[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 :cheers:

.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]

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]

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

[Hackman]

Whoa! That's not right. I think I did the math wrong. That should be around 12 seconds for full table traverse. I've got a 12" traverse @ 60 IPM which should come out to 12 seconds and not 5 seconds. Sorry about that.

To give you more context, I'm running a little X1 mill with 20 TPI leadscrews, so @ 60 IPM that gives me 1200 RPM. Apples to Oranges since you are using different stepper motors, drivers, and your mill is probably about 10x as big as mine, but just for fun, 1200 RPM @ 10TPI would give you 120 IPM, 1200 RPM @ 5TPI would give you 240 IPM.

[Oldmanandhistoy]

Whoa! That's not right. I think I did the math wrong. That should be around 12 seconds for full table traverse. I've got a 12" traverse @ 60 IPM which should come out to 12 seconds and not 5 seconds. Sorry about that.

To give you more context, I'm running a little X1 mill with 20 TPI leadscrews, so @ 60 IPM that gives me 1200 RPM. Apples to Oranges since you are using different stepper motors, drivers, and your mill is probably about 10x as big as mine, but just for fun, 1200 RPM @ 10TPI would give you 120 IPM, 1200 RPM @ 5TPI would give you 240 IPM.

I’ve been doing a search for “rapids” on this forum (should be in bed really but am on holiday so noprob) and I was seeing 60ipm rapids as a good average. So your 60ipm sounds good for your size of machine. I have been getting up to 82”/min on the stock lead screw so 0.1” pitch sounds like a good size for my X. With my motors and a 0.1” pitch screw I might be able to reach 100”/min rapids (would need to do the maths) which sound good to me. I get good rapids for a hobby machine and higher resolution to boot.

Will go away now and do more research on good cutting speeds and do the maths.

Got any recommendations for max cutting speeds for a hobby machine?

Thanks for your help and no problem with the error it made me get off my ass and find the figure.

John

[S_J_H]

John,
I use the .2 pitch 5/8" diameter rolled ballscrews that many people use. Main reason is simply the price. They are cheap, any other lead or diameter costs more. Add a second nut and preload them and you get pretty good performance at a very good price. Much cheaper to gear them up or down if needed.

I run 1/8 steps. Direct drive is on my X axis and belt driven 2:1 reduction on the yaxis.
So resolution is .000125" for x and .0000625" for Y.

Now I have never tested table movement at such a small increment, and my machine will not come remotely close to holding such tolerances!
But I can get reliable .0002" table movement. I use .0002" movements all the time manually. That's easy to verify. This small movement does come in handy when indicating in bores or edge finding.
If I thought for one second I could somehow get better performance from 1/2 or full steps I would switch. But I have never observed any issues with torque. I can command .0002" movement and get it.

I honestly see no difference in gt2 timing belt drive vs direct drive.
So I would go with whatever is easier to mount. I like belt drives as sometimes you can be much more compact with your motor mounting.

The most important thing I find to keeping backlash low and movements free and accurate, is clean and well lubed ways and well adjusted gibs. I would say I keep the gibs a tad looser than I would for a manual machine. This is necessary for free movement at very small increments.

I like to keep my rapids at 60ipm. I just find it's a good speed for manual jogging and I have never seen it lose a step at that speed. I can go well over 100ipm but see no need to risk lost steps.
I rarely feed at over 30ipm in aluminum. Much lower in steels.
Good luck!
Steve

[Oldmanandhistoy]

John,
I use the .2 pitch 5/8" diameter rolled ballscrews that many people use. Main reason is simply the price. They are cheap, any other lead or diameter costs more. Add a second nut and preload them and you get pretty good performance at a very good price. Much cheaper to gear them up or down if needed.

I run 1/8 steps. Direct drive is on my X axis and belt driven 2:1 reduction on the yaxis.
So resolution is .000125" for x and .0000625" for Y.

Now I have never tested table movement at such a small increment, and my machine will not come remotely close to holding such tolerances!
But I can get reliable .0002" table movement. I use .0002" movements all the time manually. That's easy to verify. This small movement does come in handy when indicating in bores or edge finding.
If I thought for one second I could somehow get better performance from 1/2 or full steps I would switch. But I have never observed any issues with torque. I can command .0002" movement and get it.

I honestly see no difference in gt2 timing belt drive vs direct drive.
So I would go with whatever is easier to mount. I like belt drives as sometimes you can be much more compact with your motor mounting.

The most important thing I find to keeping backlash low and movements free and accurate, is clean and well lubed ways and well adjusted gibs. I would say I keep the gibs a tad looser than I would for a manual machine. This is necessary for free movement at very small increments.

I like to keep my rapids at 60ipm. I just find it's a good speed for manual jogging and I have never seen it lose a step at that speed. I can go well over 100ipm but see no need to risk lost steps.
I rarely feed at over 30ipm in aluminum. Much lower in steels.
Good luck!
Steve

As always a wealth of information imo you are an asset to the zone.:cheers:

Point taken about the cost over there in the US; I only wish we had sources similar (low priced) to what you have over there.

A question or two to you if I may. If you don’t have time or the energy to respond that is fine with me, I am grateful for the info you previously provided.

If you could buy 0.1 pitch screws at the same price as the 0.2” and you had no restrictions of stepper motor and drive ratings what would you have used?

With that question in mind:

Reading your reply you say you are using 2:1 belt reduction on your Y (I will be using a belt drive system similar to yours for the same reason you are) so effectively a 0.1” pitch with 1:1 gear ratio. I think I know the answer but just to be sure can you tell me why you are using that gear ratio?

Would you agree using 0.1” pitch on a Z axis would be a good way to go?

Would you also agree using 0.1” pitch on the X would be a good idea IF your stepper set up would give you over 39”/min cutting speeds and up to 83”/min rapids reliably?

You say you use your machine in manual mode also; do you use hand wheels or do you effectively jog in software?

I have been hoping to still be able to use my mill in manual mode (using hand wheels) so I was thinking the 0.1” pitch screws would better suit this aim. Would you mind giving your opinion on this?


Are you saying that there is no point/advantage using a higher resolution on a hobby machine period?

If you got this far and answered the questions, thanks a million:wee:

John

[LongRat]

It sounds to me like you have good rapids and high linear forces with the 0.1" pitch screws. You might as well go with them in that case. However I am in the UK myself and got all my stuff from Homeshopcnc in the US. With the strength of the UK£, it is CHEAP - even with shipping and import taxes. I paid 62p per inch for my screws from them! That is down the back of the sofa prices.

[Oldmanandhistoy]

It sounds to me like you have good rapids and high linear forces with the 0.1" pitch screws. You might as well go with them in that case. However I am in the UK myself and got all my stuff from Homeshopcnc in the US. With the strength of the UK£, it is CHEAP - even with shipping and import taxes. I paid 62p per inch for my screws from them! That is down the back of the sofa prices.

Hi again LongRat,

Did you buy 1 preloaded ball nut/axis or two standard nuts/axis?

I think the way things are looking unless some one convinces me other wise (still time yet) I will be going with the 0.1” pitch screws. I have never imported anything but from what I am hearing I might very soon. My dilemma is this, if the 2.5mm screws I can get over here are the best choice and as this is a one off pay out then I maybe better off long term buying them. But if there is not a good enough reason/advantage to them then the Homeshopcnc screws will very likely be the ones I would buy.

What is going on with UK prices when you can import cheaper from the US? Either I am missing something or some one is adding a very large mark up.

The price for 1” aprox of screw over here is £1.28 +P&P and a preloaded ball nut will set me back £61.69 +P&P. Maybe Mr Marchantdice would like to jump in and out of this thread also with a quick sales pitch?

John

[S_J_H]

John,

If you could buy 0.1 pitch screws at the same price as the 0.2” and you had no restrictions of stepper motor and drive ratings what would you have used?

Hmmm,... If my motors could reliably rapid at 80IPM with a .1" lead and price was not a consideration I would say yes, I would have gone with that pitch. I have not seen any negatives from the .2 pitch though.
You may find your cutting limitations will be spindle HP and rpm and overall rigidity and not the ballscrew pitch. That is the case on my x3.

Reading your reply you say you are using 2:1 belt reduction on your Y (I will be using a belt drive system similar to yours for the same reason you are) so effectively a 0.1” pitch with 1:1 gear ratio. I think I know the answer but just to be sure can you tell me why you are using that gear ratio?

Well... when I did my x3 conversion I decided to space out the bearing block and drill the Z column so I could get near 8" of Y travel. But things get sloppy way out at the extreme Y axis travel because of the design of the Y axis dovetails. So the gibs need to be tight. I "thought" a reduction in lead pitch might help provide more torque to the Y axis , and since It was a simple matter of selecting timing pulleys I did just that.
Honestly, I can not say the 2:1 reduction has any real benefit though.

Would you agree using 0.1” pitch on a Z axis would be a good way to go?

I do use .1" pitch on the z-axis plus 2:1 belt driven reduction. I use the factory x3 acme .1" pitch screw on the z-axis. No back drive possible for superior holding power. The only negative is slow rapids because of the inefficient acme thread. But it has good plunging power.

I have been hoping to still be able to use my mill in manual mode (using hand wheels) so I was thinking the 0.1” pitch screws would better suit this aim. Would you mind giving your opinion on this?

Manual mode, I jog and do all manual work with cnc. No hand wheels on my mill.
I probably use the mill in a "cnc manual mode" 90% of the time vs running G-code parts programs.
IMHO forget keeping hand wheels, You'll never need them!! Yes, it does take a little getting used to running the mill with a keyboard, jog dial, MDI, or a gamepad "manually". But you can do things way faster than cranking hand wheels and with power feed on all 3 axis.

are you saying that there is no point/advantage using a higher resolution on a hobby machine period?

No I am not. I think having resolution of at least .0002" is good to have. I am saying not to expect the mill to produce parts with a .0002" tolerance.
All of the above just my $.0002
And, wow, nice job on the scraping of the ways on your mill!
Steve

[Oldmanandhistoy]

All of the above just my $.0002
And, wow, nice job on the scraping of the ways on your mill!
Steve

Might be just $0.0002 to you but it means a great deal more to me thanks again. I will go and take another look at your X3 conversion thread and see what motor/drive combo you are using .I do believe it was a Xylotex and 420+ oz in motors. Also from what you have said about your Z axis, I will go see if I can use my X screw for the Z its 1.5mm (0.059”) so a direct drive there might be ideal. The scraping is time consuming but definitely IMHO worth all the effort thanks.

I have enough info now to make an informed decision so thanks to all posters for your help.

John