[kostas1]

Hi guys, I bought from Chai 2 ballscrews, one 1900mm long and the other 1200mm long. Right out of the box, I noticed that the 1200mm ballscrew is bent about 2-2.5mm and it causes vibrations to motor and the sound when turning is not so smooth especially when the nut approaches the BK/BF supporting units that hold the ballscrew.

I know that this is bad news and probably will need to replace it with a new one, but could this problem affect the axis accuracy? During some tests with the digital gauge, I notice there is a 0.05mm loss when the axis travles from one end to the other and this loss gets bigger or smaller.

Thanks in advance.

[RCaffin]

Bent ball screw?
Ask for refund.

> 0.05mm loss when the axis travles from one end to the other
That may be within spec for the ball screw. None of them are perfect; they all have tolerances. But you can correct in Mach by mapping.

Cheers

[rowbare]

It isn't unusual for a screw to be slightly bent after coming half way across the globe. It isn't out that much and you should be able to straighten it easily enough without any special tools.. There are plenty of instructions/videos online dealing with straightening screws.

Also check your bearing blocks. People have reported that the pockets are sometimes too deep and that the bearing can float.

bob

[ger21]

and the sound when turning is not so smooth especially when the nut approaches the BK/BF supporting units that hold the ballscrew.

This sounds like your bearing mounts and nut are out of alignment. With a bent screw, it should get smoother as it approaches the bearing blocks, as it's moving away from the bend and toward the blocks, which should be in perfect alignment.

[RCaffin]

> you should be able to straighten it easily enough without any special tools.
True - if you have a press able to do controlled movements. Been there, done that - except that I had a very skilled tool maker do it for me.
But it is still 2nd best, and when you buy a ball screw you are entitled to a straight one.

Cheers

[kostas1]

The strange thing is that while testing the Z axis travelling from one end to another with some MDI commands, I put the gauge in one end and I notice that every time the z axis goes and returns, gauge's indication value increases. This means that there is something wrong in the Y axis setup. Could this be the Y axis bent ballscrew?

While testing, in the first 3 attempts gauge indication can be 0.00mm and then it starts increasing to 0.03, then to 0.05, 0.09, 0,11mm etc. I zero the gauge and after several tries, I get the increasing indications again. Maybe there is a loose thing there, not sure or maybe the bent ballscrew?

[RCaffin]

the first 3 attempts gauge indication can be 0.00mm and then it starts increasing to 0.03, then to 0.05, 0.09, 0,11mm etc

If you do this for many, many cycles you might find that there is a linear shift in the zero with the number of cycles. The results you quote suggest this.

If that is happening, then you may be losing a step pulse at the turn-around. A common cause for this is having the Step line 'upside down' or inverted. What this means is that the Dir line changes state while the Step line is 'active', which is going to upset the driver. Try inverting the Active state for the Step line - if it is Active HI change it to Active LO, & v/v.

Cheers


Cheers

[kostas1]

If you do this for many, many cycles you might find that there is a linear shift in the zero with the number of cycles. The results you quote suggest this.

If that is happening, then you may be losing a step pulse at the turn-around. A common cause for this is having the Step line 'upside down' or inverted. What this means is that the Dir line changes state while the Step line is 'active', which is going to upset the driver. Try inverting the Active state for the Step line - if it is Active HI change it to Active LO, & v/v.

If it loses one step in turn around, then this is 0.05mm, which is close to my measurements. Soyou say that this may confuse the driver. Is this some error in wires? These are all shielded but all in one cable chain, so is this an issue?

Also, could it help if I run the drivertest.exe? Maybe my pc cannot send pulses correctly.

[RCaffin]

If it loses one step in turn around, then this is 0.05mm, which is close to my measurements. So you say that this may confuse the driver.

OK, The solution is extremely simple. It's a software config change in Mach3.

Mach3, Config, Ports&Pins, Motor outputs, Z axis, Step Low Active - click on this box to change the cross to a tick (or vice versa), then click on OK.
Then you MUST save the settings! Config, Save Settings.

Now try that and see IF it solves the zero drift. Let us know.

Cheers

[kostas1]

OK, The solution is extremely simple. It's a software config change in Mach3.

Mach3, Config, Ports&Pins, Motor outputs, Z axis, Step Low Active - click on this box to change the cross to a tick (or vice versa), then click on OK.
Then you MUST save the settings! Config, Save Settings.

Now try that and see IF it solves the zero drift. Let us know.

Cheers

According to this thread, all the step low active is enabled.
Is this really how they should be set?

Also the dir low active according to this link, should be set to enabled except from the Z axis.

I have them all disabled.

[ger21]

Step Active Low is specific to the drives that you are using. Different drives have different requirements.

Dir Active Low just changes the direction that the motor spins.

[kostas1]

Step Active Low is specific to the drives that you are using. Different drives have different requirements.

Dir Active Low just changes the direction that the motor spins.

Thanks Gerry, my drivers are G203V and bob is C11 from cnc4pc.Gecko states that "Stepping occurs on the positive edge of the STEP pulse", so I suppose that this setting is wrongly set to red X. I should try setting this to true. Maybe this is why I cannot get a single piece be accurate as I describe in this thread

I have many things to try this weekend in my machine. Unfortunately, I can operate my cnc only on weekends.
I hope to come back on Monday with some fixes on my issues.

[kostas1]

Sorry guys but I noticed something odd in Mach3 -> Config -> Motor Tuning -> Step pulse setting.
If I change this setting for one axis or for the spindle, then all the other axis are set to this same value.
Can't I set a diferrent value for each axis or spindle? I need the axis to be set to 1 or 2us and the spindle to 3-5 us.

[ger21]

There is only 1 pulse width for all axis + spindle. Gecko specifies a minimum of 2us, but they should work fine set at 5us.

[RCaffin]

> I have them all disabled.
If you have them all disabled, you won't get any movement. So that can't be right.

Cheers
Roger

[kostas1]

> I have them all disabled.
If you have them all disabled, you won't get any movement. So that can't be right.

Cheers
Roger

SOrry,I meant I have the dir low active set to red X

[RCaffin]

SOrry,I meant I have the dir low active set to red X

OK then, if you do not have active LO you must have Active HI.
Geckos usually clock the DIR signal at the +ve transition:
" Stepping occurs on the positive edge of the STEP pulse. The DIRECTION input must be true 200nS before and after this STEP pulse edge."

On the other hand, I believe that if you do a dynamic change of direction, Mach can issue both the Step pulse (going HI) and the new DIR signal at the exact same time. This could lead to your losing one STEP pulse with each DIR change. We have seen this one before now.

Try changing your Mach system to 'Exact Stop' in Config and see if you still get any creep.

Cheers
Roger

[kostas1]

OK then, if you do not have active LO you must have Active HI.
Geckos usually clock the DIR signal at the +ve transition:
" Stepping occurs on the positive edge of the STEP pulse. The DIRECTION input must be true 200nS before and after this STEP pulse edge."

On the other hand, I believe that if you do a dynamic change of direction, Mach can issue both the Step pulse (going HI) and the new DIR signal at the exact same time. This could lead to your losing one STEP pulse with each DIR change. We have seen this one before now.

Try changing your Mach system to 'Exact Stop' in Config and see if you still get any creep.

Cheers
Roger

Hi, I checked the step active low and dir active low settings in Port&Pins ->Motor Output tab and they are set to red 'X'
I tried to set step active low to green 'tick' for the Y axis and the axis could not move.

What is the wrong with this?

[RCaffin]

Hi kostas

Um - that is a bit strange, if it worked with a red cross there.
The Y Step pulse width may be not quite enough for proper operation in this mode. This is a known problem as well, and Mach has a setting for it.

Try this with the Y axis set to a green tick for Step Active Low:
Config, Motor Tuning, click on Y axis, look at 'Step Pulse 1-5 uS'. This is probably set to 0. Change it to 2 and then click on Save Settings.
Now click on X axis and make sure that Step Pulse is set to 2. If not, repeat the above.
Click on OK, then go back to Config and click on 'Save Settings'.
Now try.
Note that you MUST click on 'Save Settings in two places.

Cheers
Roger

[kostas1]

Hi kostas

Um - that is a bit strange, if it worked with a red cross there.
The Y Step pulse width may be not quite enough for proper operation in this mode. This is a known problem as well, and Mach has a setting for it.

Try this with the Y axis set to a green tick for Step Active Low:
Config, Motor Tuning, click on Y axis, look at 'Step Pulse 1-5 uS'. This is probably set to 0. Change it to 2 and then click on Save Settings.
Now click on X axis and make sure that Step Pulse is set to 2. If not, repeat the above.
Click on OK, then go back to Config and click on 'Save Settings'.
Now try.
Note that you MUST click on 'Save Settings in two places.

Cheers
Roger

Hi Roger,
in fact the us step pulse all this time was from beginning set to 5us and the dir pulse us to 1us.
Will try your suggestion this weekend though. Maybe I need to send you my config.xml.