Cutting force milling aluminum / Force sensor concept
metron9
Posts: 1,100
Since i do not have a math degree the documents I have found that completly explain cutting forces on milling materials is of no use.
Hands on experimennts are what i plan to do.
What I want to be able to do is record the force used in various cuts using various end mills etc..
My initial concept:
Make a slide table using a DC motor / acme screw to cut at various speeds in one direction.
The motor current would be recorded as one of the data sets for stepper motor size requirements.
The actual cutting force on the connection between the table and the acme screw is quite different than the force needed to move the table and my concept here is to design a pressure gauge data output device using a dial caliper as the sensing device and replaceing the needle on the dial caliper with a simple digital encoder wheel. A 1/10,000 dial caliper has a .01 per revolution so quadrature·encoder wheel with 10 divisions would get you to .001 and allow the actual forces to be recorded.
the calculation for Lbs per sq inch could be arrived at by loading the table manually with a known force. Since I don't know the parameters of the maximum force yet and I assume a maximum of .050 flex to the maximum force used I may have to calibrate 50 points where .001 is x lbs and .050 flex is x-lbs since the spring will be a non linear compression factor but that is simple to do.
From looking at force and pressure gauges they can be quite expensive and I dont know how fast they can react to give the required data feedback.
So the finished product would be a 2 line graph showing the force being exerted on the screw by the motor current feedback and the actual cutting forces during the cutting process. A spindle speed of 3000 rpm / 50 rps would require 18,000 samples per second but I would collect just 360 per second and average that data stream to output one datapoint per second, 360 datapoints to achieve a burst of data for a 360 degree rotation of the cutting tool.
These things keep me up at night but it's the kind of thing I hope to explore when·I actually do have the means of making precision parts necessary for implementing a tool such as this.
Suggestions on pressure devices that would fit the bill and not break the bank?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
Hands on experimennts are what i plan to do.
What I want to be able to do is record the force used in various cuts using various end mills etc..
My initial concept:
Make a slide table using a DC motor / acme screw to cut at various speeds in one direction.
The motor current would be recorded as one of the data sets for stepper motor size requirements.
The actual cutting force on the connection between the table and the acme screw is quite different than the force needed to move the table and my concept here is to design a pressure gauge data output device using a dial caliper as the sensing device and replaceing the needle on the dial caliper with a simple digital encoder wheel. A 1/10,000 dial caliper has a .01 per revolution so quadrature·encoder wheel with 10 divisions would get you to .001 and allow the actual forces to be recorded.
the calculation for Lbs per sq inch could be arrived at by loading the table manually with a known force. Since I don't know the parameters of the maximum force yet and I assume a maximum of .050 flex to the maximum force used I may have to calibrate 50 points where .001 is x lbs and .050 flex is x-lbs since the spring will be a non linear compression factor but that is simple to do.
From looking at force and pressure gauges they can be quite expensive and I dont know how fast they can react to give the required data feedback.
So the finished product would be a 2 line graph showing the force being exerted on the screw by the motor current feedback and the actual cutting forces during the cutting process. A spindle speed of 3000 rpm / 50 rps would require 18,000 samples per second but I would collect just 360 per second and average that data stream to output one datapoint per second, 360 datapoints to achieve a burst of data for a 360 degree rotation of the cutting tool.
These things keep me up at night but it's the kind of thing I hope to explore when·I actually do have the means of making precision parts necessary for implementing a tool such as this.
Suggestions on pressure devices that would fit the bill and not break the bank?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
Comments
If you wish to experiment further you could construct a specialized load cell to measure the cutting side force on a member supporting the milling head. Load cells can be bought or homebuilt. They use, for the most part, strain gages as the sensing element and as such the millivolt output signal must be amplified and converted to digital if it is to be read into a Stamp for further processing. Let me know if you wish to·persue the use of·strain gages.
cheers, David
I wish to persue strain gages.
Just thinking out loud again, what if you designed a nice little base for a mill vice that had readouts similar to a DRO that showed the force being used during a cut.
Perhaps connect to an E-stop to safetly shut down the mill if forces greater than allowed were to be reached.
I am sure something like this is in some of the big CNC mills but it would be a nice thing to have for teaching how to mill. A fly cutter hitting a vice for example the strain gage would shut down before the cut got too deep and perhaps save a person from a loud banging noise.
Heck, I don't even have·my mill yet...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
0-1000lbs compression and tension.
I don't see a datasheet though. How much deflection do you get with one of these? For example at 1000lbs of force is there a few thousands of an inch, less, more?
Reason i ask is if one used this on a vice mill platform for example you would not want the vice to be able to move even a half a thousands.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
en.wikipedia.org/wiki/Moment_arm
You will have to keep clear in your mind where your tangential cutting force really is. Its vector or "aiming point" will change somewhat as your tool moves through the piece.
Hmmmm....
I guess I'm doing a terrible job of explaining this. Normally I'd have to draw a bunch of vectors and use trigonometry to derive what your cutting force will be when given such-and-such force data from such-and-such a location. Frankly, I don't see how you can do this without some math. For example, you need to be aware that forces will add as vectors, so if you've got 100 lbs of force measured along the X axis and 100 lbs of force measured along the Y axis at the same time, then you've got a force vector equal to 141 pounds aiming between the X and Y. Your cutting tool will experience that 141 pounds of force in many ways: as a bending moment and as a twisting moment within its shaft. If you are also cutting in the Z-direction at the same time, then you've got yet another vector to add into the analysis.
What you need to worry about I guess depends on what your goals are. But the dynamics of "simply" cutting metal can get pretty complicated! I'm guessing there are machinist's handbooks that can give you ballpark numbers for feed rates, depth of cuts, specific materials, etc. so you might want to start with those. Keep in mind that if you use a force sensor to dumbly shut down a machine when it experiences too much force, it might not work very well unless it can automatically back off from the workpiece. If you just shut down all the power when it gets in trouble, that cutter is going to chatter and jam and probably inflict all kinds of damage to the machine as the machine literally grinds to a halt.
Turns out that the forces in milling typically are in the tens of pounds even though a Kirk vice can cost over a grand and has 8000 lbs of clamping force.
So, I plan to use a VFD 3 phase device that converts 220 single phase . My initial thought is to insert a current monitor on the 220V line. The question would then become would that even be usefull for say a 1/4 inch endmill since the power difference between light cutting and too much force using a 2hp 3 phase motor would perhaps be quite small like measuring a milliamp plus or minus on a 220 volt line. Perhaps on the large cutters it would be usefull so I think playing with a load cell is going to be my first attempt after finding and reading a datasheet of course.
Good news, the rigger called and the mill will be here Monday, ill take pictures.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
It looks like the one mentioned above has a "Full Scale Deflection" of 0.003 inch.
If the VFD is not too stupid, it has an output (or can be configured to have one) of the actual load.
But that doesn't give you the cutting forces. You will have to scale with motor efficiency (frequency dependent at least) and subtract zero load (frequency + temperature of the spindle dependent).
Why do you want to measure the cutting force? Spindle load (% of max output) is enough in many cases. Do you want to avoid breaking small mills?
Nick
With machine work, every material and every cutting head requires some adjustment to get good results. And just talking about aluminum or steel is rather naive. You have to specify alloy. High strength aluminum alloys are more difficult to machine than cheaper, low strength aluminum. Similar issues show up with steel, brass, copper and more exotic non-corrosive nickel alloys.
Nonetheless you can get a project started and work through it safely if you get the right RPM for cutting heads via trial and error and go very slowly with feed rate until you have some idea of what the metal will tolerate. Also, even if you can't get machinist data, you certainly get some idea on line of how easy it is to work any specific alloy. Always try to get to know your material first as you have a better idea of what to expect from it. You may even decide to change to a different alloy that is both cheaper and easier to work.
I believe it would be much easier to just use a rpm display. If the duration of one revolution is measured in milliseconds, you can reduce typical rpm error of +/- 10's to 100's of revolutions to just +/- 1 revolution.
You would need to input tool size, and then, load can be calculated by measuring the rpm loss of a previously measured (calibrated) torque/rpm value.
If you really want a load cell, some of the larger dyno's I have worked with used a hydraulic load cell. A simplistic version can be found at the very bottom of this web page http://www.nakka-rocketry.net/hydlc.html built from a small clutch slave cylinder and an electronic pressure sensor. A small inline needle valve can be used to smooth out the data before the pressure sensor.