Axis run too far

Hello, I've build my own CNC machine. But there is a problem with my Y axis.


For example I want draw a line of 600 mm, the machine draw 605, 620 or something like that. Alway too far. What could be the reason for this behavior?
The axis move relativ smooth in my opinion, but there is a strange sound like this:
dl.dropboxusercontent.com/u/33977923/Y_Noise.mp3

From my point of view it sounds like a step loss, but then the drawn line must be shorter.
Here the technical data of my machine:
Y spindle: 16x4mm
Y Motor: 0,91 Nm; 2A
Driver: Chinese TB 6560 3 Axis, 75% Current; fast declay; 1/2 step
PC: latency 16000 ns

regads
Sebastian

can be a number of things . but i'll place them in order

Mechanical Binding
steppers at too high microstepping for the rated spec of stepper motor
steppers under powered
velocity too high

start by reducing the stepper driver micro steps by 1/2
over running of distance can be a number of things , mostly mechanical , such as backlash and sloppy movement
to check distance make a g1move say 1" and without making any changes set this as your touchoff point 0 then measure from this point and say move 6" and measure

adjust the scale setting in your ini to adjust measured distance , but this needs to be done a few times to get it right
measuring over a long a distance as resonable , will get better results

okay thanks.

The micro step is already 1/2. Actual the velocity is 10 mm/s and the acceleration is 100 mm/s*s. The other axis work well with this settings. On short distance there is no over runnung of the y axis. The LinuxCNC example looks very well.

I' ve tryed to comtensate the over running with the pitch settings. But if the over runned distance always change like discribed it is not a proper method.

As cncbasher mentioned try making the long moves at a very slow speed and see if the axis still overshoots or it overshoots a different amount than at fast speed.

Rick G

seppel wrote:

okay thanks.

The micro step is already 1/2. Actual the velocity is 10 mm/s and the acceleration is 100 mm/s*s. The other axis work well with this settings. On short distance there is no over runnung of the y axis. The LinuxCNC example looks very well.

I' ve tryed to comtensate the over running with the pitch settings. But if the over runned distance always change like discribed it is not a proper method.

this is nearly always a mechanical problem such as backlash or binding , or as a result of missing steps if the stepper motor is not capable of enough torque
is the screw the size and pitch you think it is ?

as Rick says slow things down and repeat the same check and see if the overshoot changes , but remember to always check from a forward move to discount as much as possible
any mechanical backlash

also try disconnecting the motor and turning the threaded rod by hand the full length , it may be possible to feel any tight spots that you may miss otherwise

sometimes it may be your power supply is not capable of supplying enough Amps , so try removing one axis and see if it improves .
your now at just a process of elimination , tedious and boring we know ..

as you say it's happening on long distances then, this may just be a tuning situation just because speeds work ok on one axis , does not mean it is good for all
as each axis has it's own problems such as weight and resistances to movement

Hello, today I've disconnected the motor and turn the whole way by hand. There are no spots or anything like that, it feels very smooth. Then I've tried to draw a line of 300mm by different velocities and settings. The result is a decreased over runnung by a low velocity. Then I changed to 1 step mode. There the way is to short.

In my opinion the motor is too small, because I can stop it by push my fingers together on the engine coupling. And because of it's the strange sound.

But why it runs too far and why the other axis are much better?

Might try a few things...

Check your motor wiring.

Substitute say the x and y drivers connected to the x and y motors and see if the problems changes axis.

When the machine is on but not making a move can you turn the coupling by hand?

After a move that is too long if the machine is commanded to return to the start position does it return exactly to where it started or does it over / under shoot?
How about...
F10
G1 x-5
G1 x0 ;MARK SPOT
G1 x100
G1 X0
G1 X200
G1 X0
G1 X300
and then try the same with F5
then G0

Rick G

Yes I can turn the coupling by hand if the machine is on.

After a over run the machine doesn't return to the home position. It runs again too far, so it reach the endswitch.

By now I think the chineseboard is the nonconforming unit. I run it with a 34V supply. I have research a high voltage increase step loss rate by this board.

it is not so much the Voltage but the Amps that matter , as most boards are current limited anyway , however do note that the limits on these Chinese TB656O'S are around the 34V region
but only around 2.5 A at absolute max , and in reality more like 1.9A .. are the stepper motors you use Nema 23's ?

what Amp rating is your power supply ? you need at least 10A in most cases if your using 3 Axis , the more the better .
also check the tb6560 for your default current setting , you may find that this is low

Yes it is a equivalent nema 23 with 2A. The power supply is a 350W device.

Could I measure the current when the machine is on and not turning in DC?