I am in the process of building a cnc router for sale and was wonder what you prefer and why. I have built machines with both and know the pros and cons of each system. I am giving two choices that are close in price. They are as follows, single nut rolled ball screws or precsion rolled acme screws with antibacklash wear compensating nuts. So what do you like?

060112-1952 EST USA

eman5oh:

You need to define displacement vs force for each excluding thrust bearings, and whether there is any backlash.

Also you need to define static breakaway and dynamic running torque for the nut alone, and with a maximum realistic load force.

.

060112-1952 EST USA

eman5oh:

You need to define displacement vs force for each excluding thrust bearings, and whether there is any backlash.

Also you need to define static breakaway and dynamic running torque for the nut alone, and with a maximum realistic load force.

.


I did not want to get into the design details, just looking for what people are looking for in say a 3 ft x 3 ft light duty cnc router. I know that the ball screw has higher load rating as well as much greater effecancy. The draw backs are that it will have a little backlash and is more effected by getting dust inside the nut as well as being a little noisy while running. The Acme be quite running and have little to no backlash. It also handles dirt and dust well. Its main draw back is that it has lower load rating and with too much force you will overcome the backlash spring. You also need more motor to turn them do to lower effecancy ratings.Taking these things into consideration what would one want on the same machine know the above.

I'd go to precision acme with zero backlash nuts. I have a K2 machine that had nonprecision ballscrew and nuts, accuracy was not good until upgraded to the precision parts. Precision acme should also be much less costly than any ballscrew. I also have a machine with 1/2-10 acme on it with 200oz direct drive motors, it will readily snap a 1/4 inch mill without even slowing down. I doubt your machine will even get close to the load capacity of a properly sized acme screw. I have talked to people using the delrin zero backlash nuts on metal cutting heavy machines and they are still not getting backlash at heavy loads. Acme may need more force to turn but it needs less force to hold.

I'd go to precision acme with zero backlash nuts. I have a K2 machine that had nonprecision ballscrew and nuts, accuracy was not good until upgraded to the precision parts. Precision acme should also be much less costly than any ballscrew. I also have a machine with 1/2-10 acme on it with 200oz direct drive motors, it will readily snap a 1/4 inch mill without even slowing down. I doubt your machine will even get close to the load capacity of a properly sized acme screw. I have talked to people using the delrin zero backlash nuts on metal cutting heavy machines and they are still not getting backlash at heavy loads. Acme may need more force to turn but it needs less force to hold.
Did you go with precision ball screws or acme screw the second time around? You may be surprised but the price difffernce is not that much and in some cases the acme screws can cost more. This is some what offset by the fact the acme screws are not a hard and are easier to machine.

You can probably cut your cost by at least half going with a slighty lower precision screw, such as the Keystone that MSC sells. Imo, even the slightly less accurate screw with anti-backlash nuts are preferable to ballscrews with backlash.

Also, if you use multiple start acme screws, the efficiency can increase up to 50%.

I agree with you on that one. Lead error can be delt with in software by screw mapping, much better than software backlash comp.

060113-1013 EST USA

Why the assumption that a single nut ball screw has to have backlash?

.

Taking a low cost Thomson or equivalant 5/8-5 ball nut and screw off the self it will have .001 -.004 backlash in it. At least that has been my experiance with them. I know that you can reload the nuts with bigger balls and what not, but I am just looking at off the self parts.

I'm building another machine and will use Acme with zero backlash nuts. My K2 machine had Nook standard ballscrew, the ball nut had 0.007 backlash (which was within their standards) and 0.005 to 0.007 in 12" on the screw (also within standards). These standard ballscrews are not sold as suitable for cnc use, and they are not, they are motion grade ballscrews. My standard grade, super cheap Enco acme with zero backlash nuts give me less than 0.005 in 12" for 5% of the cost of the precision ballnut/ballscrew assembly from Nook anyway. IMHO
A single ballnut can have very low backlash, but either you or the factory is going to match the ball bearings to achieve it at a much greater cost than a simple acme zero backlash nut (ie dumpster sells these nuts for like $19.00). Nook wants $100 to load a single ballnut, $175 will buy a double ballnut for zero backlash.

I ended up ordering some 1/2-8 2 start Keystone Acme screw class 2C and nuts from Dumpster. I will let you all know how these work out. For me retrofit mill I used Thompson ball screws .004 lead accuacy, with a single ball nut. I reloaded them with new balls and they ended up with .0005 to .0025 depending where you checked them. It works pretty well, but I would like to find some bigger balls for it. Any have a sorce for these. The ones I used were .125 +/- .0001.

Well i will be using rolled ballscrews with a single nut for my CNC router. One thing that is appreciated, is that the backlash is constistent. THis means i can use software compensation. Or I might just buy another nut and make it antibacklash.
I know the accuracy is not nearly as great as some of the acme out there, but i think a lot of it comes down to what you want to do.
For my router, i am looking for speed and ruler accuracy, so the .004" per foot is fine

So far I have foun the lead error to be well under the .004 that they spec. Where I have checked it it is .0003-.0005 which is damm good IMHO. I found the back lash to very a bit though. I measued as good as .0005 and up to .0025 with reloaded nuts on my mill. My router has the stock balls and is around .002-.004. I did not bother to reload them as I use the router for wood and even .004 is unnoticable in the finished parts.

Yea ive noticed that too, the lead accuracy is actually pretty good. THe .004 is more an insurance if you ask me.
I've got some beautiful ground ballscrew, and its on a whole different level than the rolled. But, i think the rolled is pretty good myself...

I've also some got 5/8 2 start ACME with teflon coating. That means a .25 lead. With a delrin nut, it is amazing how much the nut will travel just by giving it a spin. If i had to use acme, it would certainly be multiple start! Most of our routers dont need the mechanical advantage of a 1/2-10 anyways

For this one I went with the 1/2-8 2 start acme. I was up in the air on this one so set up this poll and still ended up unsure. The deciding factor was mostly that the acme is not as hard to turn the ends to accept bearings and is a little cheaper. Turning the ends of the rolled ball screws is a real PITA.

Rolled ballscrews are a horrible thing to machine, I agree. Just last week i thought "Well i only need to turn down 1/2", i dont need to heat it first". Well i did make it, but it was an ugly proces. You really want carbide to do it...
Or be proper and heat those ends :)

If you use multistart ACME screws, do you plan to tap your own anti-backlash nuts, or would you buy them pre-made? If you tap them yourself, how do you plan to accomplish this?

I have actually already made some very nice delrin nuts for them. Rather than spend 100+ dollars for a 2 start acme tap.... here is what i did:
First, i cut off a piece of the rod 5" in length.
Using a lathe, i turned a taper on the first 10 threads
Using a milling machine and a 3/16 end mill, mill 4 flutes deeper than the depth of the threads. These are not very exact, the flutes are not exactly 90deg apart.


And then i just drilled a hole in the end for a piece of 1/4" steel to slip in, as a handle.
Tap tap tap! And one thing i quickly learn, tapping ACME is hard
and it makes sense why, since your removing so much more material than 60deg threads.

The milling makes for pretty sharp edges, ideal. I had to run the tap over several times to make hte nut very smooth.

I've only done this with delrin though, would not even bother with metal. Its just so hard to tap ACME (I think that roton said its 5x harder than unified threads). Plus, metal would wear off the teflon coating

I bought the nuts from here http://www.gonebowlin.com/dumpsterCNC they were cheap enough that I won't bother to make them. If I were to make them I would cut the threads on a lathe. Doing two starts is no more difficult that duing one just takes twice as long as you are cutting two threads.

when i first read this i was wondering why you went with 2 start for two reasons; 1) wouldn't precision ground acme screws and nuts be cheaper/more available in a regular one start and 2) the faster the helix the more leverage the cutting force has on turning the screw. If you need a faster action, couldn't you just change the gear ratios or even overdrive it - isn't that all you are doing with the 2 start anyway?

i'd say multistart threads are a little tricker than single start to cut, you've got to figure out how to index the work and the cutting tool geometry is a little tricker with the higher helix angle, still you are right they are a diy item.

I haven't made an acme tap, but if I was charge with the task I'd consider making it progressive - ie a set of 3 or maybe more with each one progessively increasing the depth of cut. imo this would give a home made tap a fighting chance in metal, although making a homemade tap is not trival, it would required grinding after hardening to be of much use on metal.

1) 1/2-8 2 start at MSC is very inexpensive - $30 for 6 ft.

2) 1/2-8 2 start is 50% more efficient than 1/2-10, so you waste less power. But yes, you do need a little more power to use them. Also, spinning the screw 2.5x faster could result in more whipping, depending on the required speeds.

Gerry, thanks, didn't realize the 2 start was so cheap. still i thought it was precision ground acme that he was talking about, but ok I'll assume no price advantage. your point on it whipping around is also good as you say if it was going very fast. on the efficiency though, isn't just changing ratios, ie 12/5 geared down 2:1 is the same as 12/10 direct drive?

The screw itself is more efficient, so just gearing it still leaves you with a less efficient screw. Check out Nook's acme catalog for lots of great acme info. www.nookind.com

Gerry, thanks, didn't realize the 2 start was so cheap. still i thought it was precision ground acme that he was talking about, but ok I'll assume no price advantage. your point on it whipping around is also good as you say if it was going very fast. on the efficiency though, isn't just changing ratios, ie 12/5 geared down 2:1 is the same as 12/10 direct drive?
Also with direct drive you have a much simpler drive system, less parts and less cost. For my application slowing the screw speed down is important, the longer the screw the slower the speed wherer whipping will occur.