CNC Milling Machine Resolution Question.

[Oldmanandhistoy]

I recently bought a small milling machine (similar size to an X3) which I plan to CNC in the near future. The machine will be used for hobby type jobs like making me a CNC router in aluminium and steel. I have almost no experience working in metals but have a desire to learn.

So to my question; what would be a good resolution to aim for in my machine? I know this is not the same as accuracy but I believe higher resolution means higher accuracy from the machine. I also know that high accuracy comes with higher quantity components and design.

I will be using 200 steps/ revolution stepper motors with a +/- 5% accuracy per step.

I know this is a hobby machine but I would like to be at the higher end of accuracy or at least what is within my budget.

As always all constructive opinions are appreciated.

John

[HuFlungDung]

.0001" or .01mm

You most likely will want to keep 'the jaggies' under control during interpolated motion (ie., circular cuts and 2 axis straight angular cuts). The lower the resolution, the more that stair-step type errors show up.

Going the metric way actually gives you a wider tolerance range, but in a practical sense, you don't want to fight with a CAM system of any sort that must attempt to round numbers off to the nearest .0003 or .0004"

[toastydeath]

Resolution has nothing to do with the actual accuracy of the machine, despite there being some very good advantages to having very high resolution (as HuFlungDung has briefly explained).

Practically speaking, all a higher resolution does is allow you to divide the machine's inherently inaccurate state into smaller steps. These steps have a high degree of nonlinearity, and so are somewhat meaningless from an absolute sense, even over small distances. If your machine is very repeatable but inaccurate, this can give you the capacity to produce a good part after producing some (a lot of?) scrap or spending a lot of time doing setup. But more often than not, the machine won't be sufficiently repeatable and people love to spend hours changing tenths on wear offsets chasing a dream.

One of the great ironies in my particular shop is a tale of two lathes. One has a resolution of .00001", but has trouble holding .0005" for more than one or two parts. The other has a resolution of only .0002", but repeats to twenty millionths until the tool begins to wear. The servos happen to stop on the same spot every single time, and the screws, ways, and spindle are all of abnormal precision.

So, which would you take? The machine that reads in .00001" but when you tell it to go 5 inches it goes 4.9942, or the one that only reads to .0002" but will not only go 5 inches when you tell it 5 inches, but it repeats beyond what most shops are equipped to measure?

Again, resolution and accuracy are two concepts in the same ballpark, but are on different teams. Make sure your machine will repeat before you spend money making it resolve .0001".

Edit:

Or, if you want the contouring smoothness.

[Oldmanandhistoy]

Thanks for your time guys it is appreciated

Originally Posted by HuFlungDung .0001" or .01mm

0.0001" = 0.00254mm

0.01mm = 0.0003936996"

If we are talking 0.0001” resolution, is that 0.0001”/full step or per micro step?

I think I am right to say many people converting a mill to CNC use 0.2” pitch ballscrews with direct drive. So they will have a resolution of 0.001”/full step. Is this number well within the range I would need for a CNC mill? I am looking at two options that would give me around 0.001”/full step or 0.0005”/full step, is 0.0005” over kill on a hobby CNC mill?

John

[HuFlungDung]

As per your calculations, .01mm is 4 times lower resolution than a .0001 inch machine. But, .001" is 2.5 times lower resolution than the .01mm metric machine.

So you can take your best guess (as Toastydeath wrote) as to what your machine is really capable of for accuracy, and design accordingly.

It is really not feasible to use other factors for the sake of calculations used in cad and cam to create toolpaths.

.001" might be all that your machine is capable of, but to use a value that large means that the typical error is .002 because the machine does not correct for less than .001 so it must therefore allow positioning errors up to .0019" This is getting fairly gross for an error especially when trying to interpolate circles.

I believe that many guys will use microstepping to break the 200 steps/rev limitation. That being said, if you are working on a serious project for accurate contouring in metalwork, just forego using steppers and plan for servos with encoder feedback. IMO.

[Oldmanandhistoy]

I’m looking at two ballscrews one is 0.2” pitch and the other is 0.1” pitch. I would like to use direct drive with a stepper motor with 200 steps/rev and run it up to 8 micro steps/ full steps.

So all I really want to know is which pitch of the two ballscrews would be the best choice for a hobby CNC milling machine?

John

[LongRat]

Personally I would go with the 0.2" pitch screws (I have these myself). You will then be revving the steppers less to get decent rapid speeds, which will reduce the chance of you losing steps. Operating in the lower rev range for a given traverse speed will place you in the higher torque regime of the steppers. The difference in resolution will most likely make no odds after you factor in all the other factors that come into play, especially on a hobby machine. If you ever lose any steps this will almost certainly be far more harmful to your results than any effect of reduced resolution of motion.

[Hackman]

I would use the assumption that in the worst case, you will be running the steppers at 1/2 step for maximum smooth torque and go from there. The incremental torque between steps for 1/8 step is a lot lower then 1/2 step so you may not be able to reliably turn the leadscrew between microsteps when 1/8 stepping. If you can, then things will only get more precise, but in the worst case you can always use 1/2 stepping.

At 1/2 stepping the 0.2" ballscrew gives you 0.0005" per step which seems a bit coarse. The 0.1" ballscrew gives you 0.00025" per step which seems acceptable.

It's a trade off, the 0.2" ballscrews will give you faster rapids but less precision, while the 0.1" ballscrews will give you more precision but slower rapids since stepper torque drops as the RPMs increase. An argument could be made that the 0.1" ballscrews can do everything the 0.2" ballscrews can do with more precision, just slower. While the 0.2" ballscrews can never achieve the precision of the 0.1" ballscrews no matter how slow you go.

[Oldmanandhistoy]

Originally Posted by LongRat Personally I would go with the 0.2" pitch screws (I have these myself). You will then be revving the steppers less to get decent rapid speeds, which will reduce the chance of you losing steps. Operating in the lower rev range for a given traverse speed will place you in the higher torque regime of the steppers. The difference in resolution will most likely make no odds after you factor in all the other factors that come into play, especially on a hobby machine. If you ever lose any steps this will almost certainly be far more harmful to your results than any effect of reduced resolution of motion.

Thanks LongRat,

The reason I am/was considering using the 2.5mm pitch ballscrew is because the other day I temporally fitted and wired up a 3Nm stepper motor to my X axis I was using the stock screw of 1.5mm pitch. The reason I did this is because I wanted a temporary power feed to help me find out what the machine was capable of where feeds and depths of cut where concerned. Also because I am very new to milling metal I wanted to use the machine manually for a while so I would have a better idea when it came to using it CNC’d.

Anyway getting to the point; with the set up I can get over 1000mm/min with lots of torque (unable to stop the table with my body weight) and up to 2100mm/min rapids. So it did not seem logical to go to such a high pitch ballscrew. The problem I have is because I am new to working with metals in general I have very little idea to what feed speeds I will be needing for general hobby use. I see things this way if I don’t need the speeds a 5mm pitch screw gave me then I would be better off going for extra cutting force which a 2.5mm screw would give me. If you or anybody else can advise me as to what would be the max feed speeds I would need I will be in a better position to calculate screw pitch, stepper motor rating and the like.

John

EDIT: Also the extra resolution would not hurt.

[Oldmanandhistoy]

Thanks for the input Hackman,

I don’t think there is any doubt about using a 0.1” pitch ballscrew on my Z axis. It would make no sense to use 0.2” pitch and then gear down. This will also increase accuracy because of the direct drive.

Before I can decide on the X & Y I need to find out what is likely to be my max feed speed required and a nice rapid so will need to do some research on that first. Hopefully the max will allow me to use the 0.1” pitch for all axes. This will mainly depend on how fast a 3Nm (420+ oz in) stepper can drive a 0.1” pitch ballscrew on my X as I think it would be fine to use it 0.1” on my Y as it is such a small distance of travel.

As already mentioned I have has 1000mm/min (39”/min) feeds with lots of torque and rapid of up to 2100mm/min (82”/min) on my X using the stock 1.5mm (0.059”)lead screw. If these numbers are good for general hobby milling then the 0.1” pitch it will be.

If anyone notices any dodgy thinking here or has better advice please feel free to jump in. Or would like to give me some number for recommended max feed and rapid speeds for a hobby mill with max 3000RPM spindle it would be most appreciated.

Thanks all again for the help,

John

[Hackman]

Just my opinion, but I would say shoot for rapids that take about 5 - 6 seconds (EDIT! That should be 12 seconds, not 5 - 6 seconds. Blame my poor math skills) to traverse from table end to table end. No scientific explanation on that number, it just seems to be the speed on my mill where it becomes bearable to watch it doing rapids when running a long CNC program.

[Oldmanandhistoy]

Originally Posted by Hackman Just my opinion, but I would say shoot for rapids that take about 5 - 6 seconds to traverse from table end to table end. No scientific explanation on that number, it just seems to be the speed on my mill where it becomes bearable to watch it doing rapids when running a long CNC program.

Wow that’s 3500mm/min (138”/min) for my machine and that’s without acceleration and deceleration. Is this the norm for a bench top milling machine?

John