Steps per motor rev.

[beebee]

Ball screw 5/8 .25 lead.
Z axis 3 to 1
X axis 2.5 to 1
Y axis direct drive.

US Digital Encoder. E2-250-250 EG. It is 250 line count, which will provide 1000 pulses per revolution using x4 quadrature counting.
Update 1/8/2020

Gecko 320 x driver. Update 1/8/2020. Step Pulse Frequency is 0 to 300 kHz

DC brushed Servo. Max operating speed 4800
Continuous torque OZ-IN 170
Peak torque 850
Terminal voltage max 90VDC. I’m supplying the motors with 72 VDC

S/R
X...10,000.....Y......4000Y.....Z...12000
Turn..10.................2.5..............12

Did a 1 inch move per axis.

Y moved 1.75
X moved 7.00
Z moved 1.00

So the Y and X need modifications. Unfortunately I’m still at a loss to set this up.
I was setting up the soft limits then I realized that was a mistake. I have to get the
machine to move to real dimensions first.

[tblough]

X, Y, and Z should all have 1000 for encoder counts/rev.
X should have 10.0 for motor revs/in.
Y should have 4.0 for motor revs/in.
Z should have 12 for motor revs/in.

[beebee]

I’ll try the new numbers. Ty.

[frijoli]

S/R
X...10,000.....Y......4000Y.....Z...12000
Turn..10.................2.5..............12

Did a 1 inch move per axis.

Y moved 1.75
X moved 7.00
Z moved 1.00

So the Y and X need modifications.

Steps per rev should match if the encoder is the same on each axis.

[beebee]

We need to know your encoder's count per revolution. For example, with a quadrature encoder that has 1,000 counts per revolution, on the Y axis, you would enter 4,000 in the Wizard's Steps/Revolution (1,000 X 4 = 4,000). You would enter 0.25 in Overall Turns Ration (turns/in.). X would be 4,000 X 2.5 = 10,000 in Steps/Revolution. Z would be 4,000 X 3 = 12,000 Steps/Revolution.

It’s 250 so it’s 1000 on the encoder.

[Richards]

250 X 4 = 1000. X ratio = 2.5:1 so 2.5 X 1000 = 2,500 to spin one revolution. Y ratio = 1:1 so 1 X 1,000 = one revolution. Z ratio = 3:1 so 3 X 1,000 = 3,000. Assuming that you have assigned Axis 1 to X, Axis 2 to Y and Axis 3 to Z, in Wizard, Axis 1 Steps / Revolution = 2500, Axis 2 Steps / Revolution = 1000 and Axis 3 Steps / Revolution = 3000.

How far does each axis move per revolution? My 1605 ballscrews move the axis 5mm per revolution. 5mm / 25.4 = 0.19685 inches and 1 inch / 0.19685 = 5.08 revolutions per inch, which I enter in Wizard's Overall Turn Ratio. Substitute the distance moved per revolution for each axis and then divide 1 inch by those numbers.

Check repeatability. If I recall correctly, Geckodrive recommended adding capacitors on some types of encoders. If you don't have repeatability, then nothing else will work.

[beebee]

x =1.125. One turn
z=.375 One turn
It seems the motor configuration we are working on has a effect on homing.
As a example when the Y has a value of 1000 its homing switch fails to clear.
Arbitrarily if I increase the value to 2000 the homing switch functions as intended.
I did at one time have capacitors on the geckos. They were removed and when I
used Mach 3 I had no issues with the accuracy of the machine.

[Richards]

Maybe we're discussing things using different vocabularies, but I'm having some trouble following along with your numbers. As I see it, the numbers we enter into the Wizard are constants. If I enter 2,500 into Axis 1 Steps/Revolution, I'm telling Acorn that it needs to send 2,500 pulses to rotate Axis 1 one time. If I enter 1.125 into Axis 1's Overall Turn Ratio, then I'm telling Acorn that when it sends 2,500 pulses, that Axis 1 will move 1.125 inches (or millimeters depending on which system you specified when setting up the software).

With those constants set properly, the Acorn knows how to move an axis precisely, but it doesn't know how fast to move that axis. That comes later in G-code parameters (F10, F200, etc.). When moving an axis to trip a sensor (homing), the axis has to move slowly enough to allow the Acorn time to sense a change in state of the homing sensor. If moved too quickly, the state may change from Off to On and then to Off again before the Acorn has time to sense that a change has occurred. As far as I know, a homing signal is not latching, i.e., it doesn't not hold a changed state until reset. On my test bench, I use proximity sensors for homing. I also use a bolt head or magnet button as the target. The bolt head (that has been faced on a lathe) or the magnet button are large enough that the Acorn doesn't miss the change of state. If I used a blade type opto-coupler interrupter circuit, then the target blade might enter and exit the slot in the opto-coupler too quickly to be properly sensed. I also run the homing routine slowly enough to ensure that the Acorn can see the signal. Once the Acorn has seen the signal, I reverse the axis a short distance and then re-home at a slower speed to get a more precise result. Some servo's produce one index pulse per shaft rotation for very precise positioning. The ClearPath servos that I'm using now don't have that feature.

.

[beebee]

You asked how far the axis moves per revolution. Below is my attempt at the answer.
I didn’t modify the configuration at all.
As a note the pitch of my imperial ballscrew is .25

Here is what I did to get both of those number. Setup a ruler with a v tool in the z axis to a
line on the ruler. I marked the pulley. I then jogged the axis for one revolution of
the pulley. Those two numbers are the result.
x =1.125. One turn
z=.375 One turn

[martyscncgarage]

The attached files may be of help once you get the correct steps per motor rev set correctly.

This video may help as well (along with the Excel Calculator)

viewtopic.php?f=61&t=1449&p=8852&hilit=Calculator#p8852