Backlash with CNC kit

[Richards]

I'm still running my IH in manual mode but I'm wondering what kind of backlash to expect when I upgrade to Aaron's CNC kit.

The reason that I'm asking, is that after I installed some digital scales on my IH (similar to the ones that Robert Warfield shows on his excellent web site), I noticed about 0.010 'slop' in my Y-axis when cutting with the rotary table. I marked the dial indicator on the Y-axis and verified that it wasn't moving before I finally realized that the cause of the 'slop' was probably the backlash inherent in the manual machine. It seems reasonable that if there is about 0.010 backlash because of the factory screw and nut, that, given sufficient force (heavy cuts with the rotary table), that the axis would be free to move the same amount as the backlash. I tried tightening the jib screws, but they're a little small to really crank down tight.

Using the rational that the backlash is the culprit, it seems that if I installed the CNC kit with virtually zero backlash, and if the servo motors has sufficient torque to not slip, that there should be virtually no unwanted axis movement.

Is there anyone who has already installed the CNC kit who can validate my assumptions?

[philbur]

CNCing a machine is an expensive way of curing backlash. I don't know about the IH but most mills have some form of anti-backlash nut that can be adjusted. If not it's easy enough to make them.

Backlash on a well setup set of ballscrews should get you down to 0.001" or less. But then using it to cut manually could be a problem.

Regards
Phil

.

Originally Posted by Richards I'm still running my IH in manual mode but I'm wondering what kind of backlash to expect when I upgrade to Aaron's CNC kit. The reason that I'm asking, is that after I installed some digital scales on my IH (similar to the ones that Robert Warfield shows on his excellent web site), I noticed about 0.010 'slop' in my Y-axis when cutting with the rotary table. I marked the dial indicator on the Y-axis and verified that it wasn't moving before I finally realized that the cause of the 'slop' was probably the backlash inherent in the manual machine. It seems reasonable that if there is about 0.010 backlash because of the factory screw and nut, that, given sufficient force (heavy cuts with the rotary table), that the axis would be free to move the same amount as the backlash. I tried tightening the jib screws, but they're a little small to really crank down tight. Using the rational that the backlash is the culprit, it seems that if I installed the CNC kit with virtually zero backlash, and if the servo motors has sufficient torque to not slip, that there should be virtually no unwanted axis movement. Is there anyone who has already installed the CNC kit who can validate my assumptions?

[kimoyo]

Hey Richard,

I was just thinking about this yesterday, maybe you could ask Aaron about it. He uses Rockford rolled ballscrews and nuts in his kits. You have the option of preloaded or double ballnuts (without a spring) but I'm not sure of the pros and cons of each (question for Aaron). I think the preloaded have less backlash but doubles (< 0.002 backlash from website) are stronger?? Because he recommends the preloaded for x and y but double for the z axis. He doesn't say on the website what the LI uses.

I read that ground is supposed to be better than rolled but been told it depends on where its made. Adobe said he tested his Rockford's and

Originally Posted by Adobe Machine I just ran a test on Rockford rolled ball screws and their double nuts, using my DRO over 28 inches on the Z axis of my lathe that is being converted to CNC. Just for fun we loaded a 19 inch chuck and a 8 inch kurt vise in the cross slide to possibly duplicate weight and some cutting pressure.Using a simple Z program, back and forth at 51 ipm 22 times, then to home and back to 28 inches, a recorded error ( maxium) of .00087 was noted. It always returned to home at .ooo,(?), the biggest error at the 28 inch stop.The ways and ball screw were well lubricated.Unless you are making a commercial machine that must repeat at .0001- 24 hrs a day, I would say that a well set up system,using rolled ball screws and double nuts, would be fine for your application.Most errors occure due to poor bearing mounts, to include every thing that NC cams has pointed out many times concerning the correct way to engineer the mounts and pre-load.You could spend thousands on ground ball screws, but have horrible errors with out the correct bearing set up. Adobe Machine ( old as dirt)

N4NV got a set of ground ballscrews and nuts from Hiwin for under $1K. He was having backlash problems and they told him that

Originally Posted by N4NV I talked with Bobby at Hiwin today about my backlash. He said the max the ballnuts have is .0008 on the X and .0006 on the Y.

From what I was told ground would cost more than the IH machine but Hiwin has them for under $1K. I'm not sure if they are made in china but I thought ground was supposed to have no backlash?? From what they told N4NV and what adobe measured the performance seems similar.

I was going to ask Aaron (and Adobe) about this but haven't spoken with him yet but there's probably a reason he's going with Rockford rolled instead of Hiwin's ground (for example).

[kimoyo]

BTW Richards, do you have the lastest version of the IH mill?

[Richards]

I probably wrote a murky post that was not clearly understandable. Sooner or later (probably sooner), I'll be adding the CNC kit to the IH so that I don't have to spend my entire day turning cranks. I have no intention of running the mill manually once the ball screws are installed. My whole objective in running the mill manually, before converting it to CNC operation, is to thoroughly learn and practice the basics of machining. After I know what to expect, I won't be writing code that violates standard practices.

On the parts that I machine, 0.010" is much tighter than I need for anything except a few bearing pockets. In fact most cuts can be +/- 0.025 and still be tighter than necessary. However, I'd really like the pockets to be between -0.0000 and +0.002 so that the bearings fit properly. (The motor mounts that I make are not high precision ) To get around the little bit of slop that I'm seeing on the digital scales, I mill the pockets to -0.025" and then use a boring head to tweak the pocket to get the final size. That method works fine.

From what I've read, I believe that the IH with CNC controls, properly installed and properly used, will get me to the 0.000 to +0.002 tolerance level, as long as I follow standard practices and use common sense. (Rough milling to +0.020 and then a finish pass to size. Properly calibrated and setup Mach 3, etc.)

[Richards]

Thanks for the information about the Rockford screws. From what I read before buying the IH, I was fairly certain that some good quality rolled screws would be better than what I really needed. (And I totally agree that 'quality' in ball screws depends more on the manufacturer than it does on the advertising department of the vendor

Now that we know what 'can' be expected, is there anyone willing to go an record and say what accuracy they're 'getting' with their IH milled CNCed with Aarons kit?

(My mill is the latest one offered by IH. It has the ground ways and other enhancements that are listed on the IH web site. In all reality, it is a very impressive machine.)

[Randall]

There is a backlash adjustment on the maual Mill. It helps a little.
I aslo put some brass shim stock behind the mount on the end of the table X axis and that toke some out. I had some extra play that I may have induced when I took my mill apart.
On another note I have won some tapping heads on ebay and what a worthwile investment. Tapping is fun now.

[Adobe Machine]

Kimoyo:Check my thread on converting a Manual Lathe to CNC...I spent the money on the Bearings ( Barden) and the bearing mounts.So far there is no backlash in either the Z or X axis, but there have been no cutting forces applied yet,and I have only used 12 volts to the servos in a battery borrowed from my Hot Rod, instead of 78 volts which is the power supply.
If the conversion were going to be a production lathe, seeing many hours a day in use, I may have reconsidered using rolled vs ground.With the ability to use a double ball nut, good lubrication, excellent bearing support and protection of the ball screws from swarf, dust and chips, the lathe should hold .0005 or better under normal cutting conditions, as my retro Tree mill does.
There are just so many other factors in conversions to CNC , which include the ability to adjust the ways and gibbs, constant, correct lubercation of the ways, gibbs and ball screws (by correct I mean using "way lube" to avoid stick-unstick of the axis..too many arguments on this subject, but you can change the accuracy of a machine by using the wrong lubercants on the ways..really) protection of the ball screws and the sliding surfaces, spindle condition and runout, the servo (or stepper) physical connections to the ballscrews, having servos that are correct for the load and friction, and one of the biggies: Using the correct tooling and cutters ! I purchased a turning center in the mid 90's, and thought I could use the same type of tooling we used on the manual machine..WRONG..it took three months to change my stubborn mind ( logic; a cutter is a cutter ect) and get the correct tooling. As soon as you REDUCE CUTTING FORCES with good tooling, the accuracy increases 10 fold. We ( Son and I) are rebuilding a 1929 Vintage Bi Plane, and he has a small HF type lathe, was trying to turn and thread some alloy steel shafts with no luck using regular carbide insert tooling. I sent him one of my toolholders ( modified to fit the smaller tool holder ) with some positive rake diamond shaped inserts, and what a differance. Using dial indicators on the x axis and taking .005 to .009 IPR ( inch per revolution ) he was able to hold tenths and a very excellent surface finish. He used my threading kit to cut the threads and had good results.Hu flung dung ( Hu) demonstrated in one of his theards how to grind a positive rake on a used triangle carbide insert, with excellent results.
Use good, known procedures and products to build your conversion, and it will be a usefull, fun tool..leave out any of the steps and it becomes a boat anchor in the Hobby Shop. Most of all, have fun doing it !

Adobe (old as dirt )

[Cruiser]

Richards, I have a IH mill that is full cnc. Mach will allow you to counteract for any backlash and set a couple of parameters so it will be fluid. I tried both manual/incremental and MDI and very quickly found that MDI moves of .1 or 1 inch were able to show the backlash comp much better than manual. As for accuracy before i was done with all the set-up, i just had to make some chips and cut a 1.5" circle and measured it to within .oo1" of being round. with some runin and squeeking it in it'll improve. Right now i'm trying to get the vfd on line and it be giving me fits ! I'm tired now ... good nite all ....

[Loadedagain]

totally ignore the mid posts and answer the question from post 1. ballscrews with preloaded nuts (as ih suppies with their conversion kit) elimenate backlash. if you find the minimum preload does not get enough of the "slop" out... increase it!

ps: i'm a happy owner of an industrial hobbies (old style) mill with their cnc conversion... installed by me.

[BobWarfield]

Originally Posted by CNCPlastic Z axis has an additional problem. It loses position towards the table on every up or down move and has done this since day one. Move 1 inch up and 1 inch down, it is now about .002 closer to the table than it was before. Move 10 inches up and 10 inches down, its now out of whack .020. I can typically only run 2-3 machining operations before I have to reset Z to the part. When I try to run several operations I'll be buried into my fixture plate .050 deep.

Interesting. Are you running servos or steppers? Sounds like lost steps. Did you lap the machine?

Seems like the performance of the gibs and slideways on these machines has been pretty variable in the past. Some people, like myself, have had no trouble whatsoever. Others report problems like this, but it has usually been on machines purchased some time ago. They've undergone a pretty major update with the ground ways in the latest version. Not sure what I'd do if I had an oldie as temperamental as this sounds though. Keep trying to tweak it in, I guess.

Best,

BW

[BobWarfield]

That's a pity you've had so many difficulties.

I still think your Z problem sounds like losing steps to me. The motor just doesn't seem to be powerful enough, even with your counterweight. I can't picture why else your loss would be proportional to distance moved. I know you've got Mach 3 dialed in to the right motion per step by now, so what else can it be?

As to what to do about it, a more powerful motor, preferably a servo, seems like the only recourse. OTOH, if you're tired of it, you're tired of it.

What's the next mill gonna be?

Best,

BW