Bolt Together CNC Router - Newbie looking for design critique

[ger21]

Yes, it will work, and in the ideal World is a perfect solution..........but.......you need to have both stepper motors deriving their signal from the same point......easy peasy, just plug them into the same point

Slaving the axis together is easily handled in the control software.

[LeeWay]

I hardy realize there is a second motor on the back side of my plasma cutter. That is how flawlessly Mach 3 handle slaving. The only time I watch both is when I home that axis. They always hit the switches at the same time.

[handlewanker]

Hi....as the man says, it works OK for me.......can't comment on that one as I'm still building my very first bridge mill and hypothesise that if the worst case scenario could happen with that design it just might.....Murphy's law etc.

I'd say design as you think best using the suggestions as best to fit your design requirements.......the end result will no doubt work just as well as trying to create the perfect machine which is beyond the means of most of us, so a bit of compromising is not a bad thing in the end if it enables the build to be finalised.

I'm going for a bridge mill design with a moving table and that means I can use a single central ball screw without the worry of any gantry slewing effects, but mainly because I want to machine harder materials and so need a more sturdy design type.

The design you are creating could easily be done as a moving table type machine too, if the gantry sides were bolted to the side rail frame and the table made to move on two linear rails instead with a central screw drive.......but the length of the base would have to be doubled to accommodate the length of the table travel.

With that configuration you can have the side supports and bridge really strong and massive as they don't move, so no inertia loads would occur, but that's another design for whatever purpose and where the length of the base frame is not a problem.
Ian.

[LeeWay]

I'm going for a bridge mill design with a moving table and that means I can use a single central ball screw without the worry of any gantry slewing effects, but mainly because I want to machine harder materials and so need a more sturdy design type.



Ian.

If the table is very large on such a machine, you can have racking as well. I am not sure at what point it comes into play, but would probably vary depending on material choices. You can minimize racking even on larger tables. I did it on my router. It's gantry is about 30" or so. Beefing up a machine to overcome racking may not be the best choice though depending on a few things like motor size and types of screws or racks, but one thing about racks or screws driven by dual motors with one in slave is that you will see zero racking.

[handlewanker]

Hi, I quite agree, the twin screw design for keeping the gantry square is the ideal choice, but if I had that choice in the steering of my car where there were separate drive motors to steer each front wheel seperatel I wouldn't take it on the road.

The problem of the twin drive on a gantry router may only occur with a poorly constructed or outfitted set-up, and I expect there are hundreds if not a lot more routers with twin drives all functioning happily etc......but.......the potential for it to happen is a factor you weigh in the balance.

Onj the subject of a moving table design, the potential to rack is only as good as the spread of the linear bearing blocks.

For instance, if the rails are placed too far apart..... right near the outer edges of the table, and you have the linear bearing blocks too close together on the rails, so giving you a bearing layout that is under square, by that I mean the longitudenal distance overall of the blocks is less than the span of them you could, just might get a potential racking occurring, but only under heavy loading from a cutter attempting to hog material.

Whereas if the rails are spaced a small amount inboard from the table sides and the blocks are spaced further apart on the rails you have more stability, and having three blocks to each rail would be extremely stable even if costing a bit more.

It would not be a good design to have a moving table that was wider than it's length purely for that reason.

If you did get racking on a moving table design, I would be extremely worried as to where it was coming from.
Ian.

[jerseyguy1996]

If the gantry is rigid enough and you have zero slop in the linear guides (granted there is no such thing as "zero" but I'm looking at it from a theoretical stand point), then racking should never be a problem even if you are pushing just on one side. The rigidity of the gantry will transmit the force to the other side and given that the other side is constrained by the linear guide to only move in the X direction it should all move as a unit. Obviously linear guides do have some slop even if it isn't noticeable and the gantry will have some wiggle in it because nothing is perfectly rigid. I'm just wondering how far out of whack it could get before it kind of self corrects due to the difference in load experienced by the two motors as the stressed (twisted) gantry pushes back on the side that is a few steps ahead.

[LeeWay]

That is just it. With a single screw, you really have to design a solid machine and preload the bearings pretty well to keep it running square and get close enough to zero racking that it never shows. You also need a larger motor, drive and PS most likely.
With two motors, it will cost more, but you could get away with smaller motors and higher speeds. Less load on the bearings as well when it is just rapid moving. This is important in a plasma cutter, but can also be relevant in a router. It means you can get away with a lighter gantry.
I have built enough machines now to know I would not drive a central screw again unless it is in a mill or lathe configuration.

In a slaved motor setup, one motor does not get ahead of the other if they are tuned correctly. If it every does due to a fault somewhere, the results would be the same as a single screw missing steps. Neither would be good.

[jerseyguy1996]

In a slaved motor setup, one motor does not get ahead of the other if they are tuned correctly.

Can you please explain what you mean by "tuned correctly"?

[LeeWay]

On Gecko's, you can physically tune the drives. Pretty much done by ear. As for acceleration and velocity settings, that comes later, but both drives would have identical step counts, acceleration and velocity settings.
In other words, you are basically just cloning the slaved drive system from the setup of the main one.

[jerseyguy1996]

I see. Thanks!

[handlewanker]

Hi, what we're saying then is....missed steps never happen.......with independent but slaved drives you don't want one missed step ever.

Why did I read that with a complex routing job involving intricate detail the machine is homed a number of times during the program to ensure it was within the program steps and not missing a step which would make accuracy non existent.....if that is the case missed steps are a real happening.

I don't think this should to an argument session as the drive is up to the builder and if problems did arise no doubt they can be sorted out.

If one of the motors fails and the other stalls with the overload, I expect there are safeguards that won't allow the "good" motor to burn out, like a cut out with over amperage demands etc.

This is one reason I would never walk away and leave a machine to it's own with such a drive system or any system for that matter.

Now why did I specifically like the moving table type machine.......well no slewing with a central screw drive was one issue but the major one was so that a really strong bridge could be achieved to do real milling with hard metals.

The bridge doesn't move, so metaphorically speaking it and the base frame could be made from a solid block of concrete with rebar included.

That would make an interesting casting all in one piece, weight no problem, in concrete or epoxy granite......LOL......boats get made out of concrete so why not a router.
Ian.

[LeeWay]

That procedure of rehoming a machine during a job is something I have never heard of. I suppose if you wanted maximum performance out of it, you might do that, but then where is the upside? You are wasting any little bit of time saved by overtaxing the steppers when you keep re homing it.

My Plasma, Router and lathe get homed once a day at start up. I am just now starting to get my mills to do homing at startup as well.

What most guys do after they setup a machine is to tune it, then calibrate it. Then see how fast you can move the machine before it will loose steps. Then back off those settings 25%. Then you have a reliable machine that will not loose steps under normal operation.

We walk away from all our machines. The plasma cutter only when it is engraving though. When it's cutting parts out,you have to watch it. Some parts will want to tip up and you can get problems if you don't catch it.
This is why some guys use a magnetic torch holder for such things.

I didn't say that missed steps never happen. They generally don't though under normal conditions. That isn't something I ever worry about. That is how rare it is in my shop.

[pippin88]

Hi, what we're saying then is....missed steps never happen.......with independent but slaved drives you don't want one missed step ever.

Why did I read that with a complex routing job involving intricate detail the machine is homed a number of times during the program to ensure it was within the program steps and not missing a step which would make accuracy non existent.....if that is the case missed steps are a real happening.

I don't think this should to an argument session as the drive is up to the builder and if problems did arise no doubt they can be sorted out.

If one of the motors fails and the other stalls with the overload, I expect there are safeguards that won't allow the "good" motor to burn out, like a cut out with over amperage demands etc.

This is one reason I would never walk away and leave a machine to it's own with such a drive system or any system for that matter.

Now why did I specifically like the moving table type machine.......well no slewing with a central screw drive was one issue but the major one was so that a really strong bridge could be achieved to do real milling with hard metals.

The bridge doesn't move, so metaphorically speaking it and the base frame could be made from a solid block of concrete with rebar included.

That would make an interesting casting all in one piece, weight no problem, in concrete or epoxy granite......LOL......boats get made out of concrete so why not a router.
Ian.

You can stall steppers with no harm done.
Stepper failure is pretty bloody rare.

Homing a machine won't tell you if you've missed steps or not (at least not in a standard Mach3 or LinuxCNC setup) - it just goes searching for the home switch and when it finds it, sets that as machine zero. It doesn't check if it has moved the correct distance from the position it thought it was in.

[ger21]

Why did I read that with a complex routing job involving intricate detail the machine is homed a number of times during the program to ensure it was within the program steps and not missing a step which would make accuracy non existent.....if that is the case missed steps are a real happening.

This is a machine that doesn't work properly. If you have to repeatedly home it to maintain position, that means it's always out of position.
If your machine is losing steps, then it's not working properly. There could be numerous causes, but the bottom line is that the problem(s) should be fixed.

Homing a machine won't tell you if you've missed steps or not (at least not in a standard Mach3 or LinuxCNC setup)

Mach3's Verify function will tell you if you've lost steps.

[handlewanker]

Hi, as I don't fancy building a router with a moving gantry, it's just academic, so if the problem does not occur during a run then it's an urban myth put to rest as a happening that might never occur in normal circumstances.

Two things you can do to not experience the problem and the first is to run with a central ball screw and the second is to extend the spacing of the bearing blocks under the gantry sides to give them more stability in the long axis, and at the same time beef up the gantry crossbeam and side supports to enable it to stay square......design allowing of course.

Adding more mass to the moving gantry part may be a step backwards, but it's in the right direction if the integrity of the machine is enhanced......there is a cut off point that is reached when the over engineering syndrome takes precedence and not priority.

While it will not stop the gantry from slewing under aggressive machining, it will right itself when the loading is removed, and no harm done except to the accuracy of the last machined event when the gantry slewed.

In the design we are discussing, there is no option but to have twin screws, so that is a consideration for the design aspect of having the gantry down on the rails and the table below the rails to enable short side supports etc.

If the Y axis rails were to be set above the bottom clearance of the gantry crossbeam you could have the toothed belt coupling across the back of the gantry to each screw without interfering with a job on the table.

In effect that would enable as well two linear rails to each side of the Y axis....one above upside down and one below the gantry crossbeam, right side up so giving double the guidance to the gantry.

This will enable a really heavier gantry crossbeam and side supports to be used as it is driven by two screws and twin stepper motors as opposed to a single screw and motor on "normal" gantry router designs with central screws.

I think if you're going to design a machine it's always best to practice overkill as most people will attempt to drive it beyond the limit anyway.
Ian.

[ger21]

Hi, as I don't fancy building a router with a moving gantry, it's just academic, so if the problem does not occur during a run then it's an urban myth put to rest as a happening that might never occur in normal circumstances.

Lost steps are not specific to moving gantries and dual driven machines. They can happen to any machine, IF it's not running correctly.

[handlewanker]

Hi, my whole case in point......if it can happen, as you say, it will be bad news with an independent electrically synchronised twin screw drive.....a single screw drive is self righting once the load is removed whereas if the independent twin drive misses a step on one motor it is continuously there until the end of the job or until you notice it......and that means if the gantry is under twisting loads it creates a big loading on the screws and motors too.

There'll be a lot of effort expended initially to get the gantry square to the table in two planes, so the last thing you need is to have the gantry being forced out of line.

So that is probably enough scare mongering for this session........the drive will prove itself or need to be re-thought out in the event it has problems.

BTW......I am over the Moon.......just down loaded Sketchup, the free version and have been having a glorious time creating abstract shapes and pulling and pushing the walls out etc........once I get to learn the basics I might just be able to make sensible models........the camera mode that allows you to walk around the model is so awesome.
Ian.

[ger21]

Hi, my whole case in point......if it can happen, as you say, it will be bad news with an independent electrically synchronised twin screw drive

I should have phrased that differently.
A properly designed and set up stepper machine should never lose steps. With steppers, you need to allow quite a bit of headroom to prevent the chance of lost steps. If you don't, then you can run into issues. If you stay well within their safe operating range, you should never lose steps.

It shouldn't be an issue that you ever need to worry about.

[handlewanker]

For those indulging in moving gantry routers....a huge sigh of relief has arisen.....LOL.

Mine will be a moving table design for other reasons, so one can but hope, but I agree, if the build is properly managed, it will fly, and the Wright bros proved that to be true.
Ian.

[jerseyguy1996]

Hi, my whole case in point......if it can happen, as you say, it will be bad news with an independent electrically synchronised twin screw drive.....a single screw drive is self righting once the load is removed whereas if the independent twin drive misses a step on one motor it is continuously there until the end of the job or until you notice it......and that means if the gantry is under twisting loads it creates a big loading on the screws and motors too.

There'll be a lot of effort expended initially to get the gantry square to the table in two planes, so the last thing you need is to have the gantry being forced out of line.

So that is probably enough scare mongering for this session........the drive will prove itself or need to be re-thought out in the event it has problems.

BTW......I am over the Moon.......just down loaded Sketchup, the free version and have been having a glorious time creating abstract shapes and pulling and pushing the walls out etc........once I get to learn the basics I might just be able to make sensible models........the camera mode that allows you to walk around the model is so awesome.
Ian.

Let me know if you have any questions regarding sketch up. Once you get handy with it you will find that you can easily visualize how everything is going to go together....create assemblies from individual components....slide those assemblies around to make sure that you have appropriate clearances from other parts when the device is running....etc.