Chinese 6040 CNC gantry desktop router/mill rework

Aschiral

Registered
Registered
Joined
Dec 21, 2018
Messages
4
This is my first thread here, I think I have it in the correct forum but please move it if I've erred!

A few years ago I purchased one of the common 6040 desktop milling machines from Ebay. Despite differing brand names and Ebay sellers I think there's just one or two factories in China that crank these things out. I was aware (or thought I was) of the limitations of these machines: Small working area, low rigidity, and overall cheap build quality. My purpose for this machine was simply as an engraver, primarily in aluminum, to make decorative engravings on firearms and other small items as need be. Engraving doesn't involve tremendous cutting forces, and positioning accuracy isn't critical due to the decorative nature of the work. I encountered a slew of problems but trucked on through them to press the 6040 into service. I'm in a slow period at the moment where the 6040 isn't being used and I can take it offline and try to improve it. If anybody is considering one of these desktop mills, I would advise against them. For the time and money I have in it I would have been better off selecting a higher quality machine from the start.

The most immediate problem I encountered was connecting the control box to a computer. The connection is to be made by parallel port, which is a rare feature on modern computers. To run this machine I purchased a refurbished business surplus workstation that came with windows XP. This was important because Mach 3 only has parallel port drivers for 32 bit systems and if I recall correctly XP was the last supported operating system. The computer didn't come with a parallel port so I bought a card to put in a PCI slot. The first control box I tried would not communicate with the PC no matter what I tried. I read that one solution was to use a Smoothstepper motion control board so I ordered one. As it was shipping I tried a different control box (a shipping mishap left me with three of them - long story) and I was able to get motion. I decided to keep the Smoothstepper anyhow because it was highly spoken of, so the new connection went by ethernet cable to the Smoothstepper, then via parallel port to the functional control box.

I then had difficulty controlling the motion due to incorrect pinout settings in the software. The manual was poorly written and of no help, but I muddled through with trial and error. I still do not have optimized profiles for the steppers but what I have worked and I needed the 6040 to be running so good enough. Mach3 was giving me many other issues as well, and in talking with their customer support I discovered that the copy of Mach3 included with the 6040 was pirated. After ridding myself of the pirated version and installing a legitimate copy most of my problems disappeared, although the stepper settings are still iffy and I need to optimize them.

Another problem that immediately presented itself was the nature of the machine's bed. Poor quality aluminum extrusions meant there was no way to tram the spindle to the bed due to uneven and erratic warping. To fix this problem I purchased an aluminum tooling plate, drilled and tapped holes on my Bridgeport, and mounted it in its place. This greatly increased the usefulness of the 6040 by increasing rigidity and allowing me to tram the spindle.

The leveling feet of the 6040 had been damaged in transit, and after having already had one lengthy battle with the Chinese seller I didn't attempt to get replacements sent out. Instead, I purchased a scrap cast iron surface plate and drilled through holes to mount the 6040. Here I believe I made a mistake: In an effort to damp vibrations, I put plastic vibration damping pads between the machine frame and the base, mounted on the bolts I used to join the two. In hindsight that seems to me counterproductive: The mass of the surface plate (I haven't weighed it but I can barely lift it by myself and I'm a big guy) and the properties of its cast iron provide plenty of vibration damping, and the gap between the 6040 frame and the surface plate/bed allowed for flexing and didn't take advantage of the rigidity the surface place could have provided.

I dropped the whole kit and caboodle into a concrete mixing tub, filled a bucket with coolant and a fountain pump, and it was good enough to press into service. I was getting acceptable results although the machine ran very slowly. It filled my business need and so I left it alone.

After being in use for a couple years I noticed a drastic decrease in the quality of engravings it was outputting. Worryingly, the 6040 seemed to be less rigid and I could flex the machine dramatically by applying pressure to the spindle with my hand. Since my need for engravings has tapered off I've decided to investigate what's going on and see what I can do to turn this into a decent little milling machine. I have several CNC conversions planned and no machine building experience so this is a good first project.
 

Attachments

  • engraving setup.png
    engraving setup.png
    2.3 MB · Views: 337
  • 20151220_195429.jpg
    20151220_195429.jpg
    3.2 MB · Views: 308
  • semper fi.png
    semper fi.png
    1.5 MB · Views: 294
To begin, I disassembled the machine bit by bit checking for loose screws, wear, or damage. I discovered the problem to be the coupling between the ball screw and stepper motor on the long axis (I call it Y). The coupling was bent, causing a severe runout of the ballscrew that made the gantry wobble in the X direction as it travelled along Y. This wobble loosened the socket head cap screws that hold the linear guide rails for the Y axis, meaning that the squareness of the frame was being maintained by two pieces of thin sheet metal only, as well as allowing the gantry to rock back and forth both under load and when just moving. This was responsible for the poor engraving quality and surface finish. I ordered some higher quality couplings that will provide a better connection.

One thing I don't like is how the ball screw is supported. On one end it has a small bearing, but on the end with the stepper it is coupled directly to the stepper's output shaft with no other support. It seems to me that a superior mounting system would have bearing supporting both ends, but I'm not sure how to accomplish that here without significant rework or buying a longer ball screw.

Another issue I aim to correct is the nature of the linear guide rods. They're not supported along their length, merely attached to the frame at both ends. This allows them to flex more than I would like them to. Supported guide rods are superior and I have ordered some. I plan to retain the original unsupported guide rods but also add two supported guide rods. I stoned the surface rust off the base and laid out the components how I plan to mount them. The parts I've ordered are expected in Wednesday the 26th so that will be my next update.
 

Attachments

  • coupler.jpg
    coupler.jpg
    143.1 KB · Views: 259
  • motor mount view.jpg
    motor mount view.jpg
    202.6 KB · Views: 251
  • y layout.jpg
    y layout.jpg
    195.4 KB · Views: 348
"Rework"? Sounds more like a 'Re-Engineer' and 'Re-build'. :grin:
Looking forward to your progress.
The Cast Iron surface plate is cool.
 
Yes the cast iron surface plate is pretty nifty. Until I saw it I had no idea such a thing existed, having only seen granite surface plates. I'm fortunate enough to live reasonably close to HGR industrial surplus and occasionally wander through their cavernous facility to window shop and impulse buy. When I saw it I knew I would find a use for it and bought it for pretty much scrap weight price. It's around 160-170 lbs and is a real delight to pick up and move :eek 2:.
The couplers arrived early and I went ahead and mounted one to see if that improved the runout. It certainly reduced how far out the ball screw is but to my eye it's still off somewhat. I don't know how one would go about indicating a ball screw. I also don't like how it's supported on one end merely by the stepper motor. I ordered a pillow block bearing to match the shaft. I think I'll have to find someone with a 4 jaw chuck lathe and have them turn another inch or so of the screw down to the 8mm shaft diameter to allow use of a bearing - as it stands right now the coupling eats up the tiny rod end. At that point I can move the pillow block around before bolting it down, with an indicator on the current rod end to get it running true, and then install the coupler and stepper.
The alternative would be to leave it as is and count on the four linear guide rails to maintain the gantry's position and let the coupler flex to take up any runout. Less work but I'm not sure I'll be happy with it like that and I don't feel like doing this again so I'd like to get it right.
 

Attachments

  • thicc plate.jpeg
    thicc plate.jpeg
    1.4 MB · Views: 244
  • newcoupling.jpeg
    newcoupling.jpeg
    1.5 MB · Views: 310
Back
Top