Author Topic: To the dark side CNC  (Read 24235 times)

Offline cfellows

  • Rest In Peace
  • Full Member
  • *
  • Posts: 1700
Re: To the dark side CNC
« Reply #60 on: August 12, 2015, 09:10:33 PM »
Hello Chuck

I like your 4th axis set-up. It has got me excited and thinking about making one for myself.

I see you are using a HUGE stepper motor and I can understand the need for plenty of torque; but you are driving the spindle through a comparatively small toothed belt. What about backlash in the belt drive and possible belt stretch ?   Is this a real or imaginary problem ?

Regards

Mike

Mike, the stepper motor isn't that big.  It's a NEMA23, I think under 200 oz in.  There is no measurable backlash or belt stretch that I can see.  I also use timing belts on my milling machine steppers.  The motors are NEMA23, 425 oz in torque.  The backlash on both X & Y is about .001" and I think that is mostly in the ball screws.  So, in my experience at least, backlash issues with timing belts are mostly a myth.
So many projects, so little time...

Offline cfellows

  • Rest In Peace
  • Full Member
  • *
  • Posts: 1700
Re: To the dark side CNC
« Reply #61 on: August 12, 2015, 09:16:58 PM »
Stuart: If the learning curve goes over the vertical that's good, as you're then on the way back to horizontal! Despite a number of CAM/CNC issues with the parts shown, the majority of the time taken was in the actual design, not related to the machining at all. I'll be interested to see what you do regarding tooling plates. Early on I bought a large (20" x 14" x 1"), and expensive, piece of ground aluminium jig plate. I designed a tooling plate with lots of tapped holes, and reamed holes for dowel pins. So far I haven't got around to actually making it. For smallish parts I use a machine vice. A lot of parts rest direct on the machine table, and for those where I am profiling edges I use a sacrificial lump of ordinary aluminium plate.

Chuck: The helical gears look good. As a result of discussions on another forum I am looking at the mathematics of helical gears. I'd be interested in knowing a bit more about the setup used to machine them. Did you use an involute cutter, set over at the helix angle, similar to using a universal mill and dividing head, or a small slot drill making a series of passes for each tooth? If you used an involute cutter how did you calculate the equivalent number of teeth used to select the cutter? If you used a slot drill how did you model and/or calculate the tooth shape?

I'm a bit confused about the use of the 4th axis for thread milling. Some while ago I bought a couple of thread milling cutters from Maritool, but haven't had a chance to use them yet. However, I was under the impression that I could use the mill in 3-axis mode to generate the necessary helix for cutting both internal and external threads. Have I misunderstood?

Andrew

Andrew, I did use an involute cutter set at the helix angle for cutting the helical gear.  I've attached an excel spreadsheet that has a lot of the tables and math used for helical gears.

My thread milling operation used a 60 degree milling cutter, not an actual thread milling cutter.  I learned later that there is an easier way to cut threads without a 4th axis, also using just a 60 degree milling cutter.  Of course, for internal thread, it may require a really small cutter, so a thread mill might be better.  I have a friend that runs a local machine shop and he swears by thread milling, but I have no experience with them.
So many projects, so little time...

Offline jadge

  • Full Member
  • ****
  • Posts: 503
  • Cambridge, UK
Re: To the dark side CNC
« Reply #62 on: August 12, 2015, 09:22:52 PM »
Dave: I have a Tormach PCNC1100 Series 2, imported into the UK from the US. I bought the mill and the 4th axis kit with 8" table together, as I figured that shipping the rotary table at the same time as the mill wouldn't be a huge extra cost. Before buying I spent a lot of time looking at various options including hobby machines, a new Haas, and secondhand commercial CNC mills like Bridgeports. The Haas was out of my price bracket, especially when including 'essential' items that were priced as options. The larger Bridgeport CNC mills wouldn't fit in my workshop, and I hated the idea of having to debug 1980s electonics if there were problems.  :ThumbsDown: The Tormach seemed a good compromise; capable of handling the size of parts I knew I would be machining, a 4th axis was available, and it would cope with tougher materials. I was also impressed by the no nonsense design notes and white papers on the Tormach website. Overall I have been very pleased with my purchase.

The calculations for ridge height versus stepover for a ball nose mill should be simple geometry, although it is built in to my CAM program so I have never actually sat down and done the maths. The finish cut on the bevel pinion shown previously was done with a 4mm ball nose cutter and a step over of 0.1mm (4 thou). Total machining time was 2½ hours, including all the roughing. As another example here is a spline cutter machined from gauge plate using only a 6mm ball nose cutter:



Stepover on the finishing cut was 0.4mm (16 thou) and that gave a pretty good finish; although the ridges can be seen they can't be felt. Total machining time was about 4 hours.

I wonder what speeds and feeds people are running at? For the bevel pinion I was running at 5000 rpm and 400mm/min, and for the spline cutter 3000rpm and 150mm/min, both cutters being 3 flute uncoated carbide.

Andrew
« Last Edit: August 21, 2017, 10:48:23 PM by jadge »

Online Vixen

  • Full Member
  • *****
  • Posts: 3075
  • Hampshire UK
Re: To the dark side CNC
« Reply #63 on: August 12, 2015, 09:35:16 PM »
Chuck

Now you have me confused. You said in your reply  'the stepper motor isn't that big.  It's a NEMA23, I think under 200 oz in' . 

This morning I found your utube video <a href="https://www.youtube.com/watch?v=1YeG42yBHwo" target="_blank">http://www.youtube.com/watch?v=1YeG42yBHwo</a> which shows a massive Nema 34 1600 oz/inch stepper.

What stepper do you use for your helical gear work?

Cheers

Mike
It is the journey that matters, not the destination

Sometimes, it can be a long and winding road

Offline Hugh Currin

  • Full Member
  • ****
  • Posts: 720
  • Box Elder, SD, USA
    • www.currin.us
Re: To the dark side CNC
« Reply #64 on: August 13, 2015, 12:46:04 AM »
The calculations for ridge height versus stepover for a ball nose mill should be simple geometry, although it is built in to my CAM program so I have never actually sat down and done the maths.

Andrew

It was simpler than I thought. Checks against at least one on line calculator.


Thanks.

Hugh
Hugh

Offline Dave Otto

  • Full Member
  • *****
  • Posts: 4693
  • Boise, Idaho USA
    • Photo Bucket
Re: To the dark side CNC
« Reply #65 on: August 13, 2015, 01:12:09 AM »
Thanks Andrew

Someday I would like to have a more capable CNC mill in the home shop. Mine is a converted Bridgeport clone, knee mill that I converted myself. I have been pretty happy with it so far but the lack of quill travel leaves a lot to be desired. My machine at work is much more capable and I'm welcome to use it after hours and on weekends which is nice.

I did include a 4th servo amp in the box and all the connectors to add a 4th down the road; I just need to figure out what I want to build and start collecting parts.

Dave


Offline Dave Otto

  • Full Member
  • *****
  • Posts: 4693
  • Boise, Idaho USA
    • Photo Bucket
Re: To the dark side CNC
« Reply #66 on: August 13, 2015, 01:21:14 AM »
I have just had a first go with Fusion 360. Previously I have done 2D with Draftsight and CamBam so I was keen to see how Fusion worked for a sculpted part.

Here is my first part as a rendering and as a first cut. I found building the model quite easy (normal software caveats apply) and rendering is just icing on the cake. CAM for 3D I think I have more to learn but the software does what it says it will and no surprises.

You can clearly see the track of the 6mm ball ended cutter even after 3 operations. I think I either need a tighter finishing pass for the near horizontal surfaces or perhaps use a bull nose cutter rather than a ball, to leave small steps rather than grooves. Anyway a first step and one I am quite pleased with, the part is to be used in a dog guard in the car.

Mman

Mman

When doing surfacing tool paths most CAM programs will let you leave a stock allowance for cleanup on the finish pass. Depending on the material being cut I will leave anywhere from .002" to .02" or more. Also there usually a setting for the tolerance; for the roughing passes I will set it at .001" and .0001" for the finish.

It appears that your roughing passes are cutting deeper than your finish pass; this could be for a number of reasons. tool flex, or maybe backlash in Z. I would start with trying to leave more for the clean up pass.

Its great to see some CNC discussions here on the MEM forum.

Dave 

Offline jadge

  • Full Member
  • ****
  • Posts: 503
  • Cambridge, UK
Re: To the dark side CNC
« Reply #67 on: August 13, 2015, 09:30:47 AM »
Chuck: Thanks for the spreadsheet, it is pretty neat. I note that the spreadsheet uses the standard divide by cos of the helix angle cubed to get the equivalent number of teeth to select a cutter. It is fairly simple to derive the equation from first principles using the radius of curvature of an ellipse. However, what confuses me is that some sources, eg, Machinery's Handbook, give a revised formula with an additional term involving the cutter pitch diameter and normal DP. I'm not sure how this extra term arises. Apparently it is most useful for gears with high helix angles and low tooth count. That seems to apply to one of your helical gears, but presumably the gears meshed without any problems?

I have also seen another variation involving dividing the actual number of teeth by the square of the cosine of the helix angle and the sine of the helix angle. The example given was for 45° which of course works as sin(45) = cos(45). However, for zero helix angle, ie, a spur gear, the equation blows up as sin(0) = 0. There is also an equation involving the tangent of the helix angle and the helix angle itself, but I haven't had time to look into that yet.

Andrew

Offline cfellows

  • Rest In Peace
  • Full Member
  • *
  • Posts: 1700
Re: To the dark side CNC
« Reply #68 on: August 13, 2015, 02:47:34 PM »
Chuck

Now you have me confused. You said in your reply  'the stepper motor isn't that big.  It's a NEMA23, I think under 200 oz in' . 

What stepper do you use for your helical gear work?

Cheers

Mike

Sorry Mike, the dividing head in this video is the second 4th axis I built.  It has a NEMA34, 1600 oz in stepper.  I rarely use it because it is so big and harder to mount in the mill.  The original 4th axis, which I use most of the time, has the NEMA23 stepper.  The spindle is an ER20 Collet chuck with a 3/4" spindle diameter.  Here is a video of it in action.

<a href="https://www.youtube.com/watch?v=L18a_wfaRBE" target="_blank">http://www.youtube.com/watch?v=L18a_wfaRBE</a>

So many projects, so little time...

Offline cfellows

  • Rest In Peace
  • Full Member
  • *
  • Posts: 1700
Re: To the dark side CNC
« Reply #69 on: August 13, 2015, 02:51:31 PM »
Chuck: Thanks for the spreadsheet, it is pretty neat. I note that the spreadsheet uses the standard divide by cos of the helix angle cubed to get the equivalent number of teeth to select a cutter. It is fairly simple to derive the equation from first principles using the radius of curvature of an ellipse. However, what confuses me is that some sources, eg, Machinery's Handbook, give a revised formula with an additional term involving the cutter pitch diameter and normal DP. I'm not sure how this extra term arises. Apparently it is most useful for gears with high helix angles and low tooth count. That seems to apply to one of your helical gears, but presumably the gears meshed without any problems?

I have also seen another variation involving dividing the actual number of teeth by the square of the cosine of the helix angle and the sine of the helix angle. The example given was for 45° which of course works as sin(45) = cos(45). However, for zero helix angle, ie, a spur gear, the equation blows up as sin(0) = 0. There is also an equation involving the tangent of the helix angle and the helix angle itself, but I haven't had time to look into that yet.

Andrew

My spur gears seem to mesh just fine for model engine work.  Didn't realize there was more math for special cases...
So many projects, so little time...

Online Vixen

  • Full Member
  • *****
  • Posts: 3075
  • Hampshire UK
Re: To the dark side CNC
« Reply #70 on: August 13, 2015, 03:50:57 PM »
Hello Chuck

Thanks for taking the time to explain about your two different 4th axis units.

I have seen adverts for some quite low torque 4th axis units originating from China, they are described as suitable for engraving (wood plastic?). When I saw your massive 1600 oz/inch motor, I was worried that that sort of torque was what was required for metal bashing. You have set my mind at ease, now I know you normally use the much smaller 200 oz/inch motor. I guess what we make is often governed by what we find in the 'come in handy' box.

I now feel comfortable that when I use of a microstep driver on a 200 step 3.0Nm NEMA 24 stepper (400 once/inches) through a 3.6:1 toothed belt reduction, with a 15mm wide, 5mm pitch, timing belt, it should be more than adequate. That should give a basic 0.5 degrees per step, then there is the microstepping.

I may start a new topic when I have made some progress.

Thanks again

Mike
« Last Edit: August 13, 2015, 06:31:54 PM by Vixen »
It is the journey that matters, not the destination

Sometimes, it can be a long and winding road

Offline Hugh Currin

  • Full Member
  • ****
  • Posts: 720
  • Box Elder, SD, USA
    • www.currin.us
Re: To the dark side CNC
« Reply #71 on: August 13, 2015, 06:33:26 PM »
You have set my mind at ease, now I know you normally use the much smaller 200 oz/inch motor.

Mike

Mike: Well, let me attempt to dash that comfortable feeling.

I think comparing the holding torque for a lead screw to that needed for a 4th axis shows the 4th axis drive needs considerable torque. The lead screw is very good as a gear reduction and in lowering needed motor torque. I ran a calc to quantify this as:



This is leading me to use a double step timing belt drive for a 4th axis. I can't see how to get reasonable reduction otherwise. There is a "commercial" 4th axis by Simpson that uses a double reduction. Actually configurable for double belt, 4th axis, or single belt, turning. This video giving an idea how it works. I like the concept but can't afford the item.

If anyone sees holes in this calc please let me  know. It does assume a worst case cut, pushing sideways. But I have not yet built one, and Chuck is having good luck with his. I'd sure like it if I was proven wrong and a single reduction was optimum.

A little off topic, but the 4th axis came up and I have an interest. Thanks for bearing with the diversion.

Thanks.

Hugh

Hugh

Online Vixen

  • Full Member
  • *****
  • Posts: 3075
  • Hampshire UK
Re: To the dark side CNC
« Reply #72 on: August 13, 2015, 08:54:13 PM »
Hello Hugh.

An interesting discussion.

I agree, for the linear axes and a given cutter load, the stepper torque requirement remains constant and is proportional to the ball contact diameter and the lead angle of the ball screw. The ball screw geometry was fixed during design.

For the 4th axis, given the same cutter load, the stepper torque requirement is a variable. It depends on the diameter where the cut is made and the angle at which it is applied. Clearly large diameter workpieces and radial cuts at 90 degrees to axis will generate the highest loads which require higher holding/driving  torque. They could be many times larger than that of a ball screw in one of the linear axes.

Fortunately the cutting force is under the control of the programmer. Reducing the depth of cut, the stepover and the feed rate will all reduce the cutter force. The torque required to resist longitudinal cuts of a spur gear will be significantly less than the largely radial cuts required for a worm wheel. It is for the programmer to consider the geometry of the required work and program the cutter parameters accordingly. Nothing changes, it has always been that way.

Yes, you can always use a bigger motor and higher belt reduction ratios to increase the holding/ rotating torque. 'The Simpson' offers the option to use one or more gear reduction ratios to match the job.

The demo video was rather misleading. There is no way that long tube held at one end only in the 3 jaw could resist much in the way of a side load, but then, the marker pen used to draw the fairy would not have produced much load anyway. You can draw as fast as the machine will move.

I still have that comfortable feeling knowing that Chuck's 4th axis unit works. He clearly programs the cutter parameters within the limitations of the machinery. We are dealing with model engineering not large scale industrial production, we can afford to slow down and play a little.

Regards

Mike


« Last Edit: August 13, 2015, 11:09:45 PM by Vixen »
It is the journey that matters, not the destination

Sometimes, it can be a long and winding road

Offline Hugh Currin

  • Full Member
  • ****
  • Posts: 720
  • Box Elder, SD, USA
    • www.currin.us
Re: To the dark side CNC
« Reply #73 on: August 13, 2015, 09:27:57 PM »
I still have that comfortable feeling knowing that Chuck's 4th axis unit works.

Mike

Real world performance trumps math every time. It's a good argument that Chuck's 4th axis works well. But still the analysis give me pause and I may well do something like the Simpson when I get around to building one.

Thanks.

Hugh
Hugh

Online Vixen

  • Full Member
  • *****
  • Posts: 3075
  • Hampshire UK
Re: To the dark side CNC
« Reply #74 on: August 13, 2015, 09:46:52 PM »
Hello Hugh

Interesting discussion.

Practical machinists learn to live within their machine's capabilities.

The Simpson, one or two belt reduction option is a neat trick and one that can be added very easily, just add the second belt when required.

Have you checked the lead v diameter ratio for real world ball leadscrews? The helix angle, theta, can be 45 degrees. I can push my x and y axis around when the power is removed, they do not lock like a screw thread.

Regards

Mike
It is the journey that matters, not the destination

Sometimes, it can be a long and winding road

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal