backlash issues

[camfambmw5]

Hi gang, thanks to all for the help on my previous thread :-)

Yet another question...

I am trying to apply stepping motors to a craftsman compound milling table. The obvious problem is backlash. Do any on you have a simple (or not simple I guess) solution. Can someone explain the principle behind a backlash nut??? Is this something easily applied to alleviate my problem?

Thanks again

[NeoMoses]

an anti-backlash acme nut is usually split, allowing you to 'squeeze' the nut together, along the axis of the threads, and eliminate backlash. By doing this, you increase the torque necessary to turn the screw, however.

Do you have any pictures of your XY table? with more information, I could possibly be of more help, but I've personally never seen a Craftsman XY table.

[camfambmw5]

NeoMoses,

Here is a picture of the milling table. Side note, I still don't get the principle behind the backlash nut...the physics of it.

Look forward to your comments :-)

[camfambmw5]

Sorry,
attachment didn't take :-(


if you go to sears.com and put in a search for
00927593000

You will see it

Thanks

[HuFlungDung]

Cam,

Ordinary manufacturing methods tend to overcut the inside of the nut to make sure it clears the screw thread. The various methods of backlash reduction are simply an attempt to reduce this clearance to a minimum.

The method that NeoMoses spoke of, is simply collapsing the nut, bringing the tops of the threads to contact the threads in the nut. Because the thread profile is tapered, this collapse brings the conical profile of the threads in nut and screw closer together.

The other method of backlash reduction, with two counter-posed nuts, is actually better. With this system, you actually counter-rotate the two nuts, and then shim them apart to keep them a fixed distance from each other. This results in the threads in one nut bearing against one side of the threads on the screw, while the thread in the opposing nut bear against the opposite side of the threads on the screw. You must make some allowance for clearance, so generally this is a "try it and see" method to reduce the clearance to a minimum that will allow the nut to pass without undo resistance, over the entire length of the thread.

A third method of backlash reduction would be to counterweight the machine slide to pull it continuously against one side of the screw thread. This is hardly reliable, but in some situations, it is good enough for the light pressure of a finish cut.

[camfambmw5]

HuFlunDung,
You should be a physics teacher. Your explanation is the only one I have seen that instantly cleared up the whole issue. Thanks a boat-load.

John

P.S.
Nice shot of Hong-Kong-Fooey

[samualt]

Excellent thread! Now that we know what backlash IS and how to try and correct it can someone explain what backlash DOES? That is, what are the symptoms of backlash; How do we know we have that problem in the first place?
(I'm building my first machine and would like to know what to look for). TIA.

[HuFlungDung]

Backlash is lost motion. You might be familiar with this on an old car steering wheel, where turning it a little bit one way or the other does not do anything, or worse, the car steering has a mind of its own, because the linkage is free to move because of looseness between the steering wheel and the tire.

On a machine, this means the table does not begin moving the opposite direction immediately when the screw has already reversed rotation.

If you had a highly magnified view of the screw threads inside the nut, you would see that when the screw reverses, the bearing side of the threads has to shift from one side to the other, across the clearance that was built into the screw and nut to begin with.

This is why ballscrews are used: the clearance between the screw and nut is essentially zero, because of the low friction presented by a series of tiny precision balls rolling in the gap between the screw and the nut.

Edit: you will notice this when climb milling with your router, because the cutter will have a tendency to grab onto the work and pull it against the loose side of the screw. However, the screw keeps moving away while it rotates, recreating the gap. So this makes your chip thickness somewhat uncontrollable, and spoils the finish look of the cut, or worse, the chip gets so heavy that it stalls the machine, breaks the cutter, moves the part from its clamped position. Other than that, nothing serious
Note: climb milling is the most desirable method for professionals to use, because it is better for the cutter, and produces a better finish on the work, because chips do not get dragged into the gap and smeared onto the finished surface of the cut.

Edit 2: Even if you do not climb mill, backlash will create little jogs in your cut lines, whenever cutting arcs past a quadrant line. It will also produce some inaccuracy in the length and width of your part.

[NeoMoses]

Backlash is much easier to eliminate when you're designing the machine than retrofitting another machine. For instance, samualt, just design your machine to use ballscrews and you'll eliminate 1 source of backlash.

When retrofitting, like camfambmw5 is doing, you first have to identify all of the places where backlash is. One place is the nut/screw interface. Another place, which may have more backlash than the nut, is where the leadscrew attaches to the table. On my XY vise, this is just a simple collar with a setscrew, which moves around quite a bit. To eliminate this, you may need to drill and pin the collar in place.

You may want to take a look at moglice for the craftsman milling table. It could help a lot.

Also, Here's a good thread to look at for removing backlash.

Here's a guy who converted an XY vise to CNC. Hope they help.

[chuckknigh]

There are two other methods for dealing with backlash, which work surprisingly well.

Oh, and by the way, I'm the one who is working on the X-Y vise...the second one is the method I'm using with it. I've considered injectable polymers, painting on coatings of JBWeld, to cast new threads in place, and even finding ACME nuts and building a pre-loaded nut. None of these options was as elegant as the solution I finally used, except building a preloaded nut, and parts aren't available out here in my small town. I live in a rural area...the nearest Home Depot is 30 miles away. The nearest machine shop is even further!

Anyway, option 1.

Compensate for it in software. Essentially, backlash is the amount of "wiggle" the screw has, before it starts moving the table. I was playing with another X-Y vise last time i was in Dallas, and one of them had a *FULL TURN* of wiggle, before the table started to move. That's a LOT!

TurboCNC (free CNC software) can be set to compensate for backlash. This method works best with minimal backlash -- if you have a full turn of the screw before anything happens, this is not a good solution. It is meant for "tweaking" a system that's already in pretty good shape.

The second option is to remove the backlash from your table. This is surprisingly easy on my X-Y vise.

OK, on my vise, the leadscrews are 14mm in diameter. Just for reference, 1/2" = 12.7mm.

Knowing that a tapped nylon leadscrew inherently has very little backlash, I took a rather unconventional approach.

I removed the leadscrews completely, and threaded some nylon "spacers" from my hardware store, which are 1/2" in diameter, with a hole already precisely pre-drilled through the center. AKA easy alignment. This was quite difficult -- imagine a 1/2" diameter lock nut, but with a nylon insert over an inch long. Worked great, though.

I then redrilled them for the new leadscrews, following the original hole, and tapped the holes for the right size and pitch. Any size or pitch is achievable, up to the limits of the internal dimension of the original thread, which was about 11.5mm.

Since I already had some lying around, I used 5/16"x18tpi allthread, though you could use anything you wanted, including a smaller ACME thread.

Just to be on the safe side, you might consider installing a set screw to hold this insert in place. It's a minimal modification to the original casting, and should not affect anything, in the grand scheme of things.

It was important to me that this modification is reversible (except for the hole for the set screw) and that it does not damage the integrity of the original equipment.

Now, you have to understand that this solution is not anti-backlash. It is minimal backlash. Backlash on a small scale can be corrected in software. If you wish to reduce the backlash even further, you can make a small modification to the insert.

Take it and, after threading it, use your bandsaw to "slit" into the end. 1 slit will work, 2 is better. Now, with 2 slits, the end looks like 4 wedges if viewed end on. With a file (or do it in your lathe) make a groove around the outside, into which you can fit a tight fitting O ring. This will serve to compress the treads on the end, while still maintaining some flexibility. The adjustment necessary to completely eliminate backlash on this type of system, is in the neighborhood of a few thousandths of an inch -- it doesn't have to visually compress.

I've heard this particular design works great, though I've not tried it myself.

Hope this helps!

-- Chuck Knight

[camfambmw5]

Thanks to all of my new ZONE friends :-) for there highly valued input. I have plenty to work with and am sure I will figure out a sound solution. Will keep you all posted.

PS
Does anyone live in the Chicago suburbs, particularly St. Charles area, would love to hook up sometime...maybe we could share design ideas over a beer. I have a tig welder, maybe I could trade fabrication assistance with someone that is "cnc live" alredy.I don't have a mill yet and could use a hand or two with some custom milling during construction (mounting plates, etc...)

John

[kong]

And here's a set of nuts I made this afternoon from Delrin (acetal). Similar to Hu's description, I used grub screws instead of a washer to keep them a set distance apart. Obviously on a heavy duty machine, the nuts should be bronze, but these should last me a while (fingers crossed!)