Building B-Carve

Getting a Z axis as a kit seemed logical as a start. It would save time, it would be a considerable upgrade over the X-Carve’s stock Z and could even serve as a drop-in replacement for the X-Carve in the beginning (as it did for a while). Also it was relatively inexpensive.

Even though its linear bearings (of the LM12UU kind) were one step above the V-wheels, they were no match to the profile linear blocks the X and Y were using. It felt like the Z was holding the whole machine back. I could flex the spindle with moderate force and started having doubts about their longevity. Here’s all 3 kinds for reference:

Also, it did not feel 100% homebuilt with an off the shelf assembly :slight_smile:

Time to correct all the above. Here’s what we’ll be building.

Here are the ingredients.

7 Likes

There are two reasons behind choosing the 2020 extrusions.

Reason #2 is for mounting accessories on either side (homing switches, drag chain bracket, dust shoe, mist).

Reason #1 is to bring closer the top face of the rail blocks to the top face of the ballnut mount. There’s another 1mm till they become coplanar, which can be achieved either by adding shims on the blocks or by milling a shallow channel on the bottom of the spindle plate. I chose the latter.

…But first things first.

In the previous setup the Z axis was slim and lightweight enough that I could get away with only one pair of rail blocks on the X. Not any more. The new Z weighs around 4.2kg (9lb) without the spindle mount, spindle and motor which add another ~3kg (6.6lb). I also want to have the option of a VFD in the future and that will also be heavier than the Makita.

Let’s add another pair of blocks on the X.

and make a new gantry plate.

We start the actual Z axis by aligning the extrusions.

Then slide the rails in. We will use the precision ground base of each rail (marked with the arrows) facing the inside.

These particular rails are 15mm wide. The 2020 extrusions are, well, 20mm wide. We sandwich the rails between two stacks of 2.5mm feeler gauges and tighten the first screw (not to final torque, we want the rail to be able to pivot slightly).

Before we move any further we need to add end stops. Derailing a block will cause the ball bearings to fall out.

We calibrate the first rail and tighten all its screws.

The inner rail-to-rail distance was designed to be 80mm in order to use a couple of 20-40-80 metric blocks (minimum rail-to-rail distance is 71mm, so a pair of 1-2-3 blocks on their 3" side could work as well). This along with placing the arrows facing each other ensures that the rails end up parallel. This is the most critical step of this mini build.

Tighten the screws of the second rail, add block and end stops.

We put together the ballscrew assembly.

Then place it loosely on the Z plate.

Add the spindle plate. Move the spindle plate all the way to the top bearing block, tighten the top bearing block. Move the spindle plate all the way to the bottom bearing blog, tighten the bottom bearing block. If we tighten both and skip this step, tolerances are so tight that the slightest misalignment will add friction to the ballscrew.

Assembly is pretty much done.

8 Likes

Just logged in to say this whole build is great and inspired me to tackle a new build of my own (c-beam sphinx inspired ballscrew build). Please keep posting and keep up the good work. Thoroughly enjoying :slight_smile:

@BrendanCoutts

Thanks :slight_smile:

The sphinx is a great design. Look for Kyo’s channel on youtube if you haven’t done so already. Out of curiosity, what size of ballscrew are you considering? I believe even the smallest ballnut mount (that of a 1204 screw) is too wide to fit in the 20x40mm channel of a 40x80mm C-Beam. I wish there was a 60x90mm C-Beam. I had to use three separate 30x60mm extrusions in a C shape for the X and bind them together.

The mounting screw holes on the gantry plate are M5. The mating ones on the Z plate are 5.5mm dia. This extra 0.5mm allows for something less than 1deg of adjustment, which is more than enough.

The measurement below is taken off the left Z rail and is referenced off the top X rail.

Btw, If you visualize the 3 motion axes of a cartesian machine floating in the air without any extrusions around them, it’s just them that have to be parallel (intra-axis) and perpendicular (inter-axis) to each other. Everything else is just structural and its orientation is irrelevant …as long as it assists the motion axes (rails in this case) to be square. This is not the case for the X-Carve though, as the Makerslides are motion axes AND structural elements at the same time.

Plus spindle mount, spindle, motor and dust caps.

Quick test. Jogging @ 10, 1, 0.1, 0.01mm.

PS. I’m not a native speaker and it looks very strange whenever I write “axes”. It makes me think of Vikings. Google says “Interestingly, axes is the only word in English that can be the plural of three different singular noun forms–ax, axe, and axis.”

2 Likes

These are polar axes, not cartesian.

No, wait…

3 Likes

*unsubscribe… can’t stop drooling

:smiley:

the pictures and the detailed steps of the build is top notch

Great progress Elias - that CNC look top notch :slight_smile:
I have so many ideas for upgrading mine but at the moment no time to implement…!

I like the sphinx a lot! I made one using my x-carve to cut the plates, not i’m working on mach 2. I’ve already purchased the linear rail and 1605 ballscrews. You’re right, they don’t fit inside the channel but i’m not fussed. Y axes go on the outside of the plate and X axis ballscrew sits above. I’ve cut test plates and things are looking good so far!

Speaking of time, I forgot to talk about the other parameter: $

The axis parts (not counting the motor and spindle mount) cost around ~$200 plus the 1~2month time penalty for waiting from aliexpress. Probably around ~$300 if you buy from a local distributor.

It’s a different class in precision and rigidity, but not a different class in $. For what it offers the cost is very reasonable. If your machine has the rigidity to take the extra weight, I would definitely recommend considering a similar solution.

Btw, the 65mm spindle mount is german (Isel) and at EUR39 (~$45) not that more expensive that the chinese ones. It’s a beast and also acts as a heatsink!

1 Like

Switched to slightly bigger drag chains which also unclip on one side (very helpful if you need to insert/remove something with connected ends… or want to untangle the contents to take advantage of the whole space).

Passed a pair of 6mm tubes (1/4" would work too). This is TPU tubing, because it needs to flex at a relatively small radius (that of the drag chain).

Got a Fogbuster head. Some people make their own (<$100 with tank & fittings), some people get a Fogbuster kit (>$325). I just got the head ($130)

and picked my own tank & fittings (~$50).

I did not opt for the quick mount, cause it’s not quick. It requires 4 screw adjustments.

I picked a small magic arm instead. One adjustment to rule them all.

The Fogbuster is 1/4-20-ready, no mods required.

The other side goes to a 1/4-20 T-nut which has been slightly ground to fit the 2020 slot. (Remember “Reason #2” earlier?)

Added a non-return valve on the coolant inlet to make sure it does not accidentally siphon off.

Done.

6 Likes

can’t wait to see this in action :slight_smile: and the clean up afterwards :frowning:

Here you go. Milling 10mm 6082T6 (95 Brinell hardness) @ 3cm^3/min (0.18in^3/min) MRR.

(Turn your volume down!)

Profiling… chips best seen on the shaded (left) side.

Trochoids.

A bit of both.

The coolant flow is turned rather low. What you see here is mostly wet air. I tried more coolant initially and it was not noticeably different in chip clearing (and sound). It was just flooding the T slots and soaking the sacrificial board.

Clean up is surprisingly easy. Cutting at full DoC creates long stringy chips. The chips being slightly dump make it very easy to grab a bunch at a time, like picking up a fallen nest.

6 Likes

Great thread man so I am sorry if this has been answered before

but did you come in at your 1.5k budget

and will you be releasing a parts list and maybe 3d model of the machine?

thanks

First, the easy question :slight_smile:

It was ~$1.6k with the first Z axis kit. If you deduct the cost of the first Z and then add the cost of the homemade Z, it’s ~$1.7k.

Now, the hard question :slight_smile:

I’m a big fan of open source. I use open source tools for work and fun daily. My first reaction would be: “sure, here are the plans and here’s the BOM”. But I am reluctant to do so. Not for reasons of intellectual property. No wheel was invented here.

I can think of 4 good reasons:

  1. Cost. The amounts mentioned above were the result of carefully picking what to order locally, what to order within the EU (no customs in my case) and what to order from the Far East. I was lucky not to pay a single penny in customs even from shipments from China & Hong Kong. Depending on somebody else’s sources, patience (some shipments took up to 2 months) and strictness of his country’s import processes this may well end up 30% to 50% more.

  2. Lengths. The extrusion lengths depend on the ballscrew lengths, their end machining, the side plate thickness and the selection of rail blocks. I got HGL15 blocks, but HGH15 are more widely available and are thicker. The HGH20 ones are even thicker.

  3. Alignment. Squaring and alignment is almost surgical. Good quality metrology tools are needed to achieve good results. A hair of misalignment might cause friction. A slight misalignment might cause stalls.

  4. Support. If I sell plans, people will come back for support. If I put a donate button somewhere, people will come back for support. If I give them away for free, people will come back for support. Individually. They will still pay $1.7k to $2.5k. There’s not enough hours in the day for that. I have to play princesses, bake cookies and build a flying fox.

However…

I decided to document this build process in order to share thoughts, receive feedback, get inspired and hopefully inspire others. There might be no recommended lengths, but all the part types are mentioned. There might be no plans, but all the axes (watch out! Vikings!) are depicted. Look at:

  • post #55 for X
  • post #173 for Y
  • post #198 for Z

This is all you need to “solve” this build. Or even better improve it to something that suits you. Part of “solving” such a kind of build is going through the process of drawing each axis (in pencil, no need to CAD), apply theoretical lengths of their parts… and if something turns our different in the mail to be able to compensate for the difference (e.g. by making your own spacers). Then you really own it.

I hope this makes sense.

4 Likes

oh yeah man I appreciate your in-depth post and I do agree with you

I really think if you enjoy doing this sorta thing and there is time left over after building that flying fox (which I want to see) maybe you can start a company when you build machines and sell them to the public

it takes alot of effort to source materials and supplies I have tried with machine designs of my own and alot of times I get frustrated and loose interest lol cause I am not passionate about it but you seem to really enjoy it

here is a 3d printer I designed but when it came to the parts sourcing and doing calculations on ball screws etc it kinda died lol

but yeah man keep on keeping on!!

This is a small 750W (1HP) 24Lt (6gal) compressor. Not too quiet, not too loud (65db).

Its pressure switch came with a factory cut-on/cut-off pressure setting of 6/8bar (87/116psi). The Fogbuster is designed to run in the 10~20psi (0.7~1.4bar) range. With the regulator set at the low end of this range and the switch at the factory cut-on/cut-off setting, the duty cycle was 2:1 on:off. Not good. The motor would run too hot too soon. There’s no point in forcing the compressor to work at so much higher pressure than the desired tool pressure. This results in high energy consumption, wear …and noise.

After tweaking the pressure switch, the lowest cut-on/cut-off pressures I could set were 3.5/5.5bar (50/80psi). This delivered a duty cycle of 1:2 on:off. Big improvement, but still running unnecessarily high compared to the required end pressure.

Out with the old one.

It took me forever to find a 4 port air compressor switch that allows for low(er) settings. Here’s the chosen one next to the stock one. It’s a Condor MDR 1/11 (I believe the US market equivalent is a MDR 11/11).

Fitted.

The same crimped connectors fit just fine. Plus relief valve, gauge and regulator.

The old one had only a cut-on pressure adjustment screw. The cut-off pressure was fixed at a 2bar (28psi) differential (i.e. higher than the cut-on pressure). The Condor has independent cut-on and differential screw adjustments.

After tweaking for 15mins, I ended up with a setting of 1.6/3.6bar (23/52psi). The duty cycle went to 1:3 on:off (53sec on, 160sec off). Now after one hour of operation the motor is warm but ok to touch. A cooler motor is a happy motor. Happy bill and happy ears too :slight_smile:

1 Like

New Z, new dust shoe.

The old requirements remain:

  • fixed type for independent router Z movement
  • split type for easy tool changing and probing
  • no additional tracks

Plus a new one:

  • small footprint for light weight and less interference with workholding

Btw, this was the first time I used workholding <1mm far from the toolpath (far left screws). A bit nerve-racking… but in CAD/CAM we trust.

Plus magnets. I had used 4x4mm (dia x height) before. 5x4 this time. More than half extra holding force.

Gen1 (for the X-Carve), Gen2 (for the B-Carve w/ kit Z), Gen3 (for the B-Carve with new Z) side by side.

Plus screws, right angle extrusion connectors, brush, vac port.

Front off for tool change and probe.

Front on and ready to go.

9 Likes

Took apart an old V wheel

found its center

and removed the middle ring that separates the two bearings. Here’s a non-modified next to the modified.

Placed a single bearing in the middle.

Found some acrylic coasters sitting around.

and made two pulley sides.

The small protrusions push against the middle cylinder of the bearing and also leave a 0.5mm gap on each side for the V wheel to spin freely.

3mm accessory cord. Fixed side.

Adjustable side.

Keyring carabiner (pretty much worthless for any other use).

Miniature zipline for the vac hose.

Which brings us to the end of Season 2. Thanks for watching! :slight_smile:

11 Likes

Bonus material…

First, let’s make a brake. This basic block worked ok, but it looks meh. Let’s correct this.

Stock.

Fixture.

Bolted and dialed in. Clamp too.

Stock to size.

Flip. Pockets.

Pilot hole, flip, repeat. Finish on the drill press.

Profile.

Chamfer, flip, repeat.

Off the machine.

Grit 400, grit 800, finishing.

Complete brake assembly: fox + cord + bungee + accessory cord + tent stakes.

Fixed side.

Adjustable side. Mezzo poldo (3:1 mechanical advantage) for tightening.

Fox on.

Tent stakes pitched. Plus a couple of round gaskets (not seen) to prevent metal to metal impact.

Trimmed handlebar plus grips, two belaying carabiners, cord and twin pulley (or golden heart according to Nelly).

Two clove hitches in the middle.

Pulley on. Plus backup. Everything is rated between 24 and 30kN (i.e. can lift a truck), but it doesn’t harm to have a backup.

Volunteers.

Wheeee…

10 Likes