[peteeng]

Hi SA - If you land the columns on top of the cars how do you screw to the cars? Do not use foaming adhesive. Its really strong but will expand for some time and can move things around I have found. I always have high Z machines with around 300mm under the gantry. Thats because I do boxes and moulds that are tall. If you can get gal steel in 1.6mm go with that for your laminates. As you say cheaper then Al if you can cut it easily and its stiffer. All your concept sketches are fine the devil is in the detail. Peter

[routalot]

How are you planning to attach the rails to the gantry?If bolts are the answer then access holes will be necessary or plates internally tht you can drill and tap.The ends of the rails in this version may make it difficult to slide the assembled gantry onto the rails.
I'm sure others would agree that you need to add all the details before cutting or assembling a single part.Mistakes are much cheaper on a monitor than in the workshop.

[SweetAs]

I went away last night and looked at a whole bunch of products and you are both right, I need to spend a little time on getting this design right before I go start work.

Torsion Box - I have an idea of how I am going to build this. I did look to see if there is some sort of formula/calculator used when assessing the thickness of the internal structure material, external face material and the spread of the internal structure (every 150mm/200mm etc) but was unable to find anything. Do you have any suggestions or just to over-engineer it?

In terms of mounting gantry onto the rails, I think I have a few options.



A - Mounting a wooden bracket, mortised into the vertical side plates. Not sure how this would work with a wooden type bracket mortised into the main torsion box. I would probably just be better off creating a metal bracket. I guess the big issue with that would be if Im making a torsion box for the side plates, the bracket would need a good amount of meat around the bracket and for the fixings to bite into, Im not sure how I would accomplish that short of framing that into the torsion box itself.

B - Mount a plate on the bottom of the Vertical section and into the top of the carriage. This would make that side of it easier but then Ive got no overhang to mount to for the R&P underneath and then I worry about the forces on the carriages.

C - I think this is probably the best option, I can bolt through the vertical torsion box and spread the load.

Any suggestions on the thickness of the Steel plate I should be looking at for the brackets? I have some 6mm mild steel plate in the workshop that I could use. I do have a friend who runs a metal fabrication place if I need something bending and we do have a supplier who is able to waterjet/laser cut steel plate. Maybe Ally would be an easier option but it is expensive in the UK. I should probably look at adding some triangles to the sides of the brackets for additional strength. I have a cheapo 100A MIG welder so that should work for welding them. I wonder though, has anyone used metal epoxy for creating something like this? I wonder how well it would hold up.

My main concern with these designs are the twisting forces, is that something that the supported round rails will be able to take easily? Am I right in thinking that the force would be equalized because of the same type but opposite on the other side?

Ballscrew - I just read another post, one that you both commented on where the OP was asking about using 16mm ballscrews and the importance of the different pitches. Makes sense but I didn't realize it. Now I think that I want to be looking for a 1610/1616 instead of a 1605 for the X gantry...I think that's right?

Please bear with me, I do a lot of conceptual type work with designers, architects, and engineers in the workshop. This means that a lot of the time we have an idea of how it needs to be built or look but we don't really know how to assemble it or what materials to use at the outset. We tend to make a rough plan, just get on with things, and deal with issues as they arise. I understand that could make an expensive mistake here with the machine so I am trying to get all my ducks in a row before I get at it so apologies for my enthusiasm.

Best get back to work and earn some actual money for the build lol

[peteeng]

Hi SA - Ballscrews for a router are best at 10mm pitch or more. 5mm is good for the Z axis to give it lots of force and the Z axis does not move fast. I think you need to move along and not get bogged in details at the moment. There are many issues to solve and you won't get a good idea about these until you have a full general arrangement done of the machine so you can look at the big picture. Don't worry about laminates just block geometry to understand the big relationships, especially how things will be bolted together. The Z axis and saddle is especially complex and you will spend a lot of time in this area. Perhaps a heavy aluminium angle is the go for the column car mounts. If you use round rails there is no local bending moment so webs are not needed to support the bracket. Relationships between drives and rails and motors etc will be difficult and only once you get the GA done can you figure these out. Keep at it. In general you are trying to minimise parts to minimise interfaces as these are inefficient. So your carriage bracket would be better if it was one aluminium or steel bracket vs many bits. or none at all - see machine attached just rotate bearing 90degs....

Maybe find a muse machine that gives you some guidance for solutions. May save you100's of hours of thought. Peter

Heres a neat ply machine that you could scale up. He made the mdf one then used that to make the ply one.

[routalot]

Hi All - Even Formula 1 uses plywood. Peter

https://en.wikipedia.org/wiki/Skid_block

https://www.wtf1.com/post/why-is-f1-...wooden-planks/

Not quite ordinary plywood,it is heavily impregnated with phenolic resin.Very tough and rather dense,would probably make a good home CNC machine actually.It certainly takes a tapped hole nicely. https://permalideho.co.uk/products/

I can see that consideration is being applied to the project and will mention once more that removing some pixels from a monitor display is less costly than throwing materials into the scrap bin.It really does pay big dividends to consider each detail and the practicalities of assembling the parts.Additionally,consider the weight of the various assemblies and how they will be held in place while the fastenings are secured.

[peteeng]

Hi - Densified wood has very good properties. About the same stiffness as UHPC (review that below, its not real stiff) but much stronger and easier to machine. I have vacuum impregnated lots of timbers and MDF to use as tooling board or interesting objects. Densification requires large presses... But I think aluminium laminated with plywood is stiffer and just as easy to work with and you can make this yourself. I used to get MDF dust from a cabinetmaker and vacuum cast it into blocks as tooling board for a client. Had to sort out getting all the moisture out of it before epoxy infusing it... Interesting stuff. Peter

edit - I agree getting the design into CAD is really important and unless you design it at least 20x you have not refined it enough and best to use up CAD time than real time... Reality sucks and mistakes are difficult to fix.... but CAD mistakes are very easy to fix.

Jabroc data attached. Its only 15GPa in flexure much less than I expected. So aluminium laminate much, much betterer...

[SweetAs]

Hello!! Apologies it's been such a long time. Work has been weird, as I'm sure it is for everyone who is self employed, and we had hoped to get a grant for a new machine but that fell through so back to the rebuild!

The plan is as follows...

Get back on the build, and sideline thoughts of a proper machine for the current time.

Ditch the chain drive for a belt drive (started a thread looking for T5 Steel Reinforced belt info) The chain is becoming a real pain with the variation in temps I'm having in the workshop. I hope belts will be better, just need to make sure my tensioning and sizing is right.

25mm Supported round rails for the long axis.

20mm supported round rails for the short, is 20mm over kill? I'm fully aware that belts will come with limitations so I don't want to over spend where I don't need to. I guess the same goes for the 25mm above?

I've seen these tank track type belt setups, what are peoples thoughts on those? It's where you run a small belt in a loop which connects to another stationary belt, like a R&P but with belts. Looks cool but I'm sure it would have issues with dust build up on the stationary belt.

CAD designs are underway to look at mounts, sizes etc. Definitely going to go with a 200mm square torsion box with 1.5mm steel plate all over. This will be fixed with adhesive and screwed to the Birch Ply box.

Another reason for the rebuild is we are getting more and more CNC work in and so time is of the essence at the moment, I'm thinking that this build is now going to be used to earn the money to buy a full, ATC Chinese router...maybe...possibly...if the bills stop rising for a little bit lol.

In terms of motors, Ive got 3 3nm (435oz?) to run the long (2 motors) and short (single) drive. Any thoughts on this?

The previous thoughts on keeping the sides low are right, I want to keep it as low, easy access to the bed as possible.

Thanks again for baring with me, work and life has been exceptionally weird here in the UK and it seems to be getting worse this year so I want to consolidate my business options before things turn.

[peteeng]

Hi SA - The tank type of drive as you call it has issues. Teeth on belts are designed to go around pulleys and squish in a bit. If you get some belt and interlock them on the flat you will find they have some clearance ie you can wobble them back and forth. I suggest you don't go down that path. I do suggest you look at helix rack and pinion this is the best solution for an 8x4 size machine. It has the least moving parts and is accurate. At least use the R&P on the 8' axis and you could use ballscrew on the gantry. But I'd go R&P for the X&Y and ballscrew for the Z.

I have built belt machines and they are fine at that size if the belt is 25mm wide. By the time you make pulley mounts, tensioners, return pulleys and motor mounts to suit, the cost is the same as the R&P probably more. R&P has two parts, belts have many many parts. 25mm round is probably overkill 20mm is a good size all round. I'd use square rail for Z. I'd use square rail all axes. Last machine I made I used round to see how it went and to see how the cost went. Didn't shift the total cost much and square preloaded cars are much better then the round rail cars... Plus you will probably have to have the long rails supplied in two pieces, so you need to spec a matched rail set. Peter

also consider gearing the motor. Direct drive belt and R&P result in very fast machines with small drive forces. A 1:3 gear or 1:5 reduction provides good speed but higher drive forces. Depending on what you are cutting will dictate how much cutting force you will need. Commercial machines like this are geared down for this reason. If your going to build, build a proper machine. If you can't justify buying a commercial machine for a biz then you can't justify building one. It maybe "cheaper" but your time value on money makes it much much dearer. But it is fun if your a maker... Peter

[SweetAs]

Hi Peter, Thanks for the quick reply!

Belts/R&P/Screw drive - I've been looking today and in terms of the belts, I can get T5 in 25 and 30 and T10 in 25, 32 and 50. In terms of the R&P setup, I can source the rack in a 3m length, that's not too much of an issue, Ill check the prices tomorrow, its just how I transfer the power from the stepper to the pulley that Im a little confused on. In my head I've got the stepper mounted with a pulley, a small continuous belt and then that will transfer to another pully which drives onto the rack, is that going to be the best way to do it? I've looked at the planetary gearboxes and they are quite expensive here in the UK. I did look at a pre manufactured version from AvidCNC, I think those pre manufactured units will run to about £400 for the pair once Ive paid shipping and import taxes.
In regards to the ballscrew, Ill have a look, I can get hold of a 2005 from a UK supplier or 2505 from China, not sure which one I would need for the Gantry. I guess I could look at R&P, just need to figure out a way to transfer the power to the rack.


Rails - Ok, Ill scale those down a little to the 20mm ones. Luckily I can source a full 2700mm pair of the SBR20 for £220 so that's a bonus! For the Z axis, I had thought about just picking one of these premanufactured units off of Ali, not sure yet how well made they are so I need to look into it. https://shorturl.at/gmlhl
In terms of the square, do you mean the standard square linear rail with the carriage blocks? A set of HGR20s are looking to be around £645 for a pair inc 2 blocks. Moving to square would give me some issues/work with alignment wouldnt it? The machine currently runs on a 25/50mm Ply frame which is going to be way too uneven for that, wouldn't it?

In terms of workload for the machine, its going to be 90% Softwood Ply, Birch Ply, Marine Ply and MDF. Today for example, Ive machined 12 sheets of 18mm Softwood ply for a set of "ribs" on garden pods that I build. The CNC we have now works just fine, its cutting at 2000mm/m at 4mm DOC is just mind numbing lol. Each pod takes me around 10-12 hours to machine so if I could cut that in half (maybe a pipedream) then that would be amazing and save me no end of time/money.

In terms of the cost etc, things have gone from bad to worse in the UK and no doubt they will get worse with the upcoming election next month. My workshop overheads are over 12k more this year than 2 years ago, not to mention the house bills, hence the reason for the limited upgrade. In reality, I need a full on 8x4 ATC unit, servo run, built in dust extraction etc but even imported from China thats going to set me back £12.5k (I have checked lol). Not a great deal in the big scheme of things and I can make it back but got to have it there in the first place. I dont like getting anything on Credit and regardless, the equivalent machine in the UK is going to cost me £40-50k at a bare minimum. My business is set around building things that others cant or wont so I figure, work with what you have and take it the best you can.

Thank you by the way, all the help previously and now is making this daunting project easier!

[SweetAs]

Not sure how big this will come up, but this is the workshop today lol

[peteeng]

Hi Sweet look at stepperonline - 5:1 or 10:1 about 40 pound. I have used these on two machines and they work very well. Peter

Planetary Gearbox | STEPPERONLINE (omc-stepperonline.com)

Don't use the premade z not stiff enough. Look at bst motion. This is where I get most of my motion stuff. BST AUTOMATION Store - Amazing products with exclusive discounts on AliExpress

Commercial machines cut 18mm in one pass. So if we assume you have a 24k spindle and that the Fz(chip thickness) is 0.1mm with a 2F tool then feed = 0.1x24000x2=4800mm/min so your feed speed will double. So work towards having a system that can cut at 5m/min. Even 10m/min that would get production up. Peter

go with R&P for X&Y and ballscrew for Z. D16mm is Ok for Z 5mm pitch.... all timber rail foundations have to be made flat. Whether you use round or square rail. I use epoxy putty filler and a long extrusion as a lap board. Make it as flat as possible using whatever references you have.

[peteeng]

Hi Sweet - What is your current machine specs? maybe we can speed it up in the meantime. So 2000mm/min at what rpm and what tooling type? straight or helix? compression etc. What's the top feed speed of your machine? etc. Maybe we can find a tool with a better higher feed for you... Peter

heres a plywood build to muse over - Peter

Project | *NEW* DIY Multiplex Plywood CNC Router | Hackaday.io

[SweetAs]

Hi Peter,

Ive been having a think...dangerous I know! I think Ive been coming at this from a bad mindset, at least in terms of running a business anyway.

You were right in regards to building a commercial level machine, I'm thinking about how much more money I can spend and how much more of a better/faster product I am going to be able to produce. I'm thinking of basically going back and doing a full redesign to see how much extra I can squeeze out of the project with the money I can put into it. It seems stupid to go to all of this effort, when the object is to make money from the machine, and then try to cut corners, especially when Ive got the work coming in for it at the moment.

For arguments sake, lets look at what would constitute a really good DIY machine and how much that would cost. Lets say we go with R&P drive and square rails as our starting point.

Mod 1 rack at 15x15mm for the drive, using 1:5 ratio gearboxes on the site you linked to. Linear rails are hugely expensive here in the UK, Ive been looking at some info on 15mm rails, they would work wouldn't they? I need to speak to a couple of companies which manufacture them in the UK, but I did find a well priced Chinese seller on eBay which had 3m 15 rails but it used a joiner, about a quarter of the price of what would be 2.7m HGR20s and half of the 15s. Is a join that much of an issue if its aligned correctly? A friend runs a specialist parts company making F1 parts using a 5 axis CNC. I'm sure I could get him to create some sort of alignment block before everything is still and screwed down.

The next weakest link is going to be the motors, do you think the 23s at 3kn is the best way to go or would I be better looking into 34s at 4kn or even 4.8kn? Makes sense that if I was going to replace the motors I would look at closed loop, they have some nice kits on the website that you linked to!

4kn - https://www.omc-stepperonline.com/ts...y-4-clts40-v41

That's the drive, rails and motors all sorted, so now we look at the frame....

Do I stay with the Steel/Birch composite or go to a steel box section frame with a steel plate construction. Is something like a 30mm box section with a 4 or 6mm steel plate going to be much stiffer on the gantry uprights...probably! To save on welding, I can get the box section laser cut so that it all folds in on itself and have pre drill holes ready for the steel sheet to be bolted onto the face.

Then its the gantry, same again, do I keep the steel/birch torsion box design or go for a bolted box section and plate instead? Obviously this is going to increase the amount of weight but as long as the rails and motors are getting upgraded then it should be fine?

Ive got some ideas how how to mount the rails to the frame etc, see below. All of these use a 50x100 steel box section to mount the rails and rack to. All of the birch designs use a 6mm steel plate on the outside, a 1.5mm steel plate on the inside. The plan would be to fix the steels to the birch and then screw into the birch. Ive also done either a trapezoid shape as well as one with a right angle on one side. Not sure which one is needed, guess the Trap would be stronger but again, it might be overkill.




Gan 1 - The Gantry sits on a 100x100 steel angle iron, the cars are offset slight to allow access. The angle iron is cut in line with the birch composite.



Gan 2 - The gantry sits on a 100x50mm angle iron, this then extends out of the side of the composite and the cars are bolted to it. This gives the advantage that the gantry forces go straight down.



Gan 3 - The Gantry sits on sections of C Channel and then everything is through bolted either side. These then extend out of the sides like Gan2, allowing the forces to go directly down.



What made me think of this is I use a local laser cutting/machining company for my metal work and they can cut and form box section ready for welding. My welding skills are a bit shoddy and so I thought I could get them to cut it, fold it and also laser cut holes in the box section. Then I can get them to laser cut a steel plate, with matching holes, so the whole thing just bolts together. I'm thinking of the time saved with this design, it will cost a little more in the initial outlay but if it saves me time that I can better spend on earning then that seems like a good option to me. I would probably look to make the whole gantry out of 30mm box with plates and bolts as and where I need them. I understand the more joints/fixings I add in, the more failure or issue points Im adding in but Im thinking that as its only used for wood and the odd bit of composite here and there, this design would work?



A lot to think about I know, just trying to take a step back and think about what the best option is in terms of cost, usability in the future and go from there.

Ill keep with the 3kw spindle I currently have and if it needs upgrading, Ill do that when the machine has been done.

What do you think? Am I right off the mark with this or onto a good idea?

[peteeng]

Hi Sweet - I'll try to work through your thoughts:
1) thinking is not dangerous. But once you have critically worked through the issues and decided a trajectory then its time to leap into action
2) Find a commercial machine that does what you want with good support. Commercial support has real value. Then talk to your bank or whatever and figure out the terms. I understand you don't like finance but that's how businesses grow. You have to leverage time and money. If you build your own machine I expect its a year or more before you have it earning, in that year you can earn alot with a new machine. At least go thru the exercise so you can compare your build costs. Commercial work and an ATC makes sense it saves huge amounts of time. ATC is not easy for a hobby builder to sort, again a commercial ATC machine is what you are aiming at....
3) 15mm rails will work but I use 20mm and have recently decided to go to 30mm for various reasons. Rail and rack come in sections that are matched. You need to work with a supplier that can grind match the ends. Don't need a "joiner" if matched ends are done. if you go helix rack (which will be right hand tooth) then you also need to order a short length of left hand rack to use as an alignment tool.
4) N23 or N24 size motors are fine with a 5:1 box
5) Do not spend money on hybrid steppers, I don't think the extra dollars is worth the small gain. If your balancing a build budget it will be std steppers or servos. Servos lift the game to the high level.
6) I would not use steel as the reinforcing for ply. I'd use aluminium. In this way you can use trimmers and routers to finish parts. Steel presents issues unless you can work with steel in total
7) Do not go down the steel welded path. If you weld it distorts and unless you have access to heat treat and finish machining then it becomes a mission, especially on a large machine as you need a very large mill to machine the lands
8) If you have a good sheetmetal shop that can bend accurately then look at folded sheet metal parts. These are accurate within the parameters you need. I build kit routers like this. Half sheet machines. This is a very valid approach. As with all technologies there is a learning curve. Its taken me 4 machines to learn all about sheet metal builds. Scoot my production half sheet kit is 90% sheet metal parts.
9) I recommend not to use std hollow sections they make you design non optimal parts. They are not flat or square and have cupped or crowned faces. All of which has to be fixed along the way
10) Sounds odd but having designed many many machines I start at sorting the Z axis first. Design a great Z axis that does what you want and work out from there. If you start elsewhere by the time you get to the Z you inevitably are squeezed for space and dimensions. The you have to expand all the rest of the machine when you find that the Z is too big or you compromise the Z and then you've cooked your goose. The Z axis is the area that does the work and unless thats uber stiff and sorted then the rest is just along for the rid.
11) You can build a totally satisfactory plywood machine. There is a german company that does this for cabinetry, and they do a great job. If you build stay in a medium you understand. If its a hobby fine, experiment but to make money you need to be brutal about the exercise. Your size machine is an $8k to $10kAUD exercise for me with no cost for my labour. Parts , freight etc add up. I've just built a bench for a 1/3 sheet 5 axis machine and its costing $800AUD for the materials and so it goes on...
12) In terms of control you need something that can scale ie add axes or controls easily. Look at Dynomotion controllers. Took me a couple of years to find these guys. I've used a few different controllers and systems and I think Dynomotion has a lot going for it...

CNC Wood Milling Machine | Mechanical Engineering | Am Luftschacht 3, Gelsenkirchen, Germany (cnc-holzfraese.de)

12) Do not buy anything until you have designed it. Its fine to source and cost stuff but buying too early will result in things sitting on selves... If you have stuff that maybe useful muster them into a single place and see how they go. I set up a folding table and start mustering parts and stuff on that as a project table. If its in sight its more likely to be used and if you see things often your subconscious will work on it more often.

Keep at it... Its better to create a great design in CAD then realise it then do it in bits and pieces... the puzzle will resolve in about 20 design cycles I find. Keep it really simple, work on it being simple. Get that Z axis nailed before you do anything else... and a finance plan for the bought router as a benchmark...and a written objective for the machine. Always helpful to make decisions by... Peter

frankie 1 is a majority plywood machine cuts aluminium easily...

[peteeng]

Hi Sweet - Just to add some more info. In past lives I have worked as a welder in auto sport (Ti, Al and steel) I have worked as a welder trainer, robotic welding and manual TIG welding. I have worked in yacht racing and auto racing - Carbon fibre and composites and timber. Then there's production management and design engineer roles in various companies (cnc programming, digital manufacture, bespoke machinery) So I feel I'm very well versed in possible methods for cnc machines. Since you have a cnc router and expertise in plywood I'd go plywood (plus a little bonded aluminium). If someone came to me with an open checkbook it would be carbon fibre all the way. But mould costs are more than part costs for a one off. But since its open checkbook that's the best solution in my books.

With plywood the trick is solid parts. Ply has a density of 700kg/m3 and steel is 7800kg/m3 so you can use 10x more timber before you get to steel weight. So you have a lot of stiffness available in ply, you just have to think about it as solid stuff vs thin stuff. Solid parts solve the local deflection (thin section issues) and local shear distribution issues (again shear flow in thin sections creates distortions) so lots to think through. Peter