Build Thread- Summer CNC project

[Farasien]

Hi all-

This summer’s project for me is to design and make the hardware for my HF 7X10 CNC conversion. I’ve looked at some people’s methods on how to do this, and was wondering which would be the best way to go about it. I’m hoping to make it capable of both CNC and manual operation and would like to keep it accurate, though I don’t really need sub-tenth-level accuracy for what I do (+/-5 thou is pretty much good enough for me). From what I can see, there are 3 main methods people have used…

1. Attach the main feed screw to the stepper/servo motors… The pros I’ve heard on this is its simple and it lets you go back and forth from CNC to manual with a flick of the power feed switch. The main con has to do with accuracy and premature wear of the screw due to constant use in this way.

2. Rig the saddle directly, either putting the motor directly on the Z axis shaft or by attaching a timing belt to the back of the handwheel and running that to the motor. The pro I can see here is easy CNC/manual use, but the big con there I can see could be a low level of accuracy (I haven’t measured what the backlash in my rig is, but its clearly visible).

3. Run a separate ballscrew (or other type of screw) to the saddle and run this to the end of the lathe where the motor mates with the shaft independent of the other systems on the lathe. The pro I see on this one is possible greater accuracy (with a properly done ballscrew setup) but the con I see there is it will make it all but impossible to use manually; every time you try to crank over the Z feed, you’ll have to fight against the ballnut and could cause it to wear out faster by attempting it. This one also looks to be the most expensive setup due to a new ballscrew and nut. Does anyone know a way to make it possible to use this setup manually with the existing handwheel on the apron for the Z feed?

Does anyone have any thoughts or opinions on this? So far, I’ve seen Ron Steele’s conversion, FigNoggle’s and the one being built at CNCFusion- all three look to use option #3. JFetting’s design (the free one floating out there on the net) uses option #1- his looks OK, but I have to wonder at his using a compression/clamp fitting to hold the Z axis motor on the machine- it looks fairly weak and potentially inaccurate (due to slippage through use) to me. Does anyone have pictures or links to any other conversions I might reference? Lastly, I’ve seen a few places where there are CNC’d turrets out there (I have 2 sets of plans already and seen one where the author sells the plans). Does anyone have any thoughts or advice on using/making these or know of any plans on the net?

Thanks-

Farasien

[sansbury]

I'm just starting to make a set of motor mounts for my 7x10 based on the JFettig design. I have some motors and controllers kicking around so it will be a cheap project at least initially. My X-axis feedscrew has around .008" of backlash which seems impossible to adjust out so I suspect that will be my limiting factor. I'm more interested in automating tedious work than being able to turn perfect balls and arcs so I'm going to give the cheap-and-dirty approach a workout first.

[Farasien]

Only 0.008? Mine seems a bit more than that... Any idea how much backlash the gear train has on the apron, by chance?

One thing I saw alot for the CNCing of the X2 mini mill was the use of acetal nuts on the feed screws to eliminate play... Do you think something like this might work for the X axis on the 7Xx machines?

Also, I looked a long time at JFetting's design... I think a simple fix for the issue I was seeing for it might be to make the motor mount for the Z axis a little larger (and out of a bit thicker of aluminum, or else you'd have to make a spacer plate of some kind) and mount it directly to the larger hardware inside the case behind the feed screw's drive gear. Going that route could eliminate the compression fitting method he used. While the method he designed might work, I worry about play creeping into the system over time. Each time the morot changes direction, I think it could lead to loosening of the coupling unless it was constantly checked. I personally think might eventually lead to serious accuracy problems and possibly cause faster wear to the end of the screw shaft.

If the acetal (or maybe some kind of compression fitting on the feed nut like some folks do with ballscrews? Hmm...) nut idea would work, with a beefier mount plate bolted to something a bit more stable, it might make for a pretty solid system. The only issue I see still blocking that route is premature wear of the feed screw- though this might be overcome with a half nut made of some kind of softer material or something...



-Farasien

[sansbury]

I agree that in theory the JFettig design could get loose but after building 4 machines I've found that an hour of practice is worth about twenty of theory. The JF design is cheap and easy to try and I am going to try it before wasting time solving problems that may not exist.

Also bear in mind that if the mount were to slip a little, any error would be divided by the feedscrew pitch, so a slip of say 10 degrees would be converted into an error 1/16th or 1/25th of that, which would be less than .001".

I'm measuring my backlash based on the dial. You can feel the feed freewheel when you reverse the direction and after .008" it engages again. There is a primitive backlash adjustment mechanism which allows for taking out some of the slop. I can get it under .008 but it doesn't stay for long. I am planning to manage this by taking all my cuts in one direction in cases where precision matters.

[blades]

I'm also working on a 7x10 conversion this summer. My main criteria is that I want the lathe to be useable in manual form as well as CNC (only hand-crank turning in manual, no power feed or threading). The second item that is important to me is to not grow the footprint excessively. I don't want my motors hanging off the machine, I want them tucked out of the way. My cross-slide motor will be tucked underneath, using a timing belt drive. I haven't decided on the Z axis as of yet... but I may tuck it under the chip guard and belt drive it. We'll see....