CNC Milling

This document traces the movement of one user down the learning curve. That movement centers on increasingly complex and challenging projects. It forms a book-of-knowledge regarding the CAD/CAM/CNC process. It serves as a bread-crumb pathway for the author. It is presented without warranty.

Lessons Learned

I> Making Swarf: On the Positive Side

1. Projects:

The Sherline CNC products and available OpenSource make it possible to enter the CAD/CAM/CNC environment with a minimal expenditure of funds. Coupled with over ten years of direct experience with Linux/Ubuntu and software engineering it was possible to successfully complete an array of increasingly more complex and challenging projects during the first calendar year of ownership.

Brass Credit Card: From 2022.

Brass Nameplate: From 2018.

Heart Shape Message Early prototype run.

Box/Lid: Implemented pockets and pads.

2> Projects Lost In The Woods

Bezier Surface Fishing Lure: Increasing complexity of surface/shape. Terminated

Using Bezier Surface for Fishing Lure: Implemented via Draft/Sketch/Part to saw dust.

Realistic Fishing Lure: Simple/Elegant Lures: Draft/Part: Lofting, Mirror, Rotate to saw dust.

CNC Tooling: On the Negative Side

1. Support Tooling to Consider:

2.1> Fixtures:
To support several of the projects shown above, aluminum hold down strap fixturing was created. Two types: slotted and spaced holes were milled. Both are index near the "first hole" using three "dibbits". The slotted version proved to be more useful. The raw aluminum strap stock was purchased at a bigbox store. But a better local source (a scrap metal dealer) of raw stock was found . The specific dealer, JD Rcyclers in Durham, had a great selection and good prices for aluminum, steel, and brass. Several pieces of scrap aluminum and steel were found to be "milled" to tolerances well beyond the user's current skill level. These pieces have been helpful in fixturing.

2.2> Spindle Tachometer:
A spindle tachometer was developed. The tachometer was composed of an Arduino, a light/grey sensor, wire cabling between the Arduino/sensor/PC-USB, and a a modified mill holder . The driving program was loaded to the Arduino before going to the mill. A mill holder was masked/modified to provide a clear breakpoint between the white and black demarcation. Then the modified end mill was installed (less any end mill bit) on the spindle per normal. The sensor was then placed as near as feasible and parallel to the "white" strip on the modified end mill holder. The Arduino was powered. At that point a GtkTerm was invoked and configured to "sync" with the Arduino tachometer. Once the GtkTerm - tachometer were communicating, the mill spindle could be measured for various rotary settings.

The device helped to verify specific rpm reference point on spindle rotary speed switch. Once a set reference rpm-s was visually set various rpms relative it were set. An old Starrett RPM and stop watch provided a rough validated check.

2.3> Axis Movement Controllers:

Implementing a separate third party axis movement became a goal due to the breakage of three end mills".
A F310 gamepad was purchased and installed for controlling movement in the XYZA planes while the operator watches the end mill move. The associated interface software for integrating it with LinuxCNC was "qJoypad". But the gamepad proved to be too user unfriendly. The user still had to take eyes off the mill to "find" the correct axis movement button as well as use two hands.
A simple USB numeric pad was purchased and installed. Installation involved plug/play of the USB into the computer. Ubuntu/LinuxCNC did the rest. It fits the bill completely. LinuxCNC provides a substantial array of keyboard shortcuts. But the USB key pad reduces the "working keys" to eight. The operator quickly learns the tactile location/feel for the appropriate axis movement. Only one hand is required to operate the USB keypad while the eyes can remain on the mill's bit. It has been very helpful for the XYZ axises. End mill breakage has decrease to zero following the installation of the USB numeric pad.

2.4> 3D probe experience:

Installed the DPP Extender Box probe per vendors instructions.
Probe testing of a metal medillion using the probe ngc driver accumulated LinuxCNC loadable NGC gcode. Loading the accumulated gcode to LinuxCNC provided a visualization of the medallion. Finer resolution is a possible by adjusting the probe ngc driver and the stylus. Experiments with adding a button to the LinuxCNC Axis that supports the pyCVP probe feature have produced favorable behavior. But focus returned to making saw dust/swarf over shop features.
Using the "tool height setter" mode of the probe proved to be cumbersome. This feature will be held for later usage.

2.5> Electronic Touch Pad:
A yellow/green led touch pad was created. It uses the PC 25D Parallel port as a power source supporting software was acquired from AskJerry Frankenlab CNC #20 Probes. The schematic is based upon the concept of inverting a digial signal. The actual electronic design was based upon the schematic shown and parts on hand at the shop. The other twist was the usage of the PC 25D Parallel to provide 5VDC and ground. A user can invest a great deal of time and effort searching the Internet, installing the touch pad with new LinuxCNC pyVCP buttons. But under a working approach, the user can invoke the touchplate.ngc from LinuxCNC just like any other and "monitor" the yellow-green transition at the instant of touch. Bottom-line a new user's time might be better used making saw dust/swarf. At a beginner's skill level and modest project complexity, this device has not yet been helpful.

2. Shop Lessons:

3. Focus on Doing:

Focus on making sawdust and metal swarf rather than puttering around with various accessories. Otherwise adequate CAD/CAM/CNC/skills can not be developed.

4. CNC Milling: Benchtop

1.1> The specific designs executed to a suitable level of repeatability.

1.2> Milling operations with multiple tooling changes were a challenge. One operation had three tooling changes. This three tool milling process was accomplished by breaking the milling process into three separate gcode programs with three separate the tooling height settings performed manualy.

1.3> Acquiring tool height, tool change-overs during the cutting process, and usage of tool setters on the Sherline was challenging. A good, repeatable mill tool height process finally evolved.

1.4> Moving a design from CAD to pyCAM/HeeksCNC/FreeCAD to produces gcode for the Sherline CNC mill requires only a minimum of actual adaption to perform suitably.

1.5> Sherline ships a largely useless set of software.

1.6> Prior to purchasing local CNC owners (Tormach 770 and Sherline 5800 with DRO) were visited and polled about their experience. CNC users indicated work piece hold-down and end-mill employment had been their second and third biggest challenges to far. (Tool height/tool table is the biggest.)

2.1>Collet/Chucks: The 3/4-16 nut technique for removing drill collet/chucks works well.

2.1.1> A set of suitable 3/4-16 slim nuts purchased at a local nuts and bolts store, Weller on Alwin Court just East of Atlantic Avenue in Raleigh. They also have an excellent supply of fastener hardware.

2.1.2> The technique was employed for removing the DPP probe and the extended length router bit (see wooden box/lid T19). The technique worked exactly as demonstrated by other users.

2.1.3> Sherline sells a similar product to the nut that is round and uses Tommie Bars.

2.1.4> For the collet/chuck retaining bolt, a small block of wood with a drilled hole is used in concert with light taps from a jeweler's brass hammer. Place the block over the head of the bolt such that the drill hole cups the bolt head. Then place a small piece of scrap wood on the mill plate/table such that this scrap protects where the bit will fall. Then lightly tap the block.

2.2> End Mills:

2.2.1> Spraying lightly (Ballistol Multi-Purpose) the interior threads of the End Mill Holders prior mounting has been very helpful.

2.2.2> When a stuck End Mill Holder needs a "break" to loosen,

2.2.2.1> Place a bounce pad/wood scrap directly under the Spindle housing. Then remove the end mill bit from the End Mill Holder.

2.2.2.> Place a large bolt like item parallel to the spindle shaft such that it is adjacent to the spindle housing. That bolt then becomes an excellent pry/force point for the spindle's Tommie bar.

2.2.2.3> Then place a Tommie Bar through "both holes" of the Spindle's Head Stock. Once inserted rotate the Spindle's HeadStock Tommie bar toward the pry/force point until they "touch".

2.2.2.4> Insert a bolt with washer all the way through the End Mill Holder and lock with a washer and nut.

2.2.2.5> Shim with brass stock as deemed needed to protect surfaces of housing and/or End Mill Holder.

2.2.2.6> Secure and tighten the clamping jaws of an adjustable wrench around the End Mill Holder's Bolt/Nut.

2.2.2.7> From looking down from above the spindle rotate by pulling the wrench clockwise toward the pry/force point.

2.2.2.8> Prior to using this leverage scheme and the Ballistol lubrication bending Tommie bars was a common event.

2.3> Spindle 3/4-16NF Maintenance:

2.3.1> Hanson/Iwrin manufacures 3/4-16NF die part number 6860ZR die for re-threading which might prove useful in maintaining a servicable external thread on the spindle. The die can be purchased for any hardware store or big box lumber company.

2.3.2> Follow guidance of lubrication of the spindle thread with Ballistol, sequence of half turn/back-off a quarter turn, inspect/test with an end mill holder. Avoid using anything more than channel locks and a Tommie bar. The mill's spindle is ready when the end mill holder can be installed with modest hand pressure before locking down with Tommie bars at teh sholders of the spindle.

3> To make real progress the serious user must have a good knowledge of Linux/Ubuntu/Debian and LinuxCNC.

3.1> It seems the Sherline intellectual property is hidden.

3.1.1> The distribution disk is mis-labeled or appears to just be a "brochure" for Sherline. But it is really a hidden restore disk. If your system crashes and needs a rebuild, you need that knowledge.

3.1.2> The actual Sherline CNC distribution is probably designed to be extremely limited. To increase usefulness and to reach a comfortable level of user friendliness and productivity additional third party software should be installed. Ubuntu Software Center, Calculator, System Monitor, Shutter, Meld, Geany, Disk Utilities, pyCAM, heeksCNC, FreeCAD, OpenSCAD, Blender, Veusz, Clonezilla, and Nautlis are strongly recommended OpenSource applicaitons that a serious user should consider installing

3.1.3> During early development Sherline employed a LinuxCNC setup wizard, stepconf. The stepconf wizard matches the physical hardware with the virtual software elements of LinuxCNC. Masking that complexity is what a users pays Sherlien to do. But the Sherline CNC folder is a menagerie of configuration files that some how work properly once the "ini" file is adjusted for the user's specific lathe or mill. Do not mess around with the LinuxCNC stepconf wizard until a full backup of the exisiting linuxcnc folder is secured. Only then, instead of creating a new stepconf, try modifying an existing one by selecting it and renaming the "new" machine stepconf file. That will create a separate, independent stepconf, separate launch icon, and a separate folder specifically for the "new" machine.

3.2> Sherline did not provide an installation CD or a downloadable installation ISO with their CNC mill bundles like the 8580 shipped during the end of October 2017.

3.2.1> When asked, their technical support suggested that a request be made to Sherline Sales. Sherline Sales response to the request was quick. The installation CD that was received contained a Live OS for installing Ubuntut 12.04 LTS and LinuxCNC v2.6. These are both down level and need of updating.

3.2.2> The installation CD has no instructions. But If one is familiar with Ubuntu the install can be achieved with a minimum of trial and error.

3.2.3> My original CNC 8580 was shipped during the transition period when Sherline was cutting-over from Debian 7.11 to Ubuntu 12.04. There is no easy upgrade path in this instance. So all prior user content that existed in Debian environment had to be manually migrated to Ubuntu -or- lost.

3.2.4> The installation process once the decision to proceed was made was straight forward but tedious. Below is a description of the resulting upgrade:

1> Simulator: IBM Think Center Model 8114-D1U with 1T Drive.

2> CNC PC/Controller: ARS with upgraded 1T Drive in a standalone role.

3> Contents of both machines:

Ubuntu 12.04

LinuxCNC 2.8

Geany

Meld

Shutter

Qjoypad and PyPar2 for logitech GamePad F310

XFBurn

System Monitor Profiler