes sense to source my steel from Seattle. However, since I’m new at this and I didn’t want to order 140 feet of steel from a supplier in Seattle only to discover that the fabrication was over my head, I went ahead and bought a 20′ length of 2″x2″x3/16″ steel from the local supplier. This, I figured, would be enough to build one end of the table frame. If I could keep that square and level, I’d go ahead and get the rest from Seattle.

Measure 3 times, Cut Once

First cut on 2X2 tubing. Clamped per the instruction manual.

First cut on 2X2 tubing. Clamped per the instruction manual.

I made two marks that the blade should go between. Then I remeasured and marked again. Sometimes even one or two more times, just to be sure. Steel is expensive.

I made two marks that the blade should go between. Then I remeasured and marked again. Sometimes even one or two more times, just to be sure. Steel is expensive.

After being super diligent about measuring and marking, I was thrilled at the accuracy of the cuts. Parts lined up perfectly.

After being super diligent about measuring and marking, I was thrilled at the accuracy of the cuts. Parts lined up perfectly.

Then I beveled the edges as required with my angle grinder.

Then I beveled the edges as required with my angle grinder. I only had to do 2 edges per joint since the tubing is rounded and provides a gap for the weld to penetrate into.

Tacking It Up

More measuring and marking. The magnets were pretty helpful in squaring things up and holding them in place before using the C clamps.

More measuring and marking. The magnets were pretty helpful in squaring things up and holding them in place before using the C clamps.

This is what your tack welds will look like if you turn the gas on to the wrong welding machine. I could have sworn that was the right tank! Now I listen for the hiss of gas when I pull the trigger.

This is what your tack welds will look like if you turn the gas on to the wrong welding machine. I could have sworn that was the right tank! Ground those out with the angle grinder and re-tacked. Now I listen for the hiss of gas when I pull the trigger.

The right angle magnets are good for getting things set up, but Bill wisely suggested that I clamp stuff down to keep it secure.

The right angle magnets are good for getting things set up, but Bill wisely suggested that I clamp stuff down to keep it secure and prevent warping.

It stands! After tacking everything looked square and true.

It stands! After tacking everything looked square and true.

In the next post, I’ll see if my MIG welding skills are up to the task of finishing up this end of the table. Stay tuned!

Welding Practice


I’ve spent the last couple of weeks trying to teach myself how to weld. I purchased an EVERLAST PowerARC 140 a while back because I thought I’d use stick or TIG welding to build my table. I did some practice stick welding with this unit and discovered that on a regular household 120 V circuit, I was pretty limited on which rods I could use effectively without tripping the breaker. Starting out with 1/8″ 6011, I found I needed at least 85 amps to keep a good arc going.  At that setting, if I pulled the arc long at all I’d trip the breaker. Longer arc means higher voltage. The voltage coming out of the wall is constant, so if you ask for more watts it tries to draw more current. I wish I had found this chart sooner.


This is how I decided to try 3/32″ 6013, since it should work well at lower amperage settings. It did work a lot better. I turned out some beads that looked respectable to my untrained eyes, including a butt joint with 1/8″ bar stock. The fillet welds were not turning out though.

One of my first arc strikes with 1/8" 6011. Note scratch marks and divots from over-zealous slag hammering.

One of my first arc strikes with 1/8″ 6011. Note scratch marks and divots from over-zealous slag hammering. Shame.

First bead with 1/8" 6011. This was running up against the duty cycle limit of the machine.

First bead with 1/8″ 6011. This 2 inch bead was running up against the duty cycle limit of the machine.

3/32" 6013 @ 67 amps. Sorta looks like a weld at least...

3/32″ 6013 @ 67 amps. Sorta looks like a weld at least…

Butt Joint with 3/32" 6013.

Butt Joint with 3/32″ 6013.


Fillet of Junk.

Bill is another tenant of the hangar I’m using.  A few days ago, he was working on repairing his airplane which had been hit by a snow plow last winter. He saw me scratching my head over my junky fillet welds and came over. It turns out that in addition to being an Airframe and Power Plant Mechanic, he’s a pretty experienced welder. He offered to let me watch him run a few beads.  Right away he said, “this isn’t how you want to build your table! Why aren’t you using that MIG Machine?” I told him it wasn’t mine and I didn’t know how to use it.

He turned it on and found a preset that looked right for what we were welding on and ran a couple beads for me. He made it look easy, but I was pretty blown away by how easy it actually was when I tried it. I had a couple tries where the wire seemed to come out way too fast, but I figured out I wasn’t keeping the tip down close enough to the weld. Once I got a feel for it I even ran a couple of fillet welds that turned out way better than what I’d been doing with the stick welder.

More like it.

More like it, I think.

The shop I’m being allowed to squat in has 120, 220 and 250 volt service. It doesn’t have 240 though, which is the other option with the EVERLAST PowerARC 140. I think this would be a fine stick welder if you had 240 volt service or if you have light duty welding to do using smaller electrodes. I’m not sure why my fillet welds weren’t turning out. Pretty sure it was my technique. I still want to try it with some 7018  just to see how it does, but after about an hour practicing with the Millermatic I felt confident that I could start building my table.

Blender as CAD

Hello again! As mentioned in a previous post, I decided to use Blender to design my table. Blender is not designed to do CAD. It’s more geared toward 3D modeling and animation. Blender is to Maya or 3ds Max what GIMP is to Photoshop. That said, folks are using it to create designs for 3d printing and milling. I used it to create this pendant for my fiancee last Valentine’s day. I had it printed at Shapeways.

I’m not going to get into the basics of Blender here. There are tons of great tutorials out there. Some of my favorites are at Blender Guru and Blender Cookie. I did have a difficult time getting Blender set up to use real world measurements though, so I’ll give a brief explanation of how I finally got it working.

Step 1

Go to the Scene menu and click on the Units drop down. Then click on Metric or Imperial depending on which units you’re working in. Scale of 1.ooo sets the scale to 1 foot in imperial or 1 meter in metric.

Step 2

Hit the ‘N’ key to bring up the Transform Properties menu.  Under ‘Display’, make sure that ‘Grid Floor’ is checked. Set the scale to 1 for meters of feet, 1/12 (.08333) for inches, .01 for centimeters, etc.

Easy, Right?

WRONG!! Say you’re a goofy American like me and you want to use imperial units. You’re Trying to model a leg for your table and you want it to be 12″ tall. You’d hit shift + A to add a plane and ‘e, z, 1′ to extrude the plane along the z axis 1 ft. However, due to an apparent bug, your new leg will be 3.2808 feet tall. This happens to be 1 meter. Therefore if you insist out of national pride, ignorance, or practicality on using imperial units, you must convert your z measurements into meters. For example, if you wanted to extrude a 2′ long leg you’d type ‘e, z, 2*.3048′

Alternatively, you can click the little magnet icon on the bottom of your screen and instead of entering a distance to extrude you can drag your mouse the direction of extrusion and it will snap to the nearest grid line. If you set 1/12 in your grid floor scale window, you’ll be extruding 1″ at a time.

Don’t forget to name the pieces that you model. I found it helped to name the piece its length and orientation, i.e. ’3_ft_vertical’ or ’50_in_horizontal’. This helped later when I went back to add up how much steel I was going to need.

Here’s what I ended up with for a frame:

Catching Up

Well hello. I started researching this project several months ago, so there’s a bit of catching up to do on documenting the process. As such, this post is meant to start bringing you up to speed on where I’m at and where I plan to go.

The Goal

The goal of this project is to build a CNC table. CNC stands for Computer Numerical Control. Basically, the idea is to be able to put a sheet of material (metal, wood, vinyl…) on the table and have the machine cut out a two dimensional design from a computer file. This is accomplished by moving a cutting tool over the material using a gantry (sort of like the one that moves the printer head on your inkjet printer except with another axis or two). Stepper motors or servo motors move the gantry and the tool mount. A motor controller interfaces with the computer to interpret a CAM (computer aided manufacturing) file. Don’t worry, I intend to delve into all of these components in much more detail as I come to them in the process of building my table.

The Motivation

My reasons for taking on this project are few and nebulous. I’m fascinated by the possibilities of locally manufacturing digital designs delivered via the web. Part of this is due to my living in a place where almost everything is shipped in on a barge or an airplane. I suppose this is true of America in general. The difference in Juneau, AK is that we don’t have the economy of scale in shipping that others enjoy. It seems like local manufacturing could be a boon in isolated places such as this. Raw materials are still required to be shipped in to the extent that they aren’t able to be produced or recycled locally. I still see a couple benefits to local production though. One is that products can be made more speedily than they could be ordered and shipped. Another is the benefit of just-in-time manufacturing. If less inventory is required, fewer resources will be wasted by shipping goods in quantities beyond the actual demand.  All of this is just to say that I think a machine like the one I’m building has the potential to add a lot of value to the local economy.

In addition to making stuff for locals, I want to be able to make stuff to sell to tourists. According to this report, over 1.1 million tourists came to Juneau in 2007-2008. I think a subset of these visitors see the environmental and economic value in purchasing locally made souvenirs. I have some ideas of things I could make to help these folks remember their trip to Alaska.

I am enormously fortunate to have a job where I have a fair amount of time on my hands. I feel a responsibility to make good use of this time. I’m going to have to learn a lot of new skills, from welding to gcode, in order to complete this project. If you follow this blog, I can guarantee you’ll have a front seat for a lot of mistakes. Hopefully we can prevent your making the same mistakes if you start down this road later on.

The Process

I started by looking at pre-built systems like those offered by Torchmate and Gotorch. I decided I had more time than money, and based on what I’ve read I’m hoping to be able to build my table for about half of what I would have to spend for a turnkey system.  Reading other build logs on forums such as CNCzone helped me wrap my mind around the process. I’m also looking forward to actually building the table and learning all the requisite skills. In time, we’ll test the soundness of this logic.

Once I chose building over buying, I looked around at different kits and plans. I’m planning a future post that will give an exhaustive comparison of the different offerings. For now I’ll just say I decided on a gantry kit from Shopdroids. I’m planning on using a comprehensive electronics kit from CandCNC. Shopdroids has plans and parts lists available for download after you purchase one of their gantry kits. The plans are fairly generic and it is expected that the builder will customize them for his/her needs.  I used Blender 3d modelling software as a free CAD substitute for designing my table. I had quite a bit of experience with this software from previous projects. I’ll go into detail on the workarounds required for this in a future post.

One of the obstacles I’ve had to overcome is the fact that I live in an apartment. I love almost everything about my seaside apartment. One thing it lacks, however, is a garage/workshop space. I’m insanely fortunate, however, to have been granted access to one of my coworkers’ hangars during the build process.

Once I had a workshop lined up and I’d finished my design I started gathering tools and equipment as well as getting quotes for materials. Another post to look forward to is a description of the tools and equipment I’ve acquired and how I decided what to get. Going forward you can also expect to read about how these items worked out for me.  I’m currently awaiting delivery of this tool before I can begin construction.

That’s all for now…

Don’t forget to enter your email address in the box so you get updates as they happen!