Maximising Yield – the Vacuum Table Story

For months I have been bantering around the idea of a vacuum table for the CNC router, but each time decided that screws or pins were easy enough, and the whole issue stayed in the too-hard basket.

As I have been doing quite a bit of nesting work recently, it gave me pause to thought – for a one-off, a few screws are all very well, but the combination of that, and the significant time wasting of using tabs to restrain the cut components (both drawing them, and then physically having to cut and remove them) was proving an incredible time waster.

So I finally was pushed into addressing the whole material hold-down issue.

I started doing a bit of research online, but the results were less than helpful, and I felt as a whole, a lot more complicated than necessary.  So instead, I decided to build an idea I had, and just see if it worked.

I did use the CNC for the following steps, but that is certainly not necessary, and secondly, while I am using this on the CNC router, there is absolutely no reason this cannot now be applied to other areas of woodworking.  Nor do I expect I have come up with anything novel, but in going back to first principles, hopefully I have significantly simplified the solution.

So to start, I took a thick piece of MDF (22mm or so, which I had to hand.  I would have used thicker, but the 32mm MDF I bought last time from Bunnies was only some of their promotional stock.  Not sure what they were promoting, because they don’t stock it otherwise).  With a 1/2″ ball nose router bit, I cut a matrix of tracks, 5mm deep, and about 20mm apart, both horizontally and vertically, stopping about 10mm from the edge.

Next, the edges of this board were sealed.  I know people use some edge tape, or shellac for this, but I thought PVA glue would suffice!

This board was then screwed down to the bed of the CNC machine, and a hole just big enough for the end of a vacuum hose was drilled, all the way down and right through the table.  The hose of the vacuum (connected up to a cyclone separator) was jammed into this hole from underneath.

A second thick board of MDF was laid on top of this bed, and the vacuum switched on.

Test one – does it suck? Yes it can! The first proof of concept is a winner.

Into this second board I cut the same matrix of slots.  By then flipping this board over, each of the passageways is doubled in size (adding together the bottom and top halves), and also exposes a significant area of the soft, porous core of the MDF.  Each passage is now 10mm diameter, so that gives significant passageways for the air to pass through.

The vacuum was switched on again, and the top surface of this second board (the sacrificial board) was machined away with a surfacing bit (otherwise called a spoilboard bit).  And that is what this upper board actually is – a spoilboard.  When it gets too badly cut up, it can be flattened again, and this repeated until it is too thin, when it is then thrown away and a new board takes over.  By planing away 0.5mm of the upper surface of the spoilboard, the hard, compressed (and more non-porous) upper surface is removed.

Now I have seen a number of vacuum tables, and spoilboards with a large matrix of holes drilled though it.  Don’t need it.  The core of the MDF is so porous, that the vacuum can draw air directly through the MDF.  And that in a nutshell, is my vacuum table!

Upper board (spoilboard) from underneath, and the upper surface of the lower portion of the vacuum table

Upper board (spoilboard) from underneath, and the upper surface of the lower portion of the vacuum table

vac-4

Detail of vacuum table

Vacuum connection

Vacuum connection

I’m using a basic Shopvac for this, so I do have a bit of a concern that this will shorten the life of the vac.  I possibly need a vacuum pump, but this will do in the meantime.  The cyclone separator is to try to keep as much MDF away from the vacuum, to try to stop it being killed even more prematurely.

The proof is in the trial.

With a sheet of 3mm MDF laid on top, the vacuum switched on, and voila – it sucked big time, right through the MDF.  The board to be cut was held firmly, enough to run a trial nesting job.

Without tabs.

It was a complete success.  Other than the noise of the vacuum cleaner, I could not fault the process.  The vacuum will soon find itself in the shed next to the workshop, and switched on and off with the remote power switch I happen to have in there (the actual switch is right near the CNC as it happens).

Test job, no tabs

Test job, no tabs

 

I cut out about 5 patterns in total, and each time it worked perfectly.

Next, I tried another idea.  If the only reason for the material between each piece is to support the piece as it is being cut, it is really necessary if the piece is supported by the vac?

So I ran a large job with a full sheet, no tabs, and only 2mm between each piece (or more precisely, between each path the CNC was trying to follow).  And 5mm from the edge.

The result?

vac-5

Yield

Pretty much nothing left, what is gone is the project, leaving this sad skeleton.

So there you have it – my poor mans successful attempt at making a vacuum table.

AWR

Been waiting for this one to come out – the latest edition of Australian Wood Review.  Has my first of a series of articles on CNC machining for small-scaled use.  This one is a 2 page spread as an introduction to the topic.

So this is the second magazine (this one, and ManSpace) that is on shelves currently with an article of mine.  And by Monday, the third will be out – the latest edition of The Shed.

AWR87coverweb_E537D540-09AA-11E5-89FC0635E51EDD1F

Celtic Knot

I’ve been working on becoming more familiar with Vectric Aspire, which is used to both create objects (including three dimensional ones), and turn existing designs into the code required for the CNC machine.

In this case, I wanted to work out how to make a celtic knot design.

3To start, I have defined a sheet of material 600x900x16mm.

Next, under “Gadgets” there is an option for a “Celtic Weave Creator”

4This creates a vector that can then be used to generate the 3D pattern.

5I have also placed a border around the outside.

Next, under the Modeling tag (the 3D section), click on the extrusion tool.

7By using the existing selection, and a small curve vector (not shown), the 3D object is created.

8It shows starting points, paths and how the paths are dealt with when they overlap.

10By first choosing the 3D Roughing Toolpath and a larger bit to waste away as much material as possible,

11 then the 3D Finishing Toolpath with a ZrN tapered ball carving bit for a final pass.

FWIW, I am slowly entering the CNC router bits I have into the Tool Database.  No point having tools in there that I don’t have.  I’ve been including the router bit number (all being Amana Tool CNC router bits) so I can reference back to the Toolstoday.com website to confirm spindle and feed rates, especially for different materials.

13Finally, with the path calculated, the simulator is run to see what the program predicts the outcome will be.

14Looks pretty good!  I wasn’t going for anything particularly artistic, or complicated here – just learning the basics. All very interesting stuff.

Swiss Cheese

Slowly perfecting my processes, in this case for what is called a profile cut, where the object’s outer border is defined and the CNC router cuts the shape out.

I’ve been using a 1/8″ upcut solid carbide router bit – the 46100 from Toolstoday.com for the job so far, but knowing that it is not the correct router bit for the job.  What I should be using is something like the 46184 1/8″ solid carbide compression bit, or the 46180. (46237 and 46227 are also interesting bits, being 1/16″ diameter, which would be needed if doing nested work in 3mm MDF.)

Few reasons.

1. An upcut bit produces a lot of tearout in MDF.  While I can easily fix this with a quick sand of the top surface before removing the sheet, it would be preferable to avoid that step.

2. The upcut bit, especially at speed and with a large depth of cut, tries to lift the material being cut.  MDF is not a stiff structure, especially in a nesting situation which really turns the board to Swiss cheese.  It becomes almost impossible to stop the board being lifted, so I had to increase the number of passes from 2 to 5, and even then had a few lifting problems.

Not a fault of the router bit, I’m just using it as I don’t have another one of that diameter (or smaller) to work with 6mm thick boards in a nested layout, especially where I need grooves cut in the workpiece that are 5.9mm wide so they can slot together.

I’m cutting at 80mm/sec, and with about a 2mm depth of cut (DOC).  I know the CNC machine and the router bit can easily handle a lot more DOC, but my holddowns cannot keep up.  If I had a vacuum table, or even used a fair amount of double sided tape, that would be much less of an issue.

To stop the individual pieces being cut loose and walking into the cutter, wrecking them, I added tabs to each piece generally 2-3 per piece, 4mm wide and 1mm deep.  These are easily cut and sanded away at the end of the job.

It is important to ensure the cut goes all the way through in that final pass.  I have been using 0.5mm, but am thinking 1mm would work better.  Certainly, that means the router bit is cutting all the way through and partially out the other side, but that is why the tabletop has a sacrificial layer added.

What I made this time is the tropical fish (Angel Fish)……..

CNC-1 CNC-2 CNC-3and a stegosaurus!

CNC-4 CNC-5

It is very addictive!

And just for a sense of scale, here are both projects photographed alongside a bottle of wine (not so easy to see sorry!)

CNC-6

Now I just have 148 designs to go!

Plans from MakeCNC

Surfing the curve

There is so much to learn with this CNC thing. It’s very cool. So much potential. The machine itself, the software, and the entire genre.

Oh, and we have progress on the name, I am pretty confident it will be officially tagged the Torque CNC 9060. The 9060 is the model, and refers to the working dimensions: 900x600mm.

Today I wanted to ensure a perfect working surface, so set the CNC up with the Amana RC2251 2+2 Spoilboard bit from toolstoday.com

IMG_2743

This is quite an incredible bit. 2 1/2″ diameter, and the 2+2 design is due to the two replaceable cutters (RC) that are on the sides, and the other two located underneath the cutter, performing a shaving action.

See MDF has a bit of a problem- when it is made, it is done with a lot of compression, producing a really clean, smooth top and bottom surface. But if that surface is removed, it becomes quite furry, with the MDF fibres sticking up all over the place. These second shaving cutters do just that- they shave these fibres, leaving a “real smooth shave”.

The finish of the resulting surface is superb- perfect for CNC work. For those tables using a vacuum surface, some draw the vacuum directly through the MDF, and to achieve that, the hard original outer surface also has to be removed, and again the finish achieved by the spoilboard bit gives that.

I used the surfacing passes to also check the CNC router out. This is a big cutter, so a good load test. I ran the machine with a 40% stepover (this is the amount of cutter that is cutting into the virgin material, in this case 1″, which is significant. I was only running a 0.2mm depth of cut, because the idea is to minimise how much of the spoilboard is lost each time it is planed flat. I also ran the machine at 200mm/sec. 12metres/min, which is positively hurtling along.

With the surface flat, and importantly, parallel with the overhead gantry, time to try another capability of the machine. 3D routing.

I’m using Aspire to generate the code, which is the premium software from Vectric. They have some 3D samples, so as a first step, I used one of those. In this case, a maple leaf.

Using the Amana Tool ZrN 3D carving bits from toolstoday.com the result is rather impressive.

IMG_2746

Some first impressions

Very early on in the piece, but some initial thoughts.

The CNC Router is bloody heavy – not sure if it is 200kg, but it is a substantial machine, and for this sort of tool, the heavier the better!  There is no flex, nor uncertainty in the movements.

I’m having a little difficulty with the interface side of things, but that is a Windows machine issue.  Wish these things had Mac software to drive them!  Once I have gotten the computer up and running, switched on the CNC, then plugged it into the USB, it seems to run smoothly.

The controller is particularly heavy duty, and although the control box is large, it has plenty of empty space inside.  Not a bad thing for heat dissipation.

Love the fact that the software starts and stops the spindle – that is a nice feature.

There are still some fundamental things I don’t know about the machine yet – spindle speed range for one, achievable resolution for another.  Standard maintenance practices for the machine for a third.

The machine is designed and built by Keith, from YAS Engineering, who is also the inventor behind the Torque Workcentre. Quite the mechanical genius for these sorts of things!  This is one of the smallest CNC routers Keith has made, most are made to order, and some are monsters.  This model however is being made as one you can purchase off-the-shelf as a standardised design.  I just have serial number 0001!

Again it is very early days, but the comparison thus far between the CNC Shark Pro and the YAS Engineering CNC (don’t know what it will be officially called as yet) puts the YAS machine in a completely different league.  Yes, overall it is about $3000 more, but that extra 25% price (price of the CNC Shark Pro Plus) is more than justified by the significant build quality difference, including the quality (and quietness) of the spindle.  I’d say the 250% difference between the two, but that is subjective only!  The previous job was run at 11pm at night.  The entire neighbourhood would have been banging down my door if it was the Shark running the job – that Bosch router is an absolute screamer (and is only 1/4″).  I was also running the machine at about double the feed rate of what the Shark was performing at, and I was limited by the strength of the router bit, not the maximum potential speed of the YAS CNC.

Just jotting down some thoughts as I continue to get to understand the new machine.

SSYTC080 First Test Run

The first run of the CNC Router, using an existing code provided so I could test the setup.

The bench rocks quite a bit which is apparent in the video – this will be corrected by securing the bench to the shed itself.  Lots of momentum in such a heavy tool (and gantry).  The machine itself doesn’t suffer from it because of the size and quality of the stepper motors used.

Guard removed for visibility.  This is a quick test of the CNC, so the camera was set to record without particular attention being paid to the lighting, sound or quality, so don’t assume too much on that regard!

%d bloggers like this: