Rich Harman
08-25-2012, 5:14 AM
A while ago I posted about upgrading my Shenhui's belts to higher quality 16mm AT5 belts.
The problem was that an object would end up being a different size (in the X direction) depending upon where it was cut out on the table. At the far left it was smallest and at the far right it was largest. In the middle it was - well, in the middle. The difference in size was about half a mm. This has nothing to do with step size, if the step size were wrong then it should be a consistent wrong size.
It took me a while to figure it out but it was due to belt stretch. The motor being on the right side meant that when the cutting head was at the left there was the greatest length of belt between it and the drive pulley. As it moved to the right the belt was reeled in and the amount of belt that was stretched was reduced. Or it's the other way around. I have to admit that I don't fully understand the physics of a stretched belt especially since it is unlikely that it is stretched uniformly.
Incidentally I discovered this while cutting out dozens of 7" diameter 122 tooth gears. A friend had a 120 watt Universal at his work so we cut a bunch there. The Universal did a better job but they still were not perfect. After correcting mine it actually was more precise than the Universal.
After I replaced the belt with a higher quality one (I also had to replace the drive pulley to match the new belt), I did extensive testing and found that it would cut exactly the same everywhere. Problem solved.
Now, the problem is reappearing. After being in place for a few months under what was apparently too much tension it is back to the way it was before. Obviously I had the belt too tight. I loosened it considerably but there is no change in operation. I will have to replace it again.
So the moral of the story is to make sure that your belts are not too tight.
Or if you never cut out pieces that need to precisely fit together then it may never be an issue - and if you have a machine with a linear encoder then belt stretch won't cause that kind of error.
--------------
So I have four possible solutions;
1) Adjust my files to compensate - real pain in the @ss
2) Replace belt and don't tighten so much - should have done this to begin with
3) Put a motor on both ends of the gantry - not too hard to do and has the benefit of greater accelerations.
4) Replace idler pulley with one that has a spring tensioner - it should be this way from the factory
The problem was that an object would end up being a different size (in the X direction) depending upon where it was cut out on the table. At the far left it was smallest and at the far right it was largest. In the middle it was - well, in the middle. The difference in size was about half a mm. This has nothing to do with step size, if the step size were wrong then it should be a consistent wrong size.
It took me a while to figure it out but it was due to belt stretch. The motor being on the right side meant that when the cutting head was at the left there was the greatest length of belt between it and the drive pulley. As it moved to the right the belt was reeled in and the amount of belt that was stretched was reduced. Or it's the other way around. I have to admit that I don't fully understand the physics of a stretched belt especially since it is unlikely that it is stretched uniformly.
Incidentally I discovered this while cutting out dozens of 7" diameter 122 tooth gears. A friend had a 120 watt Universal at his work so we cut a bunch there. The Universal did a better job but they still were not perfect. After correcting mine it actually was more precise than the Universal.
After I replaced the belt with a higher quality one (I also had to replace the drive pulley to match the new belt), I did extensive testing and found that it would cut exactly the same everywhere. Problem solved.
Now, the problem is reappearing. After being in place for a few months under what was apparently too much tension it is back to the way it was before. Obviously I had the belt too tight. I loosened it considerably but there is no change in operation. I will have to replace it again.
So the moral of the story is to make sure that your belts are not too tight.
Or if you never cut out pieces that need to precisely fit together then it may never be an issue - and if you have a machine with a linear encoder then belt stretch won't cause that kind of error.
--------------
So I have four possible solutions;
1) Adjust my files to compensate - real pain in the @ss
2) Replace belt and don't tighten so much - should have done this to begin with
3) Put a motor on both ends of the gantry - not too hard to do and has the benefit of greater accelerations.
4) Replace idler pulley with one that has a spring tensioner - it should be this way from the factory