• Spring 2024 meetup in Calgary - tentative date Saturday, April 20/2024. Other regions are also discussing meet ups. If you want one in your area get going on organizing it! discussion
  • We are having email/registration problems again. Diagnosis is underway. New users sorry if you are having trouble getting registered. We are exploring different options to get registered. Contact the forum via another member or on facebook if you're stuck. Update -> we think it is fixed. Let us know if not.
  • Spring meet up in Ontario, April 6/2024. NEW LOCATION See Post #31 Discussion NEW LOCATION

Canadian brothers make their own CNC

Janger

(John)
Administrator
Vendor
:cool::D:D:D:cool:
Wow. Amazing. I wonder what they spent in the end?

In the comments the author says their budget was $15K but they spent less. The controller is "masso controller". I'll have to look that up. Talk about chutzpah those guys are fantastic.

I was thinking about the alignment of Z axis to the table. Tramming the head as well as making the table flat as the x y axis moves. Sounds tricky. One idea - what if you just tried to get the X and Y axis flat but ignored the trammming. Then used the spindle to flatten the table by facing it. Would that then give you a flat and trammed spindle and table?

Action box will be selling machines before long. Will they be made in Canada?
 
Last edited:

Mcgyver

Ultra Member
Talk about chutzpah those guys are fantastic.

I agree on that, they get more done a week than I do in a year. Their "get to it", "make it happen" drive is impressive.

On the build though, meah. Some bad information and while it looks great at a superficial level, is it? I really hate the guy who, on the sidelines, watches someone run through walls and get 95% of it right and then sits from the armchair criticizing the 5%. However, for posterity and to make it an interesting discussion, today, I take one for the team, and will be that guy :D. My critique is based the claimed goal, being an alternate to a "professional grade" cnc mill, which implies a level accuracy. I don't think they could have possibly achieved anything close to that.

A major, catastrophic flaw is nothing was done to insure the accuracy of the way geometry. Parallelism of the linear bearings means nothing (except they won't bind). What matters is, are they straight? The rails are very flexible and will conform to whatever you bolt them to. You cannot make much of a machine if you are relying on the flatness of hunks of cold or hold rolled (or your floor lol) For that matter, the material will have a high level flex over that length. The fundamental accuracy of the machine is determined by the flatness of whatever the rails sit on, and achieving flatness, to the leveled needed for a reasonably accurate machine tool, is far from trivial. Its a key element in ending up with a good machine, and one the most expensive and time consuming aspects of creating a machine. They skipped/ignored that as far as I could see.

Epoxy granite is amazing stuff and holds great potential for machine tool builds, especially when combined with fabrications. EG is comparatively weak and flexibly compared to steel or cast iron, but it brings vibration damping ability to about 2x that of cast iron. btw, they should know, vibrations are damped, not dampened. No water is involved. This works because energy is absorbed as the wave goes through the boundary layer of two unlike materials - stone and epoxy. His lecture says the rock is added because its cheap filler and second it adds mass. No, its because its presence creates the boundary layer, the large surface area, between the epoxy and granite, that is absorbing the energy - basically converting motion into heat. In cast iron this happens the same way: the boundary layer between different materials - the graphite and the metallic structure. Calling it "dampening", thinking the granite is just filler and using stone rather than sand (far more boundary layer) is a not a huge fail, but still, its says they don't really undertabnd what they are doing and that qualifies as armchair comment worthy :).

He's also dead wrong with the claim that casting have to left for several years. That's a once upon a time, old wives tale. Currently engineering/science is that they don't move about on their own. (they or any material can move if you machine it, that's movement as internal stresses are changed and a new equilibrium of forces is reached, but they don't keep wandering about on their own). Manufactures do not age machine tool castings and have not do so for a long time. Once they are stress relieved they are stable. I take offense at lecturing when you don't know what you are talking about (hopefully I'm not too guilt,y too often :) )

More use of steel (for rigidity) combined with EG (for damping) would imo be better. They did some, but they didn't normalize or stress relieve the fabrication. That puppy's going to move as its machined and won't end up an accurate structure.

Same with the table. A big slab of stress relieved case iron would have been a lot better, but if using steel it should have be sent for stress relieving first. Even hot rolled; machine a slab like that on one side and its going to spring into new shape.

I've used those low budget linear rails before. They are not worthy of spot in a $15k machine tools....which I'd say is really is 40,000+ machine tool....who's time is free?

I didn't see how they got everything square. Maybe the did, but this not easy and is critical to performance. There is no mention of it in the video and it would have been challenging and a mission critical step...makes think because of its absence, like flatness, that it was skipped.

I do admire what they were able accomplish, so feel a bit like schmuck being so critical. However these finer points of machine design and manufacture are what cost the money and are what makes the machine perform (accuracy etc). They deserve mention in the context of understanding of whats going and what needs to into a machine tool.

I just scrapped out a Mazak. Japanese, best of quality cnc mill with a tool changer. With whatever wear was present it still would far more accurate than lengths of hot rolled steel would provide. Box way vs linear rail. Japanese servos and spindle. Mostly it need someone to put a bunch of time into working out and fixing/replacing the control system....but you can't give them away. I think rebuilding that, even if scraping was required, would yield a far better result for less cost.
 
Last edited:

PeterT

Ultra Member
Premium Member
I was wondering out loud about the (stone + epoxy) mixture too. I'm guessing the larger size was more convenient but maybe also result in a significantly lower counts per volume of rock to rock contact? Therefore more epoxy filled void space? Therefore cured structure under load more resembles epoxy vs aggregate? I've had some limited experience with (sand + epoxy) molds for composites work. The important part as I recall was small particle size & very consistent sorting in order to minimize net epoxy filled porosity. But that was more about mold stability basically under its own weight & heat/cool cycles vs. what a machine would require.

When you have seen commercial machines with composite fill, what kind of material & size are they using?

Boy my Excel cartoon making is getting a workout lately LOL
 

Attachments

  • SNAG-1-27-2022 002.jpg
    SNAG-1-27-2022 002.jpg
    12.3 KB · Views: 4

Mcgyver

Ultra Member
When you have seen commercial machines with composite fill, what kind of material & size are they using?

Harcrete in Hardinge mills is one example. I've seen example of it used is some quite large machines made of filled fabrications, but I couldn't name names. It is a neat thing, weld up a great big huge assembly, stress relieve it, fill it with EG and get better than cast iron performance.

imo the stones vs sand isn't a big fail, it will still work but perhaps not as well as it might. Just sort of supports that they don't full get whats going on when they talk about whats going on :). There might be practical reasons for giving up a a bit of performance using stones, but the way they explained the function of the rocks, I don't think they knew what their function is. Its not filler and mass, its boundary layer.

How did you get those chickens to lay orange eggs? :)
 

Dabbler

ersatz engineer
Many CNC machine builders use this exposy/granite mixture because of its superior stability. The only optimization that could have been made is to use different sizes of granite chips, down to particles.

It turns out that by using graduated particles you use even less expoxy, but it becomes very difficult to mix at some point.
 

Janger

(John)
Administrator
Vendor
I appreciate the comments @Mcgyver you bring up good points not being completely critical but asking how did they make the machine square flat true & plumb?

For the video in the comments the builders do talk about accuracy and what they did to address that and/but promised a future video on that topic. I wonder how good/crappy the first Haas was? In some haas factory video I watched they were facing the mounting surfaces of mill castings for the rails with even bigger mills. VM6 size machines with stairs. That machine was milling VM1-3 type of castings.
Perhaps there is a Gingery type of approach to use the mill to finish the mill that explains the chicken egg problem. How do you make a mill if you don’t already have an even bigger mill?
 
Last edited:

Janger

(John)
Administrator
Vendor
What would the approach be to stress relieve a mill sized fabrication? Big furnace ? Slow cooling? Then mill surfaces to flatness?
 

Dabbler

ersatz engineer
Castings used to be aged outdoors going through summer/winter to stress relieve (up to 7 years). Large environmental chambers can do the same much faster.
 

Janger

(John)
Administrator
Vendor

haas factory tour. How haas verifies their assemblies is shown. I found it surprising lots of dial indicators and sweeping surfaces. Old school methods still work. At one point the tour guide says they go through 33,000 ball screws a year. Divide by 3 and you get 10,000 machines a year. Of course that’s not completely accurate some machines have more or less but I imagine it’s in the right area. That’s a lot of machines!
 

Mcgyver

Ultra Member
What would the approach be to stress relieve a mill sized fabrication? Big furnace ? Slow cooling? Then mill surfaces to flatness?
yes, a big furnace. take it up to critical temp and a slow cool. Bigger than a bread box size, its something most everyone would outsource.....and there are some big furnaces out there so for someone serious about making a good machine its not really a challenge, just a cost.

After milling, to get to surfaces to commercial cnc machine levels of accuracy, you'd have to grind or scrape imo. One of the guys here worked for XLO, be curious to hear how they did it. My guess is planing and then grinding. I've looked in detail at how Standard Modern did it (spent a few days in the business studying everything) which was a combination of grinding and scraping. If one hasn't (suffered) the indoctrination of reconditioning machines, it might not be that easy to see just how big a deal it is getting things really flat, and if you want a machine to deliver X accuracy, it itself has to be some level more accurate than X

I don't see indicators as old school, but however you come at it, you have to have a comprehensive methodology to get everything that needs to be quantified, quantified. I doubt they did anything for flatness (didn't see anyway it could have been checked, none of tools present you'd need), and they were silent on squareness. I did see the comment about following up on this, I'm guessing it didn't occur to them and they didn't have an answer.

What could they have done?

How do you get flatness? You need a sufficient way to reference it and compare that to the work. This could be a large camel back (I've got them up to 60", might have been enough) which they could have bought or a autocollimator. Possibly you do an electronic version of the taught wire thing or rough it out with a Starrett 199 level (rather dicey). Probably lots of other ways (not my expertise), I just know from reconditioning how important flatness is and how very difficult it is to achieve. It does strike me that the right tool for the job would be an autocollimator.

Here's an idea to get flatness without having to normalize a fabrication, mill grind or scrape. Lets say the finish of cold rolled is good enough as a surface to bolt the rails to, and it probably is. Create a fixture/frame aspect to the member that would allow you to shim along the length of the steel. Maybe a large beam or channel running the length of the member that the CR is attached on top of (via shims). Like if the member is 14" tall, have a 12" channel running down the middle of it - the steel would stiffen the EG as well. Then shim, shoot and repeat until it shot perfectly flat (like a surface plate) along its length. Pour the EG such that it came up to and under the flat, shimmed CR steel so its fully supported. Given EG is stable, the steel should stay flat. Much like when installing a big piece of equipment you get it level and aligned with shims then grout underneath. It would be a bit painstaking with the shimming, but doable I think and it avoids the more painstaking task of scraping (I'll rule grinding out, few have the capacity for a machine that size, and those that do aren't trying to make DIY mills)
 
Last edited:

LenVW

Process Machinery Designer
Premium Member
yes, a big furnace. take it up to critical temp and a slow cool. Bigger than a bread box size, its something most everyone would outsource.....and there are some big furnaces out there so for someone serious about making a good machine its not really a challenge, just a cost.

After milling, to get to surfaces to commercial cnc machine levels of accuracy, you'd have to grind or scrape imo. One of the guys here worked for XLO, be curious to hear how they did it. My guess is planing and then grinding. I've looked in detail at how Standard Modern did it (spent a few days in the business studying everything) which was a combination of grinding and scraping. If one hasn't (suffered) the indoctrination of reconditioning machines, it might not be that easy to see just how big a deal it is getting things really flat, and if you want a machine to deliver X accuracy, it itself has to be some level more accurate than X

I don't see indicators as old school, but however you come at it, you have to have a comprehensive methodology to get everything that needs to be quantified, quantified. I doubt they did anything for flatness (didn't see anyway it could have been checked, none of tools present you'd need), and they were silent on squareness. I did see the comment about following up on this, I'm guessing it didn't occur to them and they didn't have an answer.

What could they have done?

How do you get flatness? You need a sufficient way to reference it and compare that to the work. This could be a large camel back (I've got them up to 60", might have been enough) which they could have bought or a autocollimator. Possibly you do an electronic version of the taught wire thing or rough it out with a Starrett 199 level (rather dicey). Probably lots of other ways (not my expertise), I just know from reconditioning how important flatness is and how very difficult it is to achieve. It does strike me that the right tool for the job would be an autocollimator.

Here's an idea to get flatness without having to normalize a fabrication, mill grind or scrape. Lets say the finish of cold rolled is good enough as a surface to bolt the rails to, and it probably is. Create a fixture/frame aspect to the member that would allow you to shim along the length of the steel. Maybe a large beam or channel running the length of the member that the CR is attached on top of (via shims). Like if the member is 14" tall, have a 12" channel running down the middle of it - the steel would stiffen the EG as well. Then shim, shoot and repeat until it shot perfectly flat (like a surface plate) along its length. Pour the EG such that it came up to and under the flat, shimmed CR steel so its fully supported. Given EG is stable, the steel should stay flat. Much like when installing a big piece of equipment you get it level and aligned with shims then grout underneath. It would be a bit painstaking with the shimming, but doable I think and it avoids the more painstaking task of scraping (I'll rule grinding out, few have the capacity for a machine that size, and those that do aren't trying to make DIY mills)
Just reading some of your CNC discussion . . .
I used to ‘stress relieve’ huge fabricated frames for forming machines we built for production lines at Big O Machinery. CleaverBrooks in Stratford had heat treating ovens that could handle my projects fabs, typically the size of semi trailers.
The formers we built had fabricated carriages that Would cycle back and forth on linear bearings during production runs, 24 hrs/day, 7 days a week.
I had working relations with West Heights Mfg, BlackClawsonKennedy and other large machining companies with pit mills the size of a hockey rink.

@McGyvers idea of shimming CRS is an option for maintaining flatness in the assemblies, but, it is an extremely time consuming exercise.
I guess it would be a great project for the CHMWs.
 

Mcgyver

Ultra Member
@McGyvers idea of shimming CRS is an option for maintaining flatness in the assemblies, but, it is an extremely time consuming exercise.
I guess it would be a great project for the CHMWs.

Its not how I'd go at it (I'd start with a Mazak like I just threw in a bin), but it was the best I could come up that would only need very simple tackle - the only addition would be the purchase of a (used) autocollimator. Still, for anyone who's scraped even a medium size surface, shimming could be the less tedious of the two (even if its a static bearing surface, far less points per inch than if something is sliding on it).
 
Last edited:

Susquatch

Ultra Member
Administrator
Moderator
Premium Member
A major, catastrophic flaw is nothing was done to insure the accuracy of the way geometry. Parallelism of the linear bearings means nothing (except they won't bind). What matters is, are they straight?

But but but but..... They have a digital read out that says it good to a 0.00000001 inches! It says so right there on the readout! Gotta be true!

Sorry, it's a favorite bitch of mine and I couldn't help myself. My bad.

I'll throw my support for your critique in here too. Good on you for taking one for the team.

Nonetheless, I'm truly happy for the guys doing the channel. Good on them for doing that too.
 
I agree on that, they get more done a week than I do in a year. Their "get to it", "make it happen" drive is impressive.

On the build though, meah. Some bad information and while it looks great at a superficial level, is it? I really hate the guy who, on the sidelines, watches someone run through walls and get 95% of it right and then sits from the armchair criticizing the 5%. However, for posterity and to make it an interesting discussion, today, I take one for the team, and will be that guy :D. My critique is based the claimed goal, being an alternate to a "professional grade" cnc mill, which implies a level accuracy. I don't think they could have possibly achieved anything close to that.

A major, catastrophic flaw is nothing was done to insure the accuracy of the way geometry. Parallelism of the linear bearings means nothing (except they won't bind). What matters is, are they straight? The rails are very flexible and will conform to whatever you bolt them to. You cannot make much of a machine if you are relying on the flatness of hunks of cold or hold rolled (or your floor lol) For that matter, the material will have a high level flex over that length. The fundamental accuracy of the machine is determined by the flatness of whatever the rails sit on, and achieving flatness, to the leveled needed for a reasonably accurate machine tool, is far from trivial. Its a key element in ending up with a good machine, and one the most expensive and time consuming aspects of creating a machine. They skipped/ignored that as far as I could see.

Epoxy granite is amazing stuff and holds great potential for machine tool builds, especially when combined with fabrications. EG is comparatively weak and flexibly compared to steel or cast iron, but it brings vibration damping ability to about 2x that of cast iron. btw, they should know, vibrations are damped, not dampened. No water is involved. This works because energy is absorbed as the wave goes through the boundary layer of two unlike materials - stone and epoxy. His lecture says the rock is added because its cheap filler and second it adds mass. No, its because its presence creates the boundary layer, the large surface area, between the epoxy and granite, that is absorbing the energy - basically converting motion into heat. In cast iron this happens the same way: the boundary layer between different materials - the graphite and the metallic structure. Calling it "dampening", thinking the granite is just filler and using stone rather than sand (far more boundary layer) is a not a huge fail, but still, its says they don't really undertabnd what they are doing and that qualifies as armchair comment worthy :).

He's also dead wrong with the claim that casting have to left for several years. That's a once upon a time, old wives tale. Currently engineering/science is that they don't move about on their own. (they or any material can move if you machine it, that's movement as internal stresses are changed and a new equilibrium of forces is reached, but they don't keep wandering about on their own). Manufactures do not age machine tool castings and have not do so for a long time. Once they are stress relieved they are stable. I take offense at lecturing when you don't know what you are talking about (hopefully I'm not too guilt,y too often :) )

More use of steel (for rigidity) combined with EG (for damping) would imo be better. They did some, but they didn't normalize or stress relieve the fabrication. That puppy's going to move as its machined and won't end up an accurate structure.

Same with the table. A big slab of stress relieved case iron would have been a lot better, but if using steel it should have be sent for stress relieving first. Even hot rolled; machine a slab like that on one side and its going to spring into new shape.

I've used those low budget linear rails before. They are not worthy of spot in a $15k machine tools....which I'd say is really is 40,000+ machine tool....who's time is free?

I didn't see how they got everything square. Maybe the did, but this not easy and is critical to performance. There is no mention of it in the video and it would have been challenging and a mission critical step...makes think because of its absence, like flatness, that it was skipped.

I do admire what they were able accomplish, so feel a bit like schmuck being so critical. However these finer points of machine design and manufacture are what cost the money and are what makes the machine perform (accuracy etc). They deserve mention in the context of understanding of whats going and what needs to into a machine tool.

I just scrapped out a Mazak. Japanese, best of quality cnc mill with a tool changer. With whatever wear was present it still would far more accurate than lengths of hot rolled steel would provide. Box way vs linear rail. Japanese servos and spindle. Mostly it need someone to put a bunch of time into working out and fixing/replacing the control system....but you can't give them away. I think rebuilding that, even if scraping was required, would yield a far better result for less cost.
If you are going to be that picky all materials have creep and flow, the real question becomes how much and how fast. Additional all material compress under load, with some rebound, again something that is important. We won't even get into the thermal expansion/contraction differentials in a machine and materials used.

As a quick reference the CN Tower is shorter and fatter than it was new. (So are all the buildings).

As to accuracy see my post about Jack and and his wooden metal working machinery and accuracy. It can be done with simple tools and patience.

It is not about the tools but the person using them that gets the results.

Another example I knew a machinist that ground and over length shaft (11ft on a 10ft machine) to 0.0002 tolerance the entire length. How by ear and the sound the machine made as it removed material making micro adjust as the machine ran.

So can they produce an accurate machine, easier than lets say 30 years ago. Knowledge and the use of less expensive accurate measure tools makes it substantially easier. The only question becomes are you smart enough?
 

Susquatch

Ultra Member
Administrator
Moderator
Premium Member
I am not as familiar with machine castings as I'd like to be. But I do know a lot about engine castings and I'd bet a lot of that applies.

The largest casting in a vehicle is the engine block. Residual stresses in an engine block casting is a major problem that has been largely scienced to death over the decades. Scrap rates have come way down and the science has improved dramatically.

There are two major kinds of residual stresses: 1. Macroscopic residual stresses that are the result of heat treatment, and machining. 2. Microscopic residual stresses that are the result of internal factors that are caused by the metallurgy, casting pour, uneven solidification, and complex shapes, and of course things like cast-in cylinder liners.

The industry has developed many ways to deal with these issues and the progress made has been nothing less than amazing.

Things as simple as shaking the casting during the initial solidification phase to uneven pour rates to metallurgy control to huge ovens to control the casting stress development.

I don't know how practical it is to apply some of those technologies to large machine manufacturing. So I would tend to favour an efficacy (bang for the buck) approach to the process. I don't think a foundry producing mill tables or knees or bases is going to do a lot of shaking in process...... And I question their ability to control the basic metallurgy too - most likely whatever they can get their hands on goes into the pot! Last but not least, I doubt they pay any attention at all to the shape of their molds or their pours to control the solidification and localized settling processes. Most likely, it's just a big huge amateur operation.

Sooooo......., that means the only real option (assuming that they even care) for the large machinery industry is to minimize residual tensile stresses through an extensive heat treating operation. Frankly, I doubt they do that. I think they just crank them out and let their customers deal with the issues.
 
Top