Tabletop CNC

[Jasonb]

Here you go Brian, although my CNC would be classed as benchtop being a little smaller than your mill this shows the process of generating the code and cutting the cam. This is a cam about the same size as the original Bobcat that I drew in Alibre and exported as a .STEP file and opened in F360

I prefer to use a 3D file as it contains all the heights so I can just click on a part of the model and let the software determine what height it is at rather than having to look back at other drawings as a DXF is only 2D so less risk of errors.

First job is to tell it what stock I'm using, Here if have positioned the cam off centre in the stock which is less wasteful and allows me to cut it from 1/2" drill rod which is close enough to Q1

I like to use an adaptive tool path to remove the majority of the waste, this cuts at a constant cutter engagement, the blue lines indicate the movement of the cutter when actually cutting, yellow is other non engaged movement. All clim cut at 500mm/min (20ipm) and 5000rpm with a 6mm 3-flute carbide cutter

Then two contour paths to finish the round boss and then the actual profile of the cam. I do this with one pass taking off 0.2mm and the second pass taking off 0.1mm, heights simply by clicking the bottom of the round part or bottom of the cam profile, cut at 400mm/min feed for a good finish.

I won't bore you with the spot drill, 3.8mm drill and 4mm reaming paths but all very simple to add and shown in the video along with the completed part at the end.

 

[Brian Rupnow]

Brian,
If it wouldn’t be too much of a problem could you email me this drawing in a .dxf file. I’ll make a prototype in a different material just to verify the cam portion configuration.
Regards,
George

Toolsrul--Send your email address to [email protected] and I will send you a .dxf file.---Brian

 

[peterl95124]

Brian, not making solid one-piece cam shafts is eventually going to bite you, this isn't your best long term option. eventually this is what you will want, IMHO, YMMV, VWPBL, yada, yada, yada...

 

[Robin Mosedale]

Anyone using Taig mills in Europe? Is there a dealer somewhere around here. Got me interested in the Taig Mill as searching a suitable first desktop CNC mill has proven a challenging task. Been looking around for two years now it seems and haven't found a suitable used option. New machine market is also quite limited it seems.

Olli, I have one, a long bed. They are traded as Peatol lathes and mills in the UK. Do a search and the site comes up. Mine is 2 axis at the moment, and use it for 2.5D milling

 

[stanstocker]

Mr Stan - - - you mention 4 different options for controls.
Which do you recommend and then why?
Also CAM software - - - do you have any that's linux usable?

TIA

I have to warn you, I used to be really up on the whole CNC CAD/CAM world, but other things have taken my fancy in the last several years. Using CNC tools still matters to me, but building, tweaking, and discussing / arguing fine and subtle points of it all just doesn't excite me the way is used to. I was never a CAD/CAM/CNC master, all of my needs were and remain fairly simple parts, particularly in the realm of this groups focus. The vast majority of my needs were met using VCarve and Gearotic, with occasional use of various CAD and CAM packages for special needs. Alibre is about as exotic as my CAD needs get.

Sad news on the CAM front, I've never found an open source CAM package that worked well for me. However : Meshcam has worked decently under Wine, so that may work for you. Grab the 30 day free trial of CamBam as well, last time I looked it was well behaved under Wine as well. It seems that these two packages have remained popular for quite some time. The latest versions of VCarve have not worked under Wine for me, older versions (up to 10 maybe?) were fine with Wine.

For control software these days unless you need more, gRbl based solutions may work well enough for hobby class machines. It's so popular as it's the default on so many popular routers and laser cutters that most everyone supports it to some level. Critical user mass and good enough to use can sometimes trump elegant but complex and/or expensive options. The all in one gRbl boxes that are of good quality are around $200 to $250, probably less expensive that step generation hardware and stepper drivers for linuxcnc in many cases, certainly less expensive than Mach4 with the required hardware.

I used linuxcnc for quite a few years, but the lack of conversational modes and wizards was an minor but nagging issue for me. Getting advanced features working well can "always be done" if you really grok the whole linuxcnc mindset. It's a good package. I had used Mach3 and it was OK but somewhat unstable and always felt cluttered. I went to linuxcnc and as a pure motion controller being fed gcode it was perfectly satisfactory. This was in the days of parallel ports, so a while back. Mach4 has been quite good, stable, and those who do a lot of probing say it is quite good in that regard. My friends who drive larger machines tend to use linuxcnc and Mesa boards to run their converted Bridgeports and Hurco class machines. They can take advantage of, and often need to have the advanced options and hardware monitoring features of linuxcnc and complex breakout boards. I appreciate their machines and accomplishments, but I'm not looking to join their group. My playground is in a small retired guy shop, making small things.

With the demise of parallel ports on PC's you need step generator hardware for linuxcnc of one flavor(Mesa typically), and another flavor (Smoothstepper, PMDX, various others) for Mach4. Either way it's $150 to $200 for the hardware and you still need the stepper drivers and related hardware. If you change from linux to Mach or the other way around you need new step generation hardware.

For machines that need less than 3.5 amps per axis the all in one gRbl control boxes like the Openbuilds BlackBox X32 or the xPro V5 are really excellent all in one boxes, easy wiring, all the connectors included, just add a power supply and get to hooking things up. Hard core CNC folks may not think much of gRbl, but it's worked very well on my routers and on my diode laser. 4th axis support is fairly recent, which is why my 4 axis Sherline machine was on linuxcnc for a few years, then I went to Mach4. The ability to do automated gear cutting in either traditional 4 axis mode or profile milling mode was important to me when I had my clock restoration business. I have not played with the most recent gRbl on a mill, but might when I get some play time. My G0704 conversion required more power than the all in one gRbl boxes can supply, that machine has a shop built controller with a PMDX step generator / break out board running four 5 amp hybrid stepper drivers with a healthy power supply. It is perfectly reasonable to break out the step and direction signals from a gRbl controller board like an Arduino or ESP32 based SBC and use those to drive higher current stepper motors if you want to go that way.

Since retiring from the clock business and being wrapped up in our retirement place I've done very little with my CNC mills, almost all of my play has been with 3D printers, routers, and lasers. It's not always a popular view, but as I get older and have multiple interests my desire is for stuff that just works. Playing with machines and software is fun and rewarding when in the mood, but when I want to make a part I just want to be able to hit the power, feed in the gcode, find home and work coordinates, hit run and have things work. Makes me lazy I guess. I spent a good part of my life doing complex and complicated stuff, now I just want to play.

To toss a wildcard into the mix, UCCNC is or at least was quite popular with a number of commercial shops running retrofitted machines that don't incorporate proprietary controllers.

I don't think there is a "best one" motion control software for everyone. It's like linux distributions were for years - pick one and master it or flit back and forth never really feeling completely in control. If you start with the free or at least lower cost options you will get results pretty quickly, and within a few weeks of regular use you'll see if you like, tolerate, or hate the first one you try. Pretty quickly you will know which items are your big things based on how you work and on the machines you use.

Sorry not to be comfortable making some one true way sort of statement.

Cheers,
Stan

 

[Leonardo]

Hi Brian and all,

I developed a tabletop 3 axes CNC milling machine some years ago. It is a quite strong machine with the frame cast in aluminum.
I have some sets of cast parts in two stage of finish. Six of then are just the cast parts and other four are mechanized end paired.

Do you think that this might be interesting? If so, I can star a new thread giving much more information.

I am attaching some pictures to show you the finished machine.

 

[timo_gross]

Brian, take a look at the cam grinder I made, there are also others on YouTube if you search for "model cam grinder" or "grinding",
"radial 9", "ken hurst", and "ruud oort" made grinders where the cam blank and the master-cam are on different shafts, IE different (parallel) axies
"mostlyIC" (me), put the cam blank and the master-cam on the same shaft/axis,
I got the idea from Bob Hettinger, and others, but aren't on youtube, just mine,
and IMHO is simpler and easier to make, though making a quality master cam for
this type can be a trick (but just takes patience)

I will be glad to share the plans for mine (though it might take a bit for me to tidy them up and make sure they're accurate), but you can see that its a Dumore toolpost grinder, a Sherline lathe bed, an ER-16 collet holder with a long straight 3/4" shank, a micrometer head, etc, and basic machining skills.

I don't have any direct firsthand experience with CNC, but I was a professional computer software (and sometimes hardware) engineer until I retired, and can say that people without a good background in computer software, hardware, and electronics, shouldn't be trying to make a computerized anything, I've read the sort of problems such people have on various forums, its mind-boggling. And building a computerized anything is an entire all consuming hobby in and of itself, you'll never have time to do any actual model engine building. IMHO, YMMV, VWPBL, yada, yada, yada...

I wonder if a 3d printed "template cam" (how is it called? ) can work on this kind of grinder? That would solve the CNC portion of the Task, because lot of people have access to it.

 

[timo_gross]

..... people without a good background in computer software, hardware, and electronics, shouldn't be trying to make a computerized anything, I've read the sort of problems such people have on various forums, its mind-boggling. And building a computerized anything is an entire all consuming hobby in and of itself, you'll never have time to do any actual model engine building. .....

I disagree: I think to learn some computer skills is not "out of limits, even with no existing background".... but.... yes it will become an "all consuming hobby", it will eat up lots of the time and energy. ... and some people will find out that this "is not for them" and give up. ( for the hobby in a relaxed and good sense, pantograph, tracer machines, geared motion control etc. are also fun solutions and some people like this better anyway )

... It's not always a popular view, but as I get older and have multiple interests my desire is for stuff that just works. Playing with machines and software is fun and rewarding when in the mood, but when I want to make a part I just want to be able to hit the power, feed in the gcode, find home and work coordinates, hit run and have things work. Makes me lazy I guess. I spent a good part of my life doing complex and complicated stuff, now I just want to play. ...

I think this view should be more popular, I mix as I go along! Manual lathe, CNC-DIY (aka never working always project in maintenance and problem discovery) and commercial ready to run CNC.
Keeping in mind: Even "only programming" a working CNC (without home made problems) can already be "challenging" and "not fun at all" if one does not like computers. (me dreaming of a horizontal manual mill )
I got very frustrated with my hobby grade desk top CNC mill. At one point I went out and bought a machine that just works.
After the working machine arrived the "hobby machine" became fun again, because I can tinker with the not working machine and if I am not in the mood to convince it to work, I just switch it off. The "working competition" makes the part the DIY machine is just for learning and tinkering.

Greetings Timo

 

[Jasonb]

Brian, not making solid one-piece cam shafts is eventually going to bite you, this isn't your best long term option. eventually this is what you will want, IMHO, YMMV, VWPBL, yada, yada, yada...

Just add a 4th axis in that case.

 

[ajoeiam]

I have to warn you, I used to be really up on the whole CNC CAD/CAM world, but other things have taken my fancy in the last several years. Using CNC tools still matters to me, but building, tweaking, and discussing / arguing fine and subtle points of it all just doesn't excite me the way is used to. I was never a CAD/CAM/CNC master, all of my needs were and remain fairly simple parts, particularly in the realm of this groups focus. The vast majority of my needs were met using VCarve and Gearotic, with occasional use of various CAD and CAM packages for special needs. Alibre is about as exotic as my CAD needs get.

Sad news on the CAM front, I've never found an open source CAM package that worked well for me. However : Meshcam has worked decently under Wine, so that may work for you. Grab the 30 day free trial of CamBam as well, last time I looked it was well behaved under Wine as well. It seems that these two packages have remained popular for quite some time. The latest versions of VCarve have not worked under Wine for me, older versions (up to 10 maybe?) were fine with Wine.

For control software these days unless you need more, gRbl based solutions may work well enough for hobby class machines. It's so popular as it's the default on so many popular routers and laser cutters that most everyone supports it to some level. Critical user mass and good enough to use can sometimes trump elegant but complex and/or expensive options. The all in one gRbl boxes that are of good quality are around $200 to $250, probably less expensive that step generation hardware and stepper drivers for linuxcnc in many cases, certainly less expensive than Mach4 with the required hardware.

I used linuxcnc for quite a few years, but the lack of conversational modes and wizards was an minor but nagging issue for me. Getting advanced features working well can "always be done" if you really grok the whole linuxcnc mindset. It's a good package. I had used Mach3 and it was OK but somewhat unstable and always felt cluttered. I went to linuxcnc and as a pure motion controller being fed gcode it was perfectly satisfactory. This was in the days of parallel ports, so a while back. Mach4 has been quite good, stable, and those who do a lot of probing say it is quite good in that regard. My friends who drive larger machines tend to use linuxcnc and Mesa boards to run their converted Bridgeports and Hurco class machines. They can take advantage of, and often need to have the advanced options and hardware monitoring features of linuxcnc and complex breakout boards. I appreciate their machines and accomplishments, but I'm not looking to join their group. My playground is in a small retired guy shop, making small things.

With the demise of parallel ports on PC's you need step generator hardware for linuxcnc of one flavor(Mesa typically), and another flavor (Smoothstepper, PMDX, various others) for Mach4. Either way it's $150 to $200 for the hardware and you still need the stepper drivers and related hardware. If you change from linux to Mach or the other way around you need new step generation hardware.

For machines that need less than 3.5 amps per axis the all in one gRbl control boxes like the Openbuilds BlackBox X32 or the xPro V5 are really excellent all in one boxes, easy wiring, all the connectors included, just add a power supply and get to hooking things up. Hard core CNC folks may not think much of gRbl, but it's worked very well on my routers and on my diode laser. 4th axis support is fairly recent, which is why my 4 axis Sherline machine was on linuxcnc for a few years, then I went to Mach4. The ability to do automated gear cutting in either traditional 4 axis mode or profile milling mode was important to me when I had my clock restoration business. I have not played with the most recent gRbl on a mill, but might when I get some play time. My G0704 conversion required more power than the all in one gRbl boxes can supply, that machine has a shop built controller with a PMDX step generator / break out board running four 5 amp hybrid stepper drivers with a healthy power supply. It is perfectly reasonable to break out the step and direction signals from a gRbl controller board like an Arduino or ESP32 based SBC and use those to drive higher current stepper motors if you want to go that way.

Since retiring from the clock business and being wrapped up in our retirement place I've done very little with my CNC mills, almost all of my play has been with 3D printers, routers, and lasers. It's not always a popular view, but as I get older and have multiple interests my desire is for stuff that just works. Playing with machines and software is fun and rewarding when in the mood, but when I want to make a part I just want to be able to hit the power, feed in the gcode, find home and work coordinates, hit run and have things work. Makes me lazy I guess. I spent a good part of my life doing complex and complicated stuff, now I just want to play.

To toss a wildcard into the mix, UCCNC is or at least was quite popular with a number of commercial shops running retrofitted machines that don't incorporate proprietary controllers.

I don't think there is a "best one" motion control software for everyone. It's like linux distributions were for years - pick one and master it or flit back and forth never really feeling completely in control. If you start with the free or at least lower cost options you will get results pretty quickly, and within a few weeks of regular use you'll see if you like, tolerate, or hate the first one you try. Pretty quickly you will know which items are your big things based on how you work and on the machines you use.

Sorry not to be comfortable making some one true way sort of statement.

Cheers,
Stan

Very interesting - - - - thank you for your detailed information!!

Kudos for your time!!!

 

[clockworkcheval]

Hi Ollimatti, the most direct and costeffective way to get a TAIG out here in Europe was for me to get a good friend to take one along from California. The thing is so small it fits in hand-luggage and the cost is so low that in The Netherlands no import duties are required.

 

[mrehmus]

Here is a picture of my TAIG CNC mill. It has a Sherline head and motor, the controls are Flashcut CNC and the frame was heavily rebuilt to keep the Z-Axis from moving. This unit was originally owned by the editor of Strictly IC, Robert Washburn.
I also have a DynaMite mill with a servo system driven by DeskCNC electronics.
I use Vectra's software to prepare files for both systems.
Both systems work well, although the DynaMite is more accurate and faster because of the servos. The TAIG has the original leadscrews so it has about 0.006" backlash which the DynaMite does not have. Both cut good and accurate circles which I think is a good quality indicator.
I've attached an article from Model Engine Builder on cutting cams using standard shop machines. It covers making single and pairs of cams but would not work for a multi-cylinder engine camshaft.

 

[NapierDeltic]

With all respect (and delight) for the solutions above, I come to add another one.
The Electronic Lead Screw for lathe or ELS:

Please take a look at his other videos and consider he has a page on Github with (almost) any detail relating to his nano-ELS
Why I support it?
Because it is not that specialised so it offers much more gain for the bucks additionally enhancing lathe's capabilities.
Because it does not take that much time, effort, money and knowledge invested.
Because it makes use (in cam machining) of existing rigid structure of the lathe and its mechanics.

There should be various designs of ELS (DIY or marketed) including those with CAM control from a PC. (in fact it is a CNC subassembly).

What tool on an ELS Lathe could fit to cam machining remains open, but I think a dumb but solid spindle hold in tool-post would work.

 

[minh-thanh]

Brian, not making solid one-piece cam shafts is eventually going to bite you, this isn't your best long term option. eventually this is what you will want, IMHO, YMMV, VWPBL, yada, yada, yada...

I disagree . I have made both types of camshafts: from one piece, assembled the shaft with the lobe and they are fine with the engine I have made

 

[minh-thanh]

Hi Brian and all,

Do you think that this might be interesting? If so, I can star a new thread giving much more information.

I'm interested, if possible, please create a new thread

 

[Noitoen]

With all respect (and delight) for the solutions above, I come to add another one.
The Electronic Lead Screw for lathe or ELS:

Please take a look at his other videos and consider he has a page on Github with (almost) any detail relating to his nano-ELS
Why I support it?
Because it is not that specialised so it offers much more gain for the bucks additionally enhancing lathe's capabilities.
Because it does not take that much time, effort, money and knowledge invested.
Because it makes use (in cam machining) of existing rigid structure of the lathe and its mechanics.

There should be various designs of ELS (DIY or marketed) including those with CAM control from a PC. (in fact it is a CNC subassembly).

What tool on an ELS Lathe could fit to cam machining remains open, but I think a dumb but solid spindle hold in tool-post would work.

He no longer has this for sale

 

[cwelkie]

I wonder if a 3d printed "template cam" (how is it called? ) can work on this kind of grinder? That would solve the CNC portion of the Task, because lot of people have access to it.

That's what I used for the camshaft in my Cirrus engine. The cam grinder used is one I built as an accessory for my Clarkson tool grinder.

 

[NapierDeltic]

He no longer has this for sale

? But on Github there is everything you need to build one. Assembly diagram is extremely simple for someone with specific skills and firmware is there also; 3d print models and so on. Maybe a friend that can help you... And there are others available; some of Chinese origin. I am really sorry ; maybe its only about re-supply. He is active on youtube.

 

[SmithDoor]

I saw in magazine back 1960's on making a cam.
It used 6" bench grinder a cam from a car a indexing 3 jaw chuck and spring. Use like tracer. He hand crank and slowly ground the cam form. Index for next cam lob and move index setup a long for each cam lob.

This is all per CNC
The downside to CNC is programming and cost.

The CAMS I have made was by hand. They most for heavy equipment.

Dave

FYI it like the one cwelkie
Well-Known Member
Today at 5:50 AM
New
? That would solve the CNC portion of the Task, because lot of people have access to it.
That's what I used for the camshaft in my Cirrus engine.

 

[timo_gross]

I see 4th axis and electronic lead screw suggestions.
How to program all this is a mystery to me. (not willing to spend the "big bucks" ). How to do the camCAM for the 4th axis is a big bottle neck.