Open Source Ecology Forums

mrsquish

After reading about the idea of a torch table, and thinking of how it
could be implemented, I came quickly to the conclusion that a large
proportion of the work required to build the torch table is actually not
specific to cutting metal at all, it is to move a tool (in this case an
Oxy/Acetalene or Plasma torch) around 3 axis, x y and z.

So, if your going to go to the trouble of building an x, y, z
table why stop at torching? Why not build it in such a way that multiple
tools could be used with the same table? In addition to a torching some
tools that come to mind are: Welder, drill press, grinder but if you
additionally extend the use of the table to using wood well then you can
add; circular saw, router, nail gun, jigsaw? sander?

I guess what I had envisioned would be a hot swappable tool set.
Each tool would have a generic coupling mechanism that would allow for
electrical and or hydraulic power. The tool set would sit within the x,
y, z table's work area (albeit off to the side) so that the table's
movable head could move to the "tool box" drop the tool it was already
holding and pick up a new one.

The upside to designing the table like this is it's ultimate
flexibility. The x, y, z capability is built into the table, not into
the tool which means reusing the table for different purposes is easy
and cheaper than redeveloping a new complete table for that purpose.

Wolfrick

What you suggest is a good idea in principle, but there's a very serious issue you'll face if you try to convert a burn table into a general-purpose Cartesian machine for other toolheads.

The problem is precision.

What's good enough for cutting steel with a torch is not likely precise enough to suit someone who wants to, say, use a router to cut detailed wooden parts.

And if you try to increase the precision of the torch table's mechanism, it stops being a torch table and becomes a much more robust (and expensive) machining center.

Much is possible, if you're willing to live with a reduced precision. When making parts with the burn table, you're likely going to burn oversize and then grind to the final dimension for a nice fit. Doing so with a circular saw would produce less satisfactory results than just having a hand-held circular saw and using the regular manual method to produce your parts.

This is not to say you can't use a burn table to fabricate a general purpose machine center with interchangeable heads. But the level of precision and repeatability required of a burn table is less than what's required of a satisfactory general purpose machine center.

$0.02

~Rick

Edited to add: I should note that I have not yet studied the OSE/FeF burn table design in any detail.

mrsquish

Hey Wolfrick,

Thanks for your $0.02 :)

I'm going to preface what I'm saying with the fact I've never actually seen and cartesian machine of any type, so I have no practical experience whatsoever.

But from the vids I've seen of Plasma cutting tables that people have built themselves, they look pretty precise. I guess it comes down to what the torch table will be used for, if it is just for cutting 1/2 inch+ steel and an oxy/acetylene is the tool of choice then maybe, but if it's going to be used in the future for other purposes whereby a plasma torch is used on thinner material for tasks that require precision then I believe that the extra upfront expense is justified.

Could you tell me where the extra cost for precision comes in? Most of these tables that I've seen use stepper motors, which I thought were accurate enough for any purpose?

Cheers,

Adam.

Conor

I think you're absolutely right.

In fact, the RepLab development is already sort of leaning in that direction. The plan is to use the same x-y-z mechanism from the Torch Table for the CNC lathe. And some open-source laser cutters and wire extruders use the RepRap x-y-z mechanism, which was originally designed for extruding plastic.

If we could design a Cartesian table that was precise and also pretty big, it could be used for the laser cutter, plasma cutter, 3D printer, lathe etc. This would make the RepLab workshop smaller, cheaper, and cut a huge chunk off the development time for the tools, allowing us to get the project finished on time, by the end of 2012.

Call it the Fab Tab!

Wolfrick

There are three aspects of, let's call it "quality" with a Cartesian machine:

Accuracy: How close is the toolhead to the intended position?

Precision or Repeatability: How wide is the bellcurve of the distribution of the accuracy over a number of attempts?

Rigidity: How rigidly does the machine resist forces that would distort the intended perfection of the Cartesian coordinate grid from the reality of the position of the toolhead versus the workpiece.

With a torch table, you're cutting metal to create weldments, and those weldments are likely measured in 1/8ths or 1/16ths of an inch. 1/16th inch = 0.0625" or in machinists lingo, "sixty three thousandths". The toolhead is either an oxy/acetylene torch or a plasma torch, so has very little side force or force of any kind other than the mass of the torch head and the mass of the cables and/or hoses that attach the toolhead to the supporting machine.

With a machining center, say a mill or a lathe, you're applying a cutting tool to a workpiece, and this produces considerable side forces as the tool pushes against the workpiece. The rigidity of the machine resists these forces and enforces the rectangularity of the (imaginary) Cartesian coordinate system. The workpieces for typical mechanical parts created on machining centers are measured down to the thousandths of an inch (0.001 inch), and a typical tolerance for a mechanical part might be five or then thousandths (+/- 0.005"). To fit a bearing to a shaft or mate two parts, the tolerance might be half-a-thousandth, or +/- 0.0005". As you can see, the "nearest sixteenth" for the torch table is a far cry from the accuracy and precision called for with mechanical parts on the machining center.

An example of an inappropriate conversion of a tool to a multi-use roles is using a drill-press as a milling machine. It seems simple to place an X-Y table on the table of a good drill-press and chuck up an endmill and go to town.

The problem with this is that the drill-press is not designed to handle the side forces of cutting with an endmill, and will break or fail under this kind of use unless only very light cuts are attempted.

Taking a drill-press and attempting to beef it up into a milling machine is likely going to be more expensive and less satisfactory than building or buying a purpose-made milling machine.

TL/DR: A drill-press isn't a milling machine, and a torch table is not a machining center.

Yes, the torch table design could probably be tightened up and rigidified to the point where it can do SOME CLASSES of machining work, like routing wood or MDF or foam, but it will never be a _good_, _accurate_ machining center for metal unless it is limited to _extremely_ light cuts, and thus would be very slow to use.

I highly approve of the kind of thinking that led you to want to use one tool in multiple roles. Keep doing that! But at the same time, it's important to understand the limitations and dangers involved.

Conor

OK, so one Cartesian machine can't do everything. What about two Cartesian machines?

Conor

Also, I understand your reasoning for why a drill-press couldn't be used for milling, but what about the other way around: using a mill as a drill press?

mrsquish

Hey Wolfrick,

Aah, I see. Thanks for taking the time to explain it to me, I appreciate it.

Ok I now understand that it cannot do the work of a mill or lathe but given the required accuracy of a torch table, and on the assumption that the table would handle forces upto those that would be applied via an average human wielding the same tool, I would think that there are still some tools that would still be relevant?

Drill : No side forces : can be used to clean up holes made with the oxy/plasma torch, can be used on metal tubing, can be used on timber, chipboard or MDF for a variety of purposes.
Grinder : Some minimal side forces : can be used to clean up edges created by the oxy.
CutOffSaw : No side forces: can be used to cut materials that are not flat (square tubing) to relevant precision measurements without needing to measure.
CircularSaw: No side forces: can be used to cut timber to relevant precision measurements without needing to measure.

Having said that though, I think the drill is the only tool that wouldn't always require the material to be clamped in some way. The implementation of the clamp would definitely make the table pretty complex.

Wolfrick

Conor: Yes, a milling machine makes a very good, very precise (and very expensive) drill press.

If you have one milling machine, you don't also need a drill press, unless you have two or more people who will be using tools at one time. A major disadvantage of multi-role or combination machines is that they can generally be used for only one task at a time, and can require quite a bit of setup and reconfiguration between jobs to switch roles. For example, it might be better to buy a small mill, a small lathe and a drill press, instead of buying a mill/lathe/drill combo, depending on what you're going to be doing with the tools.

MrSquish: It is certainly possible to design a Cartesian machine for the requirements of multiple roles. Doing so would make a very useful multi-purpose machine for a certain group of machining functions.

But building such a machine would be more expensive than building _just_ an acceptable burn table.

One must decide what is needed, and what one wants to afford.

But a burn table could be used to make a second generation, stronger version general purpose Cartesian machine. A self-replicating machine should be able to produce parts for a better successor machine :)

Remember, early machinists often built their own tools without access to other advanced machine tools. A machinist (or blacksmith) might move to a new location with only some hammers, tongs, files, etc. and build up his machine shop from stock materials, castings and elbow grease.

We should be able to bootstrap our way into whatever level of advanced tools we find necessary, but I hope we can use technology and digital fabrication to smooth the way and skip a few steps here and there :)

The main takeaway point I'm trying to make is yes, we can have multi-role tools, but they'll be at least as expensive as the most demanding role requires.

Howdy, Stranger!

Categories

In this Discussion

Tagged

Open Source Ecology Forums