[Andrew2111111]

Hi i have an issue with mach3 as far as i can tell.
mach3 .62
Ethernet SmoothStepper
MB-02-V6 bob from Homandesigns
gecko g320x

if i run a simple gcode to go from X0 to X250 to X0
it will over run X0 IE it will be X-1
depending on speed at full speed it will be out by about 10mm.
i have changed the pin on the bob from 2,3 to 8,9 no change.
if i go to port and pins and change to active high it goes the other way IE x10.
i was looking at the G320x when it moves and both lights stay on.

if any one has an idea what this is please let me know.

Andrew

[MartinBarfoed]

It could be that your steps/mm is off. You can calculate the step/mm and get a very good first number. However, if you want to maximize the accuracy you need to calibrate the value by doing repeated moves (the longer the better), measurements of actual moved distance and then recalculations of the step/mm value.


I did this video on how I did some while ago:https://www.youtube.com/watch?v=LRagjLewSJg

[Andrew2111111]

hi Martin

even if i am out on the steps per mm i am telling it to move the same amount in both directions.
so it should go back to where it started.

Andrew

[ger21]

Yes, it should always return to the same place it started.

With a servo, you're problem likely lies in the encoder signals.

[beone]

You are losing steps somewhere. Either the stepper is overloaded at some point, sounds like going out-not back, or a coupling is slipping.
Dave

[Andrew2111111]

the weird thing if i change mach3 to active high it gains in the other direction.
i have changed the drive, the pins on bob, the encoder.

Andrew

[xjdubber]

sound like you have a problem with pulse width setup and or pulsing to fast.

[SignTorch]

have you put alignment marks on every shaft, coupler and pulley to rule out mechanical slippage?

[Andrew2111111]

i was thinking the pulse is not right in 1 direction but i don't know how to change it in mach3.


if it was slipping it would be all over the pace but i can get to do it the same amount each time.

[jfong]

What kind of encoder. Some brands have a problem with signal quality. Here is info on it.

http://www.geckodrive.com/support/ap...-encoders.html

[Andrew2111111]

i am using an AMT102-V set to 512 steps

[SignTorch]

i was thinking the pulse is not right in 1 direction but i don't know how to change it in mach3.


if it was slipping it would be all over the pace but i can get to do it the same amount each time.

I guess that's a no - I disagree - good luck....

[jfong]

i am using an AMT102-V set to 512 steps

The encoder wheel has been known to slip. The fix is to put a tiny bit of super glue on the plastic center piece.

I would contact gecko and see what other issues they have seen with this encoder.

[Andrew2111111]

i did put super glue on the center piece.
if i had slippage i would expect it to be all over the place.
the mill has keyed taper locks on the motor.

[SignTorch]

I would expect electrical problems to vary all over the place and mechanical slippage to be more predictable and constrained like you are describing

so I wouldn't even suspect or consider or chase electrical problems unless I was 1,000,000% certain that it was not mechanical

the motor (probably) could not hold position if the encoder signal is failing consistently in one direction

if the motor holds anywhere near position and doesn't ever fault or run away then that's a strong indicator that the encoder and drive are (probably) working correctly

the only observable signal difference between going forward and backward is the state of the dir pin - if that's causing a directional problem then it might be a faulty or floating pin but I wouldn't expect that to cause a consistent 1 in 250 error

you do have a decoupling capacitor on the encoder power pins, right?

[Andrew2111111]

I would expect electrical problems to vary all over the place and mechanical slippage to be more predictable and constrained like you are describing

so I wouldn't even suspect or consider or chase electrical problems unless I was 1,000,000% certain that it was not mechanical

the motor (probably) could not hold position if the encoder signal is failing consistently in one direction

if the motor holds anywhere near position and doesn't ever fault or run away then that's a strong indicator that the encoder and drive are (probably) working correctly

the only observable signal difference between going forward and backward is the state of the dir pin - if that's causing a directional problem then it might be a faulty or floating pin but I wouldn't expect that to cause a consistent 1 in 250 error

you do have a decoupling capacitor on the encoder power pins, right?

I would expect slipping to be all over the place and not repeatable.
the mill is a Bridgeport R2E3 and you cannot get to the front of the motor while it is fitted.
I will pull it off and check that it is tight and has no signs of having slipped.
I think it is a pulse issue because I can get to do it in both directions by changing the active state in ports and pins.

also what is a decoupling capacitor ?

Andrew

[jfong]

Does this happen to all three axis servo motors?

Since you are using a Ethernet Smoothstepper, I can't see it having bad step/direction signal quality unless the board is defective. Seems like you should contact Warp9 to verify the board is working correctly.

[SignTorch]

I would expect slipping to be all over the place and not repeatable.
the mill is a Bridgeport R2E3 and you cannot get to the front of the motor while it is fitted.
I will pull it off and check that it is tight and has no signs of having slipped.
I think it is a pulse issue because I can get to do it in both directions by changing the active state in ports and pins.

also what is a decoupling capacitor ?

Andrew

in a move and return motion it could slip 1 unit during the reversal - it would not necessarily be slipping all over the place in a single move with one reversal -

I don't know how much screw revolution we're talking about to be off by one unit - if it's more than a full turn then yes the slippage would be more random - if it's only a few degrees then a loose set screw on a flat for example can slip a certain amount every time it reverses - I can imagine several causes of small consistent slippage during reversals - including ordinary backlash - I don't know specifically about the Bridgeport design - I could be way off

if changing active state means changing the direction of the initial move might that explain how slippage could be reversed

if you go to x 250 then to x 240 and it's off by 1 then all of the position loss occurred near the reversal and it must be slippage

if you go to x 250 then to x 125 and you lose 1/2 unit - that would mean the error is accumulating evenly and it must be a pulse issue - and it sounds like it might be happening when dir is in one certain state

sorry - I can't get past thinking it could be slippage

a decoupling capacitor is a very small capacitor on the encoder power pins to filter out tiny power spikes that might cause the encoder signals to flutter which can cause erratic position errors

[Andrew2111111]

i did pull the cover on the lead screw back and mark the nut and lock nut which is all i can see without pulling the motor off and they did not move that i could see.
i will have to see if i can get a decoupling capacitor localy.
at the moment i am trying settings in the ess as per info i fond on warp9.
i am running a test program that pauses at each end that way i am removing the change of direction.

[Andrew2111111]

I have not tried the other servos yet.
I had been using the mill and noticed that it was out in x axis.