[Martin Long]

I have recently been using the 4Axis to engrave a scale on a 30mm dia bar, I found the when producing the numbers they appear to be, not quite right, sort of on an angle of 1-2deg. On further inspection I found quite a lot of backlash in the rotary table, so no big deal to adjust the worm and wheel to eliminate this, I thought!
The worm wheel is not truly a worm wheel, it is a helical gear, with teeth which looked like they have been cut with a rusty nail, the burs on each tooth were so big, when I used a cloth to wipe the grease away they tore the cloth into shreads, OK, made in China you may say!!!
After tidying the teeth with a needle file, cleaned up and re- greased, I re-assembled the unit into the housing, I still found that I was having trouble adjusting the backlash, I could adjust the worm to eliminate the backlash at zero position but in 180deg. position it was too tight, or running in clockwise was ok but in counterclockwise is was tight.
On the rotary table side of the wheel assy. there is a thrust surface with an eccentric oil groove, however on the opposite side there is only a check nut without any form of thrust bearing or washer, the checknut face is not only running direct onto the housing casting, it is also machined out of square to the bore. This explains why it will run in one direction, but in the other direction it has no decent bearing surface.
To overcome this I have ordered a needle thrust bearing which is 5mm thick, I can machine the checknut by 3mm to accomodate the bearing whilst still ensuring to keep the checknut within the bottom surface of the housing.
I will take some photo's and post the bearing numbers etc. as I continue to try and make a silk purse out of a sow's ear.

Martin

[jalessi]

Simple solution.

http://tinyurl.com/4qnu3f

http://tinyurl.com/3ry7la

Jeff...

[Martin Long]

Jeff,
Thanks for your constuctive comment.
US$304 = AU$460
Shipping from USA to Aust. = US$70-100
Appox. total cost AU$600
One needle roller bearing AU$20

I will try my way first.
Please understand I'm not whinning about Chinese manufacture, I'm aware of what I purchased. I'm only trying to assist others, so they don't have to waist time investigating the problem.
Martin

[jalessi]

Martin,

I really welcome your efforts, it would be nice to have a inexpensive rotary table solution.

Please keep us posted.

Thank You

Jeff...

[Martin Long]

Further to my earlyer post.
Note the worm wheel has straight teeth.
Checknut with the shinny witness mark on one small area.
Location of where the checknut sits against the housing.
I should have the needle bearing tomorrow and will post my progress accordingly.

Martin

[iGG1e]

I hope you do find a solution. I'm just about at the stage where I want to use the 4th axis and getting it to run properly seems to be first on the list!

Please keep us posted.

[PEU]

Early this year we shown the Syil X4+ w/4th axis at a local expo, we ran a demo using the 4th axis all day long and I can vouch for its repeatability, the key setting IMHO is low speed, I had it set up at 2turns per minute and never had a problem.
Also you need to true it before using it, that will improve accuracy too.


Pablo

[sharpshooter90]

Our company had a Shoptask with 4 axis CNC and the Shoptask CNC rotary table. It used a worm gear on an eccentric to adjust the backlash. It ran zero lash and was smooth all the way. The table was made in India, and the quality was surprisingly good. The shoptask design also allowed manual use and use of dividing plates as well. Look at www.shoptask.com under "tooling" part # ST32

[Martin Long]

Fixed the backlash, but still not getting the quality I would like from my engraving.
Zero's appear slightly on an angle, Im using BobCam V21 to produce the text, vectorizing, then creating G code programe in BobCad.
As a newbe to CNC there must be something I'm missing, should'nt the two axis's X and A move in an interpolation and not as separate movements (ie I&J command)?
O()
G90 G80 G40 G54
G53 Z0.
T0 M06
S1600 M03
G90 G54 X0. Y0.
G43 H0 D0 Z1.
M09
G00X13.5965A34.3373
G01Z-0.2 F50
F200
X13.7252A34.8275
X13.8275A35.3972
A36.2428
X13.9694A36.7942
A37.3517
A37.9215
A38.4974
X14.0737A39.0734
A39.6493
A40.2313
A40.8072
A41.3832
A41.953
X13.9694A42.5105
A43.0619
X13.8672A43.6072
A44.2076
X13.7252
A44.7407
X13.5965
A45.1818
X13.4481
X13.2841A45.6413
X13.1089
X12.9288
X12.7492
X12.5787
X12.4211A45.1818
X12.2812A44.7407
X12.1681A44.2076
A43.6072
X12.0241A43.0619
A42.5105
A41.953
X11.9067A41.3832
A40.8072
A40.2313
A39.6493
A39.0734
A38.4974
A37.9215
A37.3517
A36.7942
X12.0241A36.2428
A35.7036
X12.1681A35.097
X12.2812A34.564
X12.4211A34.1229
X12.5787
A33.6695
X12.7492
X12.9288
X13.1089
X13.2841
X13.4481
A34.1229
X13.5965
G00A34.3373Z1.
M05
M09
G53 Z0.
M30

[iGG1e]

Not being exactly what you'd call experience in CNC either, but isn't it down to the configuration of the CAM program and the post it uses to implement the likes I and J commands in the final G code?

Was it the bearing fix that sorted your backlash out?

[Martin Long]

iGG1e,
I will wait to see what response I get from the post regarding the program.
Yes, I have fixed the backlash, but not entirely, I think there is some eccentricity in the wheel, there is a tight spot in the wheel, when I set it up so it drives ok over 360deg, there is a small amount of backlash in some areas. If I remove the backlash totally the motor stalls in some locations, I'm hoping with time the worm/wheel will bed in at the tight spots so I can re-adjust the worm accordingly.
With the needle bearing in place it certainly drives consistant in both directions. The bearing is slightly too big in the bore by 1.4mm, I will make a sleave to centralise the bearing at a later date. I had to turn the checknut thickness down by 3.3mm to keep it below the housing face, this also left just enough meat for the locking screw to be effective.
I can post photo's if anyone wants?
Bearing no. Torrington (not sure of the spelling)
1 @ AXK4060 (Bearing)
2 @ AS4060 (Washers)
AU$21
Martin

[ataxy]

photos are always good! please do.

[neilw20]

These were done on the flat, vectorization of Autocad complex font which is a double stroke version of simplex.

Using a square ended 0.3mm solid carbide cutter fashioned from a broken 0.5mm drill, and just ground square on the end.

Start grind using a CBN wheel till it just touches on 0.5mm then advance feed by 0.1mm and wait for it to spark out.Sparks visible with 10x magnifier.
Just grind the end flat.

Here is the results in stainless steel. Cutter survived 30 hours at 15mm/min and only 3600 RPM.

The trick is clocking the surface to better than 0.025mm.

Customer very happy. Black powder coat in grooves and raise to 200 deg C with the rest of the meal.

These scales used in a sandy environment. Original laser ones disappeared after a few months of use, so better solution was found.

As for the rotary I think I will support the other end of the cantilevered worm, and if that fails generate a split worm wheel. (using the old one)

[Martin Long]

Further to my exploits of trying to reduce backlash in a X4+ 4th axis.
I have come to the conclusion that without replacing the worm and wheel (helical gear) with a proper worm and wheel, and supporting the worm shaft at both ends it is not possible to reduce backlash to an acceptable amount, suitable for an accurate CNC application.
The amount of backlash changes just by the how much pressure is exerted on the grubsrew which locks the worm shaft.
This unit is a standard manual rotary table with the handwheel removed and a stepper motor attached, as originally intended this would work fine, backlash can be managed with careful use of a hand wheel, but as a CNC this is just not working.
Can anyone recommend a proper CNC 4th axis at a reasonable cost (without backlash preferably!!!)?

Martin

[iGG1e]

Not sounding too good there Martin. Yet another thing to add to the list of disappointments.

[neilw20]

I was talking today to a person that was involved in a large multi-spindle lathe with live tooling, and they had a problem working out why, although a spindle could run very fast, and had lots of power, when it was being moved slowly with live tooling cutting grooves etc in the part, the speed of the same spindle being being fed quite slowly doing some milling operations was quite limited in feed rate.

After some dismantling and inspecting how they got rid of the backlash, THEY DIDN'T. They used a resolver to measure where it was exactly and just used the 20HP main direct drive motor to servo it to the exact position.

The exact position with no 'backlash' was achieved by a DISK BRAKE causing drag to react against the cutting force.

They just didn't worry about 'backlash'. There was none!

Put a DRAG BRAKE, and use a HUGE stepper motor. That's one way but you still need to spring load the worm for NO CLEARANCE.

I am going to mount my worm at both ends. Over hung, as it is, is fairly useless.
For light work like engraving, a drag brake will work, but you need backlash compensation on the A-axis, or always present the load in the same direction.
You don't have a problem. There is just not enough friction! :withstupi

[ataxy]

well i guess i am stuck at the same point as you i got the backlash down to .3 degree but just grabbing the chuck now makes the motor stall i am waiting to ear from keith at syil canada about the torque that the motor as because i have a spare keling 495oz motor so that might help eleviate the problem abit also i have filled the bugger with grease help a bit with the slope of the "worm" screw mine as a hard spot between 300 and 10 degree

[neilw20]

I guess the tight spot is because the worm wheel is slightly eccentric.
That's in keeping with the accuracy of the whole show.
I suppose they are more useful as a positioning table.

If you clock up a blank to within microns, and have plenty of patience, you can make your own worm wheel.
Use two discs and worm the junction, and then you can make a real preloaded worm wheel.

Use the existing one 0.3 degree error one to make the cutter.
After a few generations you might end up with something perfect.

It depends how you value your time.

[docltf]

Martin
i have a rotory that did the same as yours.the way i cleaned it up,was to run the gears in using valve grinding compound.start with 90 grit and finish with 200 grit.

bill

[neilw20]

Grinding them in will help a lot, but until you support the worm both ends it won't take much load.