I promise I will stop asking daft question soon. Ish.

I know I said I would do a cheap mdf machine, but the more I look at it the more I trouble I have. I don't think mdf will give me a decent 3x4 cutting area for doing useful work on and at the same time be able to mill new bits for itself reliably. The amount of mdf needed for a strong structure of that size would also end up as scrap when I came to upgrade. I realised that I don't have a problem with building boxes out of wood or steel I have a problem with all the cnc fetish talk of whipping, backlash, ballscrews and the unhealthy obsession with industrial strength vibration.

So steel it is for the frame and gantry.

I am intending to use 80x40x4mm box for the legs and the top of the longest axis held together with plates and M8 threaded rod. For the rest of the frame I'll use 30x30x3mm box. I'll be sheathing it in mdf so that I can use the underneath as a cabinet.

My longest axis will be 1.5m of rail and the gantry will 1.25m of rail. I am intending to construct the gantry out 30x30x3mm box backed by 3/4 inch mdf and also use 3/4 mdf for the gantry base & z-axis plates etc. Later I plan to machine new plates from 20mm aluminium to replace the mdf.

I'll be using sbr20 supported rail for all the rails and 20mm antibacklash ballscrews/couplers driven by nema 23 3.1nm motors driven by a cw8060 with a 600w 48V 12A power supply.

Each side of the longest axis will have a single sbr20 rail mounted directly to the 80x40 and ballscrew above, the gantry and z axis will have 2 rails and one central screw. I'll be ordering an extra long z axis but not using the full amount until I install the 4th axis.

The steppers etc I plan to get from cnc4you and the shiny metal bits & connectors from Chai.

Now for the questions...

1- Is 80x40x4mm & 30x30xmm mild steel good enough for the frame and gantry and will the threads/plates hold up ok without welding?

2- Will steel & mdf provide enough stiffness without cross bracing? (the frame will have a cabinet floor of 30x30x3mm near it's base)

3- Will the sbr20 mount ok directly to the 80x40x4mm or do I need a cunning plan to attach it?

4- Should I drive the two threads with a belt system or just slave an extra motor?

5- Anything horribly flawed with this masterplan?

I need this build to deal with light use, 3D carving of upto 4 inches of laminated ply/mdf, accurate cutting of inlays on veneered wood, occasional soft metal upto 20mm thick (only need that thickness for cutting new machine parts for it) and cutting/engraving on acrylic/pcbs.

Thanks for all the help!

Oh, and one last question does Chai sell the water cooled spindles?

Quote:

---

Originally Posted by D.C.

Oh, and one last question does Chai sell the water cooled spindles?

---

yes, inverters too. no mounts though :(

1) Depends how you use it, but I'd reccomend using something larger than 30x30 for the gantry - don't worry about the weight. It's easier to mount the SBR20 rails on something that is at least as wide as the rail base.
2) MDF wouldn't be my first choice, but it will still add stiffness.
3) Yes it will be fine on the 80mm surface.
4) 6 of one... could spend forever discussing this. I'd probably go for two motors, Jazz will say one motor and a long belt. Price wise there's not that much in it.
5) Use 16mm diameter, 10mm pitch (RM1610) ballscrews, not 20mm diameter. You wont gain anything by using the larger size ballscrew since those motors, especially when run from only 48V, will not be able to get good accelleration and speed with the larger ballscrew, especially if you're thinking of the RM2005 from Chai. Get RM1610. Yes the critical speed is lower, but that's only an issue on the 1.5m ballscrew and even so if comparing to 5mm pitch screws it will still go much faster as the rpm is halved for a given feedrate. For the Z-axis the best option is RM1605. Also you should use timing belts/pulleys to get the best performance...couplings are notorious for snapping or being annoying to align, whereas belts are trivial to align, allow you to change the ratio to gain speed or resolution and also help suppress resonance.

Quote:

---

Originally Posted by D.C.

Oh, and one last question does Chai sell the water cooled spindles?

---

Yes, but he's one of the more expensive sellers for spindles on eBay. I got mine from solar.jean on eBay, and know a fair few other people who have and it's been excellent.

Quote:

---

Originally Posted by Hobgoblin

no mounts though :(

---

Attachment 7119Attachment 7116Attachment 7117Attachment 7118Attachment 7115Attachment 7112Attachment 7114

:whistle:

Are they from chai jonathan? As when I asked him last week if he could supply he said no :(

Sees spindle/motor mounts and wants ;-)

If you're selling those Jonathan and can take paypal, don't be shy and send me a PM (2.2Kw water cooled spindle 80mm dia attaching to a 160mm wide Z plate BTW) :-)

I'm guessing thise mounts are from Jonathan DC

Yes they're all ones I've made. I need to make one for someone else on the forum asap, but I'm very short for time currently since I'm at university and have another project that needs finishing. However if I can make a few to the same design that would be pretty efficient...the one in the second photo from the left is easiest if made from 1/2" thick aluminium, to this drawing (can change the hole spacing if required):

http://jonathan.blissett.me.uk/image...%20drawing.png

Thanks for all the help 80x40 it is then and I'll plan for using belts.

As far as the ballscrews go, would a 20mm 5mm lead not offer better accuracy, I'm not intending to do large volume runs of stuff so speed isn't that important (as long as it doesn't case any material problems) but precision for engraving is.

And last two questions, is there a good online resource for explaining the basic maths and mechanics for working this stuff out properly so I can stop vaguely throwing combinations of parts around in the hope that someone will work it out for me? :)

Is there a good all-in-one type book (real dead tree type) that covers all the basic aspects of cnc preferably with emc2 examples?

Ta very much!

I have a book called "the CNC cookbook" which is very simple but a good intro. Available on amazon...

Quote:

---

Originally Posted by D.C.

As far as the ballscrews go, would a 20mm 5mm lead not offer better accuracy, I'm not intending to do large volume runs of stuff so speed isn't that important (as long as it doesn't case any material problems) but precision for engraving is.

---

Yes, but overall you're still better off getting 10mm pitch ballscrews since you can use pulleys and have a 1:2 (motor:ballscrew) ratio to get the same resolution as a 5mm pitch screw on 1:1. If you got 5mm pitch then the ballscrew has to spin twice as fast, so unless the critical speed is twice as great you will be limited. The critical speed is just proportional to the diameter, so using 20mm will only gain 25% speed hence there's no point. If the 20mm diameter screw was 10mm pitch I probably still wouldn't use it since the required torque to accelerate it is much greater.

Also the resolution you get from 5mm pitch is hardly ever required. I cut this with (roughly) 10mm effective pitch:
Attachment 7120
The area shown in that image is roughly 50mm wide...

Quote:

---

Originally Posted by D.C.

online resource for explaining the basic maths and mechanics for working this stuff out properly so I can stop vaguely throwing combinations of parts around in the hope that someone will work it out for me? :)
... preferably with emc2 examples?

---

Er, me? :friendly_wink: But seriously, the information on this forum is all you need. If you've not already then read as many of the build logs as you can as the same things come up regularly, so it shouldn't take long to find out what the best way to each part is. The fact you're intending to use LinuxCNC doesn't make a lot of difference to the process - in most cases it'll be just selecting a different option when you export the G-code from the CAM program.

This is vagulely seeping in now, thanks! I really need to brush up on my basic mechanics I've forgotten it all.

I was reading through the build logs and found a link to a section of this site I never new existed, the FAQ... ooops.

I'm going to order the cnc cookbook from amazon, does mycncuk have an affiliate account I can click on?

I love the finish on your ply chair, it almost looks like spalted beech.

Well I hope you find it useful - definitely a bit basic for some of the guys on here but for relative newbies like me it's good.

Glad you like the chair - that was what got me into this lark! I made it for my son and cut it all out by hand - never again! Next time I'm Cncing all the way...

By all means buy the CNC Cookbook, but you'll get a lot more useful & detailed info from here. Its a bit basic and out of date (2009) in many areas and a lot of things are glossed over, like pros and cons of microstepping for example. In fact that subject gets 1 line in the book and not even a index entry!

Quote:

---

Originally Posted by irving2008

By all means buy the CNC Cookbook, but you'll get a lot more useful & detailed info from here. Its a bit basic and out of date (2009) in many areas and a lot of things are glossed over, like pros and cons of microstepping for example. In fact that subject gets 1 line in the book and not even a index entry!

---

It wasn't a slur on the quality of the information on this site but I have zero experience with this sort of stuff and I'll be happier when I have a basic grasp of the fundamentals. In that situation I normally find an academic textbook/guide is useful as it covers all the fundamental topics in a logical order (or they should!) and gives you something to build on. At the moment I'm kinda grasping at threads of information and trying to knit them together randomly.

For example in one of the build logs a tip is given about the positioning of the screw in relation to the rail, I would like to actually know this stuff and the mechanics behind it to work it out instead of stumbling across nuggets of information and thinking 'oh yeah, of course'.

I'll see if the library has a copy if it's not worth buying and in the mean time dust off my old Sadler and Thorning 'Understanding Mechanics' that has been sitting on the shelf untouched for years.

Ah... then you wont find that level of detail in the Cookbook... or at least only superficially... its a cookbook, much like if you beat some prescribed quantities of eggs, flour, bicarb of soda and water together and put it in the oven you get a cake-like thing... it says nothing about the chemistry of cakemaking or how the different ingredients might affect the final result in any detail. So if you're looking for the maths on acceleration/mass calculations, bending moments/rigidity, stepper motor torque v leadscrew pitch, etc etc. you'll find it light on detail...

(for a test, go to the Amazon page for the book and try some search terms in the "Search Inside This Book" field)

Are there any good recomendations for a basic introduction to cnc design/use including explanations of the maths and principles then? First year undergrad level is probably about right for me.

Now there's an interesting question... none of the books I've looked at appear to have much of the maths in them. The bulk of the books are either "Build a CNC like I did" and then are essentially a set of plans (or not in one case) but with little explanation why... or are more focussed on the programming and usage but relatively little on the design aspects... something we have great debates on the forum about, between those that prefer the suck it and see based on experience of doing it several times before (and there's nothing like experience I agree) and prefer empirical evidence of what works or doesn't work and others (and I include myself in this category) who want to understand what constitutes a good design from good engineering principles and want evidence to demonstrate that, and indeed want to understand how to make it an optimal design.

Of course, those that build big commercial CNC machines must, we assume, have done the full blown structural design, but most of the low-end 'Chinese' machines are almost certainly 'empirical' in nature (and it shows in many of them). There is very little on the web about the calculations, though plenty of generic 'bending of beams' etc stuff which is 1st year undergrad (or when when I did it 30y ago). As always its in knowing how to apply the knowledge that counts. There are nuggets out there, like this thread on CNCZone and this one here, but again they don't explain too much of the 'why'.

That is one of the reasons we say that no question is stupid... and reading through build threads past and present will give a good cross-section of both the 'calculated' and 'empirical' approaches. If you have specific questions about calculations abd/or principles then ask away...

Hmmmm, perhaps you guys should put one together, then you can answer all daft questions with a BTFM and link to amazon. :)

I've thought about it.... if only I had the time....

Irving,

Your not the only to do some serious thinking about it but the time and resources to put it together proper is no small matter. Just setting up the tables and getting photos of different design set ups to show the pluses and minus of each and where they can be used and how they are applications of different formula is a nightmare. It would be extremely useful to folks wanting to build and understand as well as students of the mechanical and electrical principals just a headache in the making and very time consuming. Have a good portion of the data, just not the time to put it all together and get the other information that is need to make it work.

Michael

This will sound negative but I doubt if there are to many people who really care about the why's. I don't mean to sound negative but from what I see from looking about most forums (not just cnc) all people want is to know how to make it work, not the reasons behind how they could figure it out for themselves.
If there were more people like DC about & this sort of question got asked a lot on forums then I would have thought someone would have done something about it before now. Not saying it would be a waste of time to do, just don't think it would get the amount of use some people think it might get.

Why don't we do it as a team effort on here through a thread. Use a post per design with honest +\-. If we had one thread with pictures and info on lots of machines it would be an amazing resource for those of us still designing. Spread the effort about! Then it could be compiled...

Hmmm... I think there are lots of different ways of approaching this... my thinking was more theoretical in terms of how to actually do the formal assessment of a design, rather than the comparative assessment of a specific build v another specific build. Don't get me wrong there's a lot to be said for the crowd view but there's always some subjective opinion. I think the issue raised by DC is one of "with all the helpful but diverse input how does he make a reasoned assessment, what are the basic parameters he needs to understand", at least in sufficient detail to get a feel for things, and, IMHO, that needs an element of the 'why' about it. For some people that could be intuitive, they can look at something and see how it might behave, others need to see the numbers and yet others need pictures...

Perhaps it would be possible to do both the same way some websites are constructed where there is little detail about an item upfront for those that already know or aren't interested but with more detail deeper within the site for those that would like to know more.
Doesn't really matter how it is done as long as someone who knew what they were doing would agree to take the lead and monitor the thread & control it's direction. Would be useful a a sticky somewhere for people who were new to the site as well as helping some of us who are already members.

Perhaps doing it online in a wiki form would help, it spreads the workload out and can be imroved by more people that way.

I agree that most people just want something to work, but the number of horror stories and people getting hugely stressed about things not working and not understanding why they don't work I think shows that such a resource would be useful. While ploughing through the forums I came across a similar sentiment of "If a machinist can't fix his own machine he isn't a machinist". People don't really decide to build their own car without expecting to learn about being a car mechanic and how internal combustion engines work so why try to reduce cnc to the level of understanding that is required to assemble a table from Ikea?

Quote:

---

Originally Posted by irving2008

I think the issue raised by DC is one of "with all the helpful but diverse input how does he make a reasoned assessment, what are the basic parameters he needs to understand", at least in sufficient detail to get a feel for things, and, IMHO, that needs an element of the 'why' about it. For some people that could be intuitive, they can look at something and see how it might behave, others need to see the numbers and yet others need pictures...

---

I think this is pretty accurate, looking at buiilding a CNC machine my inuition tells me that optimum design would be to make a lightweight but very rigid moving gantry that is supported on four rails inside a structure that is a hollow cuboid. My inuition might be totally wrong, it might be that the extra friction would eliminate any benefits or one of a million other reasons why that might be an awful layout but I lack the engineering knowledge to work it out for myself.

There are loads of books for cars or photography or telescope building etc etc that are aimed at a mechanic sort of level, combination of practical skills and enough of the maths to work things out reasonably well in most circumstances without reducing things to the level of 'insert screw a into hole b' and without bogging things down with doctorate level maths/material science.