[YZF]

Hi all,

First post!

I've been working on a desktop CNC design for about a month or two now. I'm mostly doing this for fun and to learn a little more about machine design and construction. I'm hoping to be able to mill metal on this machine though it's not completely clear to me (yet) what sort of stiffness you need to achieve for that - a lot depends obviously on what sort of cuts you are taking. I'm starting to get a better feel for working with metal which I've really never done before, also did quite a lot of reading on this.

I've gotten a lot of information from these forums so hopefuly someone will find my stuff useful...

It's a combination of scrap I've had lying around, scrap from work, cheap ebay stuff. Trying to keep the costs down as much as I can. The working envelope is going to be about 100mm x 200mm x 250mm or so. Initially I was aiming larger but it gradually became smaller and smaller

The design looks something like this (not all details were drawn):


It's based around aluminum extrusions reinforced with aluminum plates.

Here's a piece of this that I'm building:


It so happened that the narrow surface of this 60x90 extrusion is quite flat allowing for the rails to be mounted directly onto the extrusion. I've made some compromises based on the stuff I could get, e.g. these rails are a different type and I've joined two shorter rails together to match the second rail I had. I hope I won't regret this some day later, travel across the joint seems reasonable, it is mostly the seals that are a bit sticky... As the rails were a different type I had 2mm to bridge between the mounting surface of one set of blocks to the other.

The Z axis is a Festo pneumatic linear actuator I got for cheap on e-bay that I've (somewhat crudely) converted into a screw driven stage. This is another example of a compromise as it looks marginal in terms of stiffness, we'll see I guess... Once I have a CNC machine I can build something better The stepper driving this is a bit on the small size but my calculations seem to indicate it should be OK. It was just a good fit and an easy mount to the actuator.

I got a variable speed router on sale that looks reasonably good:


And I'm waiting for a router mount that I ordered from K2CNC that should fit this router... (3.5")

I'll need to figure out how to align everything once it all gets put together. There are also a few other loose ends to tie up in the design.

Anyways, that's about it, what do you think? If anyone is interested in more details I'll be happy to provide...

Guy

[Kipper]

Do you have a longer extrusion available for the gantry? If you do you could flip your pink bits and gain some travel...Junk is such fun

[YZF]

That's a good point! This piece of extrusion used to be bit longer before I cut it (though not by much as far as I recall) This is your classical thinking outside the box

It's really the linear rails that are currently restricting my travel on that axis. There would also be a tiny bit of stiffness lost. The demand for longer rails on ebay is much higher A lot of sweat went into drilling and tapping all the mounting holes accurately enough with my crude setup so I'm rather attached to this particular piece...

Good upgrade idea though! I'll keep my eyes open for a good longer piece and keep that in case I want longer travel.

Thanks,
Guy

[solodex2151]

I would suggest that you ditch the pink plates for anchor counterbores or tapped end connections. Joining plates work nicely for the bottom, but are really not suited to that top connection. The grey plates on the bottom are good. I think you can get away with 1530 on the bottom sides instead of the 3030 you have now (I am assuming you are working with 15S extrusion). How are you planning on fastening the bottom frame? I would go with tapped end connections or anchor counterbores.

Congrats on the progress.

Nate L.

[YZF]

Hi Nate,

Thanks for the feedback!

It's hard to see in the view I've posted but each of the pink plates has an identical plate on the other side of the joint. These are 0.5" thick plates.
Can you explain a bit more why you think this won't work well? I could optionally throw in more plates to completely box this joint if needed. I like the adjustability I have with the plates...

Another detail not shown is a steel L bracket on the inside opposite the gray plates to prevent the setup from bending side to side.

What are tapped end connectors? Do you tap the center section of the extrusion and then connect to the T-slot? I'm also not familiar with anchor counterbores, can you explain in a bit more detail? Anything that can help in fastening the bottom frame would be most useful right now... I have not worked with these extrusions before but I've seen them used...

The sides of the bottom frame are Bosch (I'm pretty sure) 90x90 mm extrusions. The extrusions supporting the X axis (I hope this is the common convention) on the side are 45x90. The X axis itself is a 90x60mm.

Thanks!
Guy

[solodex2151]

Guy,

Anchor Counterbores use a Anchor Fastener Assembly that slips into a counterbore drilled into the extrusion (machining details 8020 Catalog Pg 386). A tapped end connection is just like the name implies. You tap the holes in the end profile of the extrusion (catalog pg 383). Then, on the connecting piece you drill an access hole(catalog pg 384) in the connecting extrusion. Put a button head cap screw in the tapped hole, slip the head in the connecting extrusion profile. Using an allen wrench through the access hole to tigthen the screw.

You should take a look at the 8020 Catalog Pg 60. 8020 did tests on all of their fastening methods and the 2 I recommended are significantly stronger than using joining plates. Not only that, but they will be cheaper than your plates too. When you use the right connections on 8020, you can treat the entire machine as one solid piece and ignore the connections for FEA purposes. This is a huge advantage.

[High Seas]

YZF/Guy
Just to add:
the 80/20 software/cd package is much more than a sales pitch. Some real good engineering data there too. Well worth a download/mail drop.
:cheers: Jim

[YZF]

Hi Jim,

Thanks for the input Jim and Nate. I had a look at the catalog and there's certainly many goodies...

I already have plates for the top connection so I think I'll go with those initially. I'll let you know how it goes. I'm still thinking about the bottom frame. What happens with those connectors if my cut isn't square? It seems the flatness and squareness of the base depend on the squareness of the cut... It may actually not matter in terms of getting the entire machine aligned but I have to think about it...

While the load ratings of some of those connectors seem really good they don't talk at all about stiffness which is really what I care about.

I'll give an update once I make some more progress... I'm thinking wiring up the motors and the drives is my next step while I consider how everything else will be put together...

Thanks,
:cheers: Guy

[solodex2151]

I already have plates for the top connection so I think I'll go with those initially. I'll let you know how it goes. I'm still thinking about the bottom frame. What happens with those connectors if my cut isn't square? It seems the flatness and squareness of the base depend on the squareness of the cut... It may actually not matter in terms of getting the entire machine aligned but I have to think about it...

I use a Bosch 4410 Miter Saw with a Freud LU90M blade and it works wonderfully. The only thing I changed was to put a shot of WD40 onto the extrusion every cut and onto the blade every 2-3 (they recommend just the blade every 4-5 cuts). Cuts are dead on. I would say you have some fair amount of leeway here, but if you are using a miter saw you should be tweaking it to square anyway. I would say you can get away with some variation, but cutting the extrusion with say an angle grinder just isn't going to work.

[YZF]

Hi,

Things are coming together slowly. A couple of photos attached.

I took a couple of hours last night to try and get the pieces I have together and roughly aligned. I worked on a flat granite surface and used an accurate square to try and get things roughly square (at least to the extrusion faces).

Unfortunately I already had to take things apart again because of a stupid mistake. I'll probably assemble the whole thing and then go from there.
I need to get more fasteners to finish installing all my plates (two missing on the left side), it would be interesting to see how ridig the setup feels with everything in place.

On the bright side I got the router mount and it fits my router nicely and it looks like mounting it on to the Z stage is going to be easy, just drilling a couple of extra holes in that back plate. Not the stiffest thing in the world this mount but I guess it'll do for now as my Z axis isn't that stiff either...

I'm still contemplating how to fully connect the bottom frame. Not happy with any of the options yet (plates, connectors) but maybe some combination will work.

The two extrusions on the bottom are flat but not perfectly parallel. I don't want a connection that will put a side force on them because I'm afraid to bend my nice flat (better than 10um) bridge into a banana. I'm looking for the stiffest connection I can get without over constraining the frame (if that is at all possible). Maybe I'm too paranoid...

Another decision needing to be made is how to mount my third axis (what I think most here refer to as X) on the frame. I'll try and post a picture tomorrow to see if I can get some ideas...

I have stepper drives (IM483) and power supplies but need to hook things up. I was contemplating mounting the drives directly on the frame close to the motors. That would simplify things a bit in terms of wiring but maybe can be a problem with stuff flying into the drives and vibration. I have an old laptop to drive this setup and was leaning towards TurboCNC for the control due to the cost (zero) and support for old hardware, maybe upgrading at some later point to something nicer... Any suggestions there?

Guy

[YZF]

Hi all,

Made some progress. The machine is basically all put together mechanically.
Photos attached. I finished the bottom frame by tapping the extrusions and then bolting with 4 M8 bolts on each side. I'm fairly happy with the end result so far, the frame overall seems very stiff. :cheers:

Not all is so great though... It's pretty clear that as I was afraid the Z axis stiffness sucks. Those 10mm steel rods just won't cut it. In combination with the router mount it all looks way too flimsy, I lean on it and can get a noticeable deflection. I'll probably start up with what I've got, should still be ok for light loads, but I'm considering building a new Z axis with linear rails or at least shortening the current Z axis by a lot to gain some stiffness.

The other discovery I made is that the scanner part I am reusing for my X axis has some weird anti-backlash spring loaded flexture/nut with a fairly weak spring. I pull on it and it doesn't take a lot of force to overcome the spring and bounce the axis by 100um (4 thou) or so... Not great.

I mounted the axis in a rather improvised manner with hardware store angle brackets 'cause I couldn't think of an elegant way to do it. Mounting seems fine though if a touch ugly.

I've some connectors on order from DigiKey to help wire things up. I have TurboCNC copied to the old laptop via a floppy drive! When it comes up it report the frequency that causes 10% jitter (or something like that), is this practically my fastest pulse rate? Can I run faster?

Anyways, the next exciting phase is getting things moving...

Guy

[YZF]

OK! Got all three axis hooked up and running under TurboCNC... :cheers:
I was a bit unsure if the laptop's parallel port could drive the opto-couplers on the stepper drive but no problem.

Some resonance (steppers!) here and there but it's all moving (~40-60IPM) and calibrated. No limits or homing yet though... Another thing I want is to hook up the router to a solid state relay for computer control, not done yet... My "table" is not quite a table yet either...

Now I have to figure out where to go from here. I guess I'll be looking for some simple G code for testing and cutting some wood.

Any advice on what router bits to start with? Router speed? Feed rates? This is all new territory for me.... Given my geometry I think I'll be going for extra long bits.

I guess I'll really want some vacuum to avoid getting dust all over the place... Thoughts on that?

Some sort of bellows to prevent dust getting on screws/guides?

[YZF]

Hi all,

I've decided to start by trying to create some sort of reasonable table. I went with a side slitting saw to try and cut off some of the stuff off my current table (some sort of lens holders). Now the saw would normally need to run quite a bit slower than the minimum speed on my router (which is about 10K RPM) but I figured if I take really light cuts, use lots of cutting fluid, "peck" at those holders, and get out of the way so if the blade flies off I'm not there - I'll probably be OK. It seemed the most efficient way of getting rid of those lens holders though it's kind of a no-no.

I wrote my first G code to do this and started attacking the first holder. To be on the safe side I put a delay in the code and moved out of range of any sharp flying projectiles. First one was cut off just fine, though with some noise, and I was celebrating my first successful cut in aluminum! All the chips looked like they're a reasonable size, the cutter seemed to maintain its sharpness and nothing was too hot.

The next cut however is where I ran into trouble as I was cutting the noise suddenly changes into a very unpleasant loud screech. No emergency-stop, took me a couple of seconds of thinking what to do and I dived for the power cable and powered down everything. Two bad things have happened, first is the cutter actually ran into something I wasn't planning to cut (yet) - I haven't planned my path properly. The second thing was that the Z axis seem to have dropped mid-cut which was why everything went crazy. The screw securing the bottom bearing to the Z axis leadscrew vibrated loose. This has also over-stressed the coupling which promply broke after I fixed everything and attempted another cut.

I need a new coupling now, need to add an e-stop, some enclosure around the machine would be nice too... Live and learn. Be careful out there!

[Kipper]

It's all good...Nobody died :beer: the estop is a great device :yes: Glad you're OK

[Eurisko]

Be careful, my friend. Add an enclosure and always wear safety glasses and hearing protection.

Crashes happen. On the large industrial CNC machines, it is unlikely for an operator to be hurt, if the machine is operated properly.

The homebuilds are a different story. The operator is within inches of a cutting tool rotating at speeds that can literally kill. Imagine a 3 inch slotting saw spinning at 10000 rpm. Now picture this same tool fragmenting and spraying metal shards at roughly 150 km/hr.

Keep it fun. And safe.

[YZF]

An update....

Added a solid state relay for software control of the router.
Added "soft" e-stop button (goes into input pin and TurboCNC monitors that pin), seems to work fine.

I am now discovering the difficulties of actually clamping something down, especially given my improvised setup. I screwed down a couple of blocks of pine as a temporary base and surfaced them with the router and a bit of G-code. Cut like butter.

Photo attached of my first "real" trial. Exported a design as an STL file and used the evaluation version of MeshCAM to generate G-code. For the most part everything worked as expected, unfortunately during the finishing pass I lost my Z axis (coupling worked itself loose). I was right by to hit e-stop but not before the machine put an ugly cut straight through my piece.

I was surprised by how long everything took. I'm taking somewhat light cuts mostly because the piece isn't that secure.

TurboCNC doesn't seem to interpolate well between linear motions so it's kind of a big stop-and-go-and-stop deal. It also promotes some burning. I'm using a 1/4" flat end router bit at ~11,000RPM. Maybe I need a stronger computer and emc? Anything else that can be done?

I've also noticed the tendency of larger pieces of wood to break off at corners or edges, i.e. if you're taking a cut around a corner the entire corner can just fly off. Does this mean I'm taking too heavy of a cut? Need more RPM?

Anyways, it's pretty exciting (besides the part where everything gets covered with wood dust)...

[TOTALLYRC]

An update....

Added a solid state relay for software control of the router.
Added "soft" e-stop button (goes into input pin and TurboCNC monitors that pin), seems to work fine.

I am now discovering the difficulties of actually clamping something down, especially given my improvised setup. I screwed down a couple of blocks of pine as a temporary base and surfaced them with the router and a bit of G-code. Cut like butter.

Photo attached of my first "real" trial. Exported a design as an STL file and used the evaluation version of MeshCAM to generate G-code. For the most part everything worked as expected, unfortunately during the finishing pass I lost my Z axis (coupling worked itself loose). I was right by to hit e-stop but not before the machine put an ugly cut straight through my piece.

I was surprised by how long everything took. I'm taking somewhat light cuts mostly because the piece isn't that secure.

TurboCNC doesn't seem to interpolate well between linear motions so it's kind of a big stop-and-go-and-stop deal. It also promotes some burning. I'm using a 1/4" flat end router bit at ~11,000RPM. Maybe I need a stronger computer and emc? Anything else that can be done?

I've also noticed the tendency of larger pieces of wood to break off at corners or edges, i.e. if you're taking a cut around a corner the entire corner can just fly off. Does this mean I'm taking too heavy of a cut? Need more RPM?

Anyways, it's pretty exciting (besides the part where everything gets covered with wood dust)...

Laptops can be very problematic with Mach3, not sure with turbo. Wood has grain and just like doing it by hand the corners can be ripped of.

11k sounds way slow to me on a 1/4" bit. If that is all you have try a slower feed rate but it will probably burn. Speeds and feeds are interelated and you can only go so slow before it has problems.

On my router I use a spoil board to 1. allow me to cut all the way thru the part, and 2. give me a place to screw into to hold stuff down.

A vacuum is a real necessity, and even after vacuuming, you could vent the vac exhaust outside depending on you setup.

Ear protection is an absolute must, because as you have noticed, the machine can run for a suprisingly long time.

Time ot put on the coffe and wake up.
Mike.

[YZF]

Thanks for the input...

The router can go faster (35K) but my maximum feed rate right now is about 70IPM on X and Y. I thought as I am feeding relatively slow for wood I should slow down my RPMs.

The laptop runs TurboCNC fine but because TurboCNC doesn't support constant velocity it has a jerky motion for complex g-code (contours and such). I've tried increasing the start velocity as much as I could, we'll see if this helps.

[YZF]

I've made a "real" table out of a scrap machined aluminum plate. I drilled holes on a 25mm grid (2 different sizes, for M5 and M8, to accomodate different clamp options) and tapped most of the holes. Broken an M5 tap in one hole and stopped tapping after I got it out :devious:

Not having any real clamps yet I bolted down a piece of wood by drilling holes in it. It was surely clamped down at that point

I switched to a 1/8th router bit and slowed my feed a bit after seeing I was missing steps at the higher feed rates. I ended up at about 45ipm, router probably at about 15K. I cut 0.7mm passes. This was a new sharp bit. It cut very nicely, no burning.

On the to do list I have:

• A proper electronics box (rather than the sprawled on the table electronics I got now)
• Dust/chip extraction
• Enclosure, preferrable perspex
• Experiment with milling aluminum
• New stiffer Z axis
• Perhaps a new X axis at some point
• New computer for running EMC
• Bellows for protecting guides/screws
• I can see I'll be busy for a while

[FandZ]

I've just read your thread and enjoyed seeing your progress. That's a great looking machine and I'm sure it'll only get better in time.

I'm still in the planning stages and am shooting for a table top CNC to do some softer metal and wood work. I plan on making a machine similar to yours using 8020 aluminum. Could you give me some advice on what power tools and/or manual tools I should have before I start my actual build. I'm really concerned about making and drilling accurately. I really want this to go right the fist time and am not afraid to spend a little extra to get it there. Just want to know what I'm up against.