Ball screw drive vibrations 1

[Hina Shah]

I am trying to use a stepper motor + ball screw drive to move my 'holder' (refer image) precisely by a few microns. The ball screw is C3 accuracy and has a diameter of 8 mm and pitch of 1 mm. There is a linear guide and block behind the ball screw. All parts are machined in Aluminium 6061-T6. We want the holder to move up and down as instructed by the controller with the least amount of vibration. In the current setup, the holder is taking approx. 20-30 mSec to stabilise. We want the holder to stabilise with 2-3 mSec of stopping the stepper motor. Any suggestions, how I can improve the design?

Extra information -
1. The stepper motor is driven in 1/8 microstepping mode by a DRV8825 using a square function (timeperiod of 1600 mSec). We give a 10 mSec delay for the motor to complete motion after each step.
2. The holder will carry a sample that has to be imaged with a camera, lens setup.

 

[drawoh]

Hina Shah,

Have you analyzed the rigidity of your structure?

Historically, manuals for stepper motors recommended sizing them to the load inertia. You do not want to be less than 10% of the load inertia. 50% is better. Given the precision of your movement, your load is a mass, and your stepper motor is a springy thing.

The most important equation in vibrations is [tt]f = 1/2[π][×]sqrt(k/m)[/tt]

--
JHG

 

[IRstuff]

That equation also applies to your stage, irrespective of the screw drive. What are your requirements for stability?

Have you made actual measurements to determine how and what part of your structure is actually ringing? Screw drives driving by steppers tend also to have excited and excitable resonances.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[handleman]

Perhaps a servo would work better.
If you're really trying to move "a few microns" at a time, you might benefit from a gear reducer as well. It would give you more steps to play with your accel/decel.

 

[drawoh]

That equation also applies to your stage, irrespective of the screw drive. What are your requirements for stability?

That equation applies to everything. Given the need for micron stability, everything in that structure is springy.

--
JHG

 

[Strong]

I assume that the vibration problem is in the vertical direction, and the Holder has the highest vibration farthest from the drive. Stepper motors produce transient torque that translates into axial linear motion of the drive and Holder vertical vibrations. There are several vibration control schemes that could be considered such as a torsional damper on motor shaft or changing the natural frequency of Holder or main structure. I suggest constrained layer damping applied to the Holder to increase damping in the bending direction.

As suggested by others, it would be useful to measure vibrations (waveform and FFT) during operation and by impact test with machine off. There may be a simple and low cost solution to an apparently complex problem.

Walt

 

[EdStainless]

So you built this from Al, a material with very low modulus and therefore low stiffness.
You will likely have to take a series of steps.
1. A gear reducer so that you can accelerate/decelerate the stage more precisely.
2. Redesign to improve stiffness via geometry, and perhaps also build from higher modulus material (steel).
3. Tune the mass of the system, which usually means adding weight near the drive, which might require a larger drive system.
But these will all make you system more stable.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[drawoh]

EdStainless,

Aluminium is one third the stiffness of steel and one third the mass. Massive chunks of aluminium can have very high vibration modes.

--
JHG

 

[Tmoose]

https://media3.giphy.com/media/l41YnJwmpR3H7hiRq/200w.webp

 

[EdStainless]

If the parts were much deeper sections in Al then it might work, but usually there are limits on the dimensions and going to a stiffer material is often much easier than creating clearance.
Al is also a fraction of the strength of steel.
I see assemblies built in PH stainless that are stronger, stiffer, and the same weight than you could do in Al without some fancy fabrication tricks.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[Hina Shah]

Thank you for the suggestions. I will look up the rigidity of structure and get some machining done in Steel, instead of aluminium.
Is there any good reading material for improving stiffness by geometry?

 

[3DDave]

While steel can be stronger, I have a tough time understanding how it can be used in structures that are both lighter and stiffer. The tensile modulus to density ratio is nearly identical between steel and aluminum. For the same volume of material steel is better, but in tension it's equal and in bending, steel can be much worse. This is the reason aircraft use aluminum for primary structure - the stiffness is much better than steel. The exception is in tube-frame aircraft where welding skills for steel are lower than for similar work in aluminum. Look at the chart of specific modulus -

https://en.wikipedia.org/wiki/Specific_modulus

 

[EdStainless]

3DD, and why not make the steel parts even thinner because they are significantly stronger?
Especially if you can use the thinner parts to move material further from the centroid to gain even more stiffness.
Now you have structure moth stiffer and lighter.
I see this done with PH Stainless often.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[Hina Shah]

I am using a NEMA 17 stepper motor. I tried the damper used commonly in 3D printers. See pictured attached. However, not much difference in the vibration is seen on the holder. It still takes 30mSec to settle down.

 

[IRstuff]

What axis is the vibration on the holder? Do you have actual measurements?

Your holder does not appear to be a rock of Gibraltar; if anything it's probably wagging like a happy dog's tail every time you move or stop the stage.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[drawoh]

Hina Shah,

Damping works on velocity. At very high frequencies, you may have high acceleration, but very low displacement and velocity. I like IRstuff's term "Rock of Gibraltar". This is what you need.

--
JHG

 

[3DDave]

That stage looks like a diving board - no wonder it vibrates. Steppers put an impulse load into the system, so that doesn't help. The energy of that impulse needs to be absorbed somewhere.

 

[drawoh]

Hina Shah,

I just took a look at your photo. That is not a damper. That is a vibration isolator or shock mount, admittedly, with damping characteristics. Is it isolating your structure from a source of vibration, or is it providing degrees of freedom for something to shake?

Anti-vibration design is not simple. You need to understand your source of vibration. You need to place anti-vibration mounts between the vibration source, and your system. This is not possible with your stepper motor at all! Ideally, the natural frequency of your system and anti-vibration mount should be way below the lowest natural frequency of your system. It should be acceptable for your system to move, because that is what it will do if it is sitting on anti-vibration mounts. Don't mindlessly stick anti-vibration mounts on things.

--
JHG

 

[Tmoose]

Attached is the OP's image of the device, marked up with my estimations of the areas that concern me, assuming the device is secured to something substantial by the multiple bolt holes in the lower most component.

The portions of the components that I suspect may be relatively flexible and almost hinge-like are circled in tan.

The mechanical connection between the the horizontal arm and the plate that runs along the guide rail is circled in red, as it concerns me as well. I envision it as two "flat" surfaces butted together and are secured by a few bolts. If either mating surface is even slightly (0.0005") convex, and the bolts are on a common horizontal centerline, the bending stiffness can be quite low, relying almost completely on the bending stiffness of the bolts.
Far better is to specify a slightly concave surface, or to machine a shallow relief in the "middle" of the contact face, in order to create a narrow band of contact along the upper and lower limits of the contact face.