6-axis Horizontal Machining Center for Education

[Goemon]

Goemon,

I think we are comparing apples and oranges here. As far as I can tell, neither Datron nor Haas sell a true 6-axis machine. Furthermore, the Datron machines are vertical, not horizontal. And the Haas machines do not have absolute encoders on their linear axes, unless you add them as an option, which starts at a whopping $15,895. You also do not get a high-speed spindle on a Haas, unless you go for the CM-1, but in that case you get a cutting envelope that is 8 times smaller, and your spindle is 3 times less powerful. And most of the rotary axes on these machines use regular servo motors, not torque motors. I would also contend that the SINUMERIK platform is a lot more powerful than the Haas platform.

The following spreadsheet outlines our design and components in details:

https://docs.google.com/spreadsheets...gid=1109769669

If you spend the time to review it carefully, you will see that very few machines are built with this type of components in that price range. In fact, as far as I can tell, there is none. For example, once you load the Haas CM-1 with all the options that would bring it closer to what we are building, retail price reaches $113,915, and you still have an envelope that is 8 times smaller, and you still do not have linear encoders. And did I mention that your piece cannot weigh more than 1.4kg once you put it on the TRT70 2-axis rotary table? I have yet to do all the maths, but it's quite likely that our rotary table will support pieces with weights in excess of 250kg...

Regarding the fact that we are planning to sell a kit, I totally agree that it will reduce the addressable market, but this is just a way to get started. If the machine works, nothing would prevent us from selling fully assembled machines if there is a market for it. And I am convinced that there is a market once people understand the benefits of a true 6-axis machine.

Regarding the focus on the education market, it's just a way to get started as well, because it will be very difficult to convince a commercial shop to make a gamble on an unproven company with an unproven design. But a school will look at this type of risk under a very different light. For progressive-thinking schools, the availability of the machine as a kit is actually a plus, because they can use it as a training opportunity. And the release of the machine's design into the Public Domain is a huge benefit, because it will allow their students to extend the machine. This is how Open Source software works, and this is why it is prevalent in academia. I do not see any reason why something similar could not be done with hardware, and it's already happening for electronics with platforms like Arduino.

Finally, as mentioned in earlier posts, I believe that an entry-level version of the kit can be sold for $80k to $90k once you get some volume (10 or 20 units) and you switch to more affordable components (NUM instead of Siemens, Fagor instead of Heidenhain, HIWIN instead of NSK). But from an engineering standpoint, it will be a lot easier to go downscale rather than upscale, because a machine's accuracy is only as good as its weakest component, and when you start from the bottom, you have no way of knowing which one that is.

It doesn't matter what the capability is. I would expect anything costing $150,000 to come ready to use. I wouldn't expect performance to be reliant on my assembly skills. I can't think of any top quality products in life that require user assembly. A kit car will never be worth the same as a Ferrari, even if it's faster....

[ishi]

It doesn't matter what the capability is. I would expect anything costing $150,000 to come ready to use. I wouldn't expect performance to be reliant on my assembly skills. I can't think of any top quality products in life that require user assembly. A kit car will never be worth the same as a Ferrari, even if it's faster....

I understand. That's why we'll provide both options.

[ishi]

I finally found the largest supplier of mineral castings:

https://www.rampf-gruppe.de/en/produ...ds-components/

IT was not easy because the stuff has many names: mineral casting, polymer concrete, resin concrete, epoxy concrete, artificial granite, epoxy granite, artificial stone, etc.

We'll see if these guys bother to return my emails...

[ishi]

I finally found the largest supplier of mineral castings:

https://www.rampf-gruppe.de/en/produ...ds-components/

IT was not easy because the stuff has many names: mineral casting, polymer concrete, resin concrete, epoxy concrete, artificial granite, epoxy granite, artificial stone, etc.

We'll see if these guys bother to return my emails...

RAMPF came back to us right away and confirmed that they can work with us. Their product line is actually quite impressive, and so is the range of services they can offer. For us, we would most likely use EPUMENT. The brochure is on the public drive:

https://drive.google.com/drive/u/3/f...kwBI-?ogsrc=32

Clearly, this would work better than natural granite (as indicated by pretty much everybody on this forum), and this supplier has presence in Europe, the US, Japan, and China. Most importantly, they can also assemble all the mechanical components, or even the entire machine. All we have to do is make sure that the machine could fit within a standard container to ensure easy shipping ($10K door-to-door pretty much anywhere in the World).

And if we ever want to make the castings ourselves, they can sell the base materials and chemicals as well...

I'm starting to like that option a lot...

[Eldon_Joh]

concerning the machine base as shown in post #62 https://www.cnczone.com/forums/uncat...ml#post2198686

I would epoxy the base together from granite straight edges, something like this: [][][]

it would cut the weight in half or less for the same stiffness and the same strength. no need for precision because two granite straight edges can be drilled and tapped for the rails to bolt to, shim the granite straight edges to the frame prior to pouring epoxy to set them in place.

[ishi]

concerning the machine base as shown in post #62 https://www.cnczone.com/forums/uncat...ml#post2198686

I would epoxy the base together from granite straight edges, something like this: [][][]

it would cut the weight in half or less for the same stiffness and the same strength. no need for precision because two granite straight edges can be drilled and tapped for the rails to bolt to, shim the granite straight edges to the frame prior to pouring epoxy to set them in place.

That could work, but I think we'll go with a mineral casting instead (Cf. previous couple of posts).

[Eldon_Joh]

That could work, but I think we'll go with a mineral casting instead (Cf. previous couple of posts).

which is fine but you can still insert a bunch of hollow sections to increase the thickness and stiffness and strength without increasing weight.

[ishi]

which is fine but you can still insert a bunch of hollow sections to increase the thickness and stiffness and strength without increasing weight.

Indeed. But I'm not too concerned about the weight of the base. In fact, we need it to be quite heavy in order to provide stability while the X or Z carriages are moving.

[Goemon]

One thought that came to mind is that, for the sort of cash you are talking about, you might be able to order the machine base and actuators from a company like Primatics. I can across them on my travels (while shopping for actuators). They specialize in high end XYZ motion systems on precision granite bases.

They could probably solve of lot of design and production issues....

[ishi]

I am working my way through the parametric modeling of our latest design, trying to externalize every single dimension as a parameter, while keeping the number of parameters to the absolute minimum. As a result, most dimensions are actually defined as equations against a handful of core parameters. That way, making changes to the design is a lot easier. But things are complicated quite a bit by the fact that I am also trying to modularize the design as much as possible, using "atomic" parts, sub-assemblies, and end-to-end assembly. This is making it a lot easier to focus on a single part at a time, but it also make the initial design a lot more laborious. Anyway, progress is being made, and here is a redesigned mount for the X-axis ball screw nut, with a much reduced weight.

[ishi]

Here is the full X-Z assembly integrated into the full design. Things are starting to take shape again...

[ishi]

One thought that came to mind is that, for the sort of cash you are talking about, you might be able to order the machine base and actuators from a company like Primatics. I can across them on my travels (while shopping for actuators). They specialize in high end XYZ motion systems on precision granite bases.

They could probably solve of lot of design and production issues....

We could, but I doubt that they would let us put the final design in the Public Domain.

Also, designing the machine is a huge part of the learning experience. Instead of outsourcing everything, I'd rather design most of the core components myself and have experts from different fields (mechanics, electronics, mechatronics, hydraulics, pneumatics, ergonomics, software, etc.) criticize the design and make recommendations for improvement. That way, I can really build some expertise. And by sharing all this feedback online, others can benefit as well. Having the full design released to the Public Domain makes it a lot easier for experts to contribute, many of the pro bono, because it benefits the community at large. And I like to believe that the lessons learned through this particular build could benefit many other builders with totally different designs.

Also, it's quite likely that our second build will be dramatically influenced by the first one, but for that to happen and lead to a much better machine, it's important that the team building the second machine is deeply involved in the design, construction, and debugging of the first one.

The downside of this approach is that everything will take a lot longer than originally expected, but I can't say that I wasn't warned by long-time members of this forum, and I really took their advice to heart. Fortunately, I am in absolutely no rush.

[ishi]

The Z-Axis Carriage will have slightly raised pads for the sliders so that we can machine and grind the mounting surfaces properly. The same was done for the X-Axis Carriage by the way.

[ishi]

Here is the all-new Y-Axis Carriage Casting, including pads for sliders, all parametric.

[RCaffin]

But I'm not too concerned about the weight of the base. In fact, we need it to be quite heavy
Chuckle. I remember we built a hydraulic-powered robot once for materials handling. The forces generated were not small, and the machine started to 'walk' across the floor. Half a dozen 18" square concrete slabs (garden footpath variety) were quickly added to the base.
Yeah, mass helps.

Cheers
Roger

[ishi]

I am starting to think that we might want to develop our own Automatic Tool Changer (ATC), because it will be a critical component of the total package, and we won't get it under Public Domain if we source it from a supplier. As far as I can tell, a traditional double arm ATC has three axes of movement: one for tool selection (usually a rotation when using a chain or carousel), one for tool exchange (rotation of the double arm), and one for tool extraction/insertion (short translation). If we're not trying to make it super fast, this all seems quite doable, and designing it ourself would allow us to optimize tool capacity and ATC placement in relation to the machine's column. My questions are the following: did anyone build such a thing on the forum? Are there any reference designs out there that I could learn from? Are there any reference materials that I could educate myself with?

[ishi]

I am starting to think that we might want to develop our own Automatic Tool Changer (ATC), because it will be a critical component of the total package, and we won't get it under Public Domain if we source it from a supplier. As far as I can tell, a traditional double arm ATC has three axes of movement: one for tool selection (usually a rotation when using a chain or carousel), one for tool exchange (rotation of the double arm), and one for tool extraction/insertion (short translation). If we're not trying to make it super fast, this all seems quite doable, and designing it ourself would allow us to optimize tool capacity and ATC placement in relation to the machine's column. My questions are the following: did anyone build such a thing on the forum? Are there any reference designs out there that I could learn from? Are there any reference materials that I could educate myself with?

This design seems to show a fourth axis of movement:

MUL Series | Automatic Tool Changers | High Speed CNC Machine Tool Changers | Quick Change Tool Changers | Gate Technologies Inc

Interesting...

And it looks like these are built using this standard precision link conveyor:

TL Series | Modular Precision Link Transfer Conveyor | Walking Beam Transfer | High Precision Indexing Conveyors | Gate Technologies Inc

And they even provide STEP files... I think we've found half of what we need...

[ishi]

Finally, here is the machine with drivetrains for all axes. Very curious to hear what the weak points might be.

***

High res pictures on public drive:

https://drive.google.com/drive/u/3/f...3dQrx?ogsrc=32

[ishi]

The TL Series Precision Link Conveyor from Gate Technologies, Inc. is quite interesting:

TL Series | Modular Precision Link Transfer Conveyor | Walking Beam Transfer | High Precision Indexing Conveyors | Gate Technologies Inc

As can be seen on the attached pictures, it's made of three modular components: a header, a body, and a footer. The body is optional, and you can use as many as you want. If you use a single, you get 30 links (for 30 tools). And if you don't use any, you get 20 links. Also, you can get it with different sizes of links, measured by link spacing of 25, 50, 75, 100, 105, 150, or 200mm. For an HSK050 or HSK063 ATC, they are using the 105mm model. I think we will do the same to sketch one a realistic ATC configuration could be.

And here is a brochure for their ATC (the one linked from their website is gone):

http://erne.com.tr/cds/catalog/MUL/M...0C008eng-1.pdf

[ishi]

Here is what the ATC conveyor could look like. It's just under 1500mm long, which means that with tools, it would fit perfectly within the machine's width. The idea is to mount this on top of the vertical column. Now, I'll try to design the double arm assembly that would bring everything together. If we can get a working design for that, we'll be in good shape...