Too many lead screws...

[MunchyMonster]

I've been lurking here for awhile and I've been slowly building up the courage to tackle building my own machine. A twist of fate allowed me to get my hands on a small (13"x20" cutting area) cnc machine for a couple of months. I grabbed some part files from a web design and started cutting them on this machine. I've done a lot of refactoring and creative jig work to get the parts so I could cut them in the small envelope I've got available, but (so far) I've been able to cut all the pieces for my machine (except for a few small parts that I should finish up shortly and will easily fit in the cutting envelope).

Since I'm getting close to having the structure done (I've been test fitting everything as I go). I need to start considering lead screws and electronics. I've already decided I'm going to buy the motors and electronics as a kit, so that just leaves the lead screws (and sundry connection bits and pieces).

I'm looking at getting ACME screws from McMaster-Carr. I've seen 1/2-10 single start, 2 start, and 5 start discussed on the forums fairly regularly. My understanding is that the more starts the "smoother" the motion will be and the more linear distance per turn. For example 1/2-10 1 start will take 10 turns to go 1 inch, but 1/2-10 5 start will only take 2 turns to go an inch. Obviously precision is going to be affected as well since each rotational step for a given stepper is a finite size, fewer turns per inch also mean the smallest step it is possible to take is going to be larger on multistart screws.

Looking around it looks like I can get 1/2-8 8 for a price I'm willing to pay. Dumpster has AB Nuts for this too. Is there any reason this size lead screw isn't used more often? Don't want to drop the bucks for screws and nuts just to find out there is a really good reason why no one uses that size.

Thanks in advance for any insight.

[DonFrambach]

I have 1/2 x 10 single start lead screws on my 14"x14" router. I had second thoughts for a while but these are working well for me. I think 1/2 x 8 single start lead screws would work fine for a small router too. (I assume you mean single not eight start in your post)

BTW I can do rapids up to 75"/minute but have set my machine to 30"/minute because that seems fast enough for me right now.

my 2 cents

Don

[prcdslnc13]

Im using 1/2-10 5 start and single start. I am in love with 5 start screws now, super fast and way easier for the machine to move. I think the reason people dont go with the 1/2 8 8 start is the more starts and less TPI you have the more resolution you loose. So if it take 10 turns to move an inch theres a lot more room for accuracy then if it only takes 2 turns per inch. With the 1/2-8 8 start 1 turn will move one inch (I think that's right correct me if I'm wrong). Thats incredibly fast, but there isnt much resolution there.

[MunchyMonster]

Nope I meant 1/2-8 8 start (1 turn per inch). I may have bitten off more than I can chew, but my grandfather taught me that it's ok so long as you keep chewing. The machine I'm building will have a usable 24x48 cutting area at least. (It looked smaller on my computer than it does in my shop )

It doesn't need to scream, but I've seen lot's of people say that they started with a 1/2-10 1 start and almost instantly wanted a 2 or 5 start. Figured I would just start with a multistart screw to begin with.

The little cnc that I've got my hands on now (twist of fate and temporary) is probably getting close to 100ipm rapids and I can tell I'll take faster if I can get it.

[prcdslnc13]

With a 1/2-10 I was maxing out at about 60 IPM. When I converted to a 5 start (a couple nights ago) I was able to get 390 IPM with 40 IPS acceleration.

[MunchyMonster]

Originally Posted by prcdslnc13 With the 1/2-8 8 start 1 turn will move one inch (I think that's right correct me if I'm wrong). Thats incredibly fast, but there isnt much resolution there.

I think you are right, 1 turn 1 inch. But if I'm not mistaken a "standard" step size is 1.8 degrees (or 200 steps per rev) that means I'm getting 0.005 inches per step using full stepping. Thats less than 1/128" (0.0078125). I'm not going to be machining metal here, I'm looking at doing wood. I can take off more than 1/128" with sandpaper and can expect changes in size along those lines due temp and humidity changes.

Am I missing something or am I underestimating the amount of precision I'm going to need to do inlay work, for example. Or are there going to be issues with turning the steppers too slow?

[prcdslnc13]

you dont ever want to run steppers in full step. Try it once and you see why, they get loud and jerky and no good for machine motion. You want to try for at least quarter step.

As for resolution your probably right. Im running roller chain on one of my axes and its pitch make that axis 2.25" per motor rev, but Im also using 1/16 microstepping on that motor.

[ger21]

1/2-10 5 start give better resolution, and more power than 1/2-8 8 start. It's what most people use.
1/2-10 single start is really way too slow.

[MunchyMonster]

Originally Posted by prcdslnc13 you dont ever want to run steppers in full step. Try it once and you see why, they get loud and jerky and no good for machine motion. You want to try for at least quarter step.

So I can buy that running at full step is going to be more jerky than I might like, but even at 1/4 step I'm only increasing the resolution. If 0.005 was good enough then at 1/4 step I'll have 0.00125 which is pretty close to 1/1000 and way better than 1/100. Realistically, when working with wood, am I ever going to need more than 1/100 precision? Like I said I'm looking at doing inlays and maybe some digital joinery (think wacky dovetails).

The other side of the coin is power. Do I really lose all that much power when going from 1/2-10 5 to 1/2-8 8? I know the 1/2-10 5 is 2:1 gear reduction so I'm doubling the force, but how does that translate to my ability to cut. I'm already thinking I'll be able to push harder than the router can cut without bogging down, am I wrong?

[michaelthomas]

My personal run in with acme screws and Dumspster AB nuts has not been stellar. You will find that a very large portion of your steppers power will be required just to turn the screw under the nuts friction, and that is without any load applied. Its actually fairly common for the resistance to overwhelm the stepper when the nut is new and has not loosened up a bit.

So my point is that when so much of the power is gobbled up in sheer friction, halving the mechanical torque with a 1/2-8 8 start compared to a 1/2-10 5 will push you much closer to your steppers limit. I believe the figure that I have read is that leadscrews and plastic AB nuts are only about 30% efficient at the most, and probably much less when new.

I ordered some Roton ballscrews and ballnuts that should be arriving today. They require roughly 1/3 the power to to perform the same work. The screw is about 10 bucks per foot for a 5/8 diameter, and the nuts are about 23.00 each. Most report a backlash of about .004 with the stock nut. You can replace the balls with oversized balls for the ballnut and decrease the backlash by about half. I am going to spring preload two nuts together, which is a very simple process, to reduce the backlash to zero. You can find a lot of info on this site about the process.

[chrisforever]

Now Ball Screw is very cheap too. I think as cheap as acme screws now.

[MunchyMonster]

Hmm...

I hadn't gone and looked at ball screws because last year they were MUCH more expensive than ACME screws. Looking on McMaster-Carr for a 6 foot length ACME (1/2-8 8 = 1tpi) is going to cost me ~$49. A Ball screw (5/8 13/64 lead = ~5tpi) will cost ~$66. The difference in cost for the nuts between the McMaster-Carr Ball screw and the dumpster AB ACME is only ~$2-3.

I might need to give this some serious thought. I just assumed I couldn't afford Ball screws.