Anti-Chatter HSM Idea...

[Colin Gilchrist]

Hey Everyone,

 

I've got an idea that I wanted to run past everyone to get your opinion. I recently contacted some guys I know at MIT to see if I can get their help to run an experiment. The trend in the marketplace for HSM machining involves producing cutting tools with variable helix and cutting tooth angles. The idea being that a non-equal spacing in the cutting flutes, coupled with a variable helix angle, will reduce the harmonic vibrations created from equally spaced teeth striking the part. The effect is that we can run higher speeds and feeds, without causing a harmonic vibration in the part and fixture.

 

My idea would be to use an electrical transducer to create a "mechanical" wave in the part and fixture, that is 180 degrees out of phase from the wave being produced by the cutter striking the work piece. Instead of using cutters with "unequal" flute spacing, we could go back to using "standard" tools that have constant helix angles, and the same angle between the teeth.

 

There are currently systems produced by several machine tool manufactures that use acoustic sensors to automatically reduce spindle speeds and feed rates when vibration is sensed inside the machine. Instead of slowing down the machining to eliminate the vibration (which causes the chatter), I think it would be possible to "tune it out", by using an anti-phase vibration frequency wave with the same magnitude, but opposite phase timing. This would lead to the vibration being "cancelled out", in a similar manner to Active Noise Control headphones.

 

It might be necessary to include several different transducers, in case there are multiple harmonic waves that need to be cancelled, or we might find that a specific transducer only covers a certain frequency range. I would imagine the frequencies involved would be in the 50 Hz - 50 KHz range.

 

This idea came to me after sitting for hours in my programming office listening to the screeching of the machines in our shop...

 

So, what would you think about a machine that had "anti-chatter" capabilities? I think we can borrow some of the frequency analysis techniques from "vibration tap testing", and apply that concept in reverse to attenuate the frequency waves produced by a tool that is "in the cut". Done right, I think it has the potential to even be something that would be available as a "field retrofit kit".

 

If they can make it work with headphones, I can't see any reason it wouldn't also work for machine tools...

 

 

[Matthew Hajicek - Singularity]

Colin, that exact thought hit me a few months ago so I looked into it.  I did find some patents along those lines, but not all mechanisms are covered obviously.  My thought was to use a pieso element that could function as both sensor and actuator, and it would be really useful in boring bars.

 

As a matter of fact I have a vague recollection of seeing something like it coming to market recently, possibly featured in MMS, but I can't find it right now.

[Matthew Hajicek - Singularity]

http://sem-proceedings.com/27i/sem.org-IMAC-XXVII-Conf-s50p001-Improving-Turning-Process-Using-Piezoelectric-Actuators.pdf

[YoDoug®]

Just a few thoughts on this. 

 

1. Not all vibration sensing options from MTB's only slow down the speed. Some will speed up slightly if applicable. Okuma's machine Navi M-i will adjust the rpm as needed to reduce the vibration sensed in the spindle. It does not change the feedrate so you do not see a cycle time change. 

 

2. Unless the machine is a fixed table adding a vibration into the table would take ever constantly changing servo tuning parameters. MTB's put a lot of effort into tuning out servo vibration because of the ill effects. 

[kccadcam]

Have you looked at Blueswarf?

You tune the tool instead of the machine.

 

http://blueswarf.com/

 

Modern Machine Shop Webinar - Machining Dynamics:

Using Science to Optimize High Performance Milling in Your Shop

http://blueswarf.com/train/webinar/player.html

 

 

http://player.vimeo.com/video/91954755

 

 

[Zoober]

We looked into the Blueswarf tap testing a couple years ago.

I could never justify the down time on the machine to tap test every tool.

We are getting a 3PM system installed on a swiss machine tomorrow. It has a 30 day "love it or leave it" policy.

So if we still have it in 30 days, it works.

One of the large dental implant companies has them on all their swiss machines. Ive seen it in action, and it works great for tool life and breakage detection on tiny tools and parts.

Their web page sucks, but here is a vid of the machine Boss system we are getting.

http://www.youtube.com/watch?v=mnWHfG0i8F0

[zero_divide]

Saw a video somewhere.

They used a spindle with constantly varying RPM to achieve that.

I think it peformed worse than the variable helix / staggered flute endmills.

 

IMO it just can not change RPM that fast to simulate the staggered flutes.

[Teh Bear without Brains kinda idiot]

Not to offend anyone but I am   very sceptic at least .

I prefer the simplest ,dumbest solutions

 

This one looks too clever to be practical

BR

[MIL-TFP-41]

Its a cool idea. Just thinking tho...It may be tough to implement. If one is dealing with tool vibration only, you have frequency and amplitude to adjust for. Frequency I can see adjusting for...amplitude would be a bit tougher.

Then how about those far to common setups where part vibration is the issue? Would one do the same thing to the tool?

[Teh Bear without Brains kinda idiot]

Metal part is not ideal body.

It will have some specific zones due to interference and difraction .,different thickness and so on.

I wanna see how you`ll manage in real time implement  contra- faze wave  that will deal with all this .

AND for what exactly?I never met a part that I was not able to get rid of extra vibrations .

And trust me I milled lot of B@stard parts like thin pipes ,camerton looking like parts .thin blades and so on .

You need to be sorta creative ,man ,that`s all .

I know lots of tricks that solve this task in simplest ways you would not believe .

My motto -be as simple as possible.

BR

[Teh Bear without Brains kinda idiot]

This idea came to me after sitting for hours in my programming office listening to the screeching of the machines in our shop...

~~~~~~~~~~~~~~~~~~~~~~~~

You have setupists or  who they are ????

It is a job of setup man to find good feed and speed.

Being machinist-programmer for more than 20 years i can not believe this is impossible.

Your guys are lazy S>O>Bs  not giving a damn for a part or tool.

you need to start from the floor and after that think about rocket science

[Colin Gilchrist]

Alex,

 

Please understand I hold you in the highest respect.

 

I do not want my operators constantly fiddling with the overrides on my machines. We are not a "job shop", we do highly engineered parts and processes. One of our goals is to minimized operator intervention. Plus, we have the desire to move some of our processes to "lights out" manufacturing, so no operator would be needed to monitor the machine while it is cutting. Of course, to get to that point we need to think about all aspects of the job, not just writing tool paths.

 

Many of the parts we cut are thin-wall aerospace components. We can always turn down the speeds and feeds. What I'm thinking about though is a system that will help compensate for the vibrations that I'm causing in the part, by the inherent nature of the tool striking the material.

 

I completely get that a good operator/setup guy can compensate for a lot. That's not what I'm after. I'd like to have a better understanding of the fundamental forces at work, and come up with a system to tune out and negate the "bad" oscillations.

 

It may end up not being feasible for a number of reasons; cost and complexity being at the forefront. I'm with you; I like simple as possible. If a device could be mounted to the machine table to produce "anti-phase" vibrations to attenuate the chatter, then why not? It may seem like a waste of energy, but I think there is a real possibility of being able to simplify the rest of the processes being used. I think having the ability to "tune out" the excess vibrations would lead to a lot of other benefits. Not only could speeds and feeds be increased, but fixtures could be simplified since rigidity would not be as necessary or important.

[Colin Gilchrist]

I like some of the ideas that others have presented. I am not a fan of trying to "tune" the tool holder. I think this adds to the complexity of the holder, which adds to the cost of the entire system. The "anti-vibration" device would have to be incorporated into each holder, thereby increasing the cost substantially. I'd rather see a "table or part mounted" system which would then be able to tune for each tool being run on the part material.

 

Matt, thanks for the link to Blue Swarf. I've seen their stuff before. They do a good job of finding the "sweet spot" for testing an individual tool, against a specific setup/part combination. I'm thinking of incorporating some of their ideas into my tests.

 

Robert, I'd be curious to get your feedback on how their systems work for you. They have the "monitoring" part of the equation down. Now, couple their sensors to a computer, analyze the vibrations, produce the same waveform, but 180 degrees out of phase, and send it to a transducer attached to the part/fixture. That's my thought on it anyway.

 

It's probably not realistic, but I'd rather have fun trying to make something new work, then being chained to the dogma of "that's just how it's done".

[YoDoug®]

Colin, Have you seen a first hand demo of Okuma Navi M-i. It uses vibration sensors to adjust the spindle speed by slight amounts to counter act the chatter from the harmonics of the flutes striking the material. It is a very robust solution with many parameters that can be set to control the outcome. 

[Joe957]

Kind of related, but I was thinking about this a few weeks ago when we had 2 machines both running similar parts across the shop from each other, and they were getting into some kind of harmonic "symphony" with one another and produced a really annoying pulsating vibration throughout the shop. I was thinking microphone and loudspeaker with a short delay to throw it off....   Changing the spindle speed ever so slightly would eliminate it though.

 

We had blueswarf in here a few years ago - got dashboards for 10 of our most common tools.  Never really got the results we were looking for, although it was a really interesting learning experience.

[Teh Bear without Brains kinda idiot]

Colin ,sometimes 5-10 % of speed change  will turn harmonics out .

As for solutions there are lot of them .

Have you ever tried Wood`s metal or Polyurethane to add rigidity?

 

 

BR

[Matthew Hajicek - Singularity]

I found what I was thinking of before.  It's an ultrasonic toolholder; not for directly countering vibrations but for adding its own:

 

http://www.mmsonline.com/articles/ultrasonic-assisted-machining-with-standard-machines-and-tools

 

http://www.acoustechsystems.com/

[Colin Gilchrist]

Colin, Have you seen a first hand demo of Okuma Navi M-i. It uses vibration sensors to adjust the spindle speed by slight amounts to counter act the chatter from the harmonics of the flutes striking the material. It is a very robust solution with many parameters that can be set to control the outcome. 

 

Hi Doug,

 

Yes, I'm familiar with the Navi M-i. It is a good system, but only available on the Okuma brand. We have a couple Okumas here in our shop, but we are by no means "Okuma exclusive". It does do a good job at reducing chatter from what I've seen of it. I believe that one of our machines has that system (or a system like it, I'll have to check).

[Teh Bear without Brains kinda idiot]

In most cases not rigid part can be milled rather rigid if you mill this wall first that surface later leave

techhological wall for support or place jacks here and there .

This is called technology and this is an art.

Instead of using HSM on whole part as it   is    you can get the job done smooth and without mad shrieks.

Yes it is additional work but it worth it.

I am rather humble person who   am I to argue I only want to say that being old school technology    engineer

I believe that sometimes (most of times)  teh   job can be done with simple things and solutions .

The simplest the best .

You need to reprogram your brain to  do   this.

Not magical it will solve all my problems something but good old xxxxing stuff called technology

BR

[YoDoug®]

Below is a slide from an Okuma presentation on Navi M-i. It is not the greatest image but it shows where the machine adjusts spindle speed from 3500 rpm to 3712 rpm to eliminate the chatter shown in the graph. This is done real time and will update as cutting conditions change, tool wear, part geometry changes, etc. You can set parameters and boundaries on how much adjustment can be made and save that data to the tool info. Now I will say it is not a magic bullet. It can make small adjustments to make small to moderate gains in performance. It is still subject to the laws of physics though and it will not make a poor setup or process into a good one.

[MLS]

Not super well versed in this but have thought about the same thing some.

Question...

Where do the harmonics reside?  I understand that they are generated by the contact of material and cutter, but in my mind they transferred and/or magnified into the spindle based on the tool, holder, spindle design, bearings, etc.

That being said, would changing the harmonics in the material based on the parameters recorded at the spindle be the same as at contact point?

Also, how much would be lost at contact point... are there lost or diffused harmonics within the transfer?

Dunno... just questions that popped into my head.
 

[taperlength]

This is a off at a tangent to the discussion but I found it interesting.

http://www.pro-micron.de/en/products/sensory-tool-holder-spike/

[mkd]

i think it is a good idea for some years now.

talking to MIT seems like the best place to start, with all the Big Brains there. pretty cool you even know anyone there

 

would this discussion be centered around the natural frequency of a CNC machine?

the simple solution used up to this pont (okuma Navi notwithstanding) is to add more mass and rigidity to the system thus raising the natural frequency (or do i have that backwards?).

at some point you reach diminishing return.

 one could build a CNC very light weight where the structure can handle the stress but deflection and vibration would be a no go.

should be a happy medium between massive structure and active vibration control.

 

i always thought weighted voice coils positioned in various planes would do the trick.

that would be a fun algorithm to develop, as to not induce vibration accidentally.

[Müřlıń®]

Harmonics is like a tuning fork....what happens if you put your finger on a tuning fork?

 

Had a kid today machining on a 25 ft jig.....tthin ribs chattered bad so he was supporting them by clamping a stiffer on the side with c clamps.

 

But it was still bad......until I told him to tape up the handles of the C clamps to keep them from creating resonance.....