[cuplinks]

Hello all,

Sorry I haven't done my research in full, so please bear with my Newb status.

I am a bespoke shoemaker in Australia, and at the moment I adjust a pair of the 1000's of lasts I have to fit my client by adding leather/putty/cork etc to get the perfect fit. I've gotten pretty good at that over the years but every now and then a client comes along that doesn't fit with the base model lasts I have in stock and the adjustment process is ridiculous.

I would love to move into the current century and use my measurements to make a 3D model and produce a pair of accurate lasts via a CNC process.

I have been dabbling for a few months in Rhino trying to get my head around CAD/CAM modelling and I can tell that it is going to take me a long time to get on top of it all. So in the mean time I have created an account here to hopefully get some ideas on how to move forward.

Basically what I am hoping to do is find a cheap machine that can carve a chunk of hardwood or polyethylene plastic into a last. Like this;



Actually that is a horrible last shape but you get the idea. Also I would probably be using hardwood as its locally source-able. I am NOT mass producing. I might make a pair of lasts a week at max at the moment. I would be looking for an accuracy of at least 0.5mm.

I have a few questions, please chime in if you have any thoughts on all or any.

1. I have been looking at CNCs on eBay and Aliexpress for something that I can get started with. Like this. Is this a good cheap starting point for what I need? I'm worried about the z-axis height.

2. I'm assuming I might need a 4-axis machine to do this? Or maybe just a 2-axis machine like a lathe/mill cnc thing? What would be the ideal rig structure?

3. Assuming I am trying to carve hardwood/polyethylene, what strength spindle will I need?

4. Also thoughts on axis motors and screws?

5. Where should I source parts or are there some prebuilt frames or kits out there that I should look at? I am based in Australia.

6. In terms of software I know there are a few shoe specific 3D options available but I am looking for any easy to use options. If you have any thoughts please let me know. Also where can I go to teach myself what I need to know?

If you read this, THANKS!

I look forward to any thoughts.

Kind regards,

James

[awerby]

This does look like a job for a 4-axis machine. A router would work, but while that one you found would certainly be a cheap starting point, it looks too flimsy to produce good surfaces in hardwoods or polyethylene. Look for something that has rails that are supported along their length, at a minimum (round rails aren't as good as the squareish ones) , and that isn't any higher than you need it to be. The extra height only makes it shake more. If you can find (or build) a machine which places the center of the A axis at the table height, rather than above it (this will require a gap in the table or placement off the end of it) that will save Z height. A machine with a fixed bridge and moving table will tend to be more rigid than one with a moving gantry, although it will have a larger footprint.

[cuplinks]

Thanks Andrew. I think I understand what you mean.

I have spent all day researching and I think I will try and build my own machine.

I think I can do what I want with a fixed gantry machine and potentially no x-axis. Just y, z, and the 4th (a) axis. (Maybe I am confusing x and y?). Will that even work with the standard software?

Anyway, with 1605 ballscrews, HIWIN linear rails, NEMA 23 stepper motors, can I get by with cheap aluminium extrusion for a fixed gantry frame?

I will probably want to buy a prebuilt z-axis and a-axis. Does anyone have any recommendations or links?

Also what type of spindle/router will I need for hardwood/HDPE?

Finally, does anyone has any recommendations on drivers and software for this build?

Regards,
James

[pippin88]

LinuxCNC (free).

2.2kw chinese spindle with VFD.

Get 1610 ball screws, not 1605

No such thing as cheap aluminum extrusion in Australia unfortunately.

Your requirements should be easy enough to meet.

You could use a 4th axis, but flipping the part (and only needing three axes) may work better.
I would expect you will be doing a bit of hand finishing.

Where in Australia are you?

[ger21]

I think I can do what I want with a fixed gantry machine and potentially no x-axis. Just y, z, and the 4th (a) axis. (Maybe I am confusing x and y?). Will that even work with the standard software?

Not sure about the software. Check out DeskProto, and see if it will work with only the 2 axis + rotary.

You want 1610 ballscrews, as 1605 will be too slow.

What do you mean by cheap extrusions? Good extrusions are not cheap. The rigidity of the gantry is the most important part of the machine.

I will probably want to buy a prebuilt z-axis and a-axis.

Quality prebuilt axis will cost far more than you think.

PRO Ballscrew Z Axis | CNCRouterParts

CNC Routers made in the USA ~ 5x10, 4x8, 4x4 - Make signs and more

Most prebuilt rotary axis that I've seen are very lightweight units, for small parts. Here's a slightly larger one.
https://www.amazon.com/Sunwin-Router...29P5SBBR1D2E5S


This is a popular spindle.
https://www.ebay.com/itm/FOUR-BEARIN...IAAOSw1vlUtyPk

[wizard]

Interesting project!

Not being a shoe maker is it fair to say that a last is a form or mold? Just want to clear that up in my mind.

Looking at these lasts pictures i can imagine you will need a larger machine than you might first suspect. These forms appear to be pretty oddly shaped which will make work holding and suppirt a chalkenge. Personally id look for example of how others have done such work on a CNC.

Spindle selection will be i teresting in and of itself. The reason here is the apparent need to support large dismeter cutters even on a 4 axis machine. If you do things like boots i could see a need for a 5 axis machine. This is one of those things where you really need to think about work holding, tool pathes and so forth before even buying or building a machine. In any event back to the spindle, on a 4 axis machine you may find yourself needing long cutters which implies large dismeter cutters for stiffness. Thus you will want to lean towards spindles with larger collets or tapers.

Work holding has me perplexed. For example you could spin between toe and heel and do a good part of a last on a 4 axis but leaving a good bit of hand work to be done. Another option would be to hold between the heel and the top of the last but that would seem to demand a 5 axis machine.

This wont be a trivial machine build especially if your needs includes very high Z clearance. How high that Z will be depends upon the lasts you try to build. The thought has occured to me that maybe a router isnt the best platform here. In that regard a machine that resembes a vertical boring mill might fit the task at hand better. On such a machine the blank would be installed toe up on the rotary table.

As to your precision spec 0.5mm is pretty course i would strive to improve that. However for things like this repeatability is far more important. To get a decent finish you will want a machine with very low backlash and minimal machining deflection. As your Z hieght increases this becomes a bigger and bigger challenge

One way or another the mechanical issues can be overcomed. The bigger challenge in my mind is software. Taking a series of measurements and turning them into a 3d cad solution is not trivial at all. If the software does exist already it could be years before something viable is created. Frankly a 3D scanner might be useful here if you can get the customer to hold the foot still. I just see this conversion of physical reality into a viable 3D model of a last to be very challenging. Frankly i would not even bother with the machine tool until after the software problem is solved. If software already exists find out what sort of machine it prefers.

Oh a note on axis naming which is always abit confusing. The Z, the axis the spindle revolves around is at the top of most routers as such i prefer to see reverse alpha labeling. This makes the gantry Y and the table axis X. This simply for communications constancy. Others have differing opinions here but some of them dont make sense to me. For examle some suggest that the axis going left and right is the X. The problem there is that is dependent upon position of the user. Others would like the long axis to be X, while that has some logic it forgets that many machines are square format. In any event just make it clear what you are talking about and most can adapt

Another thing to consider is enlisting a local machine shop to machine your 3D models. Costs will vary but it takes the cost and distraction of router build out of the equation. It would give you the non trivial time required to learn the software tools.

Finally do consider steel for this design. I dont expect a huge machine here but the high Z compared to most machines here means you need to pay attention to stiffness. Likewise choose suitable square profile linear bearings.

By the way as i write this im looking at my size 16 feet and wishing that shoes that actually fit could be had retail. Life can be hell when you are outside the norms.

[awerby]

Thanks Andrew. I think I understand what you mean.

I have spent all day researching and I think I will try and build my own machine.

I think I can do what I want with a fixed gantry machine and potentially no x-axis. Just y, z, and the 4th (a) axis. (Maybe I am confusing x and y?). Will that even work with the standard software?

[I can picture a dedicated 4th axis machine doing this without a Y axis (by convention, the longest axis is called "X"). As was suggested, DeskProto will write code for a setup like that, using an "along X" strategy, incrementing in A.]

Anyway, with 1605 ballscrews, HIWIN linear rails, NEMA 23 stepper motors, can I get by with cheap aluminium extrusion for a fixed gantry frame?

[Since a design like that would have the Z axis moving along a rail mounted to a fixed bridge, you don't need to use aluminum (which isn't cheap, compared to steel). The only reason to use it is to save weight on a moving part. Steel is stiffer as well as cheaper.]

I will probably want to buy a prebuilt z-axis and a-axis. Does anyone have any recommendations or links?

[There was a fellow who posted not long ago who was talking about building a machine based on a wood lathe frame. Your machine could be done like that; it would give you a way to mount the tailstock so it could slide back and forth. You can also look for precision automation actuators for the Z (and possibly the X). These have the advantage of being pre-built with rails and screws already in place, mounted in an extrusion and lined up correctly with a motor mount. They don't seem as common on Ebay Australia as here in the States, but you can probably find them there, or pay a lot of shipping costs: https://www.ebay.com.au/itm/Parker-P...EAAOSwzJ5XaFF3
For the A axis, you can use a rotary table like the one Sherline makes, that has a mount for a 23-frame stepper: https://sherline.com/product/3700-cn...rotary-table/]

Also what type of spindle/router will I need for hardwood/HDPE?

[A 3-phase spindle with VFD (like those already suggested) will work better than a hand router.]

Finally, does anyone has any recommendations on drivers and software for this build?

Regards,
James

[Build it first, and while you're working on it you can research the issues involved. You might end up changing your design, and need something different than you thought initially. Or you might luck out and find an actuator with a motor on it that you can use if you find the right driver for it. ]

[cuplinks]

LinuxCNC (free).

2.2kw chinese spindle with VFD.

Get 1610 ball screws, not 1605

No such thing as cheap aluminum extrusion in Australia unfortunately.

Your requirements should be easy enough to meet.

You could use a 4th axis, but flipping the part (and only needing three axes) may work better.
I would expect you will be doing a bit of hand finishing.

Where in Australia are you?

Thanks, I'm happy with a bit of hand finishing. It's what I'm doing already. I'm in Kyneton, Victoria. 1 hour north of Melbourne.

[cuplinks]

Thanks all who have commented. I have a lot to think about.

https://youtu.be/FvxN9PbueCo?t=72 Here is a video of the last Last manufacturer in Australia. They went out of business a few years back now. Notice how their machines are basically a CNC Lathe.

Maybe that is the best route to go. Retrofit a lathe with CNC components? and add a spindle on z-axis. You could either move the spindle on an x-axis or the material. And then the pathing would be based off either a continuously rotating a-axis or having it move in steps I guess.

Also I had planned to use Aluminium extrusions as opposed to steel as I don't know how to weld and like the idea of the extrusions being a bit future-proof if I decide to change the build. But I will look into it. I do know someone who could build me up a steel frame.

And in regards to software, there is CAD/CAM software out there specifically for shoe lasts that I can get hold of. They basically have a library of prebuilt models that you import and adapt to the clients measurements. You can also start from a 3D scan of the feet and build a model that way but that would be a hurdle for a future endeavour.

Thanks again all for you help. I will post more once I have something made up!

[pippin88]

That machine is 3 axis. Rotary, in and out, and along. - essentially a lathe, except the rotary component is moved in a controlled fashion rather than just at roughly X speed.

I think a dedicated build is wiser than converting a lathe.

If you only want to make lasts and nothing else at all, a similar design would be fairly easy to make.

For aluminium extrusions, I've used amesystem.com.au
They were good to deal with and actually publish a price list, which is refreshing.

Ultimately you'll have extrusions cut to certain lengths with holes in certain places, and that limits future re-use somewhat. I used extrusions for a machine and they worked ok. Moved to welded steel (not hard to teach yourself to stick weld if needed). I did build a machine for my dad out of the old extrusions, which meant a few compromises to fit what I already had, but at least they did not go to waste.

[pippin88]

cuplinks, I know nothing about the use of lasts, but would 3D printing be suitable?

You could buy a printer for not much and skip the designing and building of a machine.

The printing itself would take longer per last, but I suspect that would not be a limiting factor for your scenario.

[HVACturnedCNC]

I'm with this guy !! if all you want to do is make a bust to taylor the shoe around the bust for proper fit then I would use a 3D printer. If you wanted it extremely smooth you could almost achieve it with the printer and a slight hand sanding, although since it is a shoe I would assume smoothness of the actual bust wouldn't matter as most people wear socks. Scan the foot into 3D CAD software , insert filament and press start....seems a lot easier than trying to make a CNC

cuplinks, I know nothing about the use of lasts, but would 3D printing be suitable?You could buy a printer for not much and skip the designing and building of a machine.The printing itself would take longer per last, but I suspect that would not be a limiting factor for your scenario.

[wizard]

Thanks for the great video?

Thanks all who have commented. I have a lot to think Here is a video of the last Last manufacturer in Australia. They went out of business a few years back now. Notice how their machines are basically a CNC Lathe.

Pretty cool. I noticed they are doing work holding between heel and toe. Im not sure if that is the best way to go but obviously works for the eqipement they have.

Here is an interesting thought, see is any of those lathes are available on the used market. You might be surprised and end up saving your self a lot of time.

Maybe that is the best route to go. Retrofit a lathe with CNC components? and add a spindle on z-axis.

In this case i have to strongly suggest that you DO NOT try to retro fit a manual lathe. The X axis is constantly in movement and as such would wear most coss slide on manual lathes fairly quickly.

Now finding an old CNC lathe to retro fit is another matter. Just make sure they have linear rail bearings on all axis.

You could either move the spindle on an x-axis or the material. And then the pathing would be based off either a continuously rotating a-axis or having it move in steps I guess.

The software would be the hardest part of getting a complete system here. You have indicated that such software exists so you only real goal would be to make sure the machine you build works with it.

Also I had planned to use Aluminium extrusions as opposed to steel as I don't know how to weld and like the idea of the extrusions being a bit future-proof if I decide to change the build.

There is a huge aversion to working with steel in these forums and frankly im not sure it is justified. For one thing working with Aluminumextrusions has its own learning curve. Niether steel nor aluminum extrusions are suitable out of the box for critical machine components. Thus maching costs are often a wash. Beyond that extrusions are no more reusable than steel.

This isnt to say that steel should be used everywhere. The point is there are places in a machines structure where a welded steel frame will make a huge difference over one built from T-slots.

But I will look into it. I do know someone who could build me up a steel frame.

Getting help with something like this is always a wise move.

And in regards to software, there is CAD/CAM software out there specifically for shoe lasts that I can get hold of. They basically have a library of prebuilt models that you import and adapt to the clients measurements. You can also start from a 3D scan of the feet and build a model that way but that would be a hurdle for a future endeavour.

Im not sure why im surprised by this.

Thanks again all for you help. I will post more once I have something made up!

In my other message i didnt stress used equipment enough. I know things are different on your side of the planet but here we have a large market for used machine tools. Often said tools can be had cheap thus making a good base upon which to build something new. If you find the right fit donor machine wise you will save considerable time and money.

If you enlist a shop to build a frame for you somebody has to design that frame. It is a lot of work to do so properly. That doesnt include the rest of the design work in a unique machine build.

As for the mentioned 3D printing do consider the pissibilities here including the more advanced techniques. FDM style printers are extremely slow. There are other methods including UV curing systems out there to consider. Some are expensive which might make for a new business venture to pay for them. On the flip side you might want to look for a job shop that does 3D printing locally. If you can find a technology that produces the right structural qualities 3D printing is almost the ideal solution. This especially if you are doing one off medical builds.

The more i think about it the more 3D printing fits the bill. The only real question here is again software. You need to make sure the CAD/CAM software will output files suitable for a 3D printer.

So yeah lots of ideas to explore here.

[pippin88]

FDM is slow, but it seems unlikely that the OP will care if takes 12 - 24 hours for a printer sitting by itself to produce a last.

[cuplinks]

FDM is slow, but it seems unlikely that the OP will care if takes 12 - 24 hours for a printer sitting by itself to produce a last.

Yeah, I don't care if it is slow. Like I said earlier, if it can make a pair a week that is all the rest of my workflow can handle anyway.

My thoughts on 3D printing was that you couldn't get a high enough density plastic to print with so I ruled it out initially. The lasts take a fair bit of a beating while making the shoe. I know that HDPE is used but can you 3D print with that?

I can't find any industry examples of using a 3D printer to make lasts.

[pippin88]

Can you provide a link to a good example of lasts being used?

I think with the right settings 3D printing would be fine.

ABS is commonly used. PETG is a pretty tough plastic also.

[skrubol]

I wouldn't go 4/rotary axis. You could cut these on a standard 3-axis by using 2 setups. Start with a block of wood, cut everything accessible from the top. Fasten the 'ankle' to a spoil board and machine the bottom (with light cuts, holding it by the ankle is not going to be as secure as would be ideal.) Quicker, but with a bit of extra hand work would be to cut the model in half, machine the halves (still may require 2 setups on at least one of the halves, but there will be a lot more to hold onto.) and then assemble the halves by hand.

I would probably 3d print though. As long as the parts don't need to be heated to the point they'll soften it would be the least work. 3d prints can be quite strong and durable, especially on something as solid as a foot model. Using the right number of wall layers and a fairly dense infill would make a part you could drive a car over without it taking much damage. With the right supports post print cleanup should be just tearing the supports off and doing a little bit of sanding to take off bumps left from the supports. I assume the layer lines shouldn't really matter, but if needed you can smooth abs to gloss with acetone vapor (or body filler, sanding and coating.)
I'm not sure what's around in AU, but it shouldn't be too hard to find a service bureau that could do an FDM print at least as a proof of concept before investing in a printer.

[wizard]

Yeah, I don't care if it is slow. Like I said earlier, if it can make a pair a week that is all the rest of my workflow can handle anyway.

What you wiil care about is a print getting screwed up when 90% complete. Also realize that feet are rather big for low end printers so you are likely to need a fairly good sized machine. Off the top of my head probably 400 square millemeters. 350 mm might squeeze in most people excluding a few giants.

In any event you really want your prints to happen quickly, it isnt a good idea to leave the low end printers unattended.

My thoughts on 3D printing was that you couldn't get a high enough density plastic to print with so I ruled it out initially. The lasts take a fair bit of a beating while making the shoe. I know that HDPE is used but can you 3D print with that?

There is a huge amount of innovation and competition when it comes to printer feed stock so it is likely that the right material is already out there. Beyond that wall thickness and infill can be adjusted so in all likelyhood this will not be a problem but might require experimentation.

The question that comes to mind is this, are these lasts one time usage or for production runs. The nice thing about 3D printing is the wide array of mayerials to choose from. If you could justify based on usage your lasts could be printed in metal.

You also have the option to 3D print molds or forms to make the lasts in other materials. For example a fiber glass reinforced plaster coild be poured into a last form. 3D ptinting does always have to produce the part directly. For example some guys have been 3D printing patterns for iron or aluminum casting.

I can't find any industry examples of using a 3D printer to make lasts.

This is where partnering with a larger supplier of 3D printers could be of benefit as they can offer guidance on the lstest feed stock and print samples for you. Yeah i know it is the expensive way to do this but they may also be able to hook you up with a near by job shope.

I come back to the idea of a job shop for a couple of reasons. First one pair of lasts a week makes any machine hard to justify. Even a CNC lathe would just be sitting idle most of the time. Second; if you are already pressed for time running a lathe, 3D printer or something ekse is just going to take up even more of your time. Third; a jub shop should have access to a wider array of professional machines. This means various types of FDM feed stock plus machines running completely different processes.

I realize this adds an expense that has to be passed onto the customer. However time and materials, plus the cost of the machine must be covered if you DIY. The same issues exist if you where to find a CNC lathe to doo the job. Paying for the machine, operating it, maintaining it and suppling it with raw materials is non trivial and those costs must be passed on.

Of course if you dont have any 3d printing contractors in the area farming out the work isnt going to cut it. However a contractor would lower risk significantly. You might also want to contact conventional CNC machine shops as the lasts might be doable on a CNC lathe with live tooling support. Some of the CNC lathes out there are certainly flexible enough to do lasts with the right tooling. They would likely machine the lasts out of whatever material you requested.

[cuplinks]

What you wiil care about is a print getting screwed up when 90% complete. Also realize that feet are rather big for low end printers so you are likely to need a fairly good sized machine. Off the top of my head probably 400 square millemeters. 350 mm might squeeze in most people excluding a few giants.

In any event you really want your prints to happen quickly, it isnt a good idea to leave the low end printers unattended.

There is a huge amount of innovation and competition when it comes to printer feed stock so it is likely that the right material is already out there. Beyond that wall thickness and infill can be adjusted so in all likelyhood this will not be a problem but might require experimentation.

The question that comes to mind is this, are these lasts one time usage or for production runs. The nice thing about 3D printing is the wide array of mayerials to choose from. If you could justify based on usage your lasts could be printed in metal.

You also have the option to 3D print molds or forms to make the lasts in other materials. For example a fiber glass reinforced plaster coild be poured into a last form. 3D ptinting does always have to produce the part directly. For example some guys have been 3D printing patterns for iron or aluminum casting.


This is where partnering with a larger supplier of 3D printers could be of benefit as they can offer guidance on the lstest feed stock and print samples for you. Yeah i know it is the expensive way to do this but they may also be able to hook you up with a near by job shope.

I come back to the idea of a job shop for a couple of reasons. First one pair of lasts a week makes any machine hard to justify. Even a CNC lathe would just be sitting idle most of the time. Second; if you are already pressed for time running a lathe, 3D printer or something ekse is just going to take up even more of your time. Third; a jub shop should have access to a wider array of professional machines. This means various types of FDM feed stock plus machines running completely different processes.

I realize this adds an expense that has to be passed onto the customer. However time and materials, plus the cost of the machine must be covered if you DIY. The same issues exist if you where to find a CNC lathe to doo the job. Paying for the machine, operating it, maintaining it and suppling it with raw materials is non trivial and those costs must be passed on.

Of course if you dont have any 3d printing contractors in the area farming out the work isnt going to cut it. However a contractor would lower risk significantly. You might also want to contact conventional CNC machine shops as the lasts might be doable on a CNC lathe with live tooling support. Some of the CNC lathes out there are certainly flexible enough to do lasts with the right tooling. They would likely machine the lasts out of whatever material you requested.

Hi Wizard, thanks for your thoughts again.

The problem I have with 3D printing is of course the durability of the plastic. When making a pair of shoes I need to drive nails in the the last to hold the leather in place. It is probably hundreds of tiny pins over the course of one pair and I would be expecting the lasts to be used for dozens of pairs. The good thing about wooden lasts and even HDPE is that the nail holes don't chip away at the last. I'm worried that the layers in the 3D printed last would chip away with the nails. Also cast aluminium lasts are used for making sneakers as they can survive the vulcanisation process of putting the rubber soles on but are not useful for traditional shoemaking.

But maybe it is possible, I have found an amateur shoemaker who experimented with 3D printing on Instructables and managed to make something work with PC-10 plastic. It was a 30 hr print and he said the cost of plastic was high. These are his lasts;




Anyway, I am not likely to go against the industry standard and would prefer to use wood as I am used to it and can source it locally and relatively cheaply. (I have a contact who has a shed full of old wooden last blanks).

Finally, in regard to using 3rd parties. I have gone down this route many times in the past and unfortunately don't get the results I need. I have ordered lasts from the UK, India and China, as well as from the old company now closed in Australia and they all have problems. I send them my measurements and all appropriate data and the lasts don't fit the client. In one case I took the client in for a 3D scan of their feet and had the lasts made from that scan at considerable expense and they still didn't fit.

I know I can fit my clients by manually adjusting a last so I just need to have more control over the process to get it right. And instead of wasting money experimenting with 3rd parties its time I just sorted it out myself.

[pippin88]

Sounds like 3D printing will not be the best solution with the need for repeated nailing.

A 3 axis custom build is probably best:
Rotary (blank is held by this axis) (Call it R)
Longitudinal (call it X)
In and Out (Call it Y)

R will need reasonable torque and only a slowish rotation speed.
X will need to be a little longer than your longest last, so 400mm should be enough
Y will need enough travel for the radius of the blank, plus the length of the longest tool.