[harryn]

Hi, I decided to abandon my prior approach to a fixed gantry system and am trying again. The main problem with the last attempt is that I failed to think through the real impact of the build on the space I had available to work with.

So - trying again, but with more "experience"

This time, I would like to attempt to build the following:

Total Size
- Can be reasonably loaded into the back of my minivan by two people
- Approx. dimensions 39 in x 24 in x 84 inch (max space available ) including the dust / noise enclosure
- No table - it just sits on the concrete floor when in use
- Can store on its smallest end with the 84 inch dimension vertically up

Uses Planned
- Able to work with mostly plywood sheets (possibly up to 5 x 10 ft), 2 x 4s, and 6x6s, and down to 2 x 2 ft sheets
- Ideally, use my existing bosch 1617 evs router
- Cutting plywood sheets to size, shaping board ends, making some decorative around the house items
- Teen projects for school and education
- Wood, plastics, foam, r/c parts
- I might attempt an electric guitar body at some point - mostly for fun, not for sale


Performance Expectations / hopes
- Over an 8 ft board length, I would be very happy with consistent +/- 32 inch results, and 1/16th might be acceptable (about +/ 1mm). +/- 1/8th is not ideal, but it is still far better than I do by hand now.

- For holes that are 6 inches in diameter, I would be happy with +/- 1/32 inch between the x and y dimension, and possible happy with lesser results.

Obviously, better results are always welcome, but I am trying to be realistic so that I can actually afford to build this thing.

Idea so far:

Y axis
- Fixed gantry for the Y axis
- Possibly a bearing based solution, but the semi - DIY ones so far were not up to my expectations
- I am fairly convinced that to make skate bearings work correctly, I would need to use the swiss style skate bearings rather than the abec 7 style bearings, and those are not inexpensive at all.
- Close to 6 ft long axis

X axis
- some method of moving the wood through the unit, similar to how a planer works.
- Will take in the wood and move it through the router about 8 - 10 ft, so it has to be robust
- Ideally, easy to store in a garage with an 8 ft ceiling and limited storage space

a) Link to a sears wood planer
Sears: Online department store featuring appliances, tools, fitness equipment and more

b) Not ideal, but I suppose it is possible to use a table that is removable and attach a pair of racks to it for insertion into the router. Then I could just attach the blank wood to the table and push it in to start.

c) Very Open to ideas on this

Frame / Enclosure
- I am considering to make the final version from Al honeycomb panels like these in a 1 3/4 in thick version for light weight and stiffness. I need to see how much they actually cost first of course.


Honeycomb Sandwich Panels | Honeycomb Sandwich Panels

Thanks for looking and for any suggestions.

Harry

[OCNC]

X axis
- some method of moving the wood through the unit, similar to how a planer works.
- Will take in the wood and move it through the router about 8 - 10 ft, so it has to be robust
- Ideally, easy to store in a garage with an 8 ft ceiling and limited storage space

For moving the material I think you would want something more along the lines of a wide belt sander mechanism rather than a planer mechanism. The difference is that the platen has a conveyor belt which could be driven with a stepper motor and the pressure rolls are free wheeling. The improvement over the planer drive should be obvious. Just remember that if you're trying to move an 8' 2x around there's a lot of inertia to deal with. I think that there's a reason we don't see machines configured like this.

Chris

[harryn]

Thanks - yes you are right - it is the highest risk part of the project.

The planers and sanders have another thing going for them as well - most of the material is close to one thickness across the sheet. This will be true often, but not always on my projects. For instance, a guitar body might start out one thickness, but will not stay that way for long.

It could be that I am forced to use a some kind of standard "moving carrier table" (such as a large flat sheet of something) with rack gear mounted to it to make this all work. That would be much less convenient, but at least is a backup plan.

[jsheerin]

I'd assume it would be 39 wide, 84 long and 24 high? If so that rules out working with full sheets of wood right there... If you want to maximize the size of sheet you can work, I'd go with a moving gantry instead of fixed. The purpose of a fixed gantry is usually increased rigidity, but for the type of work you're talking about that's not required. You could easily get accuracy better than 0.01" if you just do a standard machine built from 80/20 framing and use cnc router parts components or something similar. I'd go r&p on the two long axes and acme screw on the z.

From what you've said, the only reason I can see not to do this is budget, but you haven't specified your budget. Going with the moving table schemes you're talking about would vastly complicate things and potentially limit accuracy quite a bit. If you just want to do it to do it, I totally understand that, but it seems like there's a very well established solution to do your stated types of work.

[harryn]

Hi, thanks for your post and link.

If I had room for a conventional moving gantry or fixed gantry design for full size 4 x 8 ft sheets, life would be much easier - indeed. I simply do not have the room for such a unit, and I don't have a need for a smaller capacity unit. Budget is always a factor, and I have not set one, but realistically, it will be moderate.

Since this is really for occassional use (not professional use) and one of the goals is that 1 or 2 people can load it into the back of a dodge minivan, that limits size considerable, and forces a non - conventional approach.

Please consider to follow the link to the sears wood planer in the original post. The idea is that the 7 foot dimension is the width of the exterior of the unit, and that the parts pass through the unit, more or less similar to how a planer passes wood through it. The 4 ft wide sheets will go into the router, which gives roughly 3 ft for making it a bit oversize, and the mechanical aspects.

No doubt, there will be some performance tradeoffs. The real challenge will be how to manage the wood as it goes through. For a perfectly flat sheet of plywood, it does not seem too challenging. For something more complex - say a 2.5 D object like an electric guitar body, or an RC airplane wing, then life is more interesting.

So far, the most robust approach I have come up with is to mount the work on some kind of spoil board table that has 2 racks pre-mounted on it. While not ideal, it is perhaps the lowest risk. Of course, this spoil board mount table will need to be larger than the full size sheet then - which is slightly painful to store.

This problem is exactly what failed on my first project, but in that case, it was even worse, as I had the 7 ft dimension mounted vertically and some additional space constraints.

[CarveOne]

Sounds like you are describing something like a large format CarveWright style machine. It would slide into the minivan edge-ways, leaving room for carrying other items beside it. To cut a 4x8 sheet the machine's width would need to be at least 5' wide to allow for mechanical parts. Like a planer, it could have fold up or removable extension plates for the in-feed and out-feed sides. Instead of a planer head it would have a moving Z axis head.

That is the idea behind the CarveWright machine. I almost bought one for my first CNC machine when Sears started selling their version of it, and decided that it was just too small for my aspirations. They should have introduced a 25" cut width version instead of the 13". I would have bought one.

CarveOne

[OCNC]

After looking at the CarveWright machine I'm convinced that trying to do this in the size you're considering would be a waste of your time and resources. Attached is a concept image of what might be a better solution. It consists of two 36" x 84" panels that are hinged on one of their long edges down the middle of the machine. The Y-axis is removable from the the dual drive X-axis carriages. Not shown are several removable stiffeners that would fasten to the underside of the table panels. The machine would require some kind of simple levelers and a basic leveling procedure each time it's knocked down and setup. My preference would be to drive the X-axis with rotating nuts. Knocked down and folded this should easily fit into your transport space. The only draw back that I see at this point is that you can't quite fully machine a 4' x 8' sheet of material without having to shift the material.

[louieatienza]

Another idea would be to make the table frame such that the outer stretchers fold inward on itself, possibly 8020 with hinges. Then the middle stretchers could be one piece and installed afterwards, and the table surface installed last.

You could put the drive motor for the long axis INSIDE the gantry beam, and drive both sides of an r&p via jackshafts and pulleys. I'd make the gantry raised, with the linear rails and racks to the side or underneath. This would give you the maximum working width for the table size. Putting all the motor components on the gantry beam may complicate things a little, but it would be self-contained and all wires would come from a single place. The gantry would have to cantilever past the uprights a bit to get full cutting capacity across the table.

[harryn]

Thanks for the inputs. Yes C1, the carvwright style was the design I had in mind, but could not remember the name of it.

It has some interesting mechanical concepts behind it, as well as the software for beginning users. (like me)

OCNC and Louie - you might be right that the carvewright concept might not scale up to full size sheets. I am very hesitant to make a design that needs to be taken very far apart each time it is used, or I will never use it, and certainly my kids won't use it then.

I am starting to lean toward using a table made from 1 inch honeycomb panel with rack on the sides that passes through the machine. In theory, I could use a smaller table for smaller jobs, and a second larger table for when that is needed. It isn't ideal, but a thin oversize panel can be stored more easily than an oversize machine.

It's still an open question, but there is plenty to learn in the meantime.

Last night I started playing with some scrap 1 7/8 in thick panel pieces to begin learning about how to work with this stuff. It is amazing material - incredibly stiff and lightweight. It also rings like a bell if rapped, so not exactly a vibration dampener.

I think there is a version made from polyeric material that is more sound deadening - I need to look into that version and see what is possible.

Step 1 - wear work gloves. The edges are razor sharp, and the edges of all of those honeycombs are sharp edges.

Step 2 - I am learning about close outs and bonding edges and L shapes.

a) Channel Edged

I originally thought about encasing each edge in a 1/4 in thick Al channel, and bonding those edges together. That would require a nominal 2 1/2 x 2 1/2 x 1/4 in channel without fillets (plus a small gap filler)

As luck would have it, channel that size is quite non standard, so I instead thought about taking square tubing, and having one edge milled off to make a channel. That would work, but it looks like a costly approach, and the channels would weight more than the panel pieces - by a lot.

b) Wood Trimmed Edges

More or less - make the channel, by bonding wood strips onto the panel edges. It can work, and probably will be needed for some spots.

c) Wood Filled Edges

Basically, you cut out a depth of the honeycomb from the edge, and fill it with wood or similar. As it turns out, these panel cores are exactly the right thickness for the 2 in side of a 2 x 4, so that is a viable approach.

d) Epoxy filled edges

More or less that same idea as (c) but you use epoxy to fill in the edge instead, and drill holes into the epoxy for bonding.

I can see that this material is expensive to use, so I might still be forced back to proto building in plywood, and then slowly convert over to the honeycomb.

[ger21]

I just quickly browsed through this, but imo, the only chance of getting anything workable is to take the "giant" carvewright approach.

Use a 37"x70" sanding belt for the feedbelt. The overall dimensions could be somewhere around ~48-56"(Y) x ~32"(X). The Y might need to be a bit bigger.
It would look similar to a large drum sander.

The biggest obstacle would probably be feed belt tracking. But here's an idea. Use a spoilboard "sled" that used UHMW plastic guides to keep it aligned.
3D type parts might need "runnners" on the sides of the sled for the pressure rollers to hold the sled down, and you'd need some sort of clamping system.

While it would fit in a minivan, I wouldn't be surprised if it weighed at least 150-200lbs.

[JerryBurks]

......While it would fit in a minivan, I wouldn't be surprised if it weighed at least 150-200lbs....

You can find a used ambulance gurney (a.k.a stretcher) with folding caster legs on eBay. They can handle the weight and are designed for the legs to fold away when pushed into the car. Sounds silly but maybe possible. I believe you can find that under Mortuary/Funeral supplies

[harryn]

I started reading about home made drum sanders - interesting topic. The sled will be a must at some point for the 3 D stuff.

Ambulance gurney - will look into that as well.

Thanks for the ideas.

HarryN

[harryn]

I started to experiment with the requirements of the nearly 7 ft fixed gantry portion last night. Obviously the first step is to make it stiff enough that it does not flex (much) during cutting.

Shown in the pictures are my test setup, a simple dial indicator mounted on a small drill press, to measure the deflection of the test piece. In this case, I used a 7 ft span between the 2 x 4s, and measured a piece of 2x6 in wood and a 2 x 10 pice of Al honeycomb. (that stuff is amazingly light)

The deflection force was me standing on the beam. (around 250 Lbs)

Results
- 2x6 - sagged all the way to the floor even before my full weight was on it, so in excess of 2 inches
- Honeycomb - deflected 0.35 inches

The honeycomb is definitely more stiff, but I still need to improve this by 5x to have any chance of getting reasonable results.

Also shown, is my work bench showing the typical box construction method I like to use. It is just some kiln dried 2x4s in a box shape, with cross members every 16 inches. The top and bottom are 3/4 inch ply screwed and glued to the 2x4s. I am pretty happy with the stiffness of it as a work bench, but it is pretty heavy.

Maybe I should consider to make something similar, but with Al.

[PaulRowntree]

I assume that in the gantry the beam would be vertical, not horizontal as in the test photo. THen the test force of 250 lbs is to simulate the lateral cutting forces? 250 lbs seems very high, I have seen estimates more of the order of 50 lbs being used. Still, I expect your are targeting displacements more like 0.010" or less, not 0.070" (almost 2 mm).

[harryn]

Hi Paul - yes, exactly right. The 250 lbs just happens to be a convenient mass (me).

[OCNC]

Hi Paul - yes, exactly right. The 250 lbs just happens to be a convenient mass (me).

Perhaps convenient but not at all realistic. Using a Bosch router as you currently propose you could never get a 250 lb load at the cutting bit. You'd be hard pressed to hit 100 lbs. (I suppose though that if the bit is stalled and you just crash into the work piece then maybe depending on your motors.)

Chris

[harryn]

Perhaps convenient but not at all realistic. Using a Bosch router as you currently propose you could never get a 250 lb load at the cutting bit. You'd be hard pressed to hit 100 lbs. (I suppose though that if the bit is stalled and you just crash into the work piece then maybe depending on your motors.)

Chris

Hi Chris, thanks for looking at the thread.

Between the cutting and acceleration, 100 lbs is probably realistic enough. 100 / 250 x 0.35 = 0.140 inches, which I am sure you would agree is still way too much deflection for a primary frame member.

That is just the contribution from one piece, not the sum of all of the contibutors of a total system. Of all of the parts of a cnc router that contribute to a good result, a stiff frame is probably the lowest cost area where I can actually make a difference, especially at this early stage.

This test was a good reality check for me, as it is easy for me to underestimate what is really needed to make a stiff frame.

[OCNC]

That is just the contribution from one piece, not the sum of all of the contibutors of a total system. Of all of the parts of a cnc router that contribute to a good result, a stiff frame is probably the lowest cost area where I can actually make a difference, especially at this early stage.

Hi Harry,

I've got to agree with that. I just hate to see something overbuilt if it isn't necessary for the use.

Chris

[harryn]

Hi, several years ago, someone posted a build where the entire gantry went up and for the Z motion, rather than just the Z plate area. I wondered - does anyone know who built it or have a link ?

To help with envisioning it - think of a fixed gantry / moving table machine, where the entire gantry goes up and down. The router went along the Y axis, but did not move up and down by itself on a local moving Z setup.

Thanks

Harry

[harryn]

I went back and re-learned how to calculate area moment of inertia and deflection. I also cheated by double checking my work using engineersedge.com, something we definitely didn't have in 1980. Also, some great vidoes by a prof at IU Purdue.

So, for a decent, first order of magnitued calculation:
- assume 7 ft beam, just freely supported on the ends - not being tensioned at the ends
- 8 x 8 x 1/4 inch
- 6061 Al
- Just assumed 100 lbs for the beam mass
- Assumed 100 lbs for the load on the beam in the center

First order of magnitude calc - 0.003 inch deflection

This is close enough for my router frame, so now I need to start looking around to see what this would cost me, either in Al tube or to make up a tube like structure using honeycomb.

I saw one on-line metals shop quote indicating around $ 300 - 400 for such a tube.