[DermotMcD]

We've been making carbon fibre gantries for some time… and just stopped.

In our application, we had to make a 10 metre wide machine which would move at 1 metre/sec or faster and support a weight of 100 kgs with moderate accuracy. The flex and the torsion of the gantry had to be low. Yes, we're not talking about mill accuracy but adequate for the purpose, say ± 0.3mm.

Carbon or aluminium were the choices with a monocoque carbon gantry seeming like a good choice compared with a fabricated aluminium one. The layup engineering was done by outside contractors with experience of masts and other high load carbon products.

The gantry was made from pre-preg and nomex core. My experience is that pre-preg is fast and clean and offers much better integrity than resin infusion where you can quite easily get failures in the infusion process giving bubbles and delimitation.

The mould was fairly immense and used a bladder to inflate to 3-4 atmospheres instead of vacuum bagging. The entire mould was fitted into a fabricated autoclave with heaters to get it up over 120ºC.

After some months of use, it was obvious that there was some sag in the gantry so the next versions used a heavier and heavier layup. I can't be sure, but my guess is that sag is a fact of life unless you have a truly massively thick layup.

In spite of the fairly substantial mould, there was some distortion in the finished product because of the pressure used in the bladder. Attaching parts is somewhat problematic since accurate location is difficult with the mould.

The gantry is also noisy. Because it is hollow and fairly stiff, any vibration is amplified like beating a drum skin. Unless you add something like Kevlar, this is a fact of life with carbon.

What made us stop this method of construction was mainly the space required for the mould and autoclave but also the noise and sagging. We decided that for our requirements, fast and light, that aluminium would be better. Note, that cost was not a decisive factor but there's no doubt that the alu version is going to be cheaper.

For a CNC machine, honestly, I can't see the point. You normally don't need more speed than you can get from a lead screw because cutting tools won't really go very fast. You do need to damp vibration and carbon doesn't look great there. You will need some compression strength and carbon fails there too.

There's no doubt that carbon is sexy as far as outward appearances go and in some applications where weight and strength are critical, carbon offers a lot but where weight is not a big issue, then metal looks more interesting and at a fraction of the cost and complexity.

D

[jono5axe]

Hello Dermot, thanks for that info, very interesting. My beam is for a gantry router, and it is moving at 1.0m/s and with acceleration of 2 m/s2 (designed for, at least). What was the cross-section dimensions of your beam? I have a couple of comments re your points below:

We've been making carbon fibre gantries for some time… and just stopped.

In our application, we had to make a 10 metre wide machine which would move at 1 metre/sec or faster and support a weight of 100 kgs with moderate accuracy. The flex and the torsion of the gantry had to be low. Yes, we're not talking about mill accuracy but adequate for the purpose, say ± 0.3mm.

Carbon or aluminium were the choices with a monocoque carbon gantry seeming like a good choice compared with a fabricated aluminium one. The layup engineering was done by outside contractors with experience of masts and other high load carbon products.

The gantry was made from pre-preg and nomex core. My experience is that pre-preg is fast and clean and offers much better integrity than resin infusion where you can quite easily get failures in the infusion process giving bubbles and delimitation.

The mould was fairly immense and used a bladder to inflate to 3-4 atmospheres instead of vacuum bagging. The entire mould was fitted into a fabricated autoclave with heaters to get it up over 120ºC.

After some months of use, it was obvious that there was some sag in the gantry so the next versions used a heavier and heavier layup. I can't be sure, but my guess is that sag is a fact of life unless you have a truly massively thick layup.D

I would have expected that additional laminate thickness would have only limited effect on deflection, and may possibly increase deflection by increasing weight. The cross-section needs to increase in size to reduce deflection. For example, if you use a steel RHS and get excessive deflection, increasing the wall thickness will not have much effect, but choosing a bigger RHS will certainly reduce deflection. Of course, this is problematic if you have already made a mold...

I suppose that we are not really going to achieve zero deflection on large structures, and are more interested in a consistent deflection which is not affected by tool forces.

In spite of the fairly substantial mould, there was some distortion in the finished product because of the pressure used in the bladder. Attaching parts is somewhat problematic since accurate location is difficult with the mould.

The gantry is also noisy. Because it is hollow and fairly stiff, any vibration is amplified like beating a drum skin. Unless you add something like Kevlar, this is a fact of life with carbon.

What made us stop this method of construction was mainly the space required for the mould and autoclave but also the noise and sagging. We decided that for our requirements, fast and light, that aluminium would be better. Note, that cost was not a decisive factor but there's no doubt that the alu version is going to be cheaper.

For a CNC machine, honestly, I can't see the point. You normally don't need more speed than you can get from a lead screw because cutting tools won't really go very fast. You do need to damp vibration and carbon doesn't look great there. You will need some compression strength and carbon fails there too.

From my research it seems that composites (both carbon laminate, GRP, and especially resin/aggregate castings) are better at vibration damping than steel/Al, maybe 5 times better, and this is a primary reason for trying to use it (maybe incorrect, more comments pls?). We are routering soft materials and certainly need speed, and long travels, so more speed than a long ballscrew can do (i.e. we need 1 m/s) before buckling, so use gear rack. We do need speed, and this would be typical for a lot of gantry routers/mills.

An important reason to consider composites is to keep weight down. In steel my 4000mm gantry design comes in at a ton, and therefore the performance comes down and the machine/servo spec goes up (costs go up). My composite design can be less than half that weight, and will perform to specs (? hopefully). It seems logical that as a gantry machine gets larger, and if it needs speed, then keeping gantry weight down is a primary concern.

Another reason to consider composites is the issue of how do you actually build a large steel or aluminium gantry beam? RHS is not available for that (not big enough), so you are welding, and therefore not straight anymore, and also full of residual weld stresses (RHS also has residual stresses?). I have just been quoted $4000 for having my welded steel side frames stress relieved in an oven, and there are not many large ovens around. So, maybe use thinner sections? folded and riveted to avoid weld stresses? but still you need to provide machine beds for rails etc, which may need welding on, and which have to be machined, which is more difficult for a large (long) machine bed, and has cost. So, it is interesting to consider if the composite materials can be used to eliminate weld-stresses (and the weld stress relieving process step and cost) and also eliminate machining cost by using epoxy/aggregate/fibre cast machine ways. [of course, machine ways could be bonded onto a steel/Al beam using modern composites structural adhesives....]

I don't agree that a carbon tube or rectangular section has no compression strength, this is not correct.

These are some of my thoughts, maybe sensible, maybe not...

regards
Jono

There's no doubt that carbon is sexy as far as outward appearances go and in some applications where weight and strength are critical, carbon offers a lot but where weight is not a big issue, then metal looks more interesting and at a fraction of the cost and complexity.

D

[Goemon]

We've been making carbon fibre gantries for some time… and just stopped.

In our application, we had to make a 10 metre wide machine which would move at 1 metre/sec or faster and support a weight of 100 kgs with moderate accuracy. The flex and the torsion of the gantry had to be low. Yes, we're not talking about mill accuracy but adequate for the purpose, say ± 0.3mm.

Carbon or aluminium were the choices with a monocoque carbon gantry seeming like a good choice compared with a fabricated aluminium one. The layup engineering was done by outside contractors with experience of masts and other high load carbon products.

The gantry was made from pre-preg and nomex core. My experience is that pre-preg is fast and clean and offers much better integrity than resin infusion where you can quite easily get failures in the infusion process giving bubbles and delimitation.

The mould was fairly immense and used a bladder to inflate to 3-4 atmospheres instead of vacuum bagging. The entire mould was fitted into a fabricated autoclave with heaters to get it up over 120ºC.

After some months of use, it was obvious that there was some sag in the gantry so the next versions used a heavier and heavier layup. I can't be sure, but my guess is that sag is a fact of life unless you have a truly massively thick layup.

In spite of the fairly substantial mould, there was some distortion in the finished product because of the pressure used in the bladder. Attaching parts is somewhat problematic since accurate location is difficult with the mould.

The gantry is also noisy. Because it is hollow and fairly stiff, any vibration is amplified like beating a drum skin. Unless you add something like Kevlar, this is a fact of life with carbon.

What made us stop this method of construction was mainly the space required for the mould and autoclave but also the noise and sagging. We decided that for our requirements, fast and light, that aluminium would be better. Note, that cost was not a decisive factor but there's no doubt that the alu version is going to be cheaper.

For a CNC machine, honestly, I can't see the point. You normally don't need more speed than you can get from a lead screw because cutting tools won't really go very fast. You do need to damp vibration and carbon doesn't look great there. You will need some compression strength and carbon fails there too.

There's no doubt that carbon is sexy as far as outward appearances go and in some applications where weight and strength are critical, carbon offers a lot but where weight is not a big issue, then metal looks more interesting and at a fraction of the cost and complexity.

D

It sounds like the outside engineers didn't do enough work on the design specs. Carbon fiber, steel, and aluminum can all be stiff and strong enough if the part is properly designed and enough material is used. They can also all be too weak and flexible if wall thickness is inadequate or the design is suboptimal etc.

Personally, I wouldn't be relying on the typical process of laying fabric horizontally in a mold and attempting to achieve the desired stiffness by using enough layers or adding a filler material for the core with parts as large as you describe. It's not that it isn't possible to achieve the desired stiffness like that. It's just not efficient or cost effective imo.

In my testing I have seen a 10x increase in strength and stiffness using the same amount of material by adding certain design features. The basic idea is that cf is least stiff when laid flat and stiffness increases dramatically when it is curved.

The center of my gantry has two layers of thick wall carbon fiber tubes. Currently, six tubes run parallel to the rails on the gantry face plate and dozens of shorter tubes are mounted at 90 degrees and embedded in high temp resin. I also added carbon nanotubes to the resin so it is quite literally bullet-proof. Nanotubes apparently add 30% in stiffness. They certainly make a noticeable difference.

If I was setting up a commercial cf gantry beam manufacturing processes, I would probably choose to use a filament winding process like Compotech for the efficiency and design flexibility. I think the winding process can be more easily optimized for maximum stiffness than using regular prewoven fabrics. I bet that hand laying woven prepreg into a mold that big took forever. A machine winding yarn or even prewoven prepreg around a mandrel lends itself to automation better.

Anyone wanting to get an understanding of how to design cf parts for maximum stiffness should take three identical pieces of cf fabric wetted out with resin. Let one cure as a single flat layer. Cure the second curved half way around a tube. Roll the third into a full tube and let them all fully cure. The flat piece will easily bend or tear by hand. The curve will bend with far more difficulty. The tube won't bend by hand at all.

[Goemon]

Very interesting thread for me, as I am just about to build a 4000mm long composite gantry. I have been procrastinating back and forth about it, and maybe just to use steel, but am pretty committed by now to doing it in composite. I therefore list my plan below, for the purpose of getting everyones comments. All comments and ideas gratefully accepted - to ensure I am not overlooking something....

- Composite gantry beam 4000mm long x 600mm high x 400mm deep [RHS - rectangular hollow section].
- Construction: Sandwich construction - 5mm foam core, internal laminate = 2mm quadraxlial glass fibre, external laminate = 4mm quadraxial carbon fibre & glass fibre (50%/50%), with transverse internal gussets/bulkheads every 300mm (bonded in).
- Resin system = epoxy, infusion, room temp.
- Method: vacuum infusion + hotbox post cure cycle. [make up core material as C-Section, lay-up inside C-Section, bond in transverse gussets, close C-Section by bonding on last side panel, vacuum infuse external laminate in one hit].
- Thermal expansion goal is to achieve approximate equivalence with concrete [13x10-6 m/(mK)] and steel [12x10-06 m/(mK)], as side frames are steel and machine base is concrete floor. Thermal expansion co-efficient of epoxy GRP = 36x10-06 m/(mK), epoxy/carbon = 2x10-06 m/(mK).

Metal threads will be achieved by either/or
1) 50x8mm steel flatbar laminated into the core of the sandwich construction, or,
2) Drilling beam and bonding in threaded inserts (with plexus/crestabond/etc)

Datum faces (machine beds) for 35mm linear rails & 24x24mm gear rack will be produced by one of the following methods, either/or:
1) Shimming components in place and then back-filling gaps with "Chock-it" [CHOCKFAST ORANGE (PR-610TCF)], or,
2) Epoxy-concrete machine beds, molded on a 4000mm long flat table produced with self-leveling epoxy (make flat table mold using self-leveling epoxy, lay in the epoxy/sand/fibre mix on the table mold, place the gantry beam on top to bond the epoxy concrete machine beds to the beam.

??? forgotten something?...

Like I said, all comments and ideals thanks very much.

Regards,
Jono

Sounds like we are following a similar path (just on a different scale). After researching until my brain melted and poured out of my ears, I also decided to go with an epoxy granite machine base for my CNC build. The first section is curing right now.

looking at your plan, one thing to think about is going to be your choice of resin. With the sizes you are talking about, heat curing is going to be difficult with most traditional heat cure epoxy resins but there could be issues with most room temp cure resins.

Most of the lower cost hobby level resins have fairly low heat deflection temperatures. This is the temp they start to soften under load. Some are as low as 110 degrees F which is far from ideal for precision machine parts. This issue is most commonly addressed by using a quality resin and post curing at 250 or 350 degrees to raise the heat deflection temp.

obviously it's going to be hard to fit 3000mm beams in the oven but there are a few good alternatives:

I have recently started testing a new room temp epoxy called Adtech 820. It has a heat deflection temp of over 180 degrees with no post curing. It's more expensive than regular laminating resins but worth it for this.

There is a low (ish) cost heat cure resin called "Max HTE" available on eBay. If curing ovens are not available for larger parts, the curing process can be completed using Ir heat lamps or even by leaving the parts outside on a sunny day. This resin has a deflection temp of 250 - 350 degrees if post curing is completed.

Another option is to just build a temporary curing oven. It's much easier than it sounds. You can use an electric heating element for the heat source (or two - three for larger ovens). The same kind people buy to build diy BBQ smokers. I am using the aluminum extrusions (that I decided not to use for my CNC build) for the frame with aluminum sheets and high temp insulation for the walls. Then you just need a temp controller and oven thermometer.

- - - Updated - - -

Very interesting thread for me, as I am just about to build a 4000mm long composite gantry. I have been procrastinating back and forth about it, and maybe just to use steel, but am pretty committed by now to doing it in composite. I therefore list my plan below, for the purpose of getting everyones comments. All comments and ideas gratefully accepted - to ensure I am not overlooking something....

- Composite gantry beam 4000mm long x 600mm high x 400mm deep [RHS - rectangular hollow section].
- Construction: Sandwich construction - 5mm foam core, internal laminate = 2mm quadraxlial glass fibre, external laminate = 4mm quadraxial carbon fibre & glass fibre (50%/50%), with transverse internal gussets/bulkheads every 300mm (bonded in).
- Resin system = epoxy, infusion, room temp.
- Method: vacuum infusion + hotbox post cure cycle. [make up core material as C-Section, lay-up inside C-Section, bond in transverse gussets, close C-Section by bonding on last side panel, vacuum infuse external laminate in one hit].
- Thermal expansion goal is to achieve approximate equivalence with concrete [13x10-6 m/(mK)] and steel [12x10-06 m/(mK)], as side frames are steel and machine base is concrete floor. Thermal expansion co-efficient of epoxy GRP = 36x10-06 m/(mK), epoxy/carbon = 2x10-06 m/(mK).

Metal threads will be achieved by either/or
1) 50x8mm steel flatbar laminated into the core of the sandwich construction, or,
2) Drilling beam and bonding in threaded inserts (with plexus/crestabond/etc)

Datum faces (machine beds) for 35mm linear rails & 24x24mm gear rack will be produced by one of the following methods, either/or:
1) Shimming components in place and then back-filling gaps with "Chock-it" [CHOCKFAST ORANGE (PR-610TCF)], or,
2) Epoxy-concrete machine beds, molded on a 4000mm long flat table produced with self-leveling epoxy (make flat table mold using self-leveling epoxy, lay in the epoxy/sand/fibre mix on the table mold, place the gantry beam on top to bond the epoxy concrete machine beds to the beam.

??? forgotten something?...

Like I said, all comments and ideals thanks very much.

Regards,
Jono

Sounds like we are following a similar path (just on a different scale). After researching until my brain melted and poured out of my ears, I also decided to go with an epoxy granite machine base for my CNC build. The first section is curing right now.

I am copying this guy's basic design for my base:

Casting Machine Bases in Composite Epoxy | Hackaday

looking at your plan, one thing to think about is going to be your choice of resin. With the sizes you are talking about, heat curing is going to be difficult with most traditional heat cure epoxy resins but there could be issues with most room temp cure resins.

Most of the lower cost hobby level resins have fairly low heat deflection temperatures. This is the temp they start to soften under load. Some are as low as 110 degrees F which is far from ideal for precision machine parts. This issue is most commonly addressed by using a quality resin and post curing at 250 or 350 degrees to raise the heat deflection temp.

obviously it's going to be hard to fit 3000mm beams in the oven but there are a few good alternatives:

I have recently started testing a new room temp epoxy called Adtech 820. It has a heat deflection temp of over 180 degrees with no post curing. It's more expensive than regular laminating resins but worth it for this.

There is a low (ish) cost heat cure resin called "Max HTE" available on eBay. If curing ovens are not available for larger parts, the curing process can be completed using Ir heat lamps or even by leaving the parts outside on a sunny day. This resin has a deflection temp of 250 - 350 degrees if post curing is completed.

Another option is to just build a temporary curing oven. It's much easier than it sounds. You can use an electric heating element for the heat source (or two - three for larger ovens). The same kind people buy to build diy BBQ smokers. I am using the aluminum extrusions (that I decided not to use for my CNC build) for the frame with aluminum sheets and high temp insulation for the walls. Then you just need a temp controller and oven thermometer.

[jono5axe]

looking at your plan, one thing to think about is going to be your choice of resin. With the sizes you are talking about, heat curing is going to be difficult with most traditional heat cure epoxy resins but there could be issues with most room temp cure resins.

Most of the lower cost hobby level resins have fairly low heat deflection temperatures. This is the temp they start to soften under load. Some are as low as 110 degrees F which is far from ideal for precision machine parts. This issue is most commonly addressed by using a quality resin and post curing at 250 or 350 degrees to raise the heat deflection temp.

Resins.
I do plan to use ambient temp resin, but was referring to POST-CURE process when I mentioned HOT-BOX, (as opposed to an oven). Any composite structure/part must be post cured to achieve stability and meet performance specification, even ambient temp resin systems. I would have to get my spec sheets out, but this would normally be a post cure (epoxy) at say 60-80 degrees Celsius for a number of hours. Failing to post cure is bad......

Again, would have to check specs on particular products, but most ambient epoxy resin systems (I don't know about hobby, I am referring to commercially available, boatbuilding/marine/industrial/etc) are good up to 120 - 150 degrees Celsius before softening (glass transition temp). To restate the obvious, we are choosing epoxy because of it's high bond strength and especially it's low shrinkage properties, which we are hoping to give us cnc machine structures with dimensional stability and time stability. Hense the need for full cure cycles / post cure processes.

obviously it's going to be hard to fit 3000mm beams in the oven but there are a few good alternatives:

I have recently started testing a new room temp epoxy called Adtech 820. It has a heat deflection temp of over 180 degrees with no post curing. It's more expensive than regular laminating resins but worth it for this.

There is a low (ish) cost heat cure resin called "Max HTE" available on eBay. If curing ovens are not available for larger parts, the curing process can be completed using Ir heat lamps or even by leaving the parts outside on a sunny day. This resin has a deflection temp of 250 - 350 degrees if post curing is completed.

Another option is to just build a temporary curing oven. It's much easier than it sounds. You can use an electric heating element for the heat source (or two - three for larger ovens). The same kind people buy to build diy BBQ smokers. I am using the aluminum extrusions (that I decided not to use for my CNC build) for the frame with aluminum sheets and high temp insulation for the walls. Then you just need a temp controller and oven thermometer.

Ovens, Hot-boxes.
I own a composites manufacturing business and we build temp ovens (hot boxes) all the time. These may be as simple and small as a cardboard carton and a fan heater to cure a small polyester part or it many be a larger timber framed box for larger, odd shaped items. typically 50x50mm timber frame with 125micron plastic sheet stapled on the outside, 2mm eva/foam/foil sheet stapled on the inside (foil facing in) with the 50mm air gap in between. Single phase 2.4kW fan heaters (several) will get a hot box like this up to 60 degrees without much difficulty, if box is well made (without leaks). Use IR pistol thermometers to monitor temperatures all over the part surfaces and hot-boxes throughout the process. Make sure the part is in the top area of the hot-box, not sitting on the floor.....

Obviously, best results from a structural performance viewpoint would be gained from using a temperature cure epoxy system and an autoclave, but I do not believe that this is necessary for building a cnc gantry beam, even quite large ones. Ambient cure and a hot-box should do the job, along with good (common sense) beam design..... for me, time will tell.


Beam design.

Regarding all the various comments/posts on here about carbonfibre gantry structures, I would like to point out that a beam is a beam is a beam (no matter what it is made of) and that good beam design does not get thrown out the window just because we are considering using carbonfibre. The process can possibly be demystified by designing a beam as normal and substituting the different (carbon, GRP) typical material data. Typical values are probably good enough (for each laminate orientation). Deflection values should be the focus. It is also important to remember that a gantry is a torsional member more that a bending beam...........especially for larger z-axis machines.

Personally, I tend to think that there should be some glass fibre in the laminate for the purpose of:
1) protecting the outer surface of the carbon from damage during operation (impact/wear/etc).
2) additional laminate thickness, for damping/robustnest.
3) less brittle structure (carbon is brittle, fails catastrophically)
4) modifying thermal expansion co-efficient.
other? maybe.

Spiral winding? ...no.
Structural part = structural material (biaxial, triaxial, quadraxial, unidirectional). Put the fibre orientation where it is needed (where the stress orientation is).

Additional note: a part made and fully cured in a mold will be as straight as the moid, but a part made not in a mold will likely have some straightness deviation from fabrication and cure.

Sounds like we are following a similar path (just on a different scale). After researching until my brain melted and poured out of my ears, I also decided to go with an epoxy granite machine base for my CNC build. The first section is curing right now.

I am copying this guy's basic design for my base:

Casting Machine Bases in Composite Epoxy | Hackaday

My machine base is steel, only my gantry is (possibly) being made in composite. The main frames are welded steel RHS. The main horizontal side rails are 250x250x4000RHS bolted on (no welding here) and are filled with epoxy/aggregate mix, for stability. Gantry and Z-axis in composite (if I can get my #### together).

Regards,
Jono

[jono5axe]

Goemon, I had another thought - re post-cure and epoxy agregates.

My earlier comments on post-cure relate to laminates, and I am not so sure about the situation with epoxy/aggregate mixtures. I would expect the same to apply in general terms, but possibly to be less important as the quantity and mass of the aggregate fill is so great that it modifies the resin behaviour to an extent (shrinkage, geltime, cure, exotherm, etc). I would expect that post-cure would still be desirable for high tolerance machines, but possibly not very important otherwise. Mild elevated temperatures and time will achieve complete stability sooner or later.

Jono

[drieslaas]

Very interesting thread for me, as I am just about to build a 4000mm long composite gantry. I have been procrastinating back and forth about it, and maybe just to use steel, but am pretty committed by now to doing it in composite. I therefore list my plan below, for the purpose of getting everyones comments. All comments and ideas gratefully accepted - to ensure I am not overlooking something....

- Composite gantry beam 4000mm long x 600mm high x 400mm deep [RHS - rectangular hollow section].
- Construction: Sandwich construction - 5mm foam core, internal laminate = 2mm quadraxlial glass fibre, external laminate = 4mm quadraxial carbon fibre & glass fibre (50%/50%), with transverse internal gussets/bulkheads every 300mm (bonded in).
- Resin system = epoxy, infusion, room temp.
- Method: vacuum infusion + hotbox post cure cycle. [make up core material as C-Section, lay-up inside C-Section, bond in transverse gussets, close C-Section by bonding on last side panel, vacuum infuse external laminate in one hit].
- Thermal expansion goal is to achieve approximate equivalence with concrete [13x10-6 m/(mK)] and steel [12x10-06 m/(mK)], as side frames are steel and machine base is concrete floor. Thermal expansion co-efficient of epoxy GRP = 36x10-06 m/(mK), epoxy/carbon = 2x10-06 m/(mK).

Metal threads will be achieved by either/or
1) 50x8mm steel flatbar laminated into the core of the sandwich construction, or,
2) Drilling beam and bonding in threaded inserts (with plexus/crestabond/etc)

Datum faces (machine beds) for 35mm linear rails & 24x24mm gear rack will be produced by one of the following methods, either/or:
1) Shimming components in place and then back-filling gaps with "Chock-it" [CHOCKFAST ORANGE (PR-610TCF)], or,
2) Epoxy-concrete machine beds, molded on a 4000mm long flat table produced with self-leveling epoxy (make flat table mold using self-leveling epoxy, lay in the epoxy/sand/fibre mix on the table mold, place the gantry beam on top to bond the epoxy concrete machine beds to the beam.

??? forgotten something?...

Like I said, all comments and ideals thanks very much.

Regards,
Jono

I am tempted do do a quick FEA on this beam, to see how it would compare with a steel beam ito stiffness and mass, and also to see what the effect would be to use more UD on the outside to decrease the carbon and use it more efficiently. That quad carbon/glass hybrid fabric seems appropriate to use in something like a boat hull, but in a highly predictable structure like a gantry beam, I suspect you can do better by using the fibres carefully.

[jono5axe]

I am tempted do do a quick FEA on this beam, to see how it would compare with a steel beam ito stiffness and mass, and also to see what the effect would be to use more UD on the outside to decrease the carbon and use it more efficiently. That quad carbon/glass hybrid fabric seems appropriate to use in something like a boat hull, but in a highly predictable structure like a gantry beam, I suspect you can do better by using the fibres carefully.

We will real-world-know about the gantry beam performance pretty soon, as the buiuld is coming along.

I have started a build thread at http://www.cnczone.com/forums/cnc-wo...neering-3.html

It is going a bit slower than planned, but it is coming together. I need to post some more progress pics.....

J.

[Goemon]

I am tempted do do a quick FEA on this beam, to see how it would compare with a steel beam ito stiffness and mass, and also to see what the effect would be to use more UD on the outside to decrease the carbon and use it more efficiently. That quad carbon/glass hybrid fabric seems appropriate to use in something like a boat hull, but in a highly predictable structure like a gantry beam, I suspect you can do better by using the fibres carefully.

The design is key when you need max stiffness from a CF part. People desperately want a simple direct comparison with linear materials like steel but, unfortunately, it doesn't work in the same way. It's not a linear material.

The choice of fabric and the orientation of the lay-up is a key part of the design. As you said, you have predictable directional forces with a gantry beam - which is ideal for CF part designs. You can specifically increase strength and stiffness in the directions needed. A high modulus unidirectional fabric is a good option for this.

Also, I can't stress enough the importance of the shape to increasing stiffness. This is true with steel too - e.g. You get better stiffness with an I-beam than a flat sheet. The shape matters even more with carbon fiber though, espiecially if your cf part has less weight than the steel one.

Pound for pound, cf parts have greater strength and stiffness than steel but... if you want the cf part to weigh less and have the same (or more) stiffness, design is key. Obviously, if you have unlimited funds, you can just increase wall thickness of a part until it is strong enough but I am going to assume that most people in the diy section don't have a limitless supply of cash....

[ack1]

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

[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.

[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.

[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.

[rocrat]

Goemon , you keep going my friend , you are on the right path and will have a superior product when done .
I am a marine project manager with an engineering background and fabricate for fun . Carbon has surpassed metals in just about every aspect in the marine environment,this including masts and the rigging used to hold it up as well as hull structures and sails .
A carbon mast is both lighter and stiffer then its aluminum equivalent and localized reinforcing can be introduced into a design to either stiffen and area or to increase strength . Like with all new technologies you will get a fair amount of naysayers finding all the excuses in the world why it won't work ,but the early adopters usually get to enjoy the advances and understanding of the new tech .
You will definitely produce a far stiffer component then a metal equivalent with a substantial weight saving ,which will give you the possibility of faster acceleration in the Y axis from point to point . This can add up to much faster job times if taken advantage of the weight saviing on a complex component .
In the marine industry we deal with compression , tension ,torque and flex and carbon has been the best material so far to address all these loads . As you have continually pointed out ,the ensign is critical to achieving your goal of stiffness , and carbon can get you there the easiest .
Just remember you may want to protect the bottom part of the gantry from micro impact from flying Swarf chips as this could eventually lead to a delamination failure from the continuous impact in the same area . Perhaps a sheet of stainless foil used for heat treating would work well here , it's available from McMaster and can be bonded to the gantry with ease .
Please continue to describe the design and build , as I think the tech explains itself perfectly .
Rocrat

[1Jumper10]

Really interesting discussion. I've never worked with composite fabrication but I've been interested in the subject for awhile. Specifically as it relates to constructing a carbon fiber gantry like what Goemon is undertaking. Dermot McD - keeping in mind I know next to nothing about CF, it would seem like you could eliminate (partially) sag in your 10M gantry by incorporating a truss or arch shape?

[rocrat]

Jumper , sag can easily be taken out of a 10 m gantry by the inclusion of a tension member in the form of a short compression post hung off the gantry underneath in the middle and a tensioning cable or dyneema line attached at either end of the gantry and supported on the post in the middle . By tensioning the line you can pull out the sag or even put prebend in the opposite direction . Obviously it could not be lower then the tooling to stop any collision .
If the section is deep enough this could even be done inside the tube if hollow and it would not add any substantial weight . Come to think of it , if done inside the tube the post is not even necessary as you could slot the lower surface of the tube and pass the line through the slot and over a fair surface like a piece of round bar fastened over the slot .

[rocrat]

Has anyone thought of using a marine spar manufacturer . I have just had a 30 m tube come out of the oven at Southern Spars in South Africa . Next in is a rectangular box section for a longeron on a catamaran . This longeron will overhang the forward beam by 3m or 10' unsupported and has a tip deflection of less than 1/2" or 12mm with a 5000Kg 12000 lbs upward load .
I'm sure the same section would not have any deflection in the middle if supported both ends like in a gantry on a cnc machine .
It's all about the sectional dimension and the laminate spec . There are also various modulus's of carbon that can factor into the engineering for the amount of stiffness . The mast tube I spoke of above uses high modulus carbon in both unidirectional and woven twill to stop any torsional movement in the section and to engineer a deflectional stiff mast .
Perhaps a mast section would work a charm with its oval type shape , this would put the linear rails on a different axis from each other which may support the up force of the tool on the material better then in a regular flat surface . It may apply the load to the rail in a more even spread over both bearing rows . It may be a good exercise to calculate the load spread on a linear rail .

[hanermo]

One can make a huge oven using plywood.
Plywood resist high temps quite well.

Upto 140+ C == 300 F has zero effect on wooden structures .. common in saunas lasting 30+ years.

Boat guys make 30 m long ovens, from tarps over pvs, double skin, and they work fine, with heat losses but they are one-use structures.

[Goemon]

One can make a huge oven using plywood.
Plywood resist high temps quite well.

Upto 140+ C == 300 F has zero effect on wooden structures .. common in saunas lasting 30+ years.

Boat guys make 30 m long ovens, from tarps over pvs, double skin, and they work fine, with heat losses but they are one-use structures.

Does plywood burn? It kinda seams like it would be a fire hazard. It's not the oven temps that would bother me. It's the heating element being anywhere near wood that would scare me.

I prefer my method of using aluminum sheet metal, extrusions and high temp insulation. It' just seems safer and more efficient.

As an fyi, I have seen people use cement board for the oven walls. It's cheap like plywood and maybe a little safer for an oven build. I have read about people using old steel school lockers as their oven enclosure.

Whatever is used, good high temp insulation is key. It will be harder and more expensive to maintain or reach 350 degrees with a small element if insulation is poor. Also, I hate it when the oven heats up the whole room and makes working near it unbearable. Or worse, where you need to use too many element and it constantly trips the breaker....

High temp insulation is very cheap. I got 24" x 48" x 1" pieces off eBay for $6 and it's good up to 1200 degrees.

For medium sized ovens ovens that require less than 48" of space, some BBQ smokers make perfectly adequate curing ovens as they come in larger sizes than regular semi-portable ovens. They are also designed to maintain a steady 250 degrees for long periods.