GME's New 80/20 CNC Build - My Design

[GME]

Even have milled surfaces for the rail mounting.


https://us.misumi-ec.com/vona2/detai...1503634&Inch=0

Yes but, their profiles are different in sections.

Yup, a nice service. Understandably, you pay extra for milling.

As David noted, the Mitsumi extrusion profiles are very different than 80/20's, even the 80/20 metric. I like 80/20's layout better, but that's just me. I've seen a lot of different cross sections on extrusions in other parts of the world. Most appear thinner than 80/20's, which I wouldn't go for. 80/20 has a couple of "lite" versions (read that as "thinner"). I don't buy them. Actually, I believe CNCRP uses a lite version in their machines - at least that's what their instructions depict. You can tell by the holes in the 4 corners as viewed from the end. You can also tell by the sort of two X shapes in the center 1/3 quadrants. That where you can thread the holes to put screws in from the end. On the full thickness heavy stuff, it's just a regular hole. In the 15 series stuff, the hole is the correct size to take a 5/16" tap. I've tapped my share. It's a nice feature.

Not suggesting there is anything wrong with the "lite" stuff. There's a market for it. It's just that the heavy stuff is stiffer and heavier, both of which are pluses when building a CNC. Also, as you would expect, the heavy stuff is more expensive.

Gary

[davida1234]

I like 80/20's layout better, but that's just me.

I personally don't like their profiles.

Their cross bracing is thinner than 80/20.

I don't even know how an engineer could put that kink in there.
Right where it transfers bending loads onto the skin of profile:

[wmgeorge]

Long story short, it appears 80/20 does the processing.

What I am saying Gary is there can be different standards, one for Retail and another for CNCRP otherwise there would be hundreds of customers upset like you were when you received Your order. How else would you account for the customer satisfaction they have?

Mitsumi extrusion profiles can be had in many different configurations. BTW The ones I ordered when I was going to build my large scale laser were cut right to spec. They are used in many different applications without issues. The ones that are milled are an option for folks like you using the linear rail to insure the rail is mounted flat and true.

Speaking of milling is 80\20 also doing that for the CNCRP pieces they cut, the round holes for the fastners?

In any event we need to let this drop and let you get back to making your machine!

[Peach52]

Sorry, but I don't understand. Your link took me to Ebay for something about suspension parts. I suspect the link was unintended. I've had CNC Zone add links to posts that I didn't input and made no sense. In fact, it happened to me with this post. I had edit to change the B in Ebay to lower case to make the link go away.

Are you indicating that you cannot see the photos in post #4? Please clarify.

Thank you,

Gary

I can't see them even if i click on the link,it says the attachement link is not good!

P.s. sorry if my english is not perfect, first langage is french

[GME]

I can't see them even if i click on the link,it says the attachement link is not good!

P.s. sorry if my english is not perfect, first langage is french

No worries about language. You've done fine.

I cannot edit post #4, so I've reproduced it below. Please confirm you can see it now.



I forgot to mention the hard stop I made to get the short pieces to precisely the same length. Here's a photo:




You may have noticed that the cross rail in the front is set back some. I did that to allow me to cut dovetails off the front of the machine. I took measurements from the Z axis assembly and spindle. I determined that the spindle will stop past the the front most extreme of the frame, but not enough to comfortably cut dovetails. The setback was to provide 2" of cushion. Here is a photo of installing the rail with setback.




After installing the cross rails, I cross measured with a tape measure to ensure that the frame was square. It was as perfect as I could get reading off a tape measure. No adjustments required.

I then proceeded with installing the risers and top Y axis rails. Here is a photo of the frame:




I cross measured again. Still as perfect as I could have hoped for.

Next up was doing the final leveling of the machine. In the photo above, you will note that the angle braces are installed. I fussed and fussed with the machinist level, but could get it level. I thought I would have to tear the whole thing down and start over when it occurred to me the angle braces might be the source of they problem. Turns out they were. I loosened them all up, and went back to leveling. I was within .0005" all around with the braces loosened. Then, taking the utmost care with positioning, I reinstalled the braces. It worked. My leveling held.

The exercise told me that I got my lengths right and the ends square on the mill. and I avoided any twist during assembly. Besides taking great pains to get things right, using something as simple as the angle to get the extrusions aligned paid off. That the extrusion ends were square was not in question. I measured each end for square before it left the mill. I also confirmed lengths before I started assembly.

Next up: Gantry interface plates.

Gary

[Ntl]

Looks great Gary looks like all the hard work is paying off!!

Dan

[GME]

Looks great Gary looks like all the hard work is paying off!!

Dan

Thanks, Dan. It's coming along nicely. It goes together way easier when all of the extrusions are cut square and to precise length.

I have the master rail installed on the X axis. I didn't photo document it, because I used exactly the same methodology I used on the Y axis master. Went easier, since any kinks in the process were already worked out on Y.

The X axis was interesting. I am using 25mm rails on Y, and 20mm rails on X. The 25mm rails were pretty stiff. The 20mm much less so. Took very little to get them to flex. Still, the method I used to set the master made flex a non issue. Just hold the rail tight to the straight edge and incrementally tighten the cap screws. I tested along the way with .001" steel shim stock. Couldn't fit it in anywhere, so I have confirmation that I'm good to something less than .001". I believe I am good to the tolerances of my straight edge, which is .0002"/foot. I am happy with that.

Gary

[Ntl]

Nice, I want to see the gantry mounted up. I like the way you did the bracing and mounting plates. I don't remember have you posted the z axis yet? I'm curious if you bought one or made your own.

Dan

[wmgeorge]

Looking good Gary!!

[GME]

Nice, I want to see the gantry mounted up. I like the way you did the bracing and mounting plates. I don't remember have you posted the z axis yet? I'm curious if you bought one or made your own.

Dan

I hate to admit it, but I bought the Z-axis. It's the CNCRP pro ball screw model with 12" of travel (the larger one). I plan to leave the generic linear rails/bearing blocks in place for now. I may upgrade to Hiwin downstream, but I haven't made a firm decision yet. I haven't posted pics of it yet. Soon!

Thanks about the gantry bracing. I designed the interface/mounting plates and and braces in Fusion 360 as a whole assembly. That's why the many threaded, blind threaded and through holes all lined up like they should. I created a job sheet cut drawings for each of the parts. Would have been tough without the drawings.

The bracing is probably overkill in the extreme, but it took little more time and not much money to over-engineer them. I'm pretty happy with the way they turned out.

I ordered an 11mm drill bit, so I could drill the tap holes for my proximity sensors. It arrived today. I also need to drill and tap the holes in the gantry interface plates for the rack and pinion drives, and fabricate a part for the R&P drive tensioning screw. Shortly thereafter, the gantry goes on.

I took the time to clean out the cutting oil from all the threaded holes. I used spray brake cleaner and Q-tips. Worked great. I needed them completely clean, or risk that Loctite wouldn't hold the screws in place when I assemble the gantry brace. Last thing I want is to have covered/hidden screws start backing out.

Gary

- - - Updated - - -

Looking good Gary!!

Thanks, Bill.

[wmgeorge]

I really don't think you can make it too strong.

[Ntl]

I'm curious how you're doing the rack and pinion, that's the weak link strength wise on my machine. Not so much on the y axis, but I get some flex from the RP drive where it mounts to the z axis in the x direction. I'm not a fan of the RP set up on my machine and would definitely have done it different if I built one from scratch.

[GME]

I'm curious how you're doing the rack and pinion, that's the weak link strength wise on my machine. Not so much on the y axis, but I get some flex from the RP drive where it mounts to the z axis in the x direction. I'm not a fan of the RP set up on my machine and would definitely have done it different if I built one from scratch.

I'm doing it pretty much the same way as on the Saturns and the CNCRP Pro kits. A better option would probably be helical racks and pinions. An even better option would probably be ball screws and rotating ball nut. Both would require some serious re-engineering. I've thought I may upgrade to helical sometime later, if I can work out a suitable R&P drive mount. Ball screws with rotating ball nut is probably outside my abilities. davida1234 has done a remarkable job with the rotating ball nut. A very talented guy. He's making and selling the parts, so that may also be an option - depending on the cost.

I never had any problems with the R&P on my old machine, but then, I didn't have the alignment problems on that machine that I saw with the Saturn. Both X and Y ran true on my old machine, and try as I might, I wasn't getting any measurable backlash. That was bound to change with wear, but no different with pretty much any drive system. My Z axis was different. It had a 5 start Acme screw. Out of the box, I had .001" backlash with it. As far as flex is concerned, I never observed any on any axis. It was an extrusion machine after all, so no surprises if there was some. I just wasn't about to observe any.

Gary

[Ntl]

On mine if I push on the spindle with the steppers locked, I get some movement in the x direction. I can see the rack and pinion drive flex. It doesn't move in the y direction or the z. I haven't looked at it very hard yet to see what exactly is allowing the RP drive for lack of a better word "flex". It doesn't seem to affect cutting, but I take pretty conservative doc and I haven't tried cutting aluminum other than some .063 sheet. I have a feeling it's because the drives are only held in place with one bolt. I have never been a big fan of the design. I just haven't put the time into coming up with a better way to mount them that is more rigid.

Dan

[GME]

On mine if I push on the spindle with the steppers locked, I get some movement in the x direction. I can see the rack and pinion drive flex. It doesn't move in the y direction or the z. I haven't looked at it very hard yet to see what exactly is allowing the RP drive for lack of a better word "flex". It doesn't seem to affect cutting, but I take pretty conservative doc and I haven't tried cutting aluminum other than some .063 sheet. I have a feeling it's because the drives are only held in place with one bolt. I have never been a big fan of the design. I just haven't put the time into coming up with a better way to mount them that is more rigid.

Dan

To use a drive like Fineline and CNCRP uses, you are pretty much stuck with one bolt to hold it in place. Otherwise, you can't use a tensioning spring. It has to pivot on something. Maybe a larger diameter shoulder bolt? Unfortunately, that would mean redesigning and cutting out new drives. I doubt you can "fix" the existing drives.

Gary

[GME]

Time for an update. My latest task was setting the second rail on my gantry. I placed shop towels on my Y axis rails and placed the gantry on them. I decided to put my rails on without having to fight gravity, so I laid it with the rail surface facing up. Attached are two photos of the setup I used to set the rail:







I started by loosely installing the second rail on the gantry extrusion. Then I added the linear rail bearing blocks, added grease and attached the Z axis mounting plate. I believe I may have mentioned it before, but in case I didn't, I am using the CNCRP ball screw Z axis. The plate mounted to the bearing blocks is theirs. You will see a piece of steel flat bar on the mounting plate, with a magnetic attached to the steel. I held the steel with a couple of screws, utilizing some of the 8mm holes already drilled in the plate. You will also note a dial test indicator wit the probe touching the lower edge of the linear rail. The indicator is a 0-.03" with .0005" graduations. I started at the right side and worked to the left.

On the right side, I let the mounting plate generally set the spacing between the two linear rails. On of the recognized methods for setting the second rail is to set up the two rails and plate as I have done and simply move the plate along and tighten the rails as you go. While recognized as a legitimate technique, it is not a preferred one. As I set my rail, I learned why. Anyway, like I said, I started out by letting the plate set the spacing. I then added the test indicator, zeroed it, and tightened the first mounting screw to the left. I moved to the next screw to the left, and fiddled with the alignment until I got it locked down at zero. The reason why just moving the plate, and tightening as you go isn't a best practice has to do with slight rail movement as you tighten the screw. As you tighten the mounting screw, the rail moves. Makes sense, if you think about it. You are using a socket head cap screw that bears on the counterbore in the rail. As you tighten the screw, the screws bearing surface is turning against the rail. That cause the rail to move.

See the clamp in the photo? The claimp was how I was able to control the rail movement. I would move the rail about .0005 more than I wanted to get the rail to zero. When I tightened the screw, the rail moved to the zero position. It was a trial and error process for each screw. For some, it a .0005 adjustment. The next screw may take .00075, and the one after that, .00025. You just have to play with each one, and figure out what works for that screw. Very time consuming, but it get you to near perfection.

When I reached the left end, I move the mounting plate back to about the middle. I moved the test indicator to right of center on the mounting plate and set it to zero again. I moved the mounting plate to the next screw that hadn't been tightened. Of course, I checked the indicator along the way to see if my methodology was working. It was. I held zero all the way to the next untightened screw. I proceeded with the remaining screws on the right as I had going left.

After tightening all of the screws, I again set the test indicator to the left of center on the mounting plate, zeroed it, and ran along to the left. I was pleasantly surprised to see that everything held. I'm not going to represent that I was reading absolute zero all the way along. I had hoped it would, but didn't really expect it. Nevertheless I was pleasantly surprised to find that the indicator needle only shifted less that the thickness of the line for zero, which couldn't have been more than +/- .00015". I called it good. No way I was going to get any closer than that.

I have no prior experience with profile linear rails, so I was surprised at how much force it took to move the mounting plate from one side to the other. This was something I played with a bit before I started locking down the second rail. I left the left rail loosely attached to the gantry and noted the force taken to move the plate. After tightening the second rail, I loosened the screws holding the mounting plate to the bearing blocks. I could detect no change. That was satisfying. Since I don't have a device for measuring how much weight was required to move the plate (not even a hanging fish scale), it took a digital scale I have, turned it on, pressed it against the mounting plate and started moving. The scale read about 7 pounds all the way across the length of the gantry. I should note that the movement was very smooth. No hesitations or jerkiness. It just seem stiff.

At bit of information about my profile rails. I am using the Hiwin HGH-CA series. The HG is the series, and the third letter, H, means it is a high height profile. The CA means it's for heavy loads. There is also a HGH-HA, which is for super heavy loads. A fuller description of my system, is HGH20CAZAC. This means 20mm, heavy load, normal precision and a medium (versus light) preload.

My next task is to drill and tap my gantry interface plates for Pepprl & Fuchs 12mmx1 proximity sensors. While they are on the mill, I also need to drill and tap 3/8" holes for my rack & pinion drives and tensioning spring mounts. Once done, I'll be mounting the gantry and aligning the second Y axis linear rail.

Gary

[nlancaster]

Nice work, I am preparing to upgrade my CNCRouterparts Standard machine with hiwin rail, Can I ask why you with the HG instead of EG 20mm rail?

[ger21]

I have no prior experience with profile linear rails, so I was surprised at how much force it took to move the mounting plate from one side to the other.

That's probably due to the medium preload. Once there is more weight hanging from it, it will appear to require less force. I think CNCRP, and most other DIY users use light or zero preload rails, as they are more readily available. But the medium are a better choice, as they'll be more rigid.

[GME]

Nice work, I am preparing to upgrade my CNCRouterparts Standard machine with hiwin rail, Can I ask why you with the HG instead of EG 20mm rail?

Thanks. Nothing profound in my choice. I liked the little extra steel in the HG series. EG would have been a little better to minimize lever effect of the Z axis, I suppose, but that wasn't really a consideration. The overall height difference between the HG and EG series for 20mm is only 2mm. Truth be told, I am building a beefier than normal machine, and wanted beefier profile rails.

I see no reason why the EG series wouldn't work just as well.

Gary

[GME]

That's probably due to the medium preload. Once there is more weight hanging from it, it will appear to require less force. I think CNCRP, and most other DIY users use light or zero preload rails, as they are more readily available. But the medium are a better choice, as they'll be more rigid.

Thank you, Gerry. I appreciate the affirmation, and others will appreciate the additional information.