Aluminum vs. Carbon-Fiber x-Beam Test

[@swisshoumi]

Super short/generalized explanation for the creep with the CF tube: In static loads metalls, up until the yield point, behave pretty much linear elastic, meaning that the deformation only depends on the internall stress and will reset to its original shape when the load is removed. So temperature (within reason) doesn't affect the deformation. Polymers are fairly different in this regard, they behave in a visco-elastic fashion, meaning that the deformation depends on the internall stress, time the stress is applied for and the temperature of the material. The deformations occuring because of the time and temperature effects also do not reset, this is what you call creep.

[@mckenziekeith7434]

So the epoxy, which is a polymer, behaves in a visco-elastic fashion? But what about the graphite fibers?

[@swisshoumi]

@@mckenziekeith7434 not a 100% certain but since carbon fibre is basically just thin strands of Graphen they should also behave in an linear elastic fashion. This doesn't really have an affect on the part strengt since the fibers need the polymer to transfer loads from one fiber to the other. Especially when its anything else than straight pulling. (you can't really push a rope)

[@mckenziekeith7434]

@@swisshoumi in a composite, and assuming that the load is aligned with the fibers (which is not always the case) the fibers bear the load, both for compression and tension. The binder (resin, in this case) mainly just holds the fibers in place during compression loading. The compression strength of epoxy with graphite fiber is much greater than the compression strength of epoxy alone. And the compression strength of epoxy with graphite fiber is much greater than with glass fiber, also. In the case where composites are subjected to load aligned with fibers, I am not sure if the visco-elastic properties of the resin can explain creep. However, if the load is not aligned with the fibers then it would make sense, because the fibers would be trying to displace the resin matrix as they are loaded.

[@seancollins5769]

I am an engineer, if you aren't an engineer you are definitely a scientist! Fantastic work.

[@premiumbasics]

Thank you! :)

[@antonmaier2263]

@@premiumbasics if your using CNC for something like this, you should definitely do topological optimising

[@Dave-wv9vc]

Something you didn't go over was the relative flatness of your x-beam vs the CF tube. In my experience CF tubes vary a lot on this.

[@premiumbasics]

You are right, the CF beam was not very flat and you can see that partially in the Video. Our X-Beam does not wiggle on a flat surface.

[@mattx3020]

extruded tubes yes, this looks more molded which is easier to hold bc you can keep fibers oriented in correct directions

[@Alhussainba]

@@mattx3020 Extruded carbon fibre tubes ? how does that happen ?

[@mattx3020]

@@Alhussainba its like a playdough extruder, the carbon fibers all tend to orient the same direction, similar to cf 3d printer filament.

[@MirageC]

Awesome video! fantastic demonstration and very rigorous approach! I love it! One thing I would like to see is how behaves the Carbon Fiber tube in a 50-55C environment which is where most desktop printers with enclosure will operate. Oh, and also cost should be factored in the equation I think. Kudos to you for this great video! Thank you!

[@premiumbasics]

Thank you! Love your Videos btw! Yes most will print at 60° max chamber temperature, maybe I do a test in the future with these temperatures. The CF tube is around 34$, our X-Beam will be around 100$ for the anodised 350mm version. As said, the CF tube is probably the better option for most.

[@macrone419]

That's the resin for sure. High Temperature resins don't show deformation up to 150 °C

[@daltonhernandez721]

I actually am an engineer and during my college internship worked with carbon fiber composites for the aerospace industry for about a year. The main problems that stand out to me unless I am completely missing something is the resin used in the carbon fiber tube. When laying up carbon fiber there are two types of resin used thermoset and thermoplastic (think heat activated epoxy that permanently sets and 3d printer filament). Both have uses, benefits, and negatives associated with them but in this application it seems clear you would want a thermoset so that the temperature would have little to no effect on the tube. It seems that this company used thermoplastic probably because of cost, layup, and curing which all tend to be lower and easier with thermoplastics.

[@premiumbasics]

Thank you for sharing your knowledge and experience with us :)

[@kmyerslp85]

Would love to see the results with the linear rail on the side vs the top.

[@any1isno1]

You and the team put quiet a lot of effort in this video. Thank you! Can't wait to hold the first 2.4 beam in my hands.

[@247printing]

Very interesting and very well done! >80°C is a lot (but not too far from being realistic). The effects from the softened resin on the CF X-Beam are not really promising - was this the Mellow CF-Beam?

[@premiumbasics]

Thanks! Yes it is the Mellow CF-Beam, bought it mid 2022.

[@Vez3D]

right.. not a lot of printers will have ambient temp at 80c. The Cf tube we make for Vz are not for high temp. It is always best to use Alu tubes if you plan on going high temp.

[@jonny7fly]

The profiles are produced in a pultrusion process. The orientation of the fibers can be adjusted in certain ways. Mellow's profile looks like a 0°/90° fabric. This means that the 5 kg weight is supported by only 50% of the fibers, which results in the relatively low strength. There are either Unidirectional profiles (Which are not suitable for a 3D printer) or differently wound profiles made in a pullwinding process. I can't explain the permanent deformation because thermosets shouldn't get that soft at 80°C yet. I guess the quality of the resin is very poor. Thermoplastic matrix materials are not yet used very often.

[@lukeglazebrook]

My chamber never gets out of the 40's when printing ABS in either my Prusa enclosure or RatRig enclosed. 80c seems a tad high for me I can imagine other problems at 80c airtemp

[@wbhandy]

Thanks for showing! 2 things I’d like to see: a comparison to the stock 2020 extrusion, and input shaper results for each one. I tested the FYSETC aluminum lightweight x beam agains the stock one in a similar fashion as you did. I found the stock x beam deflected about half as much as the FYSETC lightweight beam. The inputs shaper results of the FYSETC were also worse.

[@antds857]

I am interested in the deformation of a normal aluminum extrusion. Could you check how much that deforms?

[@premiumbasics]

I’ll try to get my hands on one and report back :)

[@sixteenornumber]

you can save another ~5% of the total weight on the aluminum one by going with titanium screws. Aluminum m3x6 would also be a good option here.

[@premiumbasics]

Thanks for the tip :)

[@hardwareful]

The Al X-beam with a steel rail bolted to it looks like a bi-metal element to me, so some deformation is to be expected, but with the rail on a vertical face, that shouldn't corrupt the bed leveling, just distort the print a very tiny amount, no?

[@premiumbasics]

This is what I thought too, probably a tiny bit of deformation on larger prints.

[@Chris-oj7ro]

This is almost definitely what is happening. The aluminum cooled quickly (which is the reason why it's used for heatsinks) and it shrank while the steel cooled much more slowly (steel holds onto heat much longer) and it basically stayed expanded. This caused the assembly to raise up. I'm sure the effect would be much less drastic if the environment cooled slowly like I would expect a heated chamber to do.

[@hardwareful]

@@Chris-oj7ro Oh yes, that's a compounding effect. Keep in mind though, AISI 52100 for example has a CTE of 11.9 ppm/K, while Al alloys are around 21-24 ppm/K. Now it's hard to bolt Al and steel together so firmly that there is no stick-slip movement, but it will definitely build up stresses.

[@ibbanez5451]

yes you should do a test with 60C on that CF tube. The main reason why it deforms - its resin, and resin become soft in 80C like pla in 40C (its just a similar point, these are different materials and resin is not thermoplastic like pla, so just keep that in mind) also, the problem with high accelerations and direct extruders are: heavy top, lightweight bottom, so not only deflaction but also twist comes to the game and things are worse and worse. best idea is a thing like bambulab did. motor between rods with reduce twist (yeah I gladly see that test on 2 steel rods becouse some guys might be suprised!)

[@darionettuno2882]

the next step could be an extrusion in which the track is an integral part of the extrusion, perhaps made of 7075 aluminum and hard anodized

[@CaptnWaffles]

Regarding the deformation of the composite, aluminum hard points are typically embedded in the layup for bolted connections. In this case where the part has been manufactured, you should slide in aluminum spacers and use longer screws to grab both sides of the tube at each bolted connection. I expect you'll see slightly less deflection although it comes with a weight penalty.

[@valdolandar534]

Your work is excellent and I am very grateful to you, but I have missed the third contender, the original kits aluminum bar

[@WhiteG60]

would be interesting to add some threaded holes on the rear of the bar, too, to add something like the titanium backers to it and see how much that would stiffen things up for the aluminum on. Also would be interesting to see how these compare to the stock extrusion.

[@premiumbasics]

Might do that once I get my granite CNC running :)

[@catnipaddict126]

So did you do it?

[@petermeier3153]

Nice video. How much would the Alu beam weigh if you designed it in such a way, that it gives roughly the same deflection as the CF beam (at 21C)? How does it compare to a standard 20x20 Alu extrusion? Also it would be interesting to test the CF tube at 50-60C again.

[@premiumbasics]

Hmm, maybe I try that once I get my granite CNC running. This video took ~40 hours to make, so I won’t test that in the near future I guess :/

[@andreaspoulsen8017]

@@premiumbasics Love to see a 5-15% slight increase in weight for aluminium. If your doing a full speed printer it could be better to add a couple of grams to make sure that it will hold up through the different forces. A couple of grams might do it. What about reducing the weight by removing some of the screw holes. It's just a geuss from my side but maybe that would be possible without loosing stiffness, seems like there is to many screew holes for the mgn rail right now. Will also remove some time from amnufactoring. Also would love to see this in different lenghts, for example a lenght for an ender 5 s1 or ratrig 500x500mm machine would be awsum.

[@l3d-3dmaker58]

probably the issue is the resin, either cheap one or polyester instead of higher temp epoxy

[@tysonl79]

The obvious hysteresis was interesting. I was thinking the same thing. Softening of the epoxy allowing the rail to slip relative to the tube at the screws and not return to its original zero position when unloaded.

[@kylegoldston]

I'm not an engineer or expert but hands on. There are high temperature resin systems and that tube didn't use them. A part that would pass this test could be made with pre-preg materials and standard industrial practices for them, within the Aerospace industry. You'd need an autoclave, data logging, refrigerated storage and a bunny suit clean room to get it done with consistent results over any significant production schedule or period of time.

[@mcivi2081]

Excellent Job. Just one cmment: You are measuring deflection of the beam + the linear rail in the z axis. As the weight of the printhead is more or less fixed. I don't see the importance if this piece of data. Also, the printhead can be as light as 200 grams. What would be nice to see is the deflection in the Y Axis. This is because in that beam the load will vary depending on the acceleration and deceleration. The linear rail should be also much stiffer in the Y axis than in the Z axis.

[@njcebin]

try it again when the rail is horizontal

[@dayumcraig7657]

whats the thermal expansion ratio of carbonfiber tubes compared to linear rails? I thought Titan is the jack of all trades cuz its lightweight ,stiffer and hast nearly the same thermal expansion rate.

[@twistedsocal]

The resin that is used with most commercially available carbon fiber is not meant for high temps, basically it's a plastic that impregnates the fibers and plastic doesn't like heat in general. There are high temp pre preg and epoxy resin options available but cost more. Are more difficult to work with and in general are not going to be found unless special ordered. Also, probably would help to have unidirectional carbon fiber but they never do that cause it doesn't look as cool ie the 2x2 twill pattern

[@3D_Printing]

Although strength will be reduced, weight will be reduced if the linear rail is machined to reduce the weight ; the outside edge are the main parts ??

[@spanierjuan]

A lot of carbon Fiber stuff is done with Epoxy resin as binder for the fibers. Most of the epoxy resins are only stable up to 60 Degrees celsius. Starting at that point, Epoxy resin starts to decompose. Cracks can be formed. There are some type of resins that are more stable to higher temperatures, like water based Epoxy resins. Which go up to 90 Degrees. But water based ones do have a lot more pores. Not the best choice if you want to have high strength. And other High temperature resins are quite expensive. For a heated chamber you should know to which temperature the resin is resitant .

[@CaptnWaffles]

***laughs in SR-71***

[@spanierjuan]

​@@CaptnWaffles sorry... I don't get the SR-71 comment.

[@p_r9077]

Great work and video. Do you plan any torsion resistance test?

[@premiumbasics]

Thanks! I don‘t plan to do it at the moment, will focus fully on the granite CNC next.

[@CanuckCreator]

Interesting results, do you happen to have the numbers when compared to just a standard aluminum extrusion profile?

[@premiumbasics]

Not yet, need to get my hands on one and will report back once I tested it :)

[@mechandphysics]

Have you considered putting holes for the Lightweight Labware extrusion backers on these? I'm sure that would drastically increase the stiffness while keeping the part lighter weight than just a 2020 extrusion.

[@timg.413]

Bought the 0.1 rail and ordering this one now. Great products.

[@premiumbasics]

Thank you! :)

[@navigas93]

I am so humble! We do not deserve this great information for free. (of course it won't be free - when I order a beam when it's back in stock). I have been doing a lot of CF. And the resin has a lot to say. The "optimal" Carbon fiber tube would be a molded (could be made flatter than a beam is machinable), prepreg tube with a high temp resin matrix. This is absolutely doable. And it would not get the same sag, and the modulus of stiffness would also be higher over the temp-range. But is it easily available - no. Is it worth the hassle over buying your beam - probably not. But doing as above, gluing the rail with a high performance 2K glue and removing the mounting HW afterwards. It would be possible to achieve lighter weight, the stiffness would still need to be tested though. Awesome work!

[@premiumbasics]

Thank you for sharing your experience with us! :)

[@RaphisWorkshop]

Nice to see some more swiss youtubers ;D And great work :)

[@premiumbasics]

Thank you :D

[@thenextlayer]

Email entered, I definitely want one...

[@hajrizdilaveri6876]

Great comparison! Keep it up👌

[@premiumbasics]

Thanks! :)

[@jeremyglover5541]

Well, this is one aspect. What about the ringing shown by anyone that has recorded a before and after adding these alloy lightweight gantrys? Any i have seen doing input shaping and accelerometer results with the alloy gantry has been considerably worse after the change and removed the ‘upgrade’ cf will show much better damping as well.

[@hman72y]

I don't think "damping" is what we need on a gantry because it mean we are loosing accuracy. Due to the damping you will introduce smoothing inside your move that will result in loose of precision.

[@jeremyglover5541]

@@hman72y that sounds like a thought bubble … not only was the fact it has some damping at high frequency an aside at the end of my comment. Having some damping is a hell of a lot better than having uncontrolled resonance …

[@jeremyglover5541]

@@hman72y if you think damping = bad, better not build a voron. The gantry connections and hotend/extruder mount design is specified to be built from abs with 40% infill. Plenty of damping there.

[@jeremyglover5541]

And input shaping is literally electronic damping

[@hman72y]

@@jeremyglover5541 I admit I'm more following VzBot than Voron ;) Did not know their using damping on voron as on VzBot they try to use Aluminium part

[@EGGSHL]

At first i see that the cf tube very much outperforms aluminum but as people here said, the flatness can be weird? So if we could find a way to machine carbon flat and use it in this way, it could be used reliably. Edit: carbon is actually a machinable material. I assumed coolants (or heat) could screw up the resin but i don’t know much about real cnc.

[@Hawk013]

THere is a very wide range of carbon fiber products, with a wide range of material properties. This tube is likely a cheap mass produced extrusion or sleeve molded part, using low quality (probably polyester) resin. Hand laid and vacuum cured carbon fiber using the appropriate epoxy will be significantly lighter for the same stiffness, and will be more resistant to temperature (100-300 degrees C typically I believe, depending on the epoxy), and can made very flat and straight, but you're looking a hundreds of dollars instead of 20. And yes, real carbon fiber is easily machinable with teh right tools, and will resist most coolants pretty well.

[@Z-add]

Did you do any finite element analysis on the xbeam design.

[@premiumbasics]

Yes, but the results were way different on the FEA, maybe I made some mistakes with the parameters there…

[@MuradirMC]

Heyho, I am curious about the status of these ultralight beams. Do you have an estimate of when we can order them? I am just waiting for one of those to rebuild my V2.4 😅

[@premiumbasics]

Hi, thank you for your interest in this Beam! Mamabot and I make them during our free time and had some busy months… Right now I am focusing on my Granite CNC project. Please don’t wait on our beam with your rebuild, I can’t promise any date since the upcoming months look busy too…

[@MuradirMC]

@@premiumbasics Ahh alright, thanks for the reply. Don't stress yourself too much. Take time and make it perfect :)

[@Mikeey22]

Normal resin has a low melting point. Next time you can use heat resistant resin. Also with the aluminum profile you can strengthen the side beams by keeping more material.

[@premiumbasics]

I bought the CF tube “off the shelf”, the video should bring attention to this problem for the sellers of CF tubes and to the community.

[@jeffl1356]

Nearly all the force on the beam due to acceleration is in the horizontal direction so you could design the beam to have more stiffness in that direction and less in the vertical.... or just less in the vertical to make it lighter....

[@dibonko]

Most cheap CF tubes use resin which can withtand temperature up to 60C. More expensive use higher temp resin.

[@tegheim]

It would have been interesting to see how it would behave after heated an weighted and just let it cool down with the weight still on.

[@MrBendix24]

What about slotted holes with fitted screws to compensate for the different coefficient of thermal expansion ?

[@macgyver9134]

Even with a very tight tolerance on the sloted and shoulder screws, aluminum is a soft metal and will become dented around the screws with use. This will lead to slop over time. The entire frame of a 3d printer will expand when heated too. Your best bet is to use the best alloy to minimize warping when it expands, and just do calibrations when heated.

[@polygorg]

man those are super pretty

[@premiumbasics]

Thank you! :)

[@Andi-Maringer]

Great work!!! Cant wait to get mine 🤩

[@premiumbasics]

Thank you :)

[@Z-add]

Is this using continuous fiber or chopped fiber?

[@premiumbasics]

Continuous woven CF as far as I can tell..

[@AdrianStephanFessler]

Great Work and great Video!

[@premiumbasics]

Thank you :)

[@3dicnc]

Great work! I need to get hand on one of this :D

[@premiumbasics]

Thanks! Are you happy with the beam you made? Any updates on it?

[@dmegzel]

i wonder whats the price of alu x beam?

[@premiumbasics]

The 350mm anodised version will be around 100$

[@dmegzel]

@@premiumbasics thats a nice price to be honest, what do you think of adding holes on the back for titanium backers for more stability?

[@DizzyWood_shop]

Hy. How can we get in touch with you?

[@premiumbasics]

Discord, Username: premiumbasics#0 hope that works for you.

[@palomo3d]

I think that 80 ºC is too much for a chamber, very good test, thanks

[@premiumbasics]

Yes it is more than most people print at and exaggerates the Problem. At 60°C the permanent deformation might be less…

[@SergeiSugaroverdoseShuykov]

Heating CF tube to unrealistic value of 80+ degrees knowing that it's resin can handle as much as 85 degrees at most is basically a misleading marketing, absolute majority of printers will never exceed 65 degrees chamber temp

[@lam_xyz]

Are you (a fellow) Swiss? I think I might hear a slight accent there... ;)

[@premiumbasics]

Yes :)

[@csdstudio78]

Zero surprises here! CF strength is not new, it's been used to replace aluminum in high performance applications for decades, and needs to have high temp resin for heated applications. Great video, but your valuable time would be better spent elsewhere in the development process. This type of data is so easily available.

[@premiumbasics]

Maybe not a surprise yes, but it gets the information out there that people who want to print at high temperatures should look for a CF beam with a good resin. This is something no seller (I know of) is mentioning when selling their CF tubes. I hope this video brings attention to this subject…

[@brunojacon7923]

just perfom FEA calculation and you get the results

[@premiumbasics]

Only possible if you know exactly which type of CF and resin they used. Otherwise it’s a guessing game and the results will be off.

[@TheGodOfWar1337]

I'm throwing money at my screen and nothing is happening 😄 but otherwise great job!

[@premiumbasics]

Thank you! XD

[@googleyoutubechannel8554]

Ok, thank you, was about to post the obvious fakery in this video, but you admitted the aluminum is already deflected at the beginning of the 81c test. BUT this still doesn't make up for trying to mislead by showing '0' deflection for the alu beam at the start of the 81c test. You know you should have zeroed it against room temp. 'bed leveling' isn't an excuse, if it is 1) WHY do these tests at all then?? 2) as you know it's a wildly imperfect solution at best, no printer is going to perfectly compensate for a complex catenary shape, you know this too.

[@kizzjd9578]

You didnt specify what resin was used, the tg of it orif it was even post cured so this test was a waste of time.

[@fredericchastro4263]

I'm sorry but these tests are no valid because the different structures of the beams!Try with a square aluminium tube as the CF one and the results will be way different or make a sophisticated shape as the AL beam with the carbon fiber one.

[@einsotto9976]

Seems kinda biased if the seller of a high quality product compares his own product to a low quality product. Aöso the Temp is somewhat strange. At 80°C no "normal" Filament benefits anymore from it and its to cold for performance stuff like peek. But at 80C you already get problems of demagnetization, lubes not working properly anymore, electronics gettign to hot... But its the temp where Carbon made with normal resin looses stiffness the fastest. mysterious.... Maybe next time you should take a same quality carbon fiber tube and a reasonable temp for comparison. In this form it just looks like an advertisment amde to look like a objective test

[@martinTintin476]

CF is more expensive then Aluminum

[@AndrewBoraas]

granite cnc uwu