[Goemon]

I called a carbon fiber supply house yesterday to see if it was possible to use stock beams instead of trying to cast my own. The biggest ones they had in stock were 2" x 4" , which didn't seem sufficient for my design, which has a 6-foot span to cover; I was thinking more like 4" x 8". The person I talked to suggested that instead of trying to glue up 4 of these 2x4s (which cost about $400 each), they could build a beam for me out of flat stock, using carbon angle in the corners. Buy he cautioned me that while it would be considerably lighter than a steel beam, it would be nowhere near as rigid. Apparently carbon fiber is rated at 8,000,000 psi, but steel is 30,000,000. I'm not exactly sure what these numbers represent, but it sounds like steel is about four times stiffer. However, a 4" x 8" steel beam with a 1/4" wall weighs 19.3 lbs per foot, so I'm looking at 116 lbs for the gantry beam alone, before I add end caps, a heavy ball screw and the Z axis assembly with spindle, motor, etc. (not to mention the 5th and 6th axes assembly). I suppose I could use aluminum, but I'd want to go with a thicker wall, which obviates some of the weight advantage. On the other hand, if I was building it out of carbon fiber, I could put stiffeners inside the beam, which would help rigidity. I'm still on the fence here...

It sounds like you spoke to someone who didn't know enough to be giving other people advice. Ignoring the fact that stiffness is not measured like that, his numbers are wrong. Perhaps he was quoting the tensile strength of their beams specifically (as that is measured in PSI) or just as likely, he got it the wrong way round - I.e. Carbon fiber's tensile strength (Howe much it can be stretched before snapping) is 4 x greater than steel. As mentioned here:

https://arstechnica.com/science/2015...-carbon-fiber/

Some of the cf I-beams, tubes and channels I see for sale are most certainly too thin and expensive to be a viable alternative to steel. In general, cf doesn't like long flat parts with 90 degree angles so a CNC gantry requires more design thought.

A properly designed carbon fiber part will be stronger and stiffer than the equivalent steel. "Properly designed" is the key point with carbon fiber parts. It makes far more of a difference than it does with metals. A part is not automatically stronger, stiffer or lighter just because it's made of carbon fiber. It is definitely possible to make a weaker part. An aluminum plate can be stronger than steel if it's 5" thick and the steel is 1".

For example of how important cf design is... if you cure a single flat layer of 2x2 twill carbon fiber, you can bend it and tear it with your hands. You cure that same piece of cf in a tube shape, you would struggle to bend it at all. If you stood on the end of the tube, it would probably support your weight.

Where cf has the advantage is that parts can be designed in a way that puts all the strength and stiffness where you need it based on the direction of the forces. As you can see in the video comparing the cf and steel drive shafts. The steel shaft bent and then broke with 1/3 of the force of the cf part. The CF part hardly bent at all before it broke. The cf shaft was designed to take that type of force in that direction.

The better approach with CF is to start with how much strength and stiffness you need and then specifying the part to deliver that. In a diy environment without engineering expertise (which is most of us here), more trial and error would be involved. I.e. If a part is not strong or stiff enough, you add more layers (in the correct orientation) until it is.

Personally, I would never buy ready made cf plates, tubes or beams. Most companies over here charge a fortune for very average parts. You'll pay $400 for $50 worth of carbon fiber. The cheapest 1.5" x 48" round tubes on eBay cost over $80. I can make the same tube myself for $15.

If money is no object and you want to buy ready-made, why not try getting a quote from Comotech? They export directly to America and the rest of Europe. They have specific expertise and experience in designing and building carbon fiber gantry beams for CNC machines. They have the capability to build beams of 12 feet (or even longer). They could make a gantry for you that is stronger and stiffer than a steel one or lighter for the same strength.

I hate the idea of ordering an expensive custom cf part from someone who doesn't understand what is needed in a gantry or how best to achieve it. Starting with cf angles and flat stock doesn't sound like a good design plan at all. It sounds flimsy. You don't want carbon fiber for the sake of it. You have to design the part properly to get the benefits.

If it were me, if faced with the choice of paying a fortune for their suggested approach with cf, or using regular steel or even aluminum beams, I would go with the steel or aluminum.

[ger21]

Most companies over here charge a fortune for very average parts. You'll pay $400 for $50 worth of carbon fiber. The cheapest 1.5" x 48" round tubes on eBay cost over $80. I can make the same tube myself for $15.

The need to make money always makes things cost more.
If it cost me $15 and an hour of my time to make that tube, I'd certainly be selling it for at least $80.

The real cost savings here comes at the expense of your time.

I'm building a large machine from wood, using a lot of very time intensive techniques. So, it's very inexpensive, but takes a lot of my time. I think carbon fiber construction is similar.

[Goemon]

The need to make money always makes things cost more.
If it cost me $15 and an hour of my time to make that tube, I'd certainly be selling it for at least $80.

The real cost savings here comes at the expense of your time.

I'm building a large machine from wood, using a lot of very time intensive techniques. So, it's very inexpensive, but takes a lot of my time. I think carbon fiber construction is similar.

Can't argue with that. Just like building a CNC machine, there is a huge saving if you do it yourself ( and a lot of satisfaction). I feel like cf parts and CNC machines seem to have much higher margins than most other businesses.

I don't think it's all justified by labor costs though. When I make a cf rifle stock, that is a really time consuming lay-up with a lot of difficulty plus high wastage costs. There are good reasons for the high prices charged by companies like Manners and McMillan. With things like the simple round tubes, it's just pure profiteering sometimes.


Making my own cf round tubes and plates was the first thing I learned. It's very easy and not time consuming at all.

I feel the same about some of the CNC mechanical kits where people are charging $1500 for $300 worth of aluminum extrusions, plates and screws. Some people think "great, they've done the work for me", diy fans like us think "what could I build for than cash if I do it myself and put their profit margin towards superior components".

[ger21]

My biggest issue with most CNC parts and kits for sale, is that the people selling them don't know what they are doing in most cases. Not only can you usually build the same thing for less money, in most cases, you can build something better for less money.

[Goemon]

This is really interesting. I have a Multicam router with 2x3 meter work area so the gantry is very long and flexes if i engrave too fast so ivve been thinking about how to stiffen it up.

Will be following closely!

What is the gantry made of and what are you cutting when it flexes?

Those Multicam branded machines look well made in the pics. The gantry beams look thick, it's surprising that flex would be a noticeable issue.

[ack1]

What is the gantry made of and what are you cutting when it flexes?

Those Multicam branded machines look well made in the pics. The gantry beams look thick, it's surprising that flex would be a noticeable issue.

This is my machine when i got it.


Side view of the gantry with the cap off. IF i remember correcly steel was about 8mm thick.



Sometimes i do whole alu sheets of engraved pieces and if I go too fast (gantry makes a sudden change from forward to reverse or vice versa, in the Y direction) the gantry flexes just by its own. Specially if im in the middle of the table.
The gantry span is somewhere around 2.3meter so its not surprising.
These numbers are very small but you can cleary see what i mean. Look at number 7.

[ack1]

This is an example of engraving. In the last few seconds i pan out and you see im only half way of a full sheet and it takes forever. And I cant go faster than this basically.
https://www.youtube.com/watch?v=w1hBhSXZb44

[Goemon]

My biggest issue with most CNC parts and kits for sale, is that the people selling them don't know what they are doing in most cases. Not only can you usually build the same thing for less money, in most cases, you can build something better for less money.

That's so true.

There is some Canadian company selling a "heavy duty gantry kit" on eBay right now for $4,000... you can see in the pic that it's just a couple of aluminum extrusions and two of the thinnest looking aluminum plate risers I have seen.

Now... I'm no communist... a company has to make some profit but.... for that sort of cash I would expect some quality custom design with thick steel components and I certainly wouldn't want to see a list of things it didn't include....

You can buy 1000mm aluminum extrusions for $10....

[NIC 77]

I think there may be a number of reasons why this is happening. I'm skeptical that this all because of your gantry tube flexing. Interesting that they don't have any ribs in the gantry tube. Could this be caused by servo over and undershoot? The straight lines look almost like sine waves. My first guess would be servo over / undershoot, and the lines that look like sine waves is the closed loop control trying to bring things back in order each time it misses the mark.

I've been reading some of the posts here, so here's my 2 cents.

Lots of mention of strength. Strength doesn't matter when building a CNC machine. If you're designing for strength you've already failed because stiffness is the important factor. The modulus of elasticity and the geometry of the tube determine a tube's stiffness.

The stiffness to weight ratio of carbon fiber is going to be superior to steel.

The rotor inertia of the motor, any gear reducers, or ball screws will still be need to be accounted for, and these use a significant amount of torque to accelerate, so you can get better acceleration for a given motor with a lighter gantry but the improvement will be far less than you might at first assume. Also, there may be a need to consider over and undershoot of the system if the load to motor inertias have too much of a miss match.

I can see carbon fiber gantries being beneficial in a couple instances.

If you can make molds that are easy to mass produce parts that are nice and flat for a reasonable price. Shipping would be less expensive than a comparable steel part as well. Or for the DIY person, if you can make inexpensive molds that have good mounting surfaces out of wood for example, perhaps that could be easier than working with steel or aluminum.

Also with a mold, you have more ability to change the geometry, for example, to mount your motor and run a ball screw in between the rails in such a way that you don't have a large offset from the gantry structure. Big commercial machines can do this with custom castings. Working with square steel tube sections, this is difficult to come up with good easy ways to do the same thing.

Or for gantries that have a very long span. The deflection of a beam is determined by the length cubed, so the relationship is not linear. For a long gantry on a specialized machine, the superior stiffness to weight ratio of the carbon fiber could be a good design choice.

I wish I had more experience working with carbon fiber, not for building a CNC machine, but for some other projects.

Goeman,

It's interesting what you've come up with. It looks like your gantry is more of a flat plate than a box so far. Are you going to add more pieces to it? It's definitely nice that you're pioneering this and sharing your results. I'm sure many people on this forum, including myself, are looking forward to seeing more posts from you.

[Goemon]

I think there may be a number of reasons why this is happening. I'm skeptical that this all because of your gantry tube flexing. Interesting that they don't have any ribs in the gantry tube. Could this be caused by servo over and undershoot? The straight lines look almost like sine waves. My first guess would be servo over / undershoot, and the lines that look like sine waves is the closed loop control trying to bring things back in order each time it misses the mark.

I've been reading some of the posts here, so here's my 2 cents.

Lots of mention of strength. Strength doesn't matter when building a CNC machine. If you're designing for strength you've already failed because stiffness is the important factor. The modulus of elasticity and the geometry of the tube determine a tube's stiffness.

The stiffness to weight ratio of carbon fiber is going to be superior to steel.

The rotor inertia of the motor, any gear reducers, or ball screws will still be need to be accounted for, and these use a significant amount of torque to accelerate, so you can get better acceleration for a given motor with a lighter gantry but the improvement will be far less than you might at first assume. Also, there may be a need to consider over and undershoot of the system if the load to motor inertias have too much of a miss match.

I can see carbon fiber gantries being beneficial in a couple instances.

If you can make molds that are easy to mass produce parts that are nice and flat for a reasonable price. Shipping would be less expensive than a comparable steel part as well. Or for the DIY person, if you can make inexpensive molds that have good mounting surfaces out of wood for example, perhaps that could be easier than working with steel or aluminum.

Also with a mold, you have more ability to change the geometry, for example, to mount your motor and run a ball screw in between the rails in such a way that you don't have a large offset from the gantry structure. Big commercial machines can do this with custom castings. Working with square steel tube sections, this is difficult to come up with good easy ways to do the same thing.

Or for gantries that have a very long span. The deflection of a beam is determined by the length cubed, so the relationship is not linear. For a long gantry on a specialized machine, the superior stiffness to weight ratio of the carbon fiber could be a good design choice.

I wish I had more experience working with carbon fiber, not for building a CNC machine, but for some other projects.

Goeman,

It's interesting what you've come up with. It looks like your gantry is more of a flat plate than a box so far. Are you going to add more pieces to it? It's definitely nice that you're pioneering this and sharing your results. I'm sure many people on this forum, including myself, are looking forward to seeing more posts from you.

My gantry beam is being made in two parts. I haven't finished the rear half yet so all you can see in the pic so far is the face plate which will hold the linear rails.

The face plate itself is not a flat structure. If you look at it from the rear, it has small crenellations which are almost like a couple of steps that follow the contours of the two steel plates (which are sandwiched between the layers of cf). There is around .75" of solid compressed carbon fiber as it stands today so that plus the choice (and orientation) of the materials makes the face place extremely rigid on it's own.

Even though the face plate is likely more than stiff enough for the purpose on it's own, the bulk of the stiffness will come from the rear half and the design of the filler that connects them.

Specifically, there is a layer of 1" long x .75" diameter cf tubes with .25" wall thickness made from 4x4 twill. 120 of these little tubes will be mounted at 90 degrees to the rear of the face plate and embedded / encapsulated in a mix of cf strands wetted out with epoxy (to fill the gaps between tubes). Behind the 1" tubes will be 6 additional tubes mounted parallel to the face plate that run across the whole length of the gantry beam. Aside from their length, the long tube specs are the same as the short tubes. This tube filler will be sandwiched in between another face plate of similar specs to the front plate.

I haven't decided yet but I am considering filling the inside of the tubes with urethane rubber to help dampen vibrations. I make the tubes using a roll wrapping process (which is very easy).

Obviously I will need to test my design but I am fairly confident that it will have enough stiffness and strength. My thinking is based on my general experience that rounded edges and cylindrical shapes produce maximum stiffness in cf parts. Plus, I can orient the weave in the tubes in a way that delivers maximum stiffness in the directions of the forces the gantry beam will be subjected to.

The way I see it, being able to create custom designs to put the strength where you need it most, is the key advantage to carbon fiber or fiberglass parts, so I'd be foolish not to make the most of it. Plus... I think a little extra creativity in the design is needed for long flat parts like a gantry beam. Long flat surface aren't naturally optimal for stiffness.

For anyone interested in designing their own cf gantry, those Compotech guys have a section on their website which describes their approach to engineering cf gantry beams and other cf machine parts for max stiffness. While I don't think that their cf gantry beam designs are the best they could possibly be, they have some interesting ideas that are worth borrowing, copying and steeling (like incorporating tubes into the inside walls of their beams).

As a general point, the only way to get experience with carbon fiber is to have a go at making something. It's a lot of fun and not overly difficult if the pattern is chosen wisely. By "wisely" I mean avoiding anything with small details, deep crevices and tight sharp corners until you have more experience. You would need to have a general interest though. The investment required to get started is not huge but you wouldn't want to buy all the resins, mold making materials, mold release products and polishes etc without a good reason.

[ack1]

NIC77, its accually running steppers. And the way the phenomena is most prominently present at the middle of the gantry makes me somewhat certain its flexing back and forth after a hasty change of direction.

EDIT: I feel we're a bit OT discussing my gantry by now so in closing i'd just like to add that if anyone has any tips on how to best stiffen this sucker up, shoot me a pm and i'd be most grateful.

[awerby]

If you mount a dial indicator under the middle of the gantry beam and put some weight on it, you can measure the deflection from the horizontal. That should tell you whether the irregularities in those letters are due to flex in the gantry or something else. If there's not enough deflection to account for your issues, then do the same thing with the spindle. To me, it looks more like something is flexing or oscillating side-to-side than up-and-down. It could be that something's come loose somewhere.

[wizard]

What is the gantry made of and what are you cutting when it flexes?

Those Multicam branded machines look well made in the pics. The gantry beams look thick, it's surprising that flex would be a noticeable issue.

From the pictures it looks like the Multicams are well made for the most part but those gantry beams just look spindly to me. For the most part gantry beams are the weak links in most router designs (this includes the beam supports). This is one reason why I'm following your thread closely to better understand how your beam design will work. Generally you would want a rather large cross sectional area, a big box beam if you will, to get the required stiffness. This is the same no matter what material you make the beam out of.

When it comes to a laminated carbon fiber beam I simply don't have the experience to say how one with a thin cross section will work in real life. You really can't tolerate much twist or deflection in the beam else you compromise with respect to the amount of driving force you can apply. That is you end up having to lower feed rate or more likely cut depth to avoid beam deflection.

In any event keep posting, I'm learning something new overtime I view this thread.

[ack1]

Awerby, it is flexing side to side. Well front to back if you are standing in front of the machine.

[ger21]

Glue some ribs inside the gantry tube. They can be as simple as plywood, with a good quality construction adhesive. Make sure they are a snug fit.
3-5 ribs evenly spaced should make it quite a bit more rigid.

[Goemon]

From the pictures it looks like the Multicams are well made for the most part but those gantry beams just look spindly to me. For the most part gantry beams are the weak links in most router designs (this includes the beam supports). This is one reason why I'm following your thread closely to better understand how your beam design will work. Generally you would want a rather large cross sectional area, a big box beam if you will, to get the required stiffness. This is the same no matter what material you make the beam out of.

When it comes to a laminated carbon fiber beam I simply don't have the experience to say how one with a thin cross section will work in real life. You really can't tolerate much twist or deflection in the beam else you compromise with respect to the amount of driving force you can apply. That is you end up having to lower feed rate or more likely cut depth to avoid beam deflection.

In any event keep posting, I'm learning something new overtime I view this thread.

In some respects carbon fiber is like any other material in that a thicker part will be stiffer than the same part design with fewer layers but it differs greatly from metals and woods in how the part design, choice of weave and orientation of the lay-up effects strength and stiffness.

Essentially, it is easily possible for a 2" thick cf part to be stiffer than a 3" thick part that uses the same volume of cf.

My finished gantry beam will end up being around 3" or maybe 4" thick. It is being designed in a way that maximizes stiffness and strength in the direction of the forces it will be subjected to.

I will find out for sure when I start testing but my assumption is that forces hitting the center of the gantry at 90 degrees would present the greatest stiffness challenge. I.e. You don't want it to bend in the middle. This is why I am using carbon fiber tubes to reinforce the beam at 0 and 90 degrees.

The other thing that is key to my design is the choice of using random matrix fabric for at least half of the thickness. Woven fabrics provide superior tensile strength and directional stiffness but the random matrix fiber's provide multi-directional stiffness. The mix of both types of cf produces really strong and stiff parts. I'm a huge fan of this approach. It's really cost effective too as the random matrix fabric costs 4 times less than woven 2x2 twill.

Another important area where cf differs from most metals is that it doesn't bend much before it breaks. With steel, there is a noticeable flex / bend as a warning before a part breaks. Cf parts just snap if the load exceeds their capability so it's important to build in more strength than is necessary so there is a good margin of safety. If there is any noticeable flex at all then the part is almost certainly too weak to be safe imo.

[Goemon]

Awerby, it is flexing side to side. Well front to back if you are standing in front of the machine.

Is the gantry beam hollow and if so, what is inside it now?

It goes to show that you can't always tell how sturdy something is from a pic. A thick-looking steel part could be made of thin walls welded together instead of being solid.

[louieatienza]

This is my machine when i got it.


Side view of the gantry with the cap off. IF i remember correcly steel was about 8mm thick.



Sometimes i do whole alu sheets of engraved pieces and if I go too fast (gantry makes a sudden change from forward to reverse or vice versa, in the Y direction) the gantry flexes just by its own. Specially if im in the middle of the table.
The gantry span is somewhere around 2.3meter so its not surprising.
These numbers are very small but you can cleary see what i mean. Look at number 7.

Have you checked the g-code for the job, especially when it does the number 7? Sometimes CAM puts out crappy code... and sometimes the font, if translated from a Windows or Adobe font, can look "pixelated" even if it's a single line font. Sometimes on what appears to be a diagonal, CAM will spit out something like

G1X0
Y0
X1
Y1
X2
Y2

and when you run this at speed it looks like it's a mechanical or electronics issue. Not saying that this is the case, but something to look at. Especially with some older machines.

[Khalid]

check all the bolts of Router and Spindle motor..seems looseness..

[awerby]

I think Louie might be onto something. Looking closely at the results you're getting, the oscillations are suspiciously regular. The number "1", for instance, wiggles exactly the same way each time. If this were looseness or deflection, as I originally suspected, the wiggles would be more random. Check your code, and see if it's in "exact stop" rather than "constant contouring" mode.