[joejared]

As expected, at least the Z-axis of the LC3024 uses a pair of spherical bearings instead of angular contact bearings, and unlike my dinosaur machine, it appears that all axis are designed this way. This time, and for future commercial installations of my own OEM package, there will be a drop in replacement that uses the above referenced bearings without the mickey moused spherical kit with washers etc. While the rest of the mechanics of the LC series of routers is very sound, I'm still surprised that they're employing this method, especially since all it would take is another pair of castings to just do it right. While I will not be selling my own complete package for generic machining applications, I will make these components available to anyone interested, although the bearings will need to be purchased separately.

I should have the first prototypes both assembled and installed in my current application by tomorrow, as it doesn't really take long to setup and program things like this using my own MicroMill code and enough bearings are also in stock for my current applications. A link to photos will also be available on my Facebook album for this machine. Based on a simple visual, comparison, the housing should be a bit stronger than the original, as the original housings for the spherical bearings are cast mystery metal, likely Aluminum as it is non-magnetic. I'll also be sending a set to Techno, as a suggested improvement to their otherwise excellent machines. In terms of strength, it is more likely that the motor mount counterbores will fail than the new design, which is a 1 piece unit, designed to mount with the major diameter facing the motor side.

[Pplug]

Interesting. I've been playing with my 4896 at home and I can't find any backlash on any axis. How much are you talking about?

[joejared]

Interesting. I've been playing with my 4896 at home and I can't find any backlash on any axis. How much are you talking about?

For my older machine, it was upwards of 0.015" until I fixed it. Using spherical bearings or radial bearings as thrust bearings is pretty much a kludge. Here's what I've done:




For the X and Y axis, I'll have to put a relief in for the support posts so that the bearings are appropriately positioned, but it's also very similar.

[wildwestpat]

Hi Joe

Suggest you look at the data sheets on bearing selection before embarking on redesign. With the shaft rotation speeds required the ball screw mount demands are relatively low in terms of radial thrust and reasonable life could be expected from deep grove axial ball races. Your picture shows a low cost die cast housing not the better quality cast iron variety. Replacing the housing and bearing may not get you the improvement you desire as 0.015 inch slop is indicative of a failed bearing. This is because single row angular contact bearings are designed to be used in pairs with a suitably controled pre load. The part you have sketched and the protype you have constructed require some method of restraining the outer of the races and depending upon the direction of the angle the outer race way may require the precision parts to adjust the pre load as well as the retaining thread on the screw shaft.

There are double row bearings available which being pre loaded will have zero backlash without the complications of you having to provide precision adjustment of the critical pre load required to use single row angular contact bearings.

I suspect that the bearings in your machine have at some time been subject to abuse or have clocked up several thousand hours of use. A good quality deep row bearing in a cast iron two bolt fixing cast iron pillow block can be had for about $15 and may well fit directly in place of the die cast unit.

Hope this helps avoid the pain of providing an adjustment mechanism to sort out the preloading of your single row bearings. To anyone thinking about bearings I suggest it is worth a few minutes going through the bearing selection guides on one of the major manufacturers web sites as there are a lot of factors to balance to get a satisfactory in service performance balance with cost.

Hope this helps.

Pat

[RandK]

I had a very experienced machinist friend help me measure the backlash on my older model Gantry III (39x21) machine and we got a whopping .001 on the X and Y and nothing on the Z. The backlash compensation is a bit unusual but seems to work. The X/Y alignment was spot on after several years of use and journey downstairs. The LC's are better machines than the Gantry's...

[joejared]

Hi Joe

Suggest you look at the data sheets on bearing selection before embarking on redesign. With the shaft rotation speeds required the ball screw mount demands are relatively low in terms of radial thrust and reasonable life could be expected from deep grove axial ball races. Your picture shows a low cost die cast housing not the better quality cast iron variety. Replacing the

Last I checked, Iron, even cast iron was magnetic. The spherical bearings housings are obviously cast aluminum, and considerably weaker than my design. The housing is plenty strong enough, actually stronger than the material in the plate it is mated to. I've already been using these bearings on my older machine with no problems.

housing and bearing may not get you the improvement you desire as 0.015 inch slop is indicative of a failed bearing. This is because single row angular contact bearings are designed to be used in pairs with a suitably controled pre load. The part you have sketched and the protype you have constructed require some method of restraining the outer of the races and depending upon the direction of the angle the outer race way may require the precision parts to adjust the pre load as well as the retaining thread on the screw shaft.

Of course they are, and of course they do butt against something. The clearance hole. I use angular contact bearings in pairs, and in fact, that bearing housing was not a casting. I made it myself. I have an album that covers my work on this machine. The change improved accuracy and reduced friction, so I'm happy.

[joejared]

I had a very experienced machinist friend help me measure the backlash on my older model Gantry III (39x21) machine and we got a whopping .001 on the X and Y and nothing on the Z. The backlash compensation is a bit unusual but seems to work. The X/Y alignment was spot on after several years of use and journey downstairs. The LC's are better machines than the Gantry's...

There's a conflict in statements here. Was backlash compensation in effect when you measured it? I never bothered to implement backlash compensation in my own code, favouring minimizing of actual mechanical backlash. My own machine runs at 200 IPM in a patented process, and my customers machines typically run at 300 IPM, continuous duty, often 8 hours/day or more.

I have worked on the LC4896 in Colorado, LC5050 in New Zealand, and am currently working with LC3024 for a local and actually have a table top LC3024 coming in next week for an Aussie, not counting my 1998 ISA based machine. As the machine was almost new, I had already planned on replacing the X and Y radial bearings with the angular contact ones mentioned previously, and retrofitting them into the Z axis, but apparently the newer machines required retrofitting on all axis... Easy enough.

[RandK]

No backlash compensation was configured into the controller from the factory and after measuring I added a little to correct the .001

It's clear that you are attempting to promote your products and services... Good luck with that.

[joejared]

No backlash compensation was configured into the controller from the factory and after measuring I added a little to correct the .001

It's clear that you are attempting to promote your products and services... Good luck with that.

Promote my product? Hmm. I only sell my machines in a specific industry and for my own network of customers, as it relates to manufacturing foot orthoses, but thought I'd share what I've done with one defect in Techno-isel product lines. I was under the impression that that is one of the things this site is supposed to be for, or am I wrong about that? :stickpoke A while back, I did something similar and posted in in another thread. I actually think this solution is best, simply because it didn't change the mechanics and decreases the amount of force required for travel. I'm actually hoping Techno will make something similar for my future purchases, because it is the better answer, but they're usually slow to take suggestion.

[joejared]

Followup:

I'm pleasantly surprised that techno now uses angular contact bearings on their Y-axis, at least on the LC 3024 tabletop model. Stateside, they assemble that portion , but the other axis still come from Germany. The X and Z axis still use the RS6200 radial bearings, but if nothing else, it's one less hassle. :rainfro: