Jerking of Robot Arm

May be a lot of mass on your arm,... you may have to have

1. a longer ramp up on the motion controls.
2. reduce the mass of the arm
3. have heavier driving components that drive the arm (chain, ball screw, rack/pinion, ??)
4. lubrication ?

You have a stiction problem....

Or backlash.

Someone has mistakenly used the plans for a seismic recorder, or there is a feedback gain problem.

You must program in more steps. I use 1600 steps per rotation and it smooths out the action.

I am giving you a GA for such an in-depth analysis of the problem, and asking questions that really need answering before a GA can be given.

the inside distance between bearing is 10 mm. And I agree that this distance is less to support the load. Another doubt is, will the preloading of bearings solve my problem.

I've used lamina coupling to couple the shaft & the gearbox.The manufacturer said that it could withstand misalignment of 0.2 mm.

Now in the new design I am keeping the distance between bearing as 100 mm, using angular contact & deep groove bearings.

If there is anything else to be considered,please suggest

No mention has been made of my third point, the reflected inertia of the load (that is the moment of inertia of the load as seen at the motor shaft). The ratio of the load inertia to the motor inertia is critical the smooth motion and accurate positioning of the system. Ideally, the ratio should be about 10:1.

From your drawing, it appears that the mass of your load is a considerable distance from the axis of rotation, leading to a high moment of inertia. Your use of a planetary gear reduction helps this situation, reducing the reflected inertia by the square of the gear ratio. However, unless you have made the necessary calculations or have a good estimate, your servomotor may be undersized for this application.

There are a number of online calculators provided by servomotor manufacturers that you can use to select the proper sized motor for your application.

Sudden full-speed starts and stops can cause jerky motion. Take a closer look at the motor specifications. There may be some information on how to pulse the motor at low and intermediate speeds, rather than run it in full-speed on/off mode. I think you need to ramp up the speed of the motor when you start an action and ramp down the speed when you finish an action.

Also, make sure the weight/torque loads are balanced. Unbalanced loads can cause jerky motion.

Slow everything down.

You have a pendulum:

Massive Bob?

Massless Rod.

Or Assless Rod.

That's not just a Bob, it's a Billy-Bob! "That there boy's too big for just ONE name, ah reckin'."

It sounds as if you need an improved PID loop and/or better feed back and/or maybe a lower acceleration and/or a modified ramp curve.

Could use a strain wave gear box (aka harmonic drive). No backlash and torque multiplication numbers that will make you go Hmmmmmmmm!

Have been using these in robotics for years now, nothing smoother.

Increase the number of steps, increase the gearing ratio to slow it down further, and remove all backlash.

If you can move it from side to side with your hand, thats simply no good.

It as to be so stiff as if it was made as a single piece of steel, like a stone statue.......like trying to move the arm in relationship to the body of the statue....NO MOVEMENT WHATSOEVER!!!

Anything less is useless.

Increase the number of steps

That is exactly what i stated in post #6. We were having trouble with laser triangulation readings, in the inside of the pipe, due to vibration between 200 steps per revolution,step to step. The laser head would still be vibrating due to the last step. When we re-programmed the rotation to 1600 steps per rev the readings stabilized.

The main vertical frame has little torsional (or fore/aft) rigidity. It's just two verticals with minimal cross bracing. A third vertical midway between the two but set back to form a triangular column and suitably braced to the other two would stabilize the top of the tower.

I see you have triangular bracing where the horizontal frame joins the vertical, but there is little designed in rigidity elsewhere.

Del

Sorry, i only speak two languages, English and bad English. (Bruce Willis in the Fifth Element)

You meanz thiz

Or Thiz

Peeewee

While I am very reluctant to criticize any member of this community, in this case I have to agree with you wholeheartedly. Not only do I find his idiotic comments to every discussion annoying, I feel that his intentional abuse of the English language is a gross insult to all our non-English speaking members who have made the effort to respond to our discussions using reasonably proper spelling and grammar.

So far, this individual has contributed nothing to this community!

I gave you a GA for that. I agree there are other on-line places for those antics

Reported!

try a couple low strength pneumatic dampeners on arm to carrier.

Consult a mechanical engineer. He knows how to do it.

(this is the 1 time reply as i missed most of my mech. class - i still got the books though . . .)

i see no cables** , counter weights nor other balancing/stabilizing elements**** (that could be omitted if your design is lightweight) - ? the precision** how the tool will be (auto/)calibrated before actual work task

**** should consider the deg.-s of freedom being reduced by object holder (not on your sketch) doing some simple vec3d + rotations