Stiffer means higher frequencies and higher speeds. Shorter should mean stiffer, but too short can be detrimental for other reasons. If you let any of the flute or washout (where the wheel grind continues past the flute length) to recede into the holder or collet you are creating a stress concentration on a smaller profile than the full round shank. It will take a lot less force to break the tool.
As Professor Scott Smith from UNC Charlotte likes to say, "Random processes produce random results". If you have been following our posts you will have learned that a milling tool is a flexible beam and that changing the length of that beam, even a small amount, will change its frequency and can impact the tool's performance. You will also have learned that rotating a spiral fluted endmill in a toolholder can change the balance of the assembly significantly. If you have something running really good, you want to keep it that way.
Most toolholders other than side-locks use friction to hold the endmill shanks. So you need help to get repeatable tool changes. We found these wrench-free aluminum shaft collars from Ruland Manufacturing. The aluminum won't damage the cutting edges and holds up to the heat of shrinkfits. You can also use steel set screw collars or drill stops.
Set the tool projection from the tool tip, not the back end. Overall length is not a tightly controlled spec in ISO or ANSI, so measuring from the back or using backup screws will not give you a repeatable tool stick-out. You can use a caliper depth rod or depth micrometer to set the projection, just subtract the collar thickness.
Also, use the slot in the collar to orient the tip of one tooth, like a gunsight. This will retain balance. Look through the slot and rotate the tool until the tip lines up with the slot. If you are using variable pitch endmills, identify tooth #1 and set to that each time. Put a witness mark on the end of the toolholder and use that as your target when orienting the new tool with the collar's slot. Balance the tool assembly first, place the collar on the tool while still in the holder and line up the slot to a tooth. Mark the toolholder at the slot with a center punch or drill point. For a collet chuck, torque the nut to the proper value, then mark the nut just as above.
You need to try new products to improve your processes, but the costs can be very high.
A shop manager in a magazine guest column wrote that he set aside two hours to a full shift for trial and error testing of a new cutting tool from a vendor. Let’s use the median of 5 hours. If your shop rate is $75 then the tool test costs you $450, with no guarantee of a successful outcome. Before you feel too sorry for yourself consider the costs to your vendor. According to a sales training agency, a field sales application engineer’s total costs are $85 per hour. Assuming that at least one endmill will be run to failure during the test (at their cost), a selling price of $100 per tool with average gross margins, they will have to sell you 25 endmills before they breakeven, ONLY IF their new tool is better.
A tap-test with an instant result Dashboard will take 5 minutes, costing you only $6.25 in spindle time and the vendor makes a profit on the first tool they sell you. No test tools are needed therefore not damaged in the process.
Now, someone has to pay for the tap-testing hardware and software but that can be amortized over several years and hundreds or thousands of tools and in the case of the vendor, hundreds of customers.
Look at the data above. Which tool would you pick as the most productive AND the most economical?
Here's the rest of the data. A lot emphasis is on Inches Per Minute. What is really most important is how quickly you are filling up those chip barrels. Its metal removal rate (MRR) expressed in cubic inches (or meters) per minute and how much it costs to get there.
Of course, finishing is a different animal.
Controlling chatter in milling is a significant MRO/MROP supply chain solution. By some measures, cutting tools are the largest product segment in MRO. Cutting tools represent 24.1% of industrial distribution sales with the next largest category reaching only 9.9%. Milling, which is the focus of this paper, is the largest cutting tool category at 40.2% of the market. Improving milling operations will therefore have a very large impact on the MRO spend and the overall efficiency of a manufacturing plant.
What is chatter?
Chatter is a self-excited vibration that can occur during machining and become a common limitation to productivity and part quality. The onset of chatter during machining is primarily a function of the variation in chip thickness that occurs due to vibration of the tool. The flexible tool engages the workpiece and, due to the cutting force, begins vibrating. This vibration is imprinted on the machined surface. In milling, the next tooth on the rotating cutter over- cuts this wavy surface produced by the previous tooth. This wavy surface varies the instantaneous chip thickness which, in turn, modulates the cutting force and the cutter vibration creating a feedback mechanism we hear as chatter. Efforts to avoid chatter, by slowing down the process, can lead to underutilization of the machining system by a factor of five or more. The variation in chip thickness per tooth described above leads to uneven tool wear and premature replacement or failure. Chatter in milling, and efforts to avoid it, is estimated to cost manufacturers over $100 Billion in lost productivity Worldwide.
With the advent of industrial vending machines and tool data management systems, manufacturers are getting near real time visibility into their cutting tool consumption, perhaps for the first time. What is discovered is that cutting tool use spikes dramatically, even if the part output remains relatively stable. Why is this? Cutting tool consumption is dynamic, subject to constant variation, rather than being static and controlled. An example of a static product is a cap screw. If a part requires four screws and the customer produces 1000 parts per month, the supplier will need to deliver 4000 screws each and every month. Easy to forecast, procure and deliver.
That same part may require 100 carbide endmills to machine those 1000 parts one month, need 200 the next and only 80 the month after that. Let’s assume that the customer maintains one month’s worth of inventory on hand. In order not to risk a stock-out and a stoppage of production, their stock level will be based on the worst case scenario, that is, the highest usage rate or 200 endmills. This means excessive inventory costs of as much as 60% in this example. Also, if there is a stock-out and the end user maverick buys endmills from a second supplier to fill the shortfall, the vending machine software does not know this and its optimum inventory level is never updated. Future stock-outs and production stoppages are inevitable.
The reasons for this variation can be traced back to supply chain fragmentation, lack of standards and tool control, but it is most likely due to chatter and the response to it.
A Wikipedia article entitled “Speeds and Feeds” describe the current method;
“In CNC machining, usually the programmer programs speeds and feed rates that are as maximally tuned as calculations and general guidelines (with charts and formulas) can supply. The operator then fine-tunes the values while running the machine, based on sights, sounds, smells, temperatures, tolerance holding, and tool tip lifespan".
Hardly sound repeatable.
Dr. Scott Smith, Professor and Department Chair of Engineering at the University of North Carolina at Charlotte likes to say, “Random processes produce random results”. This randomness of change at the CNC machine by each operator makes accurate forecasting and optimizing of the cutting tool inventory levels virtually impossible. Variation also comes in the form of suppliers. The cutting tool industry is highly fragmented with no one manufacturer having a greater than 20% market share. The supply channel is equally fragmented with many competing distributors, as described below:
“...a large, fragmented industry characterized by multiple channels of distribution. We believe that there are numerous small retailers, dealerships and distributors that supply a majority of the market. The distribution channels in the MRO market include retail outlets, small distributorships, national, regional and local distributors, direct mail suppliers, large warehouse stores and manufacturers’ own sales forces.”
What this means is that due to multiple sources, pricing or backorders, substitution of tools, toolholders and inserts are a frequent occurrence. Tool assembly dimensions are not tightly controlled as on-machine probes and presetting machines connected to the CNC control allow tool length or diameter variations to be automatically compensated. Therefore, tool assemblies become frequently ad-hoc, with varying individual components and/or dimensions. Though similar, slight changes in design, dimensions or geometry will create a frequency change at the tool point. Different toolholders have different stiffness and damping properties, as do machine tools and their spindles. Workpiece materials have different properties that impact tool point behavior. Due to breakdowns or bottlenecks, jobs, part programs and tooling are moved from one machine to another. Any of these changes will result is a change in the tool point dynamics and in the stable speeds or “sweet spots” for that particular application. The process moves from stable to unstable and chatter occurs. Trial and error tweaking at the CNC control, as described above, injects variation into the performance of the operation and the consumption rates of the tools fluctuate. As previously stated, there are scientific solutions to this problem. A recent review in the International Journal of Machine Tools & Manufacturer cited 174 published peer-reviewed research papers on the subject of chatter and Machining Dynamics.
Positive impacts to the supply channel using the science of Machining Dynamics include:
Lock-in Business – With Machining Dynamics analysis, tool performance can be maximized, tool life extended (since all teeth will be cutting equally) and total costs dropped, thus erecting an enormous barrier to competition.
Vending Machines – By using Machining Dynamics analysis, predictable cutter/insert consumption can be realized, reducing the size and cost of safety stocks while minimizing the risk of production stopping stock-outs and emergency deliveries. Machining Dynamics is an ideal pairing with tool vending.
Tool Trials – Machining Dynamics analysis eliminates the need for trial and error tool testing. Current methods of introducing new products require the customer to take their machine out of production and provide no guarantee of improved performance. A day or more of an application engineer’s time is also wasted on failed trials. Even if you win, significant costs have to be recovered before the new order is profitable. A quick (>3 minute) test will yield the optimized potential for the new tool and/or toolholder in that machine and in that application. Results will be known without making a single cut. If you cannot win, you can change the tool configuration to one that will.
Resharpening - Tool resharpening is a favorite cost saving measure, but because of the aforementioned science of Machining Dynamics, the change in dimensions from the subtractive grinding of the cutting tool edges to restore sharpness will change the tool point frequency and may move the process from stable to unstable. Often resharpened cutters are automatically operated at far slower speeds than new thus lowering productivity. Utilizing Machining Dynamics analysis, loss of performance can be minimized or a strategy to use only new cutters justified.
Tool Substitution – Sometimes the desired cutting tool is not available. Replacement tools often result in compromised performance, requiring significant manual adjustments at the CNC control, resulting in lost production and slower cycle times. The substitute tool can be measured with a Machining Dynamics test and optimal parameters calculated, minimizing any loss of production and downtime.
Mandated Cost Reduction – Instead of relying on pricing concessions from suppliers or reduced margins to meet required cost reduction targets, improved throughput can be achieved through Machining Dynamics analysis. Substantial hard and soft cost savings can be quantified and validated.
Gain Sharing – Significant performance improvements are not only attainable, but can be accurately predicted with quick and non-invasive Machining Dynamics tests. Terms where the financial gain from the improved output is shared can be negotiated with minimal risk to the supplier and to the customer. “Big Bites” in productivity are realized immediately, with on-going, science-based “kaizen” or continuous improvement.
Six Sigma, Lean and Green – Machining Dynamics analysis eliminates variation in tooling performance. It also reduces waste from downtime, scrap, secondary finishing, trial and error testing or tweaking and premature tool wear. More efficient tool point behavior also consumes considerably less energy.
SOURCES: 2010 Profit Report – Industrial Supply Association, Dormer/Sandvik D.World n.01-2009, http://en.wikipedia.org/wiki/Speeds_and_feeds, page last modified on 31 March 2011 at 05:43, Frost & Sullivan World Machine Tool-Cutting Tool Market Report - 2004, MSC 2010 Annual Report, Chatter in Machining Processes: A Review by Guillem Quintana and Joseph Ciurana, International Journal of Machine Tools & Manufacturer 51
Tap-testing and generating stability lobe diagrams have been the standard technique used by machining researchers at universities and large manufacturing plants for milling optimization. Their widespread adoption has not take place, primarily due to their complexity.
Stability Lobe Diagrams are being substituted with new Tool Dashboards that are both interactive and intuitive to use. With no special knowledge of chatter or mechanical vibrations, a programmer or machinist can quickly try an unlimited number of speed, feed and depth of cut combinations of a tool assembly with instant feedback as to its stability, power usage, surface accuracy and finish. All relevant calculations and conversions are available on the Dashboard. Multiple tools can be compared or tried in multiple machines with accurate predictions BEFORE a cut is made. Eliminates the need of trial and error test cuts. Develops first time right part programs and enables more accurate bidding.
You can try out a Tool Dashboard at the following site:
Hopefully, we have convinced you of the importance of the science of Machining Dynamics, that is understanding the behavior of the tool point during milling. Tools are not 100% rigid, they deflect when a tooth enters the cut and snap back when it exits. The machine, the spindle, the toolholder and the cutter are a unique system that vibrates at a unique frequency. Cutting forces and speeds can be tuned to avoid chatter and optimize performance.
Now, how do we apply this? To generate a stability lobe diagram, we attach an accelerometer to the tool tip of an assembly in a spindle. We tap it with an instrumented hammer to measure it's flexibility and damping properties. Software analyzes this data and generates a stability lobe diagram. Kits are commercially available and new software makes this a fast and easy process.
There is something new on the horizon. We can tap an artifact one time in a machine and then model any toolholder and cutter combination to accurately predict it's tool point behavior and generate a stability lobe diagram without any additional tap-tests. We will need the help of machine tool builders and tooling manufacturers to bring this fruition.
Tooth Impact Frequency is a subject you probably have never heard about before. It is simply the number of times a tooth on a milling cutter impacts the workpiece in a second. It is calculated by multiplying the number of teeth by the spindle speed in revolutions per minute and then divide that by 60. The result is expressed in Hertz (Hz).
In this example at a spindle speed of 7500 RPM, the two flute endmill has a tooth impact frequency of 250 per second with the six flute having 750 impacts per second. Why is this important? Think of the ruler analogy from a prior post. If you have a very stiff tool vibrating at a very high frequency, the tool with more teeth will give you a better chance to match up the two frequencies (tooth impact and tool point vibration) at a higher speed. Conversely, a tool vibrating at a very low frequency will need more space between teeth to try and maintain a higher speed.
There was a video recently posted that compared the performance of the same endmill in the same machine running at the same speeds, feeds and depth of cut using several different toolholders. The toolholders that showed well were no doubt pleased and those that didn't or not included in the test were not. It is not that simple. Freed from the constraints of having to advocate for a particular brand or product, we created this simple presentation to describe how milling tools behave. Milling is not a black art. Vibration phenomena are not random, but rather they can be quantitatively measured and described. It is now possible to scientifically quantify the vibration characteristics of a milling process, predict chatter, and make recommendations to eliminate it.
I bet you never considered this. This was shot in our lab many years ago (yes, we own a Haimer). What we are showing is that just by rotating the endmill in the holder a few degrees, the balance of the assembly changes dramatically. Makes sense, a carbide endmill is fairly heavy and you have removed mass in a spiral to form the flutes. Of course, the balance would change. What does this mean to you? If you balance tools, make sure you make a witness mark on the toolholder that lines up with a flute tip (tooth#1 if its variable pitch) and line that up each time, or you will need to rebalance every time you change a worn endmill. Sorry for the quality, I was holding a camcorder (what's that?) in one hand and running the balancer with the other.
Here is a tool that was tap-tested in a 10,000 RPM spindle. The fastest and deepest available lobe was just under the 10K limit. I intentionally expanded the range of the stability lobe diagram to 17,000 RPM (shown in green) so you can see the big dominant lobe at about 16K. It would cut 25% deeper than the one at 10K. BUT, I CAN'T ACCESS THAT SPEED with this machine.
If we LENGTHEN the tool the lobes would shift to the left and that large lobe, or one of the two other deeper ones, may become available under the 10K limit. We are making the tool more flexible in order to improve its performance. Its not that longer tools are more stable, per se, its that we are dealing with the constraint of available spindle speed.
Of course, other things may change, frequency-wise, if we lengthen the tool too much, but if you are able to measure it, you can dial in the tool for maximum performance.
It used to be that the re-sharpening of an endmill consisted of grinding just the primary and secondary land, as shown in figure 2 (Source: Gurhing), until the edge was restored. In addition to the dimensional change to the diameter of the endmill, this would also change the rake angle from positive to negative and reduce the space for chip evacuation. Dynamically, the tool's mass and length were changed enough to alter its frequency. To compensate, the performance of the reground endmill would have to be de-rated by as much as 50%.
With current CNC grinding equipment, the entire original geometry can be replicated. Diameter and length loss can be minimized if land wear is managed and chipping is avoided. Figure 3 shows a series of stability lobe diagrams; new in red, Regrind #1 in light blue (0.005” removed from diameter and length), Regrind #2 in purple (0.010”), Regrind #3 in green (0.015”) and Regrind #4 in burgundy (0.020” removed). As you can see picking a speed and depth of cut in the center of a robust lobe (yellow area) could maintain high performance through up to 4 regrinds.
Here is another counterintuitive example and before anyone writes, yes the endmill is a long L:D. The zero projection ER collet chuck (blue) had the lowest performance of the three toolholders tested. A couple things are going on here. With the toolholder ending at the gage line, the assembly did not allow much damping resulting in slower speed stable lobes. Even though the "beam" created by the second ER collet chuck (red) was longer it had more mass and damping. The longest chuck (purple) had the second best performance. In addition, it was discovered that the connection stiffness of the zero projection holder was less than the other two because the spanner nut and wrench didn't allow as much torque to be applied as with the conventional nut. The collet acts like a spring, the higher the torque the greater the grip. The zero projection holder performed better with smaller diameter and with shorter tools.
So many variables to consider!
We have explained how every tool/toolholder/spindle/machine combination has a stability lobe diagram and they are all different. The lower axis is spindle speed and the right axis is depth of cut. The white areas are stable and the red areas are unstable where the tool will likely chatter. The region in blue is called the "b-Limit or Unlimited Speed" range. The cutting force is low enough not to cause the tool to deflect. If the diagram is set to a full slot then any axial depth of cut under the b-limit will be stable at any speed. Used with high feed mills. If the diagram is set to the full cutting length of the tool then any radial engagement or step-over under the b-limit will be stable at any speed. Exploitable by Trochoidal HSM milling. The yellow area is the "Process Damped" region. The tool is rotating slow enough that the vibration one tooth is dissipated before the next tooth comes through. You can cut as deep as you want in this region. Variable Pitch Endmills can increase this region speed significantly. The following closer pitch tooth acts as a damper for the previous tooth.
The b-limit is determined by the stiffness of the tool assembly. The Process Damping Wavelength can be determined experimentally.
We all preach shorter is better, don't have unnecessary length for greater stability. This chart was from a aerospace maintenance facility that believed in this. They stocked assemblies in 1/4" increments. In this series, we measured each assembly with a tap test and then validated the data with cutting tests. Turns out that the most stable, the tool that removed the most material was the second to the longest. Go figure. This customer was able to eliminate 6 of the 8 tools from their library. Milling's counterintuitive behavior is what keeps us up at night.
Trochoidal milling, in its many forms and names, is a very powerful tool path strategy. To describe it simply, it utilizes the full cutting length of the endmill with a light radial engagement or step-over (<25%) and fast tool path motions to generate high metal removal rates.
There is a dynamics element to this. On a stability lobe diagram there is what we call the b-limit. This is the cutting depth, either axial or radial, that results in cutting forces light enough not to deflect the tool. High feed mills took advantage of the axial b-limit and trochoidal milling now the radial b-limit. You can run without chatter at any speed if you are under the b-limit.
Since trochoidal milling works so well, naturally we want more of a good thing. In this case, it is by increasing the flute lengths up to four times the cutter diameter or more. But there are limits.
The tool assembly is not 100% rigid and the most flexible part of the assembly is the fluted portion of the endmill. On a typical 4-flute endmill, 30% of its mass is removed from grinding the flutes. So a 0.500” diameter endmill is actually the equivalent of a 0.350” diameter rod.
We tap tested three endmills of the same style in the same toolholder in the same machining center. The endmills had flute lengths of 1.5, 2 and 2.5 times the diameter. We overlaid the three stability lobe diagrams and as you can see, the performance declines as the flute length increases. Not just the stable lobes, but the b-limit decreases as well. You can expect that this trend would continue if we tested 3X and 4X endmills. To compensate for the declining stability, you will have to keep reducing the radial engagement and speed to avoid chatter. You may reach a point with such a small step-over (<2%) and slow spindle speed that a more conventional tool path strategy with a shorter flute length may remove more material in a shorter time. If you do want to go deeper with trochoidal milling you may have to increase the diameter of the endmill and/or increase the number of flutes with a greater core diameter to increase the system's stiffness.
Of course, different configurations (tool/toolholder/spindle/machine) could measure and behave very differently than the example above.
Now let's put this together. This is how the tool's frequency impacts the spindle speed, but...
In this video, we damp the ruler with two sponges. As you can see, the ruler still vibrates. Damping doesn't eliminate vibration, it changes its frequency.
BTW it's damping, not dampening, as Modern Machine Shop's Mark Albert explained here.
This video illustrates how milling tool flexibility impacts it's frequency. The shorter, stiffer tool deflects less and therefore rebounds quickly resulting in a higher back and forth cycle frequency. The longer more flexible tool deflects more so it takes it longer to rebound. This results in a lower frequency of back and forth cycles. We measure frequencies in Hertz (Hz) or the number of back and forth cycles in one second.
To accept the science of Machining Dynamics, you have to accept the premise that all milling tools, no matter their size or projection, are flexible. This short video illustrates this with a 75-pound anvil. We attached an accelerometer on one side of the anvil and lightly tapped it with a hammer. The software screen shows the anvil deflecting in response to the hammer force. It deflects, rebounds and continues to vibrate back and forth until it returns to rest.
With IMTS 2018 just around the corner with it’s myriad of machine tools under power, I am reminded of the story of the perhaps the greatest machining demo ever. It happened at the 1900 Paris Exhibition and world’s fair.
Robert Kanigel wrote about it his massive (675 page) biography of Frederick Winslow Taylor, called The One Best Way. Taylor was considered the first efficiency expert and the creator of scientific management, a field that uses time and motion studies to increase productivity. Before that though, he was a leading cutting tool researcher at Bethlehem Steel in Pennsylvania. His Taylor curve and Taylor tool life equation are still in use today. Along with Maunsel White, Taylor invented High Speed Steel and introduced it to the world in Paris. In the relatively tiny Bethlehem booth they installed a massive lathe that chucked a 10-foot long solid steel cylinder. They would run their HSS tool for 20 minutes and then take a 10-minute break. French news reported that machine shops of Paris would see metal cut SIX TIMES faster than ever before. Crowds packed the small booth to watch the Taylor-White cutting tools in action and other exhibitors complained about them siphoning off attention.
Kanigel wrote: “No one who saw the sight ever forgot it. For many it was the defining moment of their careers, when they watched the world speed up before their eyes. When they told their colleagues about it or talked about it among themselves, technical restraint gave way to simple wonder, the very sight of billowing coils of hot blue chips burned into their brains forever.”
Note in the photo above that hanging between the stanchions at the front of the booth are not velvet rope, but ribbons of the blue chips that came off the tool, so thick that a hammer and chisel had to be used to break them to length.
Germany’s largest machine shop, Ludwig Loewe, sent engineers to view the demonstration where they collected pieces of the chips to take back to their disbelieving bosses.
Wouldn’t every IMTS exhibitor like to have the impact Taylor did in 1900?
It was common knowledge in the Silicon Valley that Steve Jobs was fascinated and respected CNC machining. At the top is an undated photo of Apple's machine shop with two Fadal machining centers visible in the background. In the early 2000's, Job's wife commissioned a San Jose CNC machine shop, owned by Gary Kidgell, to machine a chess set for her husband as a gift. He liked it so much he ordered several more. My colleague, Dr. Scott Smith of UNC-Charlotte, was consulted on the machining of the first iPod case. Scott and UNCC's Dr. Tony Schmitz have traveled to China to work on the machining of Apple products. Today, Apple is one of the world's largest consumers of CNC machining centers. A 1997 video of Jobs returning to Apple has been getting renewed interest on YouTube and LinkedIn. When I watched it I was struck how the language of machining was used by Jobs in his discussion of marketing. See for yourself in the clip below.
Sharing information about high performance milling technologies, the result of 30 years of research.