049c :: minor adjustments to belt + sticking tests

Its a national holiday, its quite boring.

Ok, after some fine adjustments to the belt retainers etc, i could reduce quite alot of the twisting action in the belts as the direction reverses.

After that, i managed to stick in the dial indicator and take some sticking and step accuracy tests :D

X-axis 1/16 step










 


 X-axis 1/8 step

X-axis 1/4 step

Z-axis 1/16step

Z-axis 1/8step

Z-axis 1/4 step












Y-axis 1/16 step












Well as we can see, Y-axis HTD3 belts have ALOT of buffer in the rubbery tension. This tension condemns the direction reversal to using about 3-5 steps to achieve directional change. As opposed to the MXL belt (from china incidently), directional change was limited to almost 1-2 steps of error.

This could only mean 1 thing : the HTD3 will have to go !!! And sadly, i do have quite alot of HTD3 left lying around, what a waste !

by running thru the above tests, i can deduce that 1/16step is going to be applied to my X axis, 1/16 step to my Z axis and i will need to change the HTD3 to MXL (once that arrives from China maybe in 2 weeks time) with 1/16 step as well. By approximating what is seen here, i can estimate that the setup is quite capable of reproducing a 20micron interval with an error rate of almost 1.5steps (which is about 10microns).

From some of the possible PCB milling scenarios i am seeing while mucking around in EAGLE/KiCAD, i could be looking at a requirement of about 2-4mils of accuracy (which translates to about 50-100microns). At this point of time, it seems that this odd looking machine could just make it without breaking a sweat.

however, to step things up to a higher level, i could order some other MXL pulley sizes maybe 14tooth (i did order some 12tooth, it seems that there is some problem getting that done).


MXL 14 tooth : 2.03mm x 14= 28.42mm
D = 28.42 / pi = 9.046mm

additional belt thickness = +1.16mm (measured)

Distance moved = pi*(D+ belt-thickness) * (0.1125/360) = 0.01002mm

belt doubling = 0.00501mm

T(MXL14) = 42 / (1.02) = 41.2N (4.2kgf) approx.

MXL 15 tooth : 2.03mm x 15= 30.45mm (I managed to order some MXL 15 as well) D = 28.42 / pi = 9.693mm

Distance moved = pi*(10.853) * (0.1125/360) = 0.01065mm

belt doubling = 0.00533mm

T(MXL14) = 42 / (1.085) = 38.71N (3.95kgf) approx.

MXL 16 tooth : 2.03mm x 16= 32.48mm
D = 28.42 / pi = 10.339mm

Distance moved = pi*(11.499) * (0.1125/360) = 0.01129mm

belt doubling = 0.00564mm

T(MXL14) = 42 / (1.15) = 36.5N (3.72kgf) approx.

 

This is getting more fun as it progresses :D

*edit* i forgot about the XL belts i have, so added herein. these China XL belts are kind of a flexible hard plastic (PU) material (non-rubber). LOOK ! the rubber buffering effect is gone !

(there could still be some error in the slight play of the wood + bearings combined, approx 1-2 steps)

XL10 tooth : 5.08x10 = 50.8mm
D = 50.8 / pi =16.17mm

D measured with belt = 17.8mm (belt adds 0.77mm)

Distance moved = pi*(D+ (beltthickness/2) ) * (0.1125/360)
= 0.01672mm

belt doubling = 0.00836mm

T(XL10) = 42 / (1.7) = 24.7N (2.5kgf) approx.

looks like this cheap XL belt has some nice plastic-stiff properties which is of great use here.

*edit* looks like things are slow these few days as i heckle with some china workshops on the subject of small tooth pulleys. they are quite annoying as they put up for sale on their site the pulleys but then refused to supply unless i buy in quantity. and the buying agent is a bottleneck as they answer 1 inquiry per day, for all the 10+ inquiries i made ... this is a waste of time.

... to be continued :D