[extrastrength]

The forum here is a fantastic resource, thanks to all the regular contributors. I've been doing research about my options for building my first 3-axis CNC router. Lest this sound like just another "What should I build?" thread, here's the skinny.

----------
Me: I've been working in CAD doing automotive design for years, and more recently gained access to a nice 4'x8' commercial CNC router and VCarve Pro. I can also TIG weld, and have a fair bit of experience in metal fabrication.

What I Want: At least a 2'x4' workspace, with enough strength to do woods and plastics. It would be a plus to do light duty aluminum work, but if the machine is at least good enough to help me build a bigger router table down the line, that's fine. I would also love to have a 4'x4' work area (half a sheet!), if possible. This will be used primarily for hobby decor production.

Budget: $1,500 tops, would like to bring it in under $1,000.
----------

I see my current access to a router as a big benefit, because I can potentially just cut MDF/HDPE parts out for a homebuild. On the other hand, if it was economical, and designs are available, I can weld up a frame and go from there. That said, I've looked at Joe's 2006 design, and it seems like the most promising and sturdy so far (within the budget). Once I convert the design to VCarve, it should be very inexpensive to build the frame.

If anyone has suggestions other than Joe's router, please bring them. Also, if anyone has pre-made VCarve/Aspire files, that might significantly influence my decision!

Thanks again,
Michael

[extrastrength]

I'm also looking at the CNC Mogul now, which seems to get positive feedback here on the forum. It certainly looks like I could put together a 4'x8' setup for $1,500 or so (425 oz/in steppers, Gecko or other controller, 1.5-2hp router), and potentially weld up a steel table somewhere down the road.

Or the Shapeoko, which I could probably build at 4'x8' for around $700. Obviously, the Shapeoko sacrifices rack and pinion, some Z-axis rigidity, limit switches, and maybe some other features. It almost seems like I can live with that for an $800 difference.

Any thoughts?

[ger21]

Any thoughts?

$1500 is a small budget for even a very small machine. A decent 2'x4' to 4'x4' machine will typically run around $2500-$3000. That's for an average machine, not a good one.
When you spend $1500 or less for a 4x8 machine, you're getting a very lightweight machine.

Opinions on the various machines and designs out there are really meaningless, unless you know exactly what you want to do with the machine, and what your expectations are.
I think the Mogul and Shapeoko are targeted towards beginners with no experience, and no expectations at all. To these people, just having a functioning CNC router is all they really expect. for people with experience looking for a certain level of performance or capability, these machines can fall very short.

[extrastrength]

To these people, just having a functioning CNC router is all they really expect. for people with experience looking for a certain level of performance or capability, these machines can fall very short.

Thanks very much for your reply.

I see an interesting trend in CNCs (routers and additive printers both), that people are taking up the banner of "affordable manufacturing for all." However you feel about it, you can't deny that it's made an impact on this hobby/industry.

Whereas six years ago, the best you could do at home (to stay in budget) was take a pile of MDF and make it into many of the specialized parts required for 3-axes (see Joe's very fine 2006 machine, which had initially caught my attention... Today, there are PURPOSE-BUILT parts developed en masse for budget builds, and many of them appear better in actual function than previous budget designs (that's what we call an advance, lol!). Prior to the last few years, machine manufacturers had no significant pressure to keep the price down, or reinvent the wheel, but it seems like the recent shift is good for us DIYers.

I'm not saying anything bad about previous open-source efforts, but pointing out that the market today is different. It is now supporting these efforts to mass produce specialized budget parts, and you can't deny that specialization in this case sure beats black pipe. When black pipe is what we have, black pipe is what we use.

In the case of the Shapeoko, the big step forward was the "Makerail" and purpose-built linear bearings. More rigidity and function than generic pipe, without much more cost.

Bear in mind that people are getting excellent accuracy, repeatability, and rigidity out of Joe's '06 design. Even to the point of machining aluminum! On a $1500 machine!

I say all this to say that field reports on the Shapeoko are returning some of the same very good results, at a lower cost, with great scalability. Accuracy of a tuned machine is stated by the design team to be .003-.004, and has been measured lower (according to keyboard jockeys, who knows?). That's plenty for me. And the design is so open-ended, I highly doubt I'd have any problem making it work my minimal aluminum jobs. I can use the machine itself, in part, to make new plates, spindle mounts (definitely putting a real router/spindle on!), and reinforced parts.

That's the kind of tweaking I'm signed up for, whether I do Joe's design, or start with a Shapeoko. That's just by virtue of my budget, unfortunately. Believe you me, I'd rather just buy a nice machine.

Anyway, I welcome other thoughts, but I hope that is insightful into my considerations now. The Shapeoko is looking pretty appealing at the moment.

[wizard]

I have to agree with most, maybe all, of what Ger21 said. I will try to address some of your concerns below.

Thanks very much for your reply.

I see an interesting trend in CNCs (routers and additive printers both), that people are taking up the banner of "affordable manufacturing for all." However you feel about it, you can't deny that it's made an impact on this hobby/industry.

Frankly the big deal here is the rapid advancement with respect to electronics, that has driven the costs of the machines down to an amazing extent. You can find microprocessors that cost $30 suitable for low end machines. On top of that a series of new microprocessors hitting the market will allow even more performance in the same general price range.

As for affordable manufacturing for all, I just see that as a trendy phrase for the DIY mentality. I've been around since 1960 and was always exposed to the DIY personality growing up. If my parents and relatives had access to such hardware back them I'm certain that at least one of them would have had a CNC machine. We all had garages full of woodworking equipment and even a few metal working tools.

Whereas six years ago, the best you could do at home (to stay in budget) was take a pile of MDF and make it into many of the specialized parts required for 3-axes (see Joe's very fine 2006 machine, which had initially caught my attention... Today, there are PURPOSE-BUILT parts developed en masse for budget builds, and many of them appear better in actual function than previous budget designs (that's what we call an advance, lol!). Prior to the last few years, machine manufacturers had no significant pressure to keep the price down, or reinvent the wheel, but it seems like the recent shift is good for us DIYers.

As they say appearances can be deceiving!!! At best the cheap machines on the market are suitable for light weight machining.

I'm not saying anything bad about previous open-source efforts, but pointing out that the market today is different. It is now supporting these efforts to mass produce specialized budget parts, and you can't deny that specialization in this case sure beats black pipe. When black pipe is what we have, black pipe is what we use.

Actually real linear motion products have been available for years now. Expensive yes and frankly these solutions still out class the many cheap designs on the market.

In the case of the Shapeoko, the big step forward was the "Makerail" and purpose-built linear bearings. More rigidity and function than generic pipe, without much more cost.

If you go to Shapeoko's web site you will see that they have basically given up on Maker Slide (rail). Shapeoko's 3 is a major overhaul of their machine. Even then they are only shipping a machine suitable for a 16" square working area. The reality is skinny framing components can and do have a significant impact on the quality of the results you can get from a CNC machine.

By the way a brief look at the web page for Shaoeoko 3 indicates a significantly improved machine so I'm not knocking their efforts here. It is obvious though that they attacked many weak spots in the machines design.

Bear in mind that people are getting excellent accuracy, repeatability, and rigidity out of Joe's '06 design. Even to the point of machining aluminum! On a $1500 machine!

Unfortunately everyone has their own metric as to what is excellent accuracy. Work in the optics industry and they would laugh at what is considered precision machining by some. In the end it really depends upon your application of the parts produced. Consider too that many will do a circuit board and claim a machine is suitable for machining copper. That may be so but there are real differences between shaving a few thousands of copper off a board and trying to machine a billet or a thick sheet of metal.

In the end these comments are only of value if you have the same expectations as the user making the claims. The other problem is how does the machine function long term. If you need daily tweaking to keep it humming along that is a big negative.

I say all this to say that field reports on the Shapeoko are returning some of the same very good results, at a lower cost, with great scalability. Accuracy of a tuned machine is stated by the design team to be .003-.004, and has been measured lower (according to keyboard jockeys, who knows?). That's plenty for me.

For some forms of wood working that can be acceptable. But you have to conside what conditions it was achieved under. A light machine can do wonders with a barley loaded spindle in balsa. Try taking a production level cut in hardwood plywood and you may get dranmaticaly different results. In one way doing light cuts might not matter as you can just let the machine run forever if you have too. However most people aren't that patient. Frankly it is better to get the work done with proper feed rates and depth of cuts for cutter life. In the end it is still a question of what is acceptable to you.

And the design is so open-ended, I highly doubt I'd have any problem making it work my minimal aluminum jobs. I can use the machine itself, in part, to make new plates, spindle mounts (definitely putting a real router/spindle on!), and reinforced parts.

Not to sound negative but that depends upon what you are making this parts out of. I really doubt that you will be machining steel plate.

That's the kind of tweaking I'm signed up for, whether I do Joe's design, or start with a Shapeoko. That's just by virtue of my budget, unfortunately. Believe you me, I'd rather just buy a nice machine.

Unfortunately with routers like this, just buying isn't all that economical. Or to put it another way the mass market isn't big enough to allow for price competitive machines relative to what the DIY crowd can make. A decent commercial machine is more than a few bucks above your budget (mine too for that matter, thus the need to look at DIY solutions.

Anyway, I welcome other thoughts, but I hope that is insightful into my considerations now. The Shapeoko is looking pretty appealing at the moment.

I wouldn't touch the current Shapeoko 2 with a ten foot pole. I'm not even sure about Shapeoko 3.

Here are my thoughts on the subject:

First take an inventory as to what you have tool wise. This can either be personally owned or something you have access to. Frankly access to machinery can have a dramatic impact on costs as opposed to paying somebody to do the work. More importantly decent equipment makes it a bit easier to build an accurate machine.

Like wise inventory any material you have suitable for the build.

If you have decent welding skills, I'm assuming you do because you mention TIG skills, consider building as much as possible out of steel. In the end steel is pretty damn cheap for machine construction. There are considerations with respect to how you flatten a steel structure and ideally you would stress relieve anything welded together. For a starter machine you might consider just making part of the machine steel, the gantry for instance is a place where a big steel box beam can be used to advantage

A well thought out design in steel (any material really) should be built such that you can upgrade controls later. If you start out with steppers and acme screws make sure an upgrade to AC servos and ball screws is possible in the future. In industry it isn't uncommon to see machines rebuilt in place though the goals are different. You should look at going cheap in the first build as a way to get started on a budget but yet maintain future plans that are your goals for an ideal machine. If you don't start out with a decent frame there is little opportunity to improve it down the road. Building a whole new machine from scratch is always possible but it is hardly inexpensive and frankly you then need room for two machines.

You seem to have plenty of design experience as such I'd recommend reading as many threads as possible in this forum to get an idea of what has worked well for people. This should prime your mind for a DIY design that fits with the materials and equipment you have to work with.

[ger21]

I see an interesting trend in CNCs (routers and additive printers both), that people are taking up the banner of "affordable manufacturing for all." However you feel about it, you can't deny that it's made an impact on this hobby/industry.

An impact on "hobby" class machines, and the industry of manufacturing these lightweight, inexpensive machines, yes. Providing "affordable manufacturing for all"? No, not really, imo. Sure, there are probably a few people making a living with some of these machines. I know of someone who runs a business using Carvewright CNC's. But those are few and far between. The vast majority of these are use to produce small "trinkets", or to augment some other hobby.

Whereas six years ago, the best you could do at home (to stay in budget) was take a pile of MDF and make it into many of the specialized parts required for 3-axes (see Joe's very fine 2006 machine, which had initially caught my attention... Today, there are PURPOSE-BUILT parts developed en masse for budget builds, and many of them appear better in actual function than previous budget designs (that's what we call an advance, lol!). Prior to the last few years, machine manufacturers had no significant pressure to keep the price down, or reinvent the wheel, but it seems like the recent shift is good for us DIYers.

In general, the members here are very different than "members" of the "Maker" movement with regards to CNC routers. You'll find very little talk of ShapeOKO type machines, and the open source software and electronics packages associated with them. I don't think I've seen anyone build here build a machine using Makerslide. CNC Zone members typically want a lot more out of their machines than the cheapest possible price. As such, many of us have very different (typically much higher) expectations.

10-12 years ago, MDF construction was the norm. The Joe's 2006 was really the tail end of widespread MDF use. 6 years ago, the majority of members here started moving towards extrusion based machines, mostly to CNC Router Parts and Joe's Hybrid machines.

Prior to the last few years, there really were no manufacturers. Funny that you mention "reinventing the wheel". I read a quote here a while ago:

"The Maker Movement - Re-inventing the wheel - The hard way"

It is now supporting these efforts to mass produce specialized budget parts, and you can't deny that specialization in this case sure beats black pipe. When black pipe is what we have, black pipe is what we use.

In the case of the Shapeoko, the big step forward was the "Makerail" and purpose-built linear bearings. More rigidity and function than generic pipe, without much more cost.

Some might argue that the Makerrail or Makerslide is NOT superior to black pipe. It depends on the application, and how they are implemented. Quite often with DIY CNC, how components are used is more important than which components are used.
A plastic V roller on a small aluminum extrusion is just not that rigid. I believe that the Makerslide was originally developed for use with lasers and printers, where NO cutting forces are involved.
Bearings rolling on pipe or steel bar stock ares till used today in industrial machinery. With a good design they are both strong and rigid.

What the Makerslide brings to the table is ease of use, at a low cost. But it's not necessarily a superior product, unless ease of use and low cost is your primary objective. My goal is to end up with the best possible machine.

I like to think that Joe based some of his 2006 machine on my router, which I designed and started building 3 years before Joe cam out with the 2006. A friend who has a 2006, and has seen a few others, claims that my machine is much better than a 2006.

Bear in mind that people are getting excellent accuracy, repeatability, and rigidity out of Joe's '06 design. Even to the point of machining aluminum! On a $1500 machine!

Nobody is getting "excellent rigidity" out of a Joe's 2006. Or a Shapeoko, or a CNC Mogul. Not even "fair" rigidity.
Keep in mind, that this is very subjective.
Rigidity is one of if not the most important factor in cut quality, and often machine quality.

Back when everyone here was building the Joe's 2006, once they finished, a lot of these same people moved on to bigger, more rigid machines. A lot of them moved to the Joes' Hybrid. You won't find much information on the Hybrid, because Joe started his own private forum when he released the Hybrid, and very few Joe's users post about there machines here these days. But if you were to join and read Joe's forum, the hybrid has evolved considerably over the years, always trying to get more rigid.

As for accuracy and repeatability. Those are easy to achieve with most any machine. But rigidity plays a huge factor here. The more rigid a machine is, the faster and deeper you can make cuts, while still maintaining accuracy and repeatability.

I say all this to say that field reports on the Shapeoko are returning some of the same very good results, at a lower cost, with great scalability. Accuracy of a tuned machine is stated by the design team to be .003-.004, and has been measured lower (according to keyboard jockeys, who knows?). That's plenty for me.

Sure, you can get .003 accuracy while cutting aluminum on a ShapeOko. But don't be surprised if your depth of cut is limited to .005-.01" per pass in order to achieve it.

Or the Shapeoko, which I could probably build at 4'x8' for around $700.

The Shapeoko 3 is about $800 for a 16"x16" machine. A machine designed to be 16"x16", scaled up to 4'x8', will have the rigidity of a wet noodle.

But again, it's all about what your expectations are. You can have a really good machine, or a really cheap machine. But not both. With CNC Routers, you really do get what you pay for. Although there are a few high priced kits out there that don't deliver. Stay away from expensive wood framed kits.

Your first post mentioned welded steel. If you can weld, a welded steel frame offers the most rigidity at the lowest cost of any DIY build methods.The tricky part is getting it flat and square. There a UK DIY CNC forum where welded steel frames are the norm. Like here, the members are more concerned about performance, than cost.

[extrastrength]

Fair enough, that's the kind of information I came here for. I suppose then that the results from both of these machines may have been overblown by some of their owners. Understandably, folks are always proud of their equipment.

A steel-framed machine sounds fantastic. I believe I can build one fairly square, I have access to a building that was designed as a frame shop, and I can TIG up the bed there.

The question then becomes: do you fellows have any suggestions for specific designs I should look at, or elements I should incorporate? Perhaps Joe's hybrid, substituting steel for most parts?

Anyway, suggestions are highly welcome.

[ger21]

The problem is going to be your budget. I don't see how you can get any quality components with what you're willing to spend.

[aarggh]

I think you'd easily look at blowing your $1000 budget just on the steppers, cabling, cable chains, controller, power supply, and linear rails. You'd still need drivescrews, and of course the materials to actually build the machine frame.

Around $2500 always seems to me to be the rough cost to build a capable machine with a decent cutting area of around 600mm x 900mm up to 1200mm x 1200mm (depending on the trade-offs made) or thereabouts. Bearing in mind scaling the machine up small percentages doesn't add too much to the overall cost. It's the initial outlay that's the killer.

cheers, Ian

[wizard]

Fair enough, that's the kind of information I came here for. I suppose then that the results from both of these machines may have been overblown by some of their owners. Understandably, folks are always proud of their equipment.

Sometimes pride is an issue and frankly building a working machine is a significant accomplishment if you haven't done it before. However I think the bigger parameter here is ones expectations. These machines might be perfect for what the user expected out of it.

A steel-framed machine sounds fantastic. I believe I can build one fairly square, I have access to a building that was designed as a frame shop, and I can TIG up the bed there.

Picture frames?

The biggest problem with welded up steel frames is that they should be stress relieved after weld up. However a lot of people do get away with skipping the stress relieve bit.

The question then becomes: do you fellows have any suggestions for specific designs I should look at, or elements I should incorporate? Perhaps Joe's hybrid, substituting steel for most parts?

I would suggest looking back through the forum and to check out as many designs as possible to wee what rocks your boat. There are so many approaches that it is really hard to say what would fit your needs. Due to your limited budget you may find that completing an entire aching to a high degree of quality cntruction may be too much. If this is the case ID sat art out with a slid table design with the thought of upgrading the rest latr.

Anyway, suggestions are highly welcome.

Your best bet at the moment is to read the forums to get a sense as to what other have done. Once you zero in on the working size and general design you then can start to design your own.

[wizard]

Err one other comment about welding.. You are a TIG welder so I kinda assumed that you already know the importance of fit up of parts. This is extremely I,portent to minimize stresses in the frame. You really want the tubing milled to precise lengths with the ends square, your ends need to be better than what the average cut off saw will provide. If not milled you need to become real handy with a file, square and a 4" grinder.

It would be wonderful if that frame shop you have access to has a large flat welding table to allow for a precise setup before welding. It really does pay to build in as much precision as possible. This will minimize the effort expended after weld up.

[extrastrength]

Alright, so I went back to the drawing board. I've started designing a steel machine, using largely C-channel, laser cut plates, and linear bearings. I'm pricing it out, and it's looking pretty good for a 4'x8' workspace.

I'll post up a DXF tomorrow, but a couple of quick questions:

1) I'm looking at building the bed frame out of C-channel, and I had thought about welding on supported linear rails (steel rod). But the thought occurred to me that I could run standard radial ball bearings on the flat three surfaces of the C-channel VERY easily. I think I've seen this done, but thoughts? I'll post my model tomorrow if it's not clear what I mean.

2) I'm thinking of using rack and pinion on X and Y, and lead screw or some such on Z. That said, it would be really nice not to design the Z-axis. Are there any fairly stout Z-axes out there (including the eBay ones)? I can always upgrade later.

Thanks.

[aarggh]

1) I'm looking at building the bed frame out of C-channel, and I had thought about welding on supported linear rails (steel rod). But the thought occurred to me that I could run standard radial ball bearings on the flat three surfaces of the C-channel VERY easily. I think I've seen this done, but thoughts? I'll post my model tomorrow if it's not clear what I mean.

I know some people do this very thing, but all I can say is YUCK!

Any steel you buy in tube, flat, or section will all vary minutely to some degree, so will require some amount of trueing up. With regards to using roller skate bearings on the edge of steel channels, the noise will be tremendous, but I always think it must be really crap dealing with the dust and debris, not to mention the layer of crud that naturally builds up on skate bearings no matter hwo clean the environment.

In a for a penny, in for a pound I reckon. Supported rails won't add that much to a build of that size I would think.

cheers, Ian

[extrastrength]

Thanks for the feedback.

"Skate" bearings had never crossed my mind... I was thinking 1 1/2" or so, precision SKF sealed radial bearings. Pretty cheap, but also tight.

Also, I was under the impression that anything we do in all steel would be loud?

I had also considered welding standard rod stock to the C-channel (or rectangular tube) sides of the bed.

Do you have a suggested source for supported rails and bearings?

[aarggh]

Thanks for the feedback.

"Skate" bearings had never crossed my mind... I was thinking 1 1/2" or so, precision SKF sealed radial bearings. Pretty cheap, but also tight.

Also, I was under the impression that anything we do in all steel would be loud?

I had also considered welding standard rod stock to the C-channel (or rectangular tube) sides of the bed.

Do you have a suggested source for supported rails and bearings?

I don't think the quality of the bearings would matter too much used in this application, one will just last longer than another potentially, but they'll all still run like crap.

A lot of people bolt down some aluminium stock to the welded frame, and then true/flatten that for the rails. I think this would be a much better approach. And before others start complaining about galvanic corrosion, steel and aluminium aren't that far apart on the scale you'd be worried about it.

For rails, ebay is your friend, I've bought tons off a great and very accomodating ebayer called econvenience I think the sellers name was, a lot use linearmotionbearings, but he does have what I would consider pretty woeful ideas of how to package goods, so I've only bought off him once.

cheers, Ian

[wizard]

Alright, so I went back to the drawing board. I've started designing a steel machine, using largely C-channel, laser cut plates, and linear bearings. I'm pricing it out, and it's looking pretty good for a 4'x8' workspace.

I'd strongly suggest tubing for some of the structural parts. It is much easier to work with than C channel. Especially for the gantry.

I'll post up a DXF tomorrow, but a couple of quick questions:

1) I'm looking at building the bed frame out of C-channel, and I had thought about welding on supported linear rails (steel rod).

That won't work for a couple of reasons. One is that linear bearing rods are hardened and sometimes chrome plated, you would most likely soften the rods too much. The second reason is that keeping that assembly straight while welding would be a bear. The third reason is that the supportes for those round linear rails are often aluminum.

Besides all of that repairing a welded up machine like that would be ugly.

But the thought occurred to me that I could run standard radial ball bearings on the flat three surfaces of the C-channel VERY easily. I think I've seen this done, but thoughts? I'll post my model tomorrow if it's not clear what I mean.

That can certainly "work" but it isn't the most robust solution. Such an arrangement might meet your expectations or it might be too sloppy to ever be good enough.

2) I'm thinking of using rack and pinion on X and Y, and lead screw or some such on Z. That said, it would be really nice not to design the Z-axis. Are there any fairly stout Z-axes out there (including the eBay ones)? I can always upgrade later.

Yes! There are all sorts of linear motion slides that you can but out there. They range from light weight air driven slides to heavy castiron slides, Dovetail ways and all that are extremely expensive new. E-Bay is an excellent place to shop for this stuff. You should also look for whole machines suitable for salvage. Like all things used you accept some risk.

Thanks.

[pippin88]

Unless you are on a very fixed income it's generally better to buy certain components rather than reinventing the wheel. Linear rails are one. You can spend countless hours making a poor substitute and end up with a useless result.

Welding creates a lot of distortion. You won't end up with straight rails if you weld rod as linear rails.

Just buy supported round rails if you have very limited budget (I'll never bother with them again though) or profile rails if you can stretch a bit further. They are available on aliexpress for good prices (15mm HIWIN from 'BST Automation' is my recommendation)

[whimsical]

pippin
Had a look at BST Automation. The li express site says genuine Hiwin rails and bearings but thier own site seems to indicate a no brand version.
What have you bought, genuine or copy. If the copy how good are they and can they be had in various preloads.

I will never deal through Ali as i have been ripped off and ALi only protect the buyer

[extrastrength]

Thanks very much for all the insightful replies.

For rails, ebay is your friend

Fantastic, I've been looking, and there really are quite a few options. I'm looking at something like this for the long axis: SBR16 Linear Rail Support Aluminm Assembles 2pcs L1500MM 4pcs SBR16UU | eBay

The question is, what diameter should I shoot for on the supported rail? The gantry might end up being 50 lbs (1/4" steel side plates, two supported rails, their crossmember, and rack/pinion).

I'd strongly suggest tubing for some of the structural parts. It is much easier to work with than C channel. Especially for the gantry.

Are you saying you would use tubing as the gantry crossmembers? I was looking at a design where they bolted in the linear rod (unsupported) as gantry members, which seems quite efficient, especially if used in conjunction with at least one supported rail. CNC Engraving Machine CNC Frame CNC DIY 3040 Aluminum Alloy Ball Screw | eBay

Yes! There are all sorts of linear motion slides that you can but out there. They range from light weight air driven slides to heavy castiron slides, Dovetail ways and all that are extremely expensive new. E-Bay is an excellent place to shop for this stuff. You should also look for whole machines suitable for salvage. Like all things used you accept some risk.

All very good suggestions. It may end up that I HAVE to design my own Z carriage, just because I'm thinking about using linear rails on either SIDE of the spindle, as well as two behind, as it is traditionally. From the standpoint of Z-axis deflection on harder materials (aluminum), that would seem to make more sense to me, especially if I just bolt them in (as above).

But, if I find one for cheap, I'll settle. Just have to find some machines in the local scrap yards.

Just buy supported round rails if you have very limited budget (I'll never bother with them again though) or profile rails if you can stretch a bit further. They are available on aliexpress for good prices (15mm HIWIN from 'BST Automation' is my recommendation)

You think profile rails are better than supported rods? Just asking, because I have no idea. All the nice machines I've ever seen use rods, from what I remember.

Out of curiosity, would that 15mm rail support a 5' long steel gantry? I see they make wider ones, but I can't find many for sale.

[aarggh]

16mm (supported rails I should specify) rails would be more than adequate for the Z, and given the size of the machine the other axis as well, it would be fine for them too, but thicker would be better if you can afford it. Going thicker will greatly strengthen the frame and increase the rigidity, and machining performance. You can also buy extra bearing assemblies for not much at all and double up there to increase the rigidity and distribute the load carrying ability.

I plan on using 4 rails per axis myself on a fairly large machine I'll be working on when my workload decreases. Some people think it's a waste, as it can be difficult to align, but if I can over-engineer it, why not I reckon.

Profile rails are far superior to round rail, and will support a vastly greater load. It can be common for machines up to 3m long to use a 16-20mm linear rail, certainly for your machine, anywhere from 12mm up or so would be adequate, 20mm would be very much more than adequate.

cheers, Ian