A test for the Shenhui machines.

[Rich Harman]

Try this:

Vector engrave a long skinny rectangle (20mmx400mm) three times. One at the left, one in the middle and one at the far right of the table.

Measure them as accurately as possible and report the results.

Results from other brand machines would be interesting too, anything with a relatively large table.

[Rich Harman]

*sigh* not one single taker on the test eh?

Well my machine was doing the following:

Rectangle on the left was 399.6mm, the middle was 400.0mm and the right was 400.5mm.

When cutting out parts that must fit together well this is a big problem.

The solution was to replace the X axis belt with an 16AT5 belt (http://shop.polybelt.com/AT5-Belt-Metric-Trapezoidal_c83.htm). This also meant replacing the drive pulley and changing the step size to 7.500 but now the measurements are 400.0mm everywhere on the table.

[Tom Schulze]

interesting. I'll have to check this out when mine comes in. If I have the same issues, I'll hit you up for help with this conversion.

[Richard Rumancik]

Rich, to what do you attribute the problem? Do you think it was a bad belt, or bad pulleys? I assume these were toothed belts. Hard to understand what kind of defect of the belt or pulley could cause a non-linear output, especially since each pulley probebly turns at least a couple revolutions to generate one rectangle.

[Tom Schulze]

I would think the original step size may be off, producing an increasing error as the head moves down the X axis.

[Khalid Nazim]

I ran the test and all rectangles were ok except the one on the left had an x-axis measurement of 399.5. With the kind of work I am doing, I am ok with this much error. Its too much work to do what Rich did to correct the problem.

Regards
Khalid

[Rich Harman]

An incorrect step size would lead to the rectangles being small or large but the same size relative to each other.

The pulleys are 20 tooth on the motor, 40 tooth secondary which turns a 20 tooth for a 1:2 ratio. It is a 3mm pitch so that is 60mm per revolution or 30mm per motor revolution. The drivers are set to 5000 steps per revolution so that makes the step size .006mm or 6.000 as entered in the program.

Belt stretch is what caused the problem. If you deflect the belt a small amount by hand (pull it away from the gantry) the laser head will be displaced several mm. Since I have replaced the belt, the same force of deflection moves the head about 1/4 as much.

Also, if you use the software "feature" that calculates the step size for you then it will almost certainly be off when used on something that is a different size. On my machine, due to the pulley ratio and steps per revolution it should always be exactly 6.000, anything else is a kludge that is masking some other problem.