Currently, my team is looking into the strength of chopped carbon prints and what is the best way to do it. We have just ordered a new ruby nozzle for out ultimaker 3 that has the ability to print chopped carbon nylon. We want to see how this goes and see if the print quality is nice enough. But when looking at the capabilities of the Markforged one what makes it special and worth our time and money to buy one over the $200 nozzle?
Have you read through these threads?
A number of them may include some of the information you are looking for.
The Mark Two can lay a continuous strand of carbon fiber, which is much much stronger than chopped fiber in the filament. Chopped fiber doesn’t do too much, as it turns out, but continuous is shockingly strong.
From what I understand chopped fiber makes parts significantly easier to print, especially with larger overhangs, than regular nylon
That is correct but the tradeoff is a decrease in layer ardhesion and tensile strength. Like with anything It depends on what you want to print when selecting a filament. As an analogy there is a reason why in cars they use glass for windshileds and rubber for tires and not the other way around. When getting a 3D printer the first question you should ask yourself is what do you want to produce Without knowing that first no one can advise you. Its kinda if I would try to suggest a vehicle for you and don’t know if you want to use it as a status symbol or a cummuting vehicle, if you drive by yourself or have a family or if you need one to deliver furniture. We print stuff for our robot and all that goes along for competing in FIRST and if I had the budget to buy a markforged I probably spend it on buying 10 cheaper printers instead so we can produce what we need quickly in pretty much the same quality. Now continous carbon fiber would be a different story. (like a Mark2) but then again do you need that strength. If you want to print something that just looks nice to put on the shelf to look at then use PLA with your Ultimaker.
Having that kinda money for one printer I’d probably go with a fully enclosed High temp one lige the Funmat that can handle PC or PEEK and has a filter if you do POM (delrin) and the kind
Chopped Fiber does a whole lot. Have you ever compared a chopped fiber part to a plain nylon part? The property changes are significant.
The biggest thing we are looking at is the strength of the chopped carbon parts so that it can be used as more structural parts of the robot. We already have 3 printers but are looking at the options for chopped carbon.
I can’t compare the strength or quality of Markforged versus other printers using Nylon with chopped carbon fiber, but I can tell you that our experience with the Markforged has been outstanding.
Last year we had about 200 Markforged-printed parts on our robot - everything from spacers to gears to structural components to complex mechanisms. And we probably printed over 1000 parts over the course of the season. Of those 1000 prints, I can recall only about 5 failing, which is an incredibly high percentage.
The dimensional accuracy is top notch. For example, we printed 20DP helical spur gears that meshed perfectly. Bearing holes for 7/8” and 1.125” bearings fit as a light press fit without any special over/under sizing. We even embedded nuts and bearings in our prints without concern of them not fitting properly.
The parts are shockingly strong and stiff in the X/Y plane. The Z plane can be more fragile, as with any additive printer. anecdotally, consider that we also own the Markforged that use continuous carbon fiber, and we only ended up using that feature on a small handful of parts. Further, we only had a single part break on the production robot the whole season, and in retrospect, it was poorly designed.
They use their own slicer and interface for printing the parts. It’s really, really simple, and anyone who can CAD a part can figure out how to use it in minutes. You don’t need to worry about figuring out the right temperature, speeds, etc. that you do with other printers. You just hit print and then you get a part - it’s that simple.
I think the extra amount that you are paying for the Markforged goes into all of those domains, resulting in a great tool if it fits your budget.
I don’t doubt the Markforged is a great printer and I have never printed on one. But our setup works well too. We have 3 printers currently 2 Anycubic Chiron that we paid about $800 for (not each for both) and one Matterhackers Pulse that we decked out at about $1200. So right now we have about $2000 in printers.
Now for starters - our setup may or may not be as accurate as yours but we can Print Nylon, HIPS, ABS PETG, PLA and Flexibles accurately enough like this HIPS print
18 gears, 4 double racks intermeshing without binding up all fully involute double helical gears and as its just the scaled up and beefed up version of last years model which lifted the robot we are pretty confident this one will too.
or this Harmonic-planetary prototype
All these and more - the whole robot you can find on grabcad at
Now I am certain you can print that stuff on your markforged too - you might have to split the rack and pinion setup into more pieces as they require a print plate of 400x400mm or larger to be printed in this configuration - also some of the pieces of our bumblebot.
Due to the current rules we will have to print everything starting at the beginning of january again and as we intend to do the whole robot that is a lot of printing and that will keep the 3 we have now and hopefully more busy 24/7 then. I would love to have at least 2x as many printers as I have now to do that. Now from the Specs your printer prints about as fast as any of ours so right now with our 2K setup we can print 3x as much - maybe you would be right to claim that there might be less problems with a printer of the quality and price of a Markforged (IDK what they cost but I would guess probably more than 5k) but for lets say 5k we could have 8 printers and that means I got at least 7 working all the time and 8 most of the time. I would expect a printer in the Markforged price range to be easier to use but ours aren’t that bad and the knowledge needed to properly use them is not beyond what teenagers can comprehend.
Now if I would buy a stainless steel nozzle or a ruby nozzle so spend $200 max extra I can do chopped carbon fiber too. I have not seen the need for it yet - but maybe we will in the future.
So now comes another point. As far as I know - and I might be wrong Markforged expects you to use the Markforged filament which as about $200/kg. We intend of having about 20-25kg of printed parts on the robot at $200/kg that would be $5000 which as filament is not KOP would blow up the budget. We use mostly hips at $10/kg and PETG at $14/kg and some bridge and CX12 and 910 for the Nylon parts at $40-$60/kg so we expect the 20-25kg of filament to come in in the $500-$700 range. depending on how much of it is going to be Nylon.
Now If I had the money to buy 8 markforgeds - maybe I would or at least get one of the continuous fiber ones and spend the rest of the budget on a mix of printers on the Anycubic/Artillery/Creality/Prusa Price range.
As for chopped carbon fiber - as stated before. Makes the part harder, stiffer, more abrasion resistant at a cost of tensile strength and layer adhesion strength. So if that is what you want/need then chopped carbon fiber is a good choice
You make a lot of great points and clearly you’ve demonstrated lots of great 3D printing expertise on this forum. The Onyx One costs $3500, and if that was my budget, I’d probably do the same thing as you - buy multiple of the less expensive, but still high quality, printers instead.
You are correct that the Markforgeds only print their proprietary filament, but it is less expensive than you might think. The spools are sold as having 800 cm^3, but I find that they actually have about 1000. They sell for $189 retail, but if you apply for their sponsorship they will sell you up to 10 spools a year at half price.
As it turns out, I have a bit more budget for another printer this year and am looking at the larger printers. I know you’ve had a lot of success with the Anycubic Chiron and I am considering that. I’m also looking at the Creality CR-10 S5, which is a bit bigger but more expensive. We’re also looking at the Modix which is a lot more expensive but even bigger and seems more industrial. ( http://www.modix3d.com/ ). Does anyone have any experience with those?
The Creality CR10-s5 is way more expensive if you want to go beyond pla. Alone the big heatbed will not get hot enough so you will replace that (I think about 250-300) plus you will have to do some beefing up. A better choice in that price range is the CR-Max - a little smaller but more modern. We got the CHIRON as it was just big enough (we were looking for 400x400x400 or larger due to the designs we have) and as we were able to get it for about 400 a piece.) In the worst case we figured its a decent price for the parts and we can build one with less $$ added than an S5 or S4 would cost. Luckily this was not necessary. And unfortunately we are a very low budget team. The only thing we did so far to the CHIRON is replace the PTFE tube with a Capricorn High temp one as we are mostly printing at 255 (HIPS) IDK if the stock one would have lasted as the printer lets you go to 260 without warning out of the box but we figured better spend the $11 for the capricorn tubing and be safe and know its rated for that (up to 300) We also printed a wire holder as we did not agree with anycubics wire managment approach. I probably will up the chiron files to grab cad too at some point.
For our next one I would like to get a delta or a corexy and enclose it pretty much airtight with a filter and an E3D copper hotend, V6 or volcano and make it a high/temp/noxious fumes type printer for POM and PC etc. POM I probably could print already if I put the printer outside for the vapors that you get when printing POM(delrin)
We also always wanted a CNC. So right now we started designing/building one. It will be powered by a handrouter someone donated to us which we will mount to an XY carriage use a Bigtreetech SKR1.3 board with DRV8825 drivers and some of the 1.7-2A NEMA 17 steppers some hardware, initially belt driven and maybe at some point rack and pinion. We are goign to use a GRBL/Marlin combo for software/firmware All in all less that $250. The dimensions will be slightly bigger than 1Mx1M so big enough to cut everything out the bottom plate on a robot frame. We will look to cut plastics, fiberglass, Hardboard, lexan, and up to 1/8 in aluminum. It is intended to compliment the 3d Printers. You might say we are not going to come close to a … in accuracy and quality but then we are looking to improve on the quality a teenager produces with a hand drill and sawsall.
You make some great points but currently we are looking at ways for stronger parts not more parts because we the machining capabilities to do most of our parts. This is the reason we are looking at doing chopped carbon parts because we want structural pieces for when we decide to use it and if we can 3d a part that is as strong as aluminum that is what we will do because of the complexity of some of the parts we made it took many hours to cam and machines mass amounts of them like we needed. Personally as the design lead of my team I would design more parts that would need to be 3d printed if I had the capabilities to 3d print strong parts
Strong is relative. And if you 3D print you have to design for 3DP. There are some things you can do with plastic you can’t with metal and vice Versa. Sometimes it also makes sens to compound them. see this latest test
This is a 3DP rack and pinion printed in hips. It is mounted on 2 1/2 in tube alluminum pieces that run through the frame as reinforcement. That way we can use it to join pieces together
Here is a very early version of bumblebot
If you zoom in you can see a couple of aluminum sticks here and there to help join it together and the complicated stuff is done by 3DP plastic and the bottom is a piece of hardboard - probably will be lexan of fiberglass in a final version. One of the ideas is to take the hardboard and wrap it in glass or carbonfiber to give the frame the stiffness instead of the aluminum rods. In which case the 3DP frame becomes the scaffolding of the Fiberglass or Carbonfiber. Goal is to cut down on the need to machine and be light weight and strong. If the robot winds up to be too light you can always add some weight
Now when it comes to material I suggest you test it in the environment and stress you intend to use it. Even Nylon is not Nylon there are hard nylons which are stiff an then there are soft ones that are quite flexi. So for gears for example you want a stiff one. If you have low RPM like on the Rack and pinion in the Video HIPS is a good choice as its pretty hard reasonably slippery so it will live doing what you saw in the video at that speed over and over (lift the robot) Without the need to lubricate. Now if you would use PLA you need to lubricate as otherwise the gears grind to dust rather quickly. Now chopped CF is good for gears - overkill for the ones in this video but probably good for High speed ones that have low to medium torque if you up the torque you probably want to drop the chopped CF and go for tensile strenght and layer adhesion. If I make a part that has to excell in resisting compression then I’d consider chopped CF if i have one that has some forces that try to rip it apart or break it in 2 I drop the CF and if necessary shove a tube of metal down the middle where the biggest stress is or a tube of Fiberglass or CF or runn straight fibers through and fill it with epoxy or PolyUrathane. So ask yourself - How strong is strong enough - what is strong - how much elongation and deflection is enough. Now if you look at our rack and pinion you will see they are double racks. Reason for that is that the biggest enemy of a rack and pinion under load is deflection. Now plastics and thin aluminum are quite flexible so there is the danger that the rack will bend away from the gear and you skip. Putting a gear on both sides and having a stiff axle with a bearing on both sides means the gear does not go anywhere and the 2 gears that have one behind it to intermesh with the other one grip the rack and the deflecting forces cancle each other out. Now with a 14.5 deg pressure angle about 10% goes on the axle as load so if I have 100lb hanging from the rack I got 10lb of axle load in the X direction and of cours the 100lb in the Z down now in the above rack held and shared by 2 gears one on each side and a 2nd set some inches away that keep the rack from twisting. the Double helix holds it in place and there is no need for a thrust bearing either. Now if I would have done that in metal I would have never done it that way. Helical and especially double helical gears are extremely difficult to machine and expensive and I probably in that weight category wouldn’t have had to worry that much about deflection if I get a beefy enough rack and it probably would have been much heavier and have a whole slew of other challenges that might have made other solutions more feasible. So the point is when you use plastic and 3DP you have to pick the right plastic and printer and design for it. It would not be the first time that I hear or read some one say “Oh yeah CF - great stuff and then it does not work as expected” Even if you work with molds in some cases Glass is better than Carbon and in some both loose to kevlar. And sometimes simple Nylon6 or 12 etc or Delrin or PEEK or PC or PP or TPU is the way to go.
A lot of the stuff you and your team do is really cool.
One thing I’m not sure I understand is how it is beneficial to print everything. In the drivetrain example you posted above, why not just print a block or something for where the wheels are, and just have regular tube structure for the frame? The assembly can be exactly the same, and it is much faster and cheaper to make this way.
Don’t get me wrong, 33 is firmly in the “print more stuff” column, but a lot of things I’ve seen over the past couple of months seem to be unnecessary and actually more of a resource draw, short of your team not being allowed to do basic bandsaw or hacksaw cuts to tubes, and basic drill press or even hand drill use.
For context, most teams that look at buying the expensive 3D printers do so because they already have access to 2/2.5D machining (usually in the form of a CNC router). Getting a small CNC mill, like a Tormach, is a significant step up from that, both in money and personnel resources. A printer on the tier of a Markforged is seen as a cheaper alternative, gaining the ability to do 90% of the parts you would need a proper CNC mill for at a fraction of the cost and effort.
If you are looking for a 3D printed part that is as strong as aluminum, at least 50% of that will be the design of the part, NOT the material strength itself. It may be thicker than the aluminum, it may be thinner, it might be a different shape altogether. I’ve been known to design a part that was a plastic shell to hold everything together, but it had steel and aluminum running through/around it for the real strength.
Strong is relative. Aluminum doesn’t have printing lines that are a bit of an inherent failure-waiting-to-happen (usually)–but if you design your parts correctly, you can make up a lot of that strength. I would also posit that you cannot fully stick with aluminum design principles when you start making your parts out of plastic, regardless of whether or not it has carbon in it. You might need it to be thicker, you might need to pre-make the holes (and I would recommend doing that wherever possible with 3D printing)–you might even need to use an insert instead of a tap to thread holes. It’s a different ballgame.
I understand I just want to avoid Chopped CF to being deified its just another material that may or may not be the best choice depending on what you are after.
Just to throw in some reasonable science behind this discussion, here is a YouTube channel I have found very helpful in decision making. Stefan does do chopped fiber testing, which includes comparisons for strength against plain Nylon.
Pay close attention to how he determines what the proper print settings are for his printer / material combination.
One of the interesting things I’ve discerned from this channel is that different filament brands & printer combinations will work very differently from a strength / layer adhesion perspective. Since the MF printers are specifically tuned to work exceptionally well with MF’s in-house Onyx filament, this may explain why teams see much better results with the base MF printer than what Stefan sees with his testing of various brands on a Prusa, even with similar printer settings / materials.
I am a subscriber - lots of good data. There are some good helpful videos coming up along the line on volumetric Extrusion. Different printers will produce different results at the same temperature but in my experience don’t vary that much when tuned properly hence the Vol. E thing. As long as the mechanics work fine. I.E. the Z-X-Y distances stay correct the rest is up to if the plastic is properly molten and at the right viscosity. Different hotends having different Melt zones and with that different speeds at which they melt plastic. Furthermore the reported temp is not the same for different printers as there is an error of as much as ± 10 deg C so a 270 on printer A could be a 250 on B cause on both it really is a 260. Hence you inspect your extrusion and establish the max Volumetric E as I describe here
Then its important to get the Extrusion multiplier correct I did a video on both of those here
By measuring a known wall thickness and inspecting for over/under extrusion you adjuest the Extrusion multiplier accordingly. At that point you will get results that lie within statistical norms of being insignificant between different printers for the same filament (there are differences between different brands of filaments So you have to establish the optimum for each).
Now if you are working within lets say the markforged Universe/environment then markforged did those calculations and tests for you and you can use their presets. For me that is not worth the extra money for someone else it might. Plus I get to teach the kids how to do that - another plus
Now Stephan has much better equipment than us and I have asked him to make some tests and have heard he will do some of them soon as he finds them interesting too. We are also working on improving our equipment and that is some of the off season projects we will tackle.
Print Quality and layer adhesion drops quickly if the mechanics are off (broken, worn, misadjusted) or some simple rules are not followed. So for example most printers that use a standard NEMA 17 stepper and a lead screw for adjusting Z and are not delta printers will produces acurate Z steps in 0.04mm increments in between you half step or use other fractional steps which are not that accurate. So for good consistent layer height on any Prusa, Creality, Anycubic etc type printer make the layer height a multiple of 0.04 so for exampe .25 is bad use .24 or .28 instead