Stepper harmonic dampers

[S_J_H]

With the recent talk of cnc bench mills I thought I would share a major problem I had with my x3 and xylotex drive. When I first built this thing it had some of the worst mid band resonance my inexperienced self had ever seen. The mill was almost unusable because of it. I had been into cars and racing prior to this hobby so I knew of many different harmonic damper designs. I decided to pattern the dampers after a TCI "rattler".
The results were amazing to me. At first it was pointed out to me that it may only be a flywheel effect. I bought into that, but then I did more testing by locking the rollers from any movement. The result was they were then 100% ineffective and the resonance returned. I also tried damping the rollers with viscous fluid. Again, it totally made them ineffective. In use they make a little rattling noise as they soak up the vibrations.

.
Here are some pics of the dampers and then links to a before and after video showing the results of the dampers installed on the motors.



The video clips are in flash player format.
The difference is astounding and so is the mills performance with them installed.

before- http://s109.photobucket.com/albums/n...=damperoff.flv
After- http://s109.photobucket.com/albums/n...t=damperon.flv
Perhaps this style of damper might also serve a purpose for some other machine, with a resonance problem, or heck even a model engine. It's fairly easy to make and scale as needed.
Steve

[quadrod]

that is just too cool

[J.Ramsey]

The TCI Rattler damper is a very good design, I've run many Late models,IMCA modified's motors with these. Never given any thought for that usage on machinery and such.Excellent idea ,I bet it would help the problems I have with some of my equipment that are single phase.

[oldtiffie]

Peace and harmony

....................................
...................................

Perhaps this style of damper might also serve a purpose for some other machine, with a resonance problem, or heck even a model engine. It's fairly easy to make and scale as needed.
Steve

Thanks Steve for the "heads up" on the sources and "cures" for vibration.

"Steppers" by their nature are "stop-go" devices that are a source of vibration. The drive belt is a fairly good dampener too providing it doesn't "flap" and add to the problem.

Its not an uncommon problem and as you say, cars etc. are one good example - that's what the harmonic balancer and flywheel are there for. Older large-mass cast iron engines acted as vibration dampeners to some extent. That they were long-stroking and slower revving helped a big bit too.

Engine mounts also act as vibration dampers and isolators as well as I've been told.

"Squealing" (and "chatter") etc. on drum - and to some extent disk - brakes are similar examples when being machined on a "brake lathe" where quite often the solution is to put a larger rubber band around the drum to absorb the "noise". Its surprising how well it works.

Once a machine gets to the frequency at which it resonates it can really do damage. Every part has a critical harmonic frequency and the net result is the action of all of them in combination. It is quite possible for effects and signs of the resonance problem of smaller parts to be "masked" by the "damping effects" of larger parts.

It is quite possible that the effect can be transmitted and show up somewhere else (those annoying noises in cars that never are where the sound is and show up at different speeds). Thin sheet material can act as a diaphragm and resonate. Solution is a often to put a diagonal crease in them (like some tool-boxes fixed to trucks). Quite often a spinning body that is out of dynamic balance can be a problem when spinning at high speed (balancing of car wheels and tyres, propeller shafts, universals etc). Grinder wheel and rotor assembles are good cases in point.

I would think that part of the noise problem on the mill CNC drive might be the light aluminium and cover discs vibrating as diaphragms - caused by the CNC motor. "Curing" the disk/diaphragm "problem" might quieten it down but may not eliminate the problem if it is in the motor rotor. The mass of the rotor acts as an absorber but the disc starts to oscillate and when that frequency gets into the audible zone it can drive you nuts. This is not unlike the vibration in a motor engine that is largely "masked" by the flywheel and the damper - doesn't necessarily cure it - just so that you can't hear or feel it.

Similar results occur with "chatter" in a lathe where there might be a thin disc in the chuck or even if the tool - especially the boring bar acts as a "reed" and vibrates. Hence slowing the job down or in some cases speeding it up can suppress but not necessarily "solve" the underlying "problem". Quite often lathe vibration can be dampened just by putting your hand on it - and you can feel it. This is quite evident when "parting off" or machining the end of a fairly thin but inadequately supported long cylinder.

If a (generally) rotating body is "spun up" slowly to high speed this effect will become evident at several points which co-incide with the harmonics of the critical frequency. Turbines and car turbo rotors and fly-wheels come to mind. Hence the need for quick "speed ups" and "slow downs".

"Resilient mounts" are often applied to machine mountings to dampen vibration to or from machines - same goes for engine mounts which operate on the same principle - they are "rubber/y" and very absorbent of vibration. Another example is the rubber mounts on bridges and especially so in building in earth-quake-prone zones. Applies to to some of the better seats in cars, trucks and aircraft.

[Bruce Griffing]

What kind of drive are you using for the stepper? That can have much more impact than anything else on resonance. Specifically - are you using full, half or partial stepping?

[Fasttrack]

WOW never would have thought about that... cool

[S_J_H]

Bruce Griffing, I am using 1/8th step mode. I have 2 xylotex drives. When using this drive with 269oz motors on my x2 mini mill I never had an issue. I'm using 425oz motors on my x3 though.

oldtiffie, great info!
I am not quite sure what you are referring to when you say light aluminum or disks may be the problem on the mill?
The only disk I have is part of the damper which are mounted to the steppers rear shaft and that is what cures the resonation.

Well it was not the noise that was my problem although that noise was extremely annoying. The problem was I simply could not run G-code with the steppers resonating that bad. It would hit that resonant band and then wham, the motors would at the very minimum lose massive amounts of steps and the part instantly ruined as the table wandered off with a mind of it's own. I observed the motors even flicker in reverse when it got real bad.
This made some very quick reaching for the E-stop!


Another interesting thing to note, the Y Axis is belt drive and the X axis is direct with an Oldham coupler. Both use the exact same bearing arrangement and dual preloaded ballnuts.
The belt driven Y axis had worse resonation than the direct driven X axis. I was always puzzled by this as I would think the belt should help a little.
If I really tightened down the gib on the X axis it would quench the resonation a little. But the gib had to be so tight that it had a major impact on backlash with all the "stiction".
Tightening the gib on the belt driven Y axis had no effect at all. It seems the belt isolated that effect. This pretty much proved to me the problem was in the motors.
I could greatly minimize the resonation in the motors by turning the xylotex drive way down which further proved it to me. But then the motors had no torque at all and the system useless.

There was not even any trial and error with the dampers. I sort of guessed at the size it should be and made one. I installed it and did not expect much to happen. I was so amazed when I tried it I had to go grab the wife and kids to come have a look,lol.

With the dampers on I have never once had the motors resonate. I can run this thing all day long now without a hitch.
Steve

[oldtiffie]

Mea culpa

oldtiffie, great info!
I am not quite sure what you are referring to when you say light aluminum or disks may be the problem on the mill?
The only disk I have is part of the damper which are mounted to the steppers rear shaft and that is what cures the resonation.


With the dampers on I have never once had the motors resonate. I can run this thing all day long now without a hitch.
Steve

Steve, I am hugely impressed.

That mill of yours is really "something else" - both the original machine and the way you have modified/upgraded it. You have every right to be proud of it and you have considerable "bragging rights" there if you ever decide to exercise them.

I apologise for the error re. the disk - my fault as it is all too evident now you have pointed it about and then after I re-visited the pics.

What about trying slit insulation (as around hot/cold pipes) wrapped around the step motors? Just fasten it temporarily at first and if it works use some spiral screw (hose type) clamps and tension them to suit. Might work - might not.

What you did with your instinctive "gut feeling" in the design and manufacture of those resonation/pulsing-absorbers was a classic in the true tradition of the original old timers who tried-modified-tried etc. until it was right. And they did most of it through having a "feel" for the job as you seem to have - its a gift. And like you they not only had a pride in their skill but enjoyed it and shared it with others (ie Apprentices).

[lazlo]

That's really slick Steve!

I had to do a web search to see how the TCI Rattler worked. How loose are the slugs in their holes? Is the whole assembly made of steel?

[Evan]

The use of dampers on steppers has been around for at least 25 years.



This is a stepper with a viscous damper that was used in a HP pen plotter made in the early 80's. I find the damper you made very interesting. It's a lot easier than trying to seal and deal with fluids.

[John Stevenson]

Steve, Good work around.
The resonance problem is more a driver problem than anything else.
It's been known about for years as Evan has pointed out but drivers then were unipolar and very limited.

When they went to bipolar it opened up new design methods and it can now be designed out.

Unfortunately the Xylotex drive are built to a price and don't have this feature. More expensive drives often do. It's a trade off.

Still a good work round to get by the problem though.

.

[oldtiffie]

The CNC has gone bonkers

Steve, Good work around.
The resonance problem is more a driver problem than anything else.
It's been known about for years as Evan has pointed out but drivers then were unipolar and very limited.

When they went to bipolar it opened up new design methods and it can now be designed out.

Unfortunately the Xylotex drive are built to a price and don't have this feature. More expensive drives often do. It's a trade off.

Still a good work round to get by the problem though.

.

John,
you mean that if the new CNC has problems it might have bi-polar disorder.

I suspect my computer has a similar problem.

Spooky.

[S_J_H]

oldtiffie, thanks for your kind words. But I am a hack at this stuff!

Lazlo, the roller slugs have about .045" clearance in the pockets. I just sort of guessed. I don't think the specs on these things are to critical or I was lucky when making it. Body of the damper is 6061 and the slugs just some saw cut mild steel round.

Evan, I considered viscous, rubber o-ring type and this one. I did a little research on this type of damper. Seems it has been around a very long time. I think it is used in aircraft engines a lot. Or was.

John,
I had read what you are saying about the drive and midband resonance. I was at my wits end with the thing and if the dampers had not worked I was going to try Gecko's or put hand wheels back on the mill.

Steve

[Swarf&Sparks]

Hmmm, Evan, how would that go filled with some of that mercury you have lying around?

[Evan]

I think that will stay where it is, in safe storage in one of my sheds. I think that lead pellets might be a workable alternative though.