[slimneill]

Background info

Father/Son team, we would be consider woodworking enthusiasts who have taken on some large custom projects for friends and family who understand that quality custom projects can't be done for the prices you find at Leon's. Some projects include custom wood furniture, entertainment centers, full kitchen renos and a spiral oak staircase.

There is no end goal of starting up a paying shop once the CNC is working, it's just another tool to improve our shops capabilities and advance the hobby. And if work comes along to help recoup some of the costs, bonus. As soon as you need it to make money, it becomes a job, which then it isn't fun anymore, right?

The old guys' been working with large CNC routers for many years in the Aerospace industry, cutting aluminum parts out of sheets for wings and fuselages. He has built up tooling over the years for a small machine shop and a decent wood shop.

The young guy, has a background in Mechanical engineering and design projects. Has a new house and lots of complex projects that would become so much easier with a new CNC machine.

We started researching DIY CNC machines a few years ago and trying to figure out:

A- What was a "must" and what was a "want"
B- What we could realistically afford on our budget
C- What we think our future uses would be and making sure that we tried to incorporate those capabilities into the machine up front.

Things that were a "MUST"
- 3 Axis Servo Motors (wanted closed loop for accuracy)
- 3 Ballscrews, don't want to deal with dual motor X axis.
- 4x8 cutting area

I'll be trying to keep the posts short and the photos of our progress plenty.

I have a separate thread which was basically the electronics portion of the project.

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

This will be the mechanical side of it and final assembly.

Here are a few pictures of the initial design concept. Not everything is set in stone. There has been a lot of reverse engineering required as I will explain in future posts.
Initial Concept was based on numerous builds we had read about, namely MadVac's design.
MadVac CNC - home made 4'x8' cnc precision gantry router

[slimneill]

The original plan was to make a 4x8 steel construction table with ballscrews and Hiwin linear rails and affordable hobby grade servos and components, not cheap but doable with proper budgeting. That plan all changed when work decided to get rid of an older Thermwood CNC machine that had been sitting in the corner for years as a pile of spare parts to the main Thermwood CNC machine. Well that machines' parts became the new basis for our machine's design. Having to redesign a lot of the machine to fit the new found freebies.

We were able to save a 5' x 10' frame from an old assembly jig from the scrap yard. It was made of 6" x 6" steel tube with six 6"x6" - 3/4" thick steel pads welded on the 4 corners and the middles then all machined flat as mounting plates for assembly jigs.
The Jig had legs on it, making the 5x10 working platform vertical. In order to convert it into a horizontal table for the basis of our CNC machine. The 2 legs had to be cut off and turned into 4 then re-welded together. Instead of scrapping the legs and starting over, we ground off all the left over welds and made it smooth again. We also saved the base feet of the jig, the feet which were 6x6 and about 48” long with ¾” plates already welded to the bottom for jacking screws

[JerryBurks]

Looks like a plan. Impressive size and structure.

I am almost sure you will regret the single x-screw at the size you are building. Your strong linear bearings will surely resist some racking but don't forget one of them must move at least a small distance to even begin building up resistance against further non-parallel movement.

I suspect you want the machine cut fast (higher forces) and when the y-axes is at their outer limits you have substantial leverage. Oh well, you got to try it out. If the machine turns out too sloppy you can still change to 2 screws. If you don't want 2 motors, a belt connection is easy enough to do.

[slimneill]

Looks like a plan. Impressive size and structure.

I am almost sure you will regret the single x-screw at the size you are building. Your strong linear bearings will surely resist some racking but don't forget one of them must move at least a small distance to even begin building up resistance against further non-parallel movement.

I suspect you want the machine cut fast (higher forces) and when the y-axes is at their outer limits you have substantial leverage. Oh well, you got to try it out. If the machine turns out too sloppy you can still change to 2 screws. If you don't want 2 motors, a belt connection is easy enough to do.

Thanks for the feedback, the CAD renderings are of an older version of the design. The gantry has been beefed up a bit since those renderings were taken, Really high cutting speeds are not a major concern for us, it’s not being used for production runs to mass produce a million widgets, it’s just for hobby use. If it takes longer to cut, so be it. It will still be faster then trying to do it by hand. The gantry is expected to be hovering near 1000lbs when all said and done. And I’m not planning to need to go 1000 IPM Rapid to go 10 feet. If it takes an extra 20 seconds to get back to home position, oh well.

We thought about ways to fight the potential racking issue. We were able to get the gear racks off of the large CNC machine. And we plan to install them on the inside of the 6x6 table and then run a solid 1” steel shaft, with spur gears fixed on either end of the shaft, from one rack to the other and mount the rack to the bottom of the gantry using 4 mounted bearings. The idea is that by locking both spur gears in unison on the racks as the ball screw pushed the cart up and down the table, any off center cutting that may cause the cart to rotate and rack will receive a reaction force in the opposite direction caused by the torsion forced created in the 1” steel shaft twisting. It will act like an anti sway bar in cars, but for the CNC gantry.

Based on some quick calculations, the 1” diameter shaft could twist half a degree if a torque of 200 lb-in were applied from one end; which is 4 times the stall torque of the motor which is 50 lb-in. I haven’t seen anyone else use this idea before (probably due to the added cost and time) but it makes sense on paper, only one way to find out….

[JerryBurks]

......... And we plan to install them on the inside of the 6x6 table and then run a solid 1” steel shaft, with spur gears fixed on either end of the shaft, from one rack to the other and mount the rack to the bottom of the gantry using 4 mounted bearings...........Based on some quick calculations, the 1” diameter shaft could twist half a degree.....

I have not seen this either, but it should do the job.

[slimneill]

This is what i was trying to describe earlier.

[JerryBurks]

I get it

One more thing that comes to mind...you have that free hanging table table substructure which is supported in the long dimension only by the 2 side tubes. I suspect this will give under varying load from above which is very annoying when you need to keep a precise depth (V-carvin, inlays and the like).

You may have calculated that already and found acceptable but if you would turn the connecting tubes 90 degrees (i.e. also lengthwise) you would increase the rigidity substantially without adding material. Keep in mind the 4-ft tubes are 8 times as rigid than the 8ft tubes for a point load. Connecting shear plates might make it even better.

[slimneill]

One more thing that comes to mind...you have that free hanging table table substructure which is supported in the long dimension only by the 2 side tubes. I suspect this will give under varying load from above which is very annoying when you need to keep a precise depth (V-carvin, inlays and the like).

You may have calculated that already and found acceptable but if you would turn the connecting tubes 90 degrees (i.e. also lengthwise) you would increase the rigidity substantially without adding material. Keep in mind the 4-ft tubes are 8 times as rigid than the 8ft tubes for a point load. Connecting shear plates might make it even better.

That is a good point. We haven’t really focused that much on the floating table top design. We just know that it will be made out of 2" x 3" tube standing on end. There have been bigger hurdles to figure out so far. I was a little worried about the sagging, but making the tubes run length wise will help that issue a lot.
There have been a few table top designs that we have tossed around. We are considering toping the steel frame table with 3/4" aluminum plate and screwing it down to the frame for firm that all up and then top it off with a few inches of plywood as spoil board. Or even a possibly a partial vacuum table. It all depends on what accuracy we are actually able to achieve. If we can get the machine supper accurate, then it may be worth investing in some plastic plywood sheeting I read about on a shopbot thread. And then just coating the table with epoxy putty and fly cutting it smooth again every so often. The plastic would be a lot less prone to thermal expansion or humidity changes in the shop.

[slimneill]

Table Legs

The height of the 4 legs turned out to only be 8” once all of the various components were layered up in the model. It was important to keep the working bed/spoil board of the table less than 34” from the ground. 34” is the working height of a nearby table saw. Seeing as this CNC machine is going to take up a lot of shop floor space, it is important to make sure that it won’t affect the other shop operations when not in use.

There are two issues with dealing with large steel structures. Keeping everything square and level while welding, and dealing with the weight once everything is assembled or welded together. Even though we plan to use large and thick steel plate for a lot of the job, which should disperse heat pretty well and eliminate shrinkage during welding, we still want to minimize as much welding as possible.

In addition, we needed to keep in mind that this machine may need to be moved in the future and the work shop isn’t tall enough to accommodate a forklift, so everything needs to be able to be dismantled and carried out by hand.

In order to fix the legs securely to the table top without drilling the 6x6 full of holes, we went with a mounting plate approach. We started with two ½” flat plates machined smooth. One plate received 4 through holes. The second plate received 4 tapped holes. The idea was to have the plate with the tapped holes welded to the underside of the table, and the plate with the through holes welded to the tops of the legs. The legs would be welded directly to the base feet.

The 4 legs were created by cutting the 2 vertical legs of the jig into 4, 8” sections with a metal band saw. Then all surfaces were machines square on the mill.

The legs components were then taken out to the work shop where the table was sitting up on car jacks. Once everything was made level and square, the legs were welded to the base feet, the plates welded to the legs and then the bolted plates welded to the bottom side of the table. The end result was a secure square table with legs that could be removed if needed.

[slimneill]

....

[JerryBurks]

........We are considering toping the steel frame table with 3/4" aluminum plate and screwing it down to the frame for firm that all up and then top it off with a few inches of plywood as spoil board.......

If you put aluminum or wood on only one side of the steel frame you need to let it float horizontally. If you bolt it down hard it will create a bi-metal effect and variable bowing. I think with about 2" center distance and 8' length it would be substantial. There are formulas around to calculate that. If I remember correctly aluminum expands 3 times as much as steel with temperature and most wood about half.

[BanduraMaker]

And I’m not planning to need to go 1000 IPM Rapid to go 10 feet. If it takes an extra 20 seconds to get back to home position, oh well.

Famous last words!

[slimneill]

If you put aluminum or wood on only one side of the steel frame you need to let it float horizontally. If you bolt it down hard it will create a bi-metal effect and variable bowing. I think with about 2" center distance and 8' length it would be substantial. There are formulas around to calculate that. If I remember correctly aluminum expands 3 times as much as steel with temperature and most wood about half.

Another good point and something to consider down the road.
I just did a quick calculation, for a 9 foot long bed and a shop temperature ranging from 60 to 80 F

Wood would expand 5.8/1000”
Aluminum would expand 26.6/1000”
Steel would expand 15.8/1000”

So:
Wood/Aluminum Delta: 21/1000” over 9ft
Aluminum/Steel Delta: 11/1000” over 9 ft

I’m not sure if you would see noticeable bowing if the difference 0.021” over 108”.
The work shop is heated in the winter and cooled in the summer, so a temperature difference of 20 degrees would be pushing it.

A good online resource for more info is: Coefficients of Linear Thermal Expansion

[slimneill]

The original plan didn’t have a center 3rd leg, we were going to use the center under the table for storage of plywood and materials etc. But we found that when you have 400 lbs worth of people jumping up and down on the center of the 6x6 beam , it was deflecting 0.030”. So to make the table as rigid as we could we made a center leg using the same construction methods used for the 2 outside legs. The horizontal foot was made from 6x6 found at the scrap yard. The vertical legs were made of 4x4, which was original diagonal bracing on the original Jig. After the 3rd leg was installed there was no deflection of the beam when people jumped on it. And the table didn’t ring half as much when it was hit with a hammer as it did unsupported in the center.

[JerryBurks]

........I’m not sure if you would see noticeable bowing if the difference 0.021” over 108”........

If my calculation is not totally off and assuming a center distance of 2" between the 2 layers, the curve radius would be 10,285", resulting in a bowing of 0.14" in the middle. That is surely noticable. If you keep the temperature stable enough, it may be O.K. but since you are still building the machine you could just as well avoid it.

[harryn]

My garage cycles 20 F (and sometimes 2X that) nearly every day, so it is something I definitely think about.

It is one of the things that pushes me toward a fully steel tube + steel rail approach (some day) although interestingly, this means that the wood will still be cut incorrectly except for a few times of day. No doubt I will mess that particular cut up myself.

Jerry, how do you deal with your steel rails + wood frame setup and this expansion ? It is not clear to me how to apply a floating rail concept + rigid mounting.

Harry

[JerryBurks]

Jerry, how do you deal with your steel rails + wood frame setup and this expansion ? ....Harry

Well, I don't. But I have a much smaller machine and the x-rails are in one plane with the extrusion table. I suppose the steel rails can anyway slide on the aluminum support extrusions. The little M5 bolts are not going to stop them from expanding. The table extrusions in turn are bolted to the wooden torsion box but the bolt holes are not so tight that they would not allow the bolts to move a few 1/1000".

I think the problem will only manifest itself if the dissimilar materials are glued, doweled or otherwise rigidly connected. That was what I meant with float. After all, the linear movement is very small.

[slimneill]

All you need to do is secure the table down tight at your "Home position" and then secure it down with either slots or looser everywhere else so that it will all expand in one direction varying amounts. Just don’t fix it rigidly at both ends of you tables and you should be fine. We are talking a couple thousands of an inch growth, not a couple inches.

In our case, if we use the aluminum in the middle of our sandwich is will expand more then the steel below or plywood above. Pinning it all down at the home position should keep the plywood in one spot on the table while still allowing the aluminum substructure to expand and still be used for fastening to or prevent leakage in a vacuum table configuration without affecting the cutting surface.

Besides, I would assume that changes in humidity in work shops is a much bigger issue for changes in cutting surface accuracy then thermal expansion ever would.

[hamzagi]

hello,

your design is very interesting,

I've posted my design but the comments were not helpful.

http://www.cnczone.com/forums/diy_cn...st_design.html

what do you think ?
the problem of racking was fixed on your machine?
do you have the last pics ?

regards

[slimneill]

A lot of progress has been made since the last post. I’ll continue from where we left off.

The 6x6 square tube frame had six 6”x6”x¾” pads that were welded on and then machined flat to make one flat plane. Across the entire table. This surface was going to be our best chance at getting the X-rails level and flat across the table. The problem was that the reference pads were only at the table ends and midpoint. The rails wouldn’t have a flat surface to rest on the rest of the way. We decided to use jacking screws between the rail’s cap screws to push up on the rail, this allowed the rail to float flat ¾” above the 6”x6” tube while still being fine-tuned to keep perfectly flat along the length of the x-rail. Once the rail was bolted down, the gap was then filled in with vibration resistant grout to lock the jacking screws in place and reduce vibration while adding mass to the machine’s foundation. We decided that it would be easier to keep the rails flat and support the load better using jacking screws then ¾” gap of epoxy.
Attachment 211490 Attachment 211492 Attachment 211494 Attachment 211496 Attachment 211498 Attachment 211500