Newbie- Uk CNC builder with pre build questions

[Sainty]

Hi

I am a joiner based in Southampton, UK. I've spent a fair bit of time over the last couple of months reading about DIY CNC, and I have to say that I'm hooked!

I'm not quite sure what specific applications I have in mind for a CNC router but I have a couple of things in mind (custom radiator cabinet grills, corian cutouts, mould/former construction, to name a few) but mostly machining and cutting 18mm MDF.

I intend to make a very cheap machine, but one which is capable of reasonably accurate/repeatable work and will be useful for creating CNC#2.

My build will be based loosely on the www.buildyourcnc.com model with improvements where I feel necessary. Mainly in the construction, as I said, I am a joiner!

Before I get too carried away, I want to check what I am doing is along the right lines.

Firstly, transmission, I intend on using 10mm threaded rod which has a pitch of 1.5mm. Is this suitable? I am hoping that this will be stable over greater distances, (approx table size of 800mm x 1400mm).

Secondly, drivers and motors. I have found these kits and plan on using kit 1. At the moment I am struggling to get my head round the electronics side of things, will these have enough power to move my router around? Also, what else will I need on the electronics side of things to get this thing moving (with the exception of a PC and software)?

Thirdly, the router. I am luck in that I have several routers available to me from a trend T3 upto a 2200w festool. Is there a recommended router size given that I want to be able to machine 18mm mdf and might progress to aluminium.

Fourthly, speed and depth of cut. Just out of interest, what sort of cutting capacity should I expect from the machine that I have described above? If I wanted to cut out a square of 18mm mdf, 300mmx300mm how long will it take? what sort of depth of cut will it achieve? Could it cut the full 18mm in one hit?

I probably have about a million other questions but that will do for now!!

Thanks in advance


Stu.

[M250cnc]

Hi Stu welcome

I am also based in the UK and have converted a mill currently working on a lathe.

The first correct thing you have done is come here for advice, this will save a lot of heartache.

I also have a lot of experience of woodwork.

Firstly you say "can you cut 18mm mdf in one pass" can you do this manually ?

If no then this will also be no with cnc.

CNC is for repetition with minimal user intervention, it is easy to spend hours programming cnc only to know that you could have done it quicker manually.

Motion Controlled products are good and i use them myself but i do no recommend kits. The MCP BOB is very basic, checkout www.cnc4ps.com for BOB's plus other stuff etc.

Speed is a combination of things, firstly stepper motors lose most of their power above 1000 rpm "Try To Keep Below 1000 RPM" so with a 1.5mm 1.5 x 1000 = 1.5 mts per min pitch thread you will be looking at a top speed of 60ipm, Not Very Fast at all.

Second and this is why i do not like kits relates to speed again the motors in that kit will be fine but i would use higher voltage power supplies and higher voltage drivers again from MCP

Think of this simple formulae Stepper Motor Speed = voltage, Stepper Motor power/torque = Amps

With lower voltage, Stepper Motors will not rev high they stall even with the motor shaft not connected to the machine, that equates to lost steps which is scrap parts and wasted time.

Now if you change to ball screws with 5mm pitch the above scenario would give you 5 x 1000 = 5 mts per min or 200ipm IPM is mainly used on here.

At 5mm which i use, i can hold a tolerance of .005mm so you could use 10mm pitch ballscrews get 400ipm with a tolerance of .01mm

Now you may think you do not need 400ipm but you need to move the table even if not cutting so that's where you need that extra speed on big tables.

The other advantage of ballscrews is friction, it is minimal the other is rigidity and lastly accuracy. The downside is cost.

A ballscrew will also come in various grades this is given in say .001" per foot so the max error over the length of the screw is known this is not available with standard studding.

So does your kit come with that option "Ballscrews" if no then the kit will have a lot of wasted parts and you will have to find a manufacturer to make the parts but this will not be cheap.

Another solution is plans, that way you can mix and match what you want i think a lot of builders here use Rockcliff, ask another question.

Hope this helps get you on the way.
Phil_H

[Sainty]

Thanks for that Phil, interesting reading.

First thing to tackle for me is the threaded rod/ball screw issue. My heart tells me that ball screws are the way to go. However this is supposed to be a budget build, if it all works, I can always upgrade. For the size of machine that I am looking at it will be in excess of £200, for the ball screw and flanged nuts. As an alternative I can get M16 (2mm pitch, 78IPM at 1000prm) threaded rod for a fraction of the price and manufacture some flanged nuts for about £50. I know that I will be at the hands of backlash but there are measures that i can take to eliminate this. If that turns out to be unbearably slow, i can always upgrade. I'm basically trying top work this all into a sub £500.00ish budget.

I'm really starting to get worried about the electrics side of this. I have no experience of the electrical side of things. Assuming that i bought the kit that I mentioned, is it a case of plug it in and go? (I wish!) How much extra work is involved? I guess i need something to house all the components. Is that an MDF fabrication? You see I really am a novice with this side of things! I guess i need an eStop and limit switches too? Am I out of my depth, I haven't picked up a soldering iron since I was at school (25yrs ago)

In respect of the driver kit do you have any recommendations for ugrading the power supply? The units in the kit are "2 X 40V, 7A non regulated power supply". They do a 75V supply, would that be more appropriate? As if to expose my lack of knowledge further, why are there two?

For fear of exposing myself as a total fool, this is how I understand the build process. You start with some form of model in a drawing package (CAD), this then gets transferred to Gcode (by software??) which a program like mach3 uses to control the router. This is done by way of three motors (X,Y and Z). The PC is connected to the drivers by way of the breakout board and printer port (can this be done via usb so that I can use a laptop?). The breakout board tells the X,Y and Z drivers what to do the process continues until a) the program is finished and the router returns to the 0,0,0 point, b) the limit switches are triggered or c) something goes horribly wrong and I hit the estop button.

Ok shoot me down if I have it all wrong and have to go back to the drawing board!

As an aside, I started to build some prototype mdf I beams today, 240mmx100mm 'I' section, 1.5m long. There is no apparent deflection when I stand on it! I think that I could probably afford to reduce the dimensions a little!

Thanks

Stu

[M250cnc]

Hi Stu

Thanks first of all it will work if everything is done correctly, this forum is full of people that will confirm that.

Ballscrews you asked for an opinion on your planned build & i gave it, you can ignore it.

Backlash it is better to design it out in hardware than to try to use software.

Put it like this when you are routing a heavy cut you will push against the direction of rotation of the cutter "Conventional Milling" this will be an effort on your part. If you were to push in the opposite direction "Climb Milling" the router would want to fly out of your hands IE no effort required but dangerous.

Now the best way of routing and milling is the climb method but to work effectively you must have no backlash. But and its a big you need less effort, the router cutter is doing most of the work the stepper motors are controlling the rate of travel.

The electrical side is easy but time consuming the control box is best made out of ali, as this is a natural conductor of heat and helps to keep the drivers cool by acting as a heat sink. There is soldering but again it is fairly simple.

The power supplies i use the 70V models and have two, if one gives 7A at 70V then two would give 14A at 70V but with a kit buy those drives and power supplies it will come with motors that are overpowered.

The cad details are as you describe but it can be simple to make your own gcode, you also need to have an understanding of gcode to allow you to fine tune or de bug the code.

In Mach you can have home switches these can be as simple as cheap micro switches these can be any where on the machine table position wise then you have soft limits in Mach so if you make a mistake in your g code Mach will not allow you to go outside those limits "Software" but here is the problem if you lose too many steps the soft limits become worthless and you can crash the machine.

Currently i have limits on the end of travel on all axis, Mach also allows them to be used as homing switches as well and the all go to one input on the break out board then into one input "Pin" in Mach

I would suggest you go to the Mach forum and do some further research Soft limits etc most people ask the same questions over and over again.

http://www.artsoftcontrols.com/forum/index.php

As to budget they can be a money pit if done incorrectly but you can build a very good machine for way less than you would pay commercially.

I would say that it is a steep learning curve much more than you would imagine my own forte is engineering, but it took quite a time before everything fell into place.

I bought a cheap machine of Asian descent, i have spent no end of time fixing its lack of inbuilt quality.

Re laptops, they can work with Mach but it must be fairly powerful and it must be dedicated to Mach, again search the Mach forum for what others have said.

It would be no problem if you wanted to talk, just PM me.

Phil_H

[ger21]

Originally Posted by Sainty Thanks for that Phil, interesting reading. First thing to tackle for me is the threaded rod/ball screw issue. My heart tells me that ball screws are the way to go. However this is supposed to be a budget build, if it all works, I can always upgrade. For the size of machine that I am looking at it will be in excess of £200, for the ball screw and flanged nuts. As an alternative I can get M16 (2mm pitch, 78IPM at 1000prm) threaded rod for a fraction of the price and manufacture some flanged nuts for about £50. I know that I will be at the hands of backlash but there are measures that i can take to eliminate this. If that turns out to be unbearably slow, i can always upgrade. I'm basically trying top work this all into a sub £500.00ish budget.

Look into multiple start acme. I believe these guys will ship to you.

http://www.roton.com/Mating_Componen...family=7060986

Keep in mind that it's unlikely that you'll get 1000rpm out of your steppers. Steppers have the most torque at 0 rpm, and torque drops off as rpm's increase. So keeping the motors spinning slower results in more torque supplied to the screws. 2 Turn per inch screws have been found by a lot of users to give the best performance for the price.

Originally Posted by Sainty For fear of exposing myself as a total fool, this is how I understand the build process. You start with some form of model in a drawing package (CAD), this then gets transferred to Gcode (by software??) which a program like mach3 uses to control the router.

For 3D carving you use a 3D CAD model, but for most router work, all you need are simple 2D drawings. You use a CAM program which converts the lines to toolpaths, and creates the g-code file. Download the Cut2D demo and you can see how the process works.

http://www.vectric.com/WebSite/Vectr.../c2d_index.htm

The PC is connected to the drivers by way of the breakout board and printer port (can this be done via usb so that I can use a laptop?). The breakout board tells the X,Y and Z drivers what to do the process continues until a) the program is finished and the router returns to the 0,0,0 point, b) the limit switches are triggered or c) something goes horribly wrong and I hit the estop button.

The breakout board doesn't tell anything anything. The control software (Mach3) controls the machine, by sending step and direction signals through the parallel port. The breakout board just routes these signals to the appropriate drives. Basically, the breakout board allows you to connect your drives and any switches or relays you might use, to the parallel port without just cutting up a cable and routing the individual wires.

Basically your a,b, and c are correct, but the machine will only return to 0,0,0 if the g-code tells it to.

There is a viable USB option for use with Mach3, but it requires additional hardware to generate the step and direction signals, as they can't be sent via USB. The SmoothStepper provides these signals, as well as emulates two parallel ports to connect breakout boards for your limit switches and other devices. www.warp9td.com

And Phil gave you a bad link, which should have been www.cnc4pc.com

[Sainty]

Thanks for all the help guys. You read a couple of build threads, look at a few websites and you think it's easy! It's not until you start to plan it all out that you get a feel for all the work and decisions involved.

Anyway, I have managed to get hold of some 16/4 trapezoidal thread and flange nuts for reasonable money and I think i will start with that.

I have knocked up a model in sketchup (not a very popular piece of software round here?) to see how things go together and it all seems to be alright,the only thing i haven't got is any sort of adjustment is things go out of square. The table is 1500mm/800mm for scale.





Criticisms (constructive of course!) always welcome.

Rgds

Stu

ps I have the sketchup model if anyone wants it

[Oldmanandhistoy]

Hi Stu,

Sorry if I missed it but have to admit I've very briefly skimmed over the info so far.

What's the main construction material for your build?

John

[Sainty]

Hi John

18mm MDF mostly. The rails are 25x25x3 steel angle for the bearings and 15x15x3 for the rails.

cheers

Stu

[Tony the Ferret]

Hi Stu
We are also UK based and have a pretty nifty software which is very forgiving for woodworkers, we supply Enroute, which is now upto V4, look at some of the other forums for user comments (mainly USA), but we can also offer you help to get your new machine "moving" contact me directly for more information
regards
Tony

[Oldmanandhistoy]

Looking at the machine you have drawn up I would say if made as is you will be disappointed with the rigidity. IMHO you need to do some major beefing up and if possible go for twin screws on your X axis.

Have you seen joescnc2006's machines here http://www.cnczone.com/forums/joes_cnc_model_2006/ will give you some ideas for improvement.

I did notice you were looking at Motion Controls electronics and just wanted to add I to have used their gear and am very happy with it. I did not buy a kit and build my own power supplies.

John

[Richster]

Hi and welcome - here's my thoughts on some of your questions as a UKer, apologies if these have been answered already - I haven't read the whole thread...

The M10 thread will give you faster rapids, but you'll be working the steppers harder so you could run into problems there. Also unless you can get hold of 10mm ID bearings, you'll need to turn down the ends of the rods. The skate bearings (cheap!) are 8mm ID, and for my money the reduction in pitch (1.25mm) is a small tradeoff. It probably makes sense to go the M8 route therefore. As others have mentioned, you might want to consider the dual leadscrew route to keep the gantry rigid.

18mm is very ambitious in one hit for a homemade machine, I doubt you could achieve this with a handheld trim router anyway! I tend not to go much more than 3 or 4mm since the power and rigidity just isn't there. The Trend T4 works well and I think is just a newer version of the T3. You may be interested to look at how I went about mounting mine, since this can be a bit of a challenge. It's hard to advise on feed rates and speeds - it's a case of trying it out and finding what your machine is comfortable with. Aluminum is a no-go, although I have had some success engraving with a v-bit.

Software wise I have Cut2D from Vetric which does a good job of converting CAD plans (DXF format) to GCode. I use Mach3 for the controller which I can thoroughly recommend.

On the electronics, the driver board and steppers are generally sold as a kit and should be enough to get things moving. You'll also need an appropriate power supply for this too. Beyond that it's up to you. There are what are termed breakout boards, which simplify and make safer the process of hooking up additional components to your PC. Or you can make your own which does much the same. I posted a circuit diagram which you might find useful, which would suit most homemade 3-axis machines. This adds e-stop, limit switches and also router power control which is a useful and neighbour-friendly feature.

It's a no to the USB port - also many people have had problems with laptop parallel ports, so you're probably best off with a desktop with onboard parallel. You can pick up a 2nd hand one for next to nothing.

Hope that helps!
Rich

Originally Posted by Sainty Hi I am a joiner based in Southampton, UK. I've spent a fair bit of time over the last couple of months reading about DIY CNC, and I have to say that I'm hooked! I'm not quite sure what specific applications I have in mind for a CNC router but I have a couple of things in mind (custom radiator cabinet grills, corian cutouts, mould/former construction, to name a few) but mostly machining and cutting 18mm MDF. I intend to make a very cheap machine, but one which is capable of reasonably accurate/repeatable work and will be useful for creating CNC#2. My build will be based loosely on the www.buildyourcnc.com model with improvements where I feel necessary. Mainly in the construction, as I said, I am a joiner! Before I get too carried away, I want to check what I am doing is along the right lines. Firstly, transmission, I intend on using 10mm threaded rod which has a pitch of 1.5mm. Is this suitable? I am hoping that this will be stable over greater distances, (approx table size of 800mm x 1400mm). Secondly, drivers and motors. I have found these kits and plan on using kit 1. At the moment I am struggling to get my head round the electronics side of things, will these have enough power to move my router around? Also, what else will I need on the electronics side of things to get this thing moving (with the exception of a PC and software)? Thirdly, the router. I am luck in that I have several routers available to me from a trend T3 upto a 2200w festool. Is there a recommended router size given that I want to be able to machine 18mm mdf and might progress to aluminium. Fourthly, speed and depth of cut. Just out of interest, what sort of cutting capacity should I expect from the machine that I have described above? If I wanted to cut out a square of 18mm mdf, 300mmx300mm how long will it take? what sort of depth of cut will it achieve? Could it cut the full 18mm in one hit? I probably have about a million other questions but that will do for now!! Thanks in advance Stu.

[M250cnc]

Hi Stu,

Looking at your pictures the way you are mounting your bearings is the worst way to do it.

As you know routing is a dusty business the dust will build up on the bearings and rails making movement very jerky. Linear rails with seals work better but cost money.

Also there must be a way built in to adjust every thing for trueness and squareness

Agree with Ger21 about smooth stepper but again they cost money and anything imported from the US can be subject to Customs charges so be careful there if you already have a laptop you can try it but they can be hit and miss apparently.

Your cutting speeds will be pretty much what you are used to doing it manually what you don't have to do is stop and start while you change to a new position which you have to do when working manually.

I will stick with my original decision that you should build from plans that are already proven rather than just making it up as you go along. They are normally reasonable cost and you can ask questions in the relevant forums

I will also reiterate that it is possible to use steppers at a maximum of 1000 rpm but that will not be achieved if you lessen the voltage and cut corners elsewhere this relates to rapid movements IE not cutting, as a ball park figure you could cut at up to 50% of your rapid speed.

Phil_H