Vertical system "jerks" and "bangs"??

I have a vertical oriented system driven by
a directly coupled 1/2" pitch ball screw to a
servo motor. We are useing a Galil 2 axis card.

The system was tuned and moves up well, but
when going down, it will at higher move rates
(greater than 90" per minute) "jerk" or "bang".
Sometimes this caused the system to error out. It happens more the higher the velocity is set to. It happens while in motion and more predominantly when trying to decel when going down. It never happens when going up. It is also more predominant as the weight on the system is increased. We have some auxiliary brakes on the system and it will work fine when we partially turn the brakes on. The jerk appears to be caused by the motor trying to stop or even momentarily reverse direction.

Is there any way to tune this out, I've played with all the pid parameters and nothing seems to help. I've even tried some of the il, fa, and fv parameters but nothing seemed to have any effect.

Any ideas would be greatly appreciated!! Thank you.

It is obvious you have some inertia issues, direct drive and fairly high lead ball screws do not help, Unless you have a very large system, I rarely use anything over .25" or .2" pitch, especially with no reduction.
Coming up with some kind of counter weight or cushion system system should alleviate it. I don't think trying to tune it out is the answer.
Al.

Thanks Al, but what makes it bad the way it is, is it just an inherant problem in all
systems that do not allow them to handle this situation? Is it something to to with
response time?

I suspect that with the set up you have at your programmed accel/decel rates you are exceeding the recommended motor/load inertia match, which is normally considered to be kept to less than 5:1.
The up side in reducing this is that any increase in reduction ratio by ballscrew lead or other means, reduces the existing inertia ratio by the square of the reduction, so it does not take much.
Generally the highest demands are during accel/decel and not neccessarily due to cutting forces etc.
By downloading one of the many system design graphic software from any of the servo manufacturers and entering your machine data can give you an insight to what might be happening, ratio wise.
The Kollmorgen one is good from MotionVillage, and there are many others.
Al.

Jerking and banging noises makes me think of some kind of binding issues to do with the slides of the machine. This could mean the center of gravity of the head is too far off center of guide bearings that are too close together on the slideways. The amount of clearance that the bearings have could also be a concern: a little too much and they will lock easily due to the available tilt in the system with an offcentered load.

We deal with accelerating and declerating mechanical systems with cams for a livelihood - and in systems without brakes or dampers.

Dealing with inertia - the ability of a body in motor or at rest to stay in that state - is a fundamental problem with ANY electro-mechanical system.

In your case, when you have the system moving UP, only the motor can apply force - all motor has to do is overcome gravity and friction from the force IT (motor) creates.

However, when system moves DOWN, it has to overcome the intertia of body and friction but now GRAVITY is adding to the speed factor, not subtracting from it as occurs when lifting.

First thing I'd try is to REDUCE the speed and/or acceleration factor of the motor during LOWERING operation if possible. Since gravity is "helping" you don't need motor to urge things along as quick or fast. IN effect, all motor has to do is brake at the desired position.

YOu might also try to counterbalance the system. This way, the upward force and downward force needed to move the system are closer match (garage doors move much smoother and quiter if the opener doesn't have to "lift" the entire door when the wind-up spring is adjusted properly).

You'd definitely want to reduce acceleration factor in a system with an external counterbalanced system. More inertia to control.

You'd be amazed at the applicability of the F=MA equation when it comes to understanding and solving problems like this....