Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: BUG on January 06, 2010, 01:16:27 PM
-
when programming small pieces (28x28) mm, the end result is a piece of (27.7x27.7) mm. But if the piece is large (150x150) mm, the end is accurate result (150x150) mm.
anyone knows what happens?
I modified the acceleration ramps and not get any results.
Thank you!!
-
Screw error, or flex in the machine.
-
but.... why the large piece it's mechanixe ok?? and why the small peace it's wrong??
-
Try the same small code but use the scale DROs to make it 5 times bigger or use the big parts code and scale it down and see what happens.
Hood
-
Hi all,
just a few questions:
- How is the drive setup of the table? Conventional screws or ballscrews? Steppers or servos?
- Is the machine new, or a used one with CNC retrofit? Possibly there is some wear of the screws causing your trouble.
- Of course, flex could be a possible cause.
Maybe, try to check wether the drive motors do the exact steps needed to move the table precisely, by marking "0°" on a flywheel or similar and make a part which needs exactly x revolutions to make the necessary travel.
Have fun
Konrad
-
The machine is a Zenbot.
Looking to analyze the information and evidence I have made.
-
Try cutting two small pieces in different places on the machine, 150mm apart, and see if they are the same size.
-
Probably screw error as Gerry said along with different backlash.
I would profile the screw some by checking the differences if in actual movements.
Say, 1" movements in one direction over 12" and then reversing direction.
Usualy you will see constant gain or loss, but it could also be more promounced in one section
of the screw. Check the backlash at different locations. Your .3mm ( .012") seems excessive for just the
screw since many are in the range of .004" per foot accuracy. So seems like something more going on here.
Timing belt tension also can have an influence. Try to isolate the contributing problem.
RICH
-
Hi all,
I just had a look on the website www.zenbotcnc.com....
The Zenbot mill is made from plastic material and furthermore is belt-driven. So, there is no threaded rod (screw) at all at the axis' drives.
Under all circumstances, most tooth-belts have remarkable backlash, depending on material and pulley geometry used.
(of course, there are "low-backlash" belts, but even the pictures on the zenbot website show a very small belt, which cannot be low-backlash.)
In other words: For this type of machine (and price...) the reported tolerance seems to be normal >:D
::) On the other hand, it seems to be a very good deal for beginners to get familiar with CNC-technology as the price is very low.
best regards
Konrad
-
I've adjusted so that the piece of 28x28mm is correct, then the large scheduled to 150x90, the result is 152x91.
Also changed the speed of 45000Hz Kernel to 25000hz and the summary is identical
-
"Under all circumstances, most tooth-belts have remarkable backlash" - That's just not true. There are MANY toothed belt systems that have essentially zero backlash, otherwise you would not see them used in CNC drives. XL belts are VERY common in CNC equipment, as are the newer, and even better, GT2 belts.
Regards,
Ray L.
-
Well...not under ALL circumstances. But in this instance, where a length of belt is used with a pinion pulley, Konrad is right. Using a toothed belt and 2 pulleys as a reduction drive is dandy to run a ballscrew, but don't expect much from a *belt n' pinion* for a metal cutting machine. Wood router....maybe. Pick and place, torch/plasma, plotter.....fine.
My 2,
RC
-
::)
maybe "under all circumstances" was a little keen ;D
In fact, I don't know any manufacturer that sells his toothbelts as "zero backlash"...
They all advertise how much the reduced the remaining problems in application.
But on the other hand, depending what is surrounding the toothbelt drive, the backlash and resilience of the belt driven stage becomes just irrelevant.
For example, using a belt as a reduction stage between drive motor and screw, is very common and does not affect the machine's precision in critical means.
Another example is using a belt-drive as "torque distribution" in machines that use two or more screws in parallel.
I like Overloaded's words "Wood router....maybe. Pick and place, torch/plasma, plotter.....fine."
Everything depends on the force applied by/to the system and the desired precision.
Konrad
PS:
* now putting the pair of gold scales back into the locker...*
-
It's an accuracy issue that you need to come to terms with. Your "system" is what it is. Usualy you would set the steps per unit based on the longest travel, say 150mm then the 28 mm error would be very small. But the accuracy may be different along the travel, in some range of travel, etc. So you have a choice, without getting into offsets / changing the piece dim's, or tool used, etc. of living with it, modifying / improving the machine or using experience to end up with what you want.
I took a quick look at the site and didn't see any closeup's of the machine, so i can't make any specific comments that may be easy and cost effective to improve on the machine.
Just make sure the steps per unit are a good comprimise.
RICH
-
You may have found the problem. This weekend more tests carried out, for sure and I explain that I get back.
Thanks
Excuse my English
-
I'm doing more tests to get the result.
But if I can say that the machine is capable of making 5 pieces of complex milling and are all identical. But not so far drawn accurately.
-
Hi, have you had any luck yet BUG?
I am experiencing the same issue