[tkms002]

OK how about this design.
This looking from the end of the gantry
Attachment 408242

[tkms002]

bigger pic

[wizard]

bigger pic

Not bad!

As for the steel beam you might try seeing if you have a local steel supplier to industry. Around here that might be Klien, Ryerson and a couple of others that escape my mind. The idea is to get the best price possible for the beam.

The primary problem with beam thickness selection is suitable thread holding capability. That is after you drill and tap the beam you want confidence that the threads will not strip out. Depending upon the rails being used you may be using 5 or 6 millimeter screws to fasten the rails to the beam. In an ideal world you would want at least 1.5 times the screw diameter in the beams thickness. Unfortunately that leads to a heavier beam than might actually be required from the deflection standpoint.

Note that all welding you do will add to beam distortion. How much depends on many factors but good fit up makes a big difference. You don't want to weld up your X rods in loose fitting holes for example. I tend to prefer plates welded into the beams with weld beads on all four walls, especially on thinner walled beams. Again though, you do run the possibility of excessive beam warpage, just have a plan to deal with this if it does happen.

As for not being bad, it is often the details that make or break a design. How that is all welded up and how the fit is made to the bearings are very significant. From all appearances you will need a machine shop to support machining the mounting surfaces for the linear bearings. If that thought has already been considered then you are OK. Otherwise I would build in a solution for mounting the gantry true to the linear bearings.

[tkms002]

Not bad!

As for the steel beam you might try seeing if you have a local steel supplier to industry. Around here that might be Klien, Ryerson and a couple of others that escape my mind. The idea is to get the best price possible for the beam.

The primary problem with beam thickness selection is suitable thread holding capability. That is after you drill and tap the beam you want confidence that the threads will not strip out. Depending upon the rails being used you may be using 5 or 6 millimeter screws to fasten the rails to the beam. In an ideal world you would want at least 1.5 times the screw diameter in the beams thickness. Unfortunately that leads to a heavier beam than might actually be required from the deflection standpoint.

Note that all welding you do will add to beam distortion. How much depends on many factors but good fit up makes a big difference. You don't want to weld up your X rods in loose fitting holes for example. I tend to prefer plates welded into the beams with weld beads on all four walls, especially on thinner walled beams. Again though, you do run the possibility of excessive beam warpage, just have a plan to deal with this if it does happen.

As for not being bad, it is often the details that make or break a design. How that is all welded up and how the fit is made to the bearings are very significant. From all appearances you will need a machine shop to support machining the mounting surfaces for the linear bearings. If that thought has already been considered then you are OK. Otherwise I would build in a solution for mounting the gantry true to the linear bearings.

I am going to be using CRS plate and CNCrouterparts bearing carriages. I think this will make bearing alignment less of an issue. I already installed them and had them working on that 3x4 be
Thanks for the feedback

[LeeWay]

In the same vein as the gantry you have chosen was not really sufficient to do the job without some error, the larger options you are looking at now are probably way too large for roller skate bearing trucks and steel flat bar as the guides.
Somewhere in the middle would probably have been fine.
There is no need for such a large beam that adds way too much weight to sub standard linear motion designs like the skate bearing on flat bar. I do use such a machine, but not as a main router. Not with such a large heavy beam either.
You will be having tolerance and wear issues far sooner and more often that you would have with a lesser weight gantry or better linear guides.
If you are going to stay with the skate bearings, I would probably just add a piece of angle iron to the back of the original gantry design for a little extra support. It will help some with both fex up and down as well as twist without adding excessive weight. Face the angle outward and you could use the outboard side to carry the echain and cable while giving the beam a larger purchase on the gantry supports.

[tkms002]

In the same vein as the gantry you have chosen was not really sufficient to do the job without some error, the larger options you are looking at now are probably way too large for roller skate bearing trucks and steel flat bar as the guides.
Somewhere in the middle would probably have been fine.
There is no need for such a large beam that adds way too much weight to sub standard linear motion designs like the skate bearing on flat bar. I do use such a machine, but not as a main router. Not with such a large heavy beam either.
You will be having tolerance and wear issues far sooner and more often that you would have with a lesser weight gantry or better linear guides.
If you are going to stay with the skate bearings, I would probably just add a piece of angle iron to the back of the original gantry design for a little extra support. It will help some with both fex up and down as well as twist without adding excessive weight. Face the angle outward and you could use the outboard side to carry the echain and cable while giving the beam a larger purchase on the gantry supports.

OK . So why do you think the skate bearing trucks are not strong enough? Even if the gantry is 150 lbs that is only 18.75 lbs on each bearing.(8 bearings supporting the load). A person skating on these bearings is applying a lot more weight than that. Maybe I don't understand something?

[LeeWay]

Not necessarily the strength of the bearings. That flat bar is cold rolled steel. It will wear depending on a lot of factors, but weight is definitely one of the main contributors.
Also a human skating has very little similarity to the forces the bearing will see on a cnc machine.
I have had to replace a few of the skate bearings over the years. Not hard or expensive, but they do wear out.

[ger21]

OK . So why do you think the skate bearing trucks are not strong enough? Even if the gantry is 150 lbs that is only 18.75 lbs on each bearing.(8 bearings supporting the load). A person skating on these bearings is applying a lot more weight than that. Maybe I don't understand something?

Ball bearings are not designed and were never intended to be used like that. They are designed to be a snug fit in a round hole, where the loading is evenly distributed around the fully supported outer race.
When the bearing is being used as a rolling "guide", it doesn't take that much force to deform the outer race, which can cause binding, and lead to failure.

It's a decent low cost system, but nowhere near "real" linear bearings.

[tkms002]

Ball bearings are not designed and were never intended to be used like that. They are designed to be a snug fit in a round hole, where the loading is evenly distributed around the fully supported outer race.
When the bearing is being used as a rolling "guide", it doesn't take that much force to deform the outer race, which can cause binding, and lead to failure.

It's a decent low cost system, but nowhere near "real" linear bearings.

Ok more context here. This will be a hobby machine, not a production machine. There must be thousands DIY routers out there with CNCRP bearing on them.
I certainly cannot afford Hiwin profile rails that are 9 feet long for a hobby machine.
Does that change anything?

[adam_m]

Just my 2 cents on the subject of CNCRP linear bearing since my machine was originally designed and I ran them for a few years on a hobby machine.

I'd just like to say that you're getting some great advice from some pretty smart guys here on the forum... sometimes it just doesn't pay to save a few dollars, you'll have to be the judge of that ultimately.

So, my personal experience was with my original gantry and it was nowhere near 150lbs, maybe 60lbs since it requires .250 CRS. The bearings although functional, as Gerry states, didn't hold up over time and use. I had to replace the bearings in the units after about 1.5 years of use, it seemed like the outer portion (retainer?) of the bearing seemed to flatten and created a bunch of slop.

Can it work, yes. How long and at what accuracy will all depend on usage?

So, even though my machine is for hobby use I ultimately replaced the CNCRP linear bearings with actual linear rails. Had I spent the additional money on rails to begin with I'd have been ahead of the game in cost and time.

Adam,

[tkms002]

Just my 2 cents on the subject of CNCRP linear bearing since my machine was originally designed and I ran them for a few years on a hobby machine.

I'd just like to say that you're getting some great advice from some pretty smart guys here on the forum... sometimes it just doesn't pay to save a few dollars, you'll have to be the judge of that ultimately.

So, my personal experience was with my original gantry and it was nowhere near 150lbs, maybe 60lbs since it requires .250 CRS. The bearings although functional, as Gerry states, didn't hold up over time and use. I had to replace the bearings in the units after about 1.5 years of use, it seemed like the outer portion (retainer?) of the bearing seemed to flatten and created a bunch of slop.

Can it work, yes. How long and at what accuracy will all depend on usage?

So, even though my machine is for hobby use I ultimately replaced the CNCRP linear bearings with actual linear rails. Had I spent the additional money on rails to begin with I'd have been ahead of the game in cost and time.

Adam,

ok so where do you get 9 foot lengths of large enough rail at reasonable prices?
I know the guys responding are being very helpful and are most knowledgeable and I greatly appreciate any and all responses. I am just trying to make sure I don't over spend for a hobby machine.
Again
Thank you all for all the great responses.

[adam_m]

I think that some rails are ground to be abutted to one another to create longer lengths, hopefully someone with personal knowledge will chime in.

I have some of this rail on my Z axis (Joomen) and they seem to be doing fine, they're not HIWIN... I did have to get them to replace three of the bearing blocks but they quickly responded, so far so good... YMMV. The rail is a THK knock off, and when I asked for specs on them they sent me THK documentation... the Chinese are just flaunting it in our face. I had some THK HSR25 bearing blocks laying around that fit the rail.

https://www.amazon.com/dp/B07287JBX6/ref=biss_dp_t_asn

I only see a 2000mm length though, maybe they have longer??... but if they can be abutted then you only have to purchase additional length rail so there is some cost savings there.

From a price standpoint, roughly CNCRP bearings are $40 ea. x2 per side $160 + CRS (Steel) $100 x2 = $360 plus shipping... let's call it $400ish. Joomen, $660 + additional rail call it $300.. Total $960. Anyway I think you get my point, you're spending about half the price on something that you may upgrade in the future.

Adam,

[tkms002]

I think that some rails are ground to be abutted to one another to create longer lengths, hopefully someone with personal knowledge will chime in.

I have some of this rail on my Z axis (Joomen) and they seem to be doing fine, they're not HIWIN... I did have to get them to replace three of the bearing blocks but they quickly responded, so far so good... YMMV. The rail is a THK knock off, and when I asked for specs on them they sent me THK documentation... the Chinese are just flaunting it in our face. I had some THK HSR25 bearing blocks laying around that fit the rail.

https://www.amazon.com/dp/B07287JBX6/ref=biss_dp_t_asn

I only see a 2000mm length though, maybe they have longer??... but if they can be abutted then you only have to purchase additional length rail so there is some cost savings there.

From a price standpoint, roughly CNCRP bearings are $40 ea. x2 per side $160 + CRS (Steel) $100 x2 = $360 plus shipping... let's call it $400ish. Joomen, $660 + additional rail call it $300.. Total $960. Anyway I think you get my point, you're spending about half the price on something that you may upgrade in the future.

Adam,

Yes, I am looking at all the angles but I can make the CNCRP bearing blocks myself. BUUUTTTT if they dont work they have no value. What about CNCRP for the Z axis? Any thoughts on that?
In a perfect world I would have enough money to get Hiwin rails. Gotta figure something else out. I see all kinds of kit and open source machines that use even weaker systems like v wheels on aluminum angle. I have even seen videos of them cutting aluminum like the Black Toe and others. Many using supported round rails as well but they are not "the best". Just wondering if I need "the best". Dang, last week I thought I had it all worked out.

[wizard]

There are bearings that are actually designed for roller duty which can solve at lest partially the issue of standard bearings of lasting in this application. Most of them are dual row bearings, some are V grooved others flat and even some with a half moon groove.

The problem is such bearings need to ride on harder steels than your run of the mill cold rolled. If you run them on cold rolled you are likely to wear the cold rolled significantly. While the bearings may do the job in a manual machine the way CNC’s work you can wear a rail out quickly. Even when you use parts all from one manufacture that is V roller bearings and the mating rails you will eventually get wear. Try Bishop Wisecaver for one example. In any event there are several types of these systems available, just realize that stiffness isn’t their strongest point.

In any event if we back up a bit you may have missed the point I was trying to make earlier. This point doesn’t in any way depend upon what type of bearing you are using, what I was getting at is that you need a robust way to mount the gantry up rights too the bearing blocks so that the gantry beam can be properly aligned. It is a detail often overlooked, but has to be considered if you have no way to machine the weldements afterward.

Ok more context here. This will be a hobby machine, not a production machine. There must be thousands DIY routers out there with CNCRP bearing on them.

Maybe. How well they will work depends upon usage and the users expectations. We here really can’t address your expectations because we don’t know what the are. I do however have a hard time calling a 9 foot long machine a hobby machine.

In any event let’s acknowledge one thing, ALL LINEAR BEARINGS WEAR OUT! That is likely something some don’t want to believe but it is true. The question in your case is how fast will they wear out.

.
I certainly cannot afford Hiwin profile rails that are 9 feet long for a hobby machine.
Does that change anything?

Not really! The real question is will the high failure rate of cheap bearings on Cold rolled be a problem for you. A good intermediate solution might be round supported rails. Often people are over eager to help you build the best machine possible. The title of your post kinda implies that you are concerned about building a really good machine so it makes sense that if you are concerned about the gantry beam that you should also be concerned about how it is mounted. He doesn’t make a lot of sense to build a really nice beamand then mount it on wobbly bearings.

[adam_m]

Yes, I am looking at all the angles but I can make the CNCRP bearing blocks myself. BUUUTTTT if they dont work they have no value. What about CNCRP for the Z axis? Any thoughts on that?
In a perfect world I would have enough money to get Hiwin rails. Gotta figure something else out. I see all kinds of kit and open source machines that use even weaker systems like v wheels on aluminum angle. I have even seen videos of them cutting aluminum like the Black Toe and others. Many using supported round rails as well but they are not "the best". Just wondering if I need "the best". Dang, last week I thought I had it all worked out.

Yeah, budget is always a killer.. you're going to have to make the choices of where you want to cut corners or extend your budget.

So, I originally had designed my Z axis to use CNCRP linear bearings but by the time had a piece of steel and bearings and router on it.. it was heavy! Would it work.. absolutely. I don't want to talk you into or out of anything, just wanted to share my thoughts and experiences with you to give you as much info. as possible before you make your decision.

The one problem I did run into when using the CNCRP bearings, was the CRS steel required for usage. Mine had a slight dip at one end of the material and there wasn't much to be done about that, it was also one of factors in convincing me to go with a different setup.

I also had bearings similar to CNCRP made from a buddy of mine... frankly after materials it would have been just as easy and cost effective to order them from CNCRP. I may still have the mastercam files for them... hmmm..

Adam,

[tkms002]

ok so how do you size the profile rails?
Is it by the load you are carrying? How do I determine how large of rails I need?
I think my z axis will weigh somewhere between 20 and 30 lbs.
The y with the z then weighs something like 150 lbs.
What size rails do I need for each axis?
Thanks

[BrokenHorn]

These are the rails and bearings I used. The site has many other options too.

They will provide up to 4000mm rails (13’+ ft).

It’s been a while since I compared everything and made the decision to use these, so I don’t remember the price difference, but I think they were slightly less than most Hiwin options I found.

I’m just getting my machine going, but happy with them so far.

https://www.anaheimautomation.com/pr...D=543&dsID=569

The 20mm rails I used all the way around my machine were probably a bit excessive, but I like having commonality of parts where I can and it made the design easier. This site has load rating info to help you determine what size you need.


Sent from my iPhone using Tapatalk

[tkms002]

How big is your machine?

[ger21]

Yes, I am looking at all the angles but I can make the CNCRP bearing blocks myself. BUUUTTTT if they dont work they have no value.

They'll most certainly work, and if I had to guess, I would think that CNCRP still sells a lot of them.
The most difficult thing about making recommendations, is that we have no idea of what you would find acceptable.
We're all making "hobby" machines, but some are more serious about their hobbies than others.

I see all kinds of kit and open source machines that use even weaker systems like v wheels on aluminum angle. I have even seen videos of them cutting aluminum like the Black Toe and others.

You can cut aluminum with ANY machine. Even little plastic machines using a Dremel for a spindle.The more rigid the machine, the faster and deeper you can cut it, and the better the finish will be.

Machines like the Blacktoe don't even cut wood all that well.

Wizard mentioned V wheels. They would be a step up from the skate bearing carriages, as CNCRP's PRO machines started with V bearings. But you really need to use the hardened V rails to take advantage of their capabilities. Shopbots have used them for decades, and the Mechmate uses them as well, although they use a larger size I think. You can get Chinese V wheels very reasonably, so the main cost is the rails. You'll also need aluminum or steel bar to mount the rails to.

ok so how do you size the profile rails?

Even small 15mm profile rails can handle incredibly high loads. But they are pretty small, and use small 4mm bolts, which can be trickier to work with.
My recommendation would be 20mm for everything. These are strong enough to literally move a full size car.
Automation Overstock offers a Chinese rail that's a very good value. low-cost BLH linear guideways equivalent to Hiwin HG series of you

A full set of 20 rails, 9' X axis, 5' Y axis, and 15" Z axis, with 12 blocks, would be about $1000. And they have long rails. The only problem is that you'll pay about $200-$250 to ship those long rails. But $1250 is still not that bad of a price, for a big machine.

How much would the CNCRP skate bearing setup cost? $400-$500??
Again, it's impossible for any of us to tell you what you'll be happy with. But if you aren't happy after going the cheap route, then you just spent $500 more than if you started with the better option.

I think my z axis will weigh somewhere between 20 and 30 lbs.

Don't be surprised if it ends up closer to 50lbs, if not more.

[adam_m]

How big is your machine?

I'm not sure who the question was directed to... but I have about 40"x50" of machinable space and I also have the Joomen 25mm and 30mm on the Y and X axis.