Chinese mills and parachutes (need advice)

[IQChallenged]

I don't mind taking the time to do the research but I'd be stupid not to exploit the research and experience of others and just place my order. I want to convert my Grizzly 3103 mill to CNC but the backlash in the stock screws is ridiculous. I really wouldn't advise anyone to buy one of these Chinese mills. Some things in life just can't be cheap and still be what they are expected to be. A parachute would be a great example. If I had it to do over again I'd just wait until I could afford an old good mill or a new little one. But here I am with a Chinese mill with tons of back lash. The screw "looks" good but I really don't have any way of knowing how good it is. The nut is steel or iron with a slot cut in it so you can adjust how much bind you want. I'm not going to waste a minute messing with that. So I'm wondering if there's a screw that's as popular to linear drive as Gecko is to drives. To be that popular they'd have to be real damned good yet affordable and I realize that term varies from one individual to another. I don't know if I must buy the screw and the nut or if I can get away with just the nut. I'll greatly appreciate any suggestions.

[Mike Everman]

Hey man,
What you're getting is a solid base, dovetail ways and a decent spindle for about two dollars a pound. Of course first thing when making it an nc is to throw out those screws, and most of the conversion kits have different grades of ballscrew as options. I can't advise on the low end, sorry.
Personally, I use Bosch-Rexroth precision rolled ballscrews. If you go to my site, you'll see my ServoNut power module. (http://bell-everman.com/actuators.html) scroll down to the picture...

Not the cheapest, but not too bad, and no where as expensive as ground, and better in my opinion. A fully internally preloaded nut, no springs, no dual nuts, just plain *****in. The rolled surface is smoother than ground. All in all, A+, and I based a successful product on it. We typically use Renishaw linear encoders (I'm not proposing that DIY'ers would) but I bring it up because we cannot stand the slightest backlash when using linear encoders.

I've tested the 16mm x 10mm Lead to 300 lb accel/decels, one cycle per sec, 24/7 for two years on the first shot of grease. I kid you not. I stopped the test without any indication that we needed to re-lube, because we were tired of listening to it.

Less often, I've made ServoNuts with their 20x40Lead, and got 2.5m/sec out of them. (5,905 ipm!)

[BillTodd]

Roller nuts?

http://www.cnczone.com/forums/showthread.php?t=13593

http://www.cnczone.com/forums/showpo...&postcount=606

[Mike Everman]

I prefer the inside out method. I'll be making one up for my lathe sometime soon.
http://www.cnczone.com/forums/showpo...&postcount=298

[IQChallenged]

Originally Posted by Mike Everman I prefer the inside out method. I'll be making one up for my lathe sometime soon. http://www.cnczone.com/forums/showpo...&postcount=298

I don't understand how the straddling rings reduces backlash.

[IQChallenged]

Originally Posted by Mike Everman you'll see my ServoNut power module.

I just can't imagine being able to stuff all that machinery in a place that was previously occupied by a simple cube of iron and a long screw. And I'd also have to run wires to that mid-engined device. Looks way over my head? Am I wrong?

Also looks like it would cost a trillion dollars. Does it not?

[Mike Everman]

If the screw itself is ground and of high precision, then the inner set is preloaded against the outer two (obviously perpendicular to the screw centerline), and the lengthwise screws tightened to lock it. This loads both the radial bearings and the thread interfaces in one shot. Wear would require re-setting.

If the screw is non-precision, then the inner should be sprung with a belleville washer set with less than half the pressure you would expect to react linearly. Ideal case is to use two row duplex angulars in each of the three positions.

[Mike Everman]

Originally Posted by IQChallenged I just can't imagine being able to stuff all that machinery in a place that was previously occupied by a simple cube of iron and a long screw. And I'd also have to run wires to that mid-engined device. Looks way over my head? Am I wrong? Also looks like it would cost a trillion dollars. Does it not?

Oh, I wasn't proposing that it would be stuffed into a retrofit, my comment was more about using those wonderful screws. Yes, just under a trillion. That unit has an NSK spindle bearing and custom brushless motor set.

[BillTodd]

Originally Posted by IQChallenged I don't understand how the straddling rings reduces backlash.

The rings or roller (which-ever floats your boat) are forced into contact with the lead screw. they roll as the screw is turned, so require no 'play' between the nut and the screw.

They are not difficult to make; I made a set for my taiwanese RF30 clone drill/mill in a couple of weeks, from design to test (it's all documented in the thread). They are great; I can't walk past the thing without turning the wheels. I intend to convert the machine to cnc at somepoint with a couple of small servo motors and a home brew driver (I'd have been doing it now were I not engaged in renovating a Hardinge HLV-H)

One advantage they have over ball-screws is the machine can be used manually or under cnc control.

[IQChallenged]

My Grizzly 3103 mill uses an acme screw with a nut as shown below for both the X and the Y. Notice the big fat stud used to secure the nut to the saddle. Problem is- I haven't found a ballscrew/nut system that mounts like that. I wonder if anyone has made such a conversion.