Hi Tom, Thanks for collaborating with our Engineers and taking the time to investigate your previous findings further. As you know, we’re always happy to jump on a call to discuss performance matters in detail. With Roto running at its rated current you were able to achieve at least 12mm/s of filament speed in PLA through a 0.6mm High Flow Revo nozzle. This is a volumetric flow of 28.86mm3/s, which exceeds our rated maximum volumetric flow rate for this nozzle-hotend-material combination (22mm3/s). Unfortunately, the issue with only using one nozzle-hotend-material combination to infer the maximum capability of an extruder is that it gives an incomplete picture as the force/speed curve for each set up varies. For example, by changing from PLA to PETG you would see a higher limit. Similarly, if you were to change to a 1.4mm High Flow nozzle you would see a higher limit again. Please check out our Shorts to see a video of Roto printing with a 0.6mm High Flow nozzle in PETG at 31mm3/s and a 1.4mm High Flow nozzle in PETG at 49mm3/s. If it suits your schedule, we’d be up for coming on The Meltzone Podcast to discuss your other points/design choices further 😊. Apologies it took us so long to reply, TH-cam wasn't liking the links we had included. Thanks, Team E3D
if you watch the video it does not relavant with flow, its simply you have a tiny motor that couldn t push as much as compaired to others(in the video) at 24V or lower.
A guy named NightHawkInLight had a somewhat recent video on how you can test multiple parameters that influence each other and test them with only a few tests. I think it would be worth a check so that you can be able to do testing of more variable while not sacrificing time
Indeed, clever evolution of concurrent testing with smart filtering of the results. Not sure how that could be accomplished given the tests can't be performed concurrently.
Design of Experiments (DoE) is the methodology he’d want to use. We call testing each and every point full factorial. With multiple variables you can quickly to an untenable amount of points to run.
processing data in spreadsheets is also pretty slow... this is where data science and the use of a python library (Pandas) comes in handy. Thomas thinks that's a lot of data, but it's definitely small as far as data science subsets go
I believe that Thomas is working on that recycle system as well. Hope it goes better than his MP-CNC where they got mad at him for making an improvement.
@@SiFireHasSpeedYikes...I actually stopped looking at the MP-CNC after Thomas' issues, not just with the creator, but the unit itself. The creator side of it sealed the deal on avoiding that thing.
Appreciate your honesty, Tom. Many would chicken out to tell a company bad news and I also appreciate E3Ds way to respond. Kind of shocking that the Smart Orbiter, as a one man show development, can not only compete but is superior in some ways
This is typical of E3D. They release their overpriced f-ed up products and later on they fix them. That is why i don't buy their products and don't trust them. They are almost like Apple.
Love the follow up Video ! If you have experiments with a lot of variables consider design of Experiments (statistische Versuchsplanung) and statistical Result testing to qunatify the results. It is a little intimitading at first but really pays of in these complex analysis.
@@Max-tj7bp You can easily analyse it with a small self-written script, even in excel if you set up your equations by hand. The mathematics behind is quite basics, if you stay at the classical fractional factorial design.
Agree with @Kaliumcyanidful on this. I was using design of experiments/DOE/DOX back in the 90s to optimise injection moulding settings. I'd rather implement it in a dedicated stats package like Minitab than Excel, or use a relevant library in R. There are also dedicated DOE packages too, which are great, but pricey, unless you can get sponsorship from them. PS Taguchi designs are interesting but in my opinion, you need to be a stats ninja to run them with confidence.
Hey Tom, thanks for taking the time and investigating the Roto Revo again, great video! I would love to see more of the smart orbiter v3.0 - it looks really promising
Absolute awesome insight! Never thought that all this factors might have such a big impact on extrusion. Also: it's a joy to see someone use excel in a proper way than just for basic calculations
This is really interesting. I would not have expect the smart orbitor 3 actually went this well. I would really like to see a video on reviewing the smart orbitor 3 in the future.
Good deep(er) dive with the comparsion to the Smart Orbiter 3 because I was having trouble justifying the way higher cost of the Revo Roto. It's very nice but couldn't understand the cost difference.
9 หลายเดือนก่อน +1
thanks for the follow up. Would love to see a dedicated video on the Smart Orbiter v3 - maybe combined with a Hermit Crab v2 CAN board would be the dream, as I'm eying with that combo :)
Take the hot end out of the test by setting a constant force on the filament. Fewer parameters to change, no filament to waste. A 'reverse extruder' on the output to adjust the back pressure.
Thought about that, too! I'd still need to run filament through the extruder though, as the force produced at standstill is quite different to the one at speed, and I'd probably also need to use fresh material each run as the extruder teeth chew up the filament quite a bit. On the upside, at least I could still re-use the chewed-on filament for simple prints in another machine.
A magnet field is caused by current not the voltage. What a driver does is open transistors to apply the voltage which results into a current raising over time with the factor e ^(t/tau) where tau=r/L. tau is the time constant. After x % of a PWM period the transistors are switched off and the energy of the magnet field getting converted into current which is sinking with the factor 1-e^(t/tau). If the PWM frequency is too high or tau too large, the driver will never be able to control the current to Imotor_max. If the voltage is too low or r too high the system will never be able to get to Imotor_max either. Hence the motor might the wrong type with a too high inductivity or coil resistance, since system parameters like the 24V can't be changed.
I don’t know what the Trinamic drivers are using internally, but there are numerous ways to implement a constant current design. They are probably using a digital averaging design for RMS current, but they could also be using a D/A voltage reference for a true set current. The problem with the Roto motor is that it has too many windings, thus it requires a higher voltage to reach the required current.
@@cnc-maker They are using PWM with a H-Bridge current measurement measuring every cycle, which is then eventually averaged with a low pass. But this isn't interesting, because the current can only be achieved when the impedance of the coils isn't to large, which is here clearly the case. My guess is E3D had another motor in the lab than in production. It's a QM problem. I had a similar problem with hydraulic valves, actuated by a magnet field of a coil. Usually the PWM has a frequency of 150Hz but an engineer gave us a cheap card with 30kHz and the range of the valve was 0%-30% only.
Woow, lots of analysis. Great Job. At one point, I was a bit lost, you needed to put a lot in the explanations. But at the end of the trending is really clear and your conclusions seem really interesting. Thank you Thomas. Have a nice day.
Was looking to "Shrink" my extruder footprint on my CR10 and then also include the Beacon Probe. I am heavily invested in Revo nozzles over all my printers so Roto looked interesting. In my case, I don't need above 6mm/s so Roto would be a good fit. But.... Then came the price and this was unfortunately the deciding factor. I am rather waiting for Orbiter 3 with the Revo integration to see what the pricing then looks like. Once again, your videos and time taken into the details are very much appreciated.
For experiments with many variables like this, you can use fractional factorial design to cut down the number of tests, which is basically the "official" version of what you did with the graphs in excel
You are missing induction aka Henry, or for coils mH(enrys). It's a key to how sudden a stepper can move on a change, but also how much it can hold as a contra.
In a bygone era (3 years ago) such considerations (very high speed and flow) were a non-consideration. For example I have put a Sherpa Mini on my old QQS-Pro. It prints at 80mm/s and the moons NEMA14 with 5:1 gearing works fine, so yeah this Revo Roto would have been a lovely project for that machine. Yeah retraction speeds are limited, but it works. Some things I have also discovered over the years; always set your TMC current to the RMS value of the max Stepper current ( multiply by 0.707) as it will inavrably push less than this in general use, and I see no need for safety margins that people sometimes recommend; temperatures of the motors are fine and have been running for years now. Also some major things about extruders; get your acceleration correct! Turn off Stealthchop and also turn off interpoleration; these settings alone will return your extruder to actually sounding like and extruder, if you're unsure. All these things alone can reduce accuracy of stepper movements; the TMC data sheets will actually tell you this, as well as some Klipper documentation out there. It's interesting as all this information is out there and written down in some shape or form, but not easy to pick out and collate in general (as in everything in electronics). As such well done for bringing some of this information with actual testing to the masses!
interesting that the revo motor is basically at Vmax the entire time, wasn't expecting that outcome, and you're right, quite surprising they didn't work with LDO on it, I wonder if Jason and Co will make a swap in LDO motor for the roto
They said they were working on a "different" motor version almost as soon as it was released because of the max temperature of the stock motor. So maybe an opportunity there to work with a different manufacturer.
the stepper on the hermera is a 24v stepper therefor its internal resistance is low. on those other ones there are for much higher voltage so there internal resistance is lower. if you want a stepper hat can push until gears breaking use lower voltage steppers and diver that can handle higher current. there is no problem over driving motors and if they get to hot, there are heat sink collars you can buy for it. the thing is if you want a fast direct drive system you need a lighter head but most of the weight of the head is the motor, but if you want a lighter motor you need a smaller one. but the problem is if you make the motor smaller and want the power you have only 2 option, first is less internal resistance for more current but most stepper driver are limited in current and cant push any harder. therefore you need a stepper with higher voltage to lower the current that is required for the same power but your system uses 24v and you using a 48v stepper because it weight less you also have less power. this may no big deal for a ender 3 or other 50mm/s printer but machines that can 200mm/s this limits you. bigger stepper are less a problem because the stepper driver sets the limit with its max current. for example you have normally a 24v stepper with 1A and your driver can 1,5A the stepper uses its 24v on full power but if you buy a 48v stepper with 1A it will use also the 24v but because of its internal resistance it will only use 0.5A. the Current is mainly the reason for the size of a motor. so bigger stepper can handle more current. but if you buy a 12v stepper with 1A and use it on a 24v system the motor uses now 24v and up to 2A but you are limited to 1.5A so in the end you overdrive the stepper a bit.
I would be quite interested to see a 48 or 60V setup, just to see if that improves performance on the Roto. You might be able to get a setup where you can pass the data lines from a 12 or 24 volt board to a stepper driver, but have the driver powered at the higher voltage from a separate power supply.
.9 stepper is the problem. Everybody thinks more precision, better. But this is an extruder we are talking about here, giving up half the power was a mistake.
Not only pushing force, but stealthchops causes so many headaches! I can only see the point on slow workhorse printers where it's used for skip detection and stuff like that
Thomas, I would like to commend your approach and digging in entirely for what appeared to be an obvious limit of flow in your previous video. Here you answer what the bottleneck is, and it's wonderful information for any printer enthousiast. To me this exposes a product that seems to fall in the category "because we can". I honestly don't like the response from E3D that they posted here, it does nothing to address the potential issues, and there are plenty: - Stepper motor needs more than 24V when running fast due to back-EMF from the inductance aspects of a stepper motor (coils don't discharge instantaneously) - Load on CPU is also way higher, as the Revo Roto extruder has 2682 steps per mm. This is such a big increase from a regular setup that this has to be investigated to find the implications when you are running Marlin on a relatively weak board (Arm M0+ for instance on a BTT SKR MINI E3 V3) - Stepper motor should run really really hot. How hot? Nobody knows. - Driver might also run hot because it has a lot of work to do. How hot? Nobody knows. Maybe you even need to turn off interpolation to alleviate some limits. But again, nobody knows. These limits that could all have become relevant, and all for just being able to have very high torque from a very weak motor in a very light package. A light package which will only shine when....you don't go slow, but start cranking the speeds. This product missed the mark. Could have gone 1.8 instead of 0.9, without any downsides, could have stayed with a Nema17, but then go full pancake and still have a pretty lightweight extruder.
I really don't like the direction E3D is going. Closed Revo ecosystem, very expensive, less performance without super expensive high flow nozzles and heaters, etc.
Personally I think that they have so much passion to tinker that they spend a lot of money on it. As they spend a lot of money on it they need to get it back somewhere... Is it sustainable? Only time will tell.
I’m not sure I agree. The closed ecosystem is a bummer, but they are partnering with other manufacturers for products that otherwise wouldn’t exist (cht, biqu, sprite upgrade?). The Revo hotends aren’t that expensive considering what they are and what they come with, though I do wish the nozzles were more affordable. They are consistently putting out great products and standing behind their products when they make mistakes (Revo heater core). Obxidian was a LONG wait, but you should check out the blog post about the process if you haven’t. It makes sense things can go wrong when pushing a new process into production at a mass scale. When we compare the level of refinement we’re seeing in the current extruders, hotends, and nozzles to the original V6 era designs, I really can’t see how the value isn’t backing up the price. As I said, I do wish the ecosystem was open, but based on my AliExpress suggestions, they aren’t going full Slice on those listings and there are tons of knock off nozzles and accessories available.
The REVO idea of uniting the nozzle and heartbreak is great at eliminating jams at the nozzle-heartbreak interface. However their flow rates, even “high flow rate”, have lagged behind many competitors. It’s telling when the printers that push the limits do not use any E3D product. Also the Obxidian fiasco was terrible. Marketing getting ahead of production by 2-3 years.
Tbh, not convinced by the Roto. However I do love the Revo system and do wish Prusa used them on the MK4 but also understand Prusas want for backwards compatibility
The concept of REVO is nice but very expensive. I even invested in a diamond nozzle.... Not worth the money if you consider mediocre performance on a "normal" brass nozzle level. 😂
It seems like were getting at the borders of E3ds abilities to design, source, and spec out to manufacturers smaller powerful systems. It may not be a physics limit, but an organizational one and or systemic in the ability to find cheaply producable components to keep driving size and weight down. A lot (most?) 3d componentry isnt developed as consumer 3d printing technology initially,. We're getting refinements of tech and designs that found other uses. For example an old M3D printer I had used motors commonly used to move the louvres on mini-split AC units for their geared extruder steppers. The more bespoke the solution is required to get these specialized designs going the higher the cost and harder it is going to be to find a manufacturer willing to make the relatively small runs that E3d can drum up.
E3D seem to be pushed towards making more high-end parts that offer specific benefits. The mid- and low-end market is quite saturated with products that all easily pass the "good enough" threshold and are often produced with maximum cost optimization in design, materials and labor. E3D, being a UK company, simply can't compete there.
@@MadeWithLayers So then how can they explain the smartorbiter v3.0? It's 1/3 the price, and the only downside is that is doesn't use Revo, which might not even be a downside. E3D has charged a premium for their products, and it has been worth it, but they need to take a look at what they are offering for the money they are asking.
Would be interesting to see how Induction heating can influence flow rate compared to various resistive heaters. Plasmics offers Induction heaters at 24 and 48 volts.
Wow. Im so glad you found that out not me. I had a tatty E3D titan Aero replica on my baby Kingroon and was looking at the Roto as its lightweight and fast but now i see its not, As luck would have it i picked up a Creality Efit extruder instead (for £30 brand new) Nothing flashy but imho, a seriously great bit of kit (considering its make) Ive not had any problems pushing it hard for 48hrs while printing the super long bedslinger parts im building for one piece RC plane wings. The quality the little KP3S is banging out now has also shocked me going on the amount of bad rap the thing gets elsewhere. Could you test one of these Efit extruders? See how it compairs the the Hemera as the New bedslingers going to be needing an extruder soon.
I really wanna make a high voltage extrusion system now. I prefer having excessive power and tuning down than knowing there's a limit. Plus having a 60 volt system is just hilarious
A bit of work making sure insulation is up to scratch. You can get away with a lot running 24 V, but 60 V is going to find ways to evade the paths you want unless you are more careful. Use resettable fuses.
PLEASE PLEASE PLEASE. do a full review of the LTO smart orbiter 3. I believe it has inbuilt tool board with accelerometers and USB connectivity? Are the hardened gears standard or an optional extra. What is the hot end max temp???? Please do a review. 🎉
Flow rate tests aren't a good way to test extruders, instead they are good for testing nozzles and heater blocks. The flaw is that higher force extruders may not really achieve a higher flow compared to what the Revo hot sides are actually rated for -- I think its more likely the extruders are just forcing colder, harder plastic through the nozzle. The Revo nozzles are pretty well known at this point at producing specific volumetric maximums in different materials, which are well-documented on E3D's website. Those flow ratings are not a function of the extruder, but rather the material and how fast that material can be heated. These days, pretty much every extruder can produce far more force on the filament and hotend than is necessary, especially at small nozzle sizes like the 0.6mm that was tested. This can be proven with intuition -- you only need light finger pressure to constantly advance filament into your standard 0.4mm hot setup! The *real* job of the extruder (outside of exotics / flexibles) is to advance and retract at very specific speeds/steps. The ultimate test data to compare extruders would be something like: 1) does it meet the target volumetric rating of the hot side in use? 2) measure precision 3) measure flexibles feeding 4) can it dial in proper amount of gear backlash? 5) long term gear-wear 6) motor temperatures 7) conditions where steps start to skip, such as current/temp/backpressure TLDR: The nozzle and heater block are almost always the problem with flow limitations, not the extruder. Need different tests for extruder performance.
PSA: The orbiter is horrible to take apart! To access the filament path you need to take out all screws (which are all different lengths and have different diameters and you need two different allen keys), and once you have it taken apart, the planetary gears fall out as well as the ball bearing for the filament sensor. Reassembling is even more annoying...
Interesting, so how does stealthchop compare on motion in 3d printers. Ive seen my printer skip with stealthchop on at 20k accel but never with it off. Would be interesting to see data on that too.
It might be interesting, but if this is a printer anywhere near a person, you wouldn't want to run with stealthchop disabled on the kinematic system motors. If you want higher acceleration, get motors that can support those accelerations at the speeds you are moving.
Using higher voltage and lower current means the copper windings can be lighter (size of copper is dictated by current). This makes for a lighter motor
Not quite - you also need more turns on the stator when using less current to achieve the same magnetic flux, so the material savings effectively get cancelled out.
The magnetic field strength(magnetic force) is directly proportional to the current but the current is proportional the the square of the weight of the copper wire. @@MadeWithLayers
As a very quick test, you should measure the DC resistance of the steppers. The Revo Roto would probably have a higher reading. Of course an inductance measurement would but better but most people tend to have DC multi meters and the measurements could possibly add to the overall conclusion you have arrived at.
The resistance is actually listed online, at 7.5 Ohm for the Revo Roto (but no inductance, which I'd assume would be rather high as well): e3d-online.com/pages/help-centre-roto
I'm not familiar with Taguchi methods, but I'm sure there are quite a few DoE optimizations that I could take. Still challenging to present such multi-dimensional data though.
Tom, can you comment on how you (and perhaps Stephan) arrive at the extrusion forces you use? I know that if I applied 5 to 10 kilograms of force to a 1.75mm diameter PLA filament in a 0.4mm nozzle heated to 210C, I would tear through a spool of filament in a matter of minutes, if not seconds. Can you clarify, please?
You're pretty far off with the numbers you "know" - there's no way a hotend could even melt that amount of filament in "seconds". Look at the speeds these extruders are achieving, they're around 10-20mm/s, using a 0.6mm HF nozzle. Assuming the 12mm/s I found for the Revo Roto, a full 1kg spool of PLA would take almost 8 hours to extrude. Not an unrealistic time to output that amount of filament tbh.
I was going to be putting revo roto's on a 4 tool, toolchanger system. Looks like I will be using a Smartorbiter v3.0 now. I cannot justify spending triple the price for a package that is barely smaller and lighter, and performs worse. The only saving grace of the revo roto, is the revo hotside, but it's not worth $200+ as well as the more expensive and more limited nozzle selection.
Well if the time it took for Hemera XS is any metric (which was obvious need from even the days of Hermes), you'll have to wait a long time for fixed Roto
Good afternoon, Assuming standard flow geometry for a 0.6mm nozzle (so a layer height of 0.3mm and a track width of 0.72mm), 140mm/s xy speed would convert to 30.24mm3/s. Here’s a video of Roto doing just that (31mm3/s) with a 0.6mm Revo High Flow nozzle - th-cam.com/users/shortsTvgeMpdz6JM. Thanks, Team E3D
It was all the way at the end, but it's definitely one of the most important parts: they're using a 0.9 degree stepper. In testing, a 0.9 degree stepper has about HALF the torque of a 1.8 degree stepper which pretty much all other extruders use. A simple stepper swap could likely massively improve the performance of the Revo Roto. I really like E3D and respect what they've done for the 3D printing space, but they don't seem to recognize that faster and high flow printing is becoming more and more the norm.
Afternoon, The design philosophy behind Roto was to have all the extrusion force you need in as lightweight a package as possible. Changing to a larger or more powerful stepper motor wouldn’t enable you to print any faster. Roto can already achieve the maximum volumetric flow requirements for every nozzle in the Revo ecosystem. If you increase flow beyond the rate that a given nozzle-hotend combination can melt it, you’ll just be pushing harder and cold extruding. The spike in force requirement in this situation is because the filament is not fully molten. Printing like this will lead to poor quality prints. If you were to put a bigger nozzle (with a higher maximum flow rate) onto the Roto you would be able to achieve higher extrusion speeds. Thanks, Team E3D
Mixed feeling: - Roto is ill-dimensioned for current 24V ecosystem standard - Roto is "future proofed" for the more desirable 48V ecosystem which is (very) slowly being introduced on the market (I am personally already transitioning to 48V using the last duet control boards) - by the time 48V ecosystem becomes a standard, the non-electrical side of the Roto might be outdated. --> maybe E3D should propose a 24V version and a 48V version of the Roto. The 24V might be a bit heavier but engineering = compromising, isn't it.
Hey Raoul, You don’t need to run 48V, 24V is more than sufficient to achieve the maximum volumetric flow requirements for every nozzle in the Revo ecosystem! Please see our pinned comment for some more information. Thanks, E3D
Interesting Test, but still I think that the Roto shines on other characteristics. Not everybody needs maximum extrusion torque. Being much lighter, with a center of gravity that is much more in the center are harder to quantify but more important for most users motion systems IMHO.
The weight is definitely an advantage over E3Ds earlier offerings, but the weight between it and the Smart Orbiter should be relatively comparable in a full setup. The Roto weighs 146 Grams + 8g Nozzle (listed) + 20g Heater Cartridge (estimation) = 174g. The Smart Orbiter v3 is listed at 175g for the whole system (as it includes everything). Center of Mass should also be mostly similar between the two. Weight and Center of Mass are only important when you are printing at high accelerations. You don't gain anything from running high accelerations if the extruder can't keep up. Especially as Revo emphasizes the ability to quickly change to the best Nozzle Diameter for each print, maxing out at ~100mm/s with a .6mm high flow nozzle is a weird choice. It is pretty likely that someone running this on a high performance printer would regularly run against the limits of this extruder. Choosing a slightly bigger and heavier motor would have been a more sensible option.
@@scratchfisch6425 Thanks for the corrections, might be comparable in weight after all... But I have to disagree with the latter. High wheight and bad center of gravity becomes a dealbreaker at high speeds, but it still affects each and every print significantly also at lower speeds. Having less energy to jerk around every corner was and always will be better.
@@rbaile508 Yes it does also that, but it also improves the quality at every speed slightly, and rreduces general wear an tear on every motion system obviously.
@@perchte I agree to a point. Running something like a Hemera at 388g will cause artifacts at much lower accelerations than a lighter extruder. Extremes are definitely bad. But basically all "current gen" extruders (BMG, Sherpa Mini, Orbiter, Revo Roto etc.) weigh less than 175g. At that weight, the remaining differences in weight/center of mass will not have any noticeable impact on print quality at reasonable accelerations (keep in mind that input shaping drastically increased maximum accelerations without noticeable ghosting) Extruders need to have the right balance of weight and power, and the roto definitely misses the mark here.
To me it sounds like someone miscalculated the winding count & wire gauge requirements for the Roto. From what you’ve presented, there are too many windings in the Roto. The higher the winding count, the higher the voltage requirement. They should be using fewer windings and a larger gauge wire, which would then work with a 24v system.
@@Papinak2 step angle has very little, if anything at all, to do with this. This was proven when he dropped the multiplier from 16x to 8x. The fact that you can’t hit Imax of the motor at 24v is the issue.
@@cnc-maker it has a lot to do with it - he reduced microstepping multiplier, which removes bottleneck on the MCU side, but it doesn't reduce frequency through motor winding, which is large contributor to inductive resistance - double the frequency reduces current to half. So motor that needs half as many steps to complete full rotation would be able to run at roughly double the speed, everything else being equal.
@@cnc-maker what do you disagree with? That more steps results in higher frequency and lower current? I don't disagree that different winding could increase the speed, but at the same time, it decreases torque - it can be compensated by higher current thanks to thicker wire, but I'm not sure how well it scales.
Yes, I was going to make that joke, but decided against it in the end. Trinamic are still sort of their own thing and the brand will most likely stick around for quite some time.
They have their gears integral to the motor, not sure if they would match up. The motor from the Orbiter should improve things on the Roto, as it's not a 0.9° unit, but the Roto is still much more geared down.
@@MadeWithLayers seriously if the motor is geared down it means it does much more rews than it wouldn't do if it was runing as direct drive . so only reason to put 0.9 degree motor in such a thing is to make it more silent and that doesn't increase the torque but makes mcu work harder for no other benefits if anything it would be much better to go with ~50-100 steps/rev motor to give it some torque and accuracy would be increased by the gear ratio anyway the faster you want to print the lower the accuracy regardless of any motor specs because of fluid dynamics
So um I got a question what if I just slap on like a giant nema like 420 and just like print with that would that be enough force to like pick up the printer?!?! 😂
As an electronic engineer I'm completely confused. According to the documentation from E3D the resistance per phase (coil) is 7.5 ohms. At 24V you'd be passing 3.2 amps through each phase (short circuit). So to not exceed the recommended 0.6A at that resistance you'd only need 4.5v. I'm not sure what the whole speech you gave about needing a higher supply voltage to drive it better. Happy to be corrected. Also happy to get out my oscilloscope and measure the voltage waveform across an extruder stepper coil.
Context of voltage is that driving the stepper motor. Almost seems like a stepper motor has mismatched specification for the application. 24V power supplies are most common. Note this is just the extruder stepper motor, not the driver motors for controlling print head movement. It's like the extruder is targeted for a specific custom printer architecture.
@@AerialWaviator what!? I have no idea what the point you're trying to make is. It's just a stepper motor attached to an extruder. It has certain characteristics. Phase resistance and max current are just two of the dimensions that vary for stepper motors. Stepper motors have a max torque output based upon those characteristics. Supply voltage has little to do with it, unless the driver doesn't have enough voltage to drive the desired max current.
As a mechatronics engineer, I'm happy to help. At standstill, the static coil resistance is the only thing determining the voltage requirement, but as the motor spins up, it also starts producing significant back EMF en.wikipedia.org/wiki/Counter-electromotive_force, essentially acting as a generator working against the driver. This is where the relatively fixed RPM limit on Revo Roto seems to stem from, as at this point the generated voltage is starting to approach the supply voltage, essentially disabling the driver from pushing any current through it.
You kept referring to “stealthchop drivers”, but stealthchop2 is just a mode/capability *of* the driver-it can be changed in firmware or on the board. It might be better to use a more modern mcu/firmware (RRF or Klipper) to power that rig so you can test the same drivers under different modes and switch back between them easily. Stealthchop reduces peak torque and, in turn, may reduce positional accuracy (especially when combined with interpolation). Its purpose is to reduce noise during linear movement, so I can’t see any reason why an extruder should ever be set to Stealthchop mode. It is never moving at a noise-producing speed for any appreciable length of time, so all it’s doing is just reducing available torque.
Are you not over complicating things, and simply running up to the limit of how fast the stepper can spin? Are they the same motor? Can they be swapped?
The planetary gear boxes make it hard to swap motors because they are made with an expected teeth count, LDO has many motors with different teeth counts because of this and other extruders. The current ones they carry are: 6, 8, 9 and 10. At least those are the ones I remember.
@@MatrixRay19 yep, 6 is custom for annex engineering and their sherpa extruders, 8 is standard, 9 is custom for galileo 2 and a longer spur gear than usual, 10 is again standard. All have a module of 2mm and the 8, 9 and 10 teeth steppers have press fit gears, the 6 teeth one has the gear integrated onto the shaft due to size constraints. There are additional variations, for example the stepper that ships with the lgx lite has a steel gear, so you can use it with the heavy duty lgx gears made entirely from steel while standard is a brass gear.
So the roto is an engineering failure. The makers didn't consider all variables and didn't do enogh testing. This thing is flawed and pretty useless as a drop in replacement since 99% of all hobby 3D printers run 24V or 12V.
Hi Tom,
Thanks for collaborating with our Engineers and taking the time to investigate your previous findings further. As you know, we’re always happy to jump on a call to discuss performance matters in detail.
With Roto running at its rated current you were able to achieve at least 12mm/s of filament speed in PLA through a 0.6mm High Flow Revo nozzle. This is a volumetric flow of 28.86mm3/s, which exceeds our rated maximum volumetric flow rate for this nozzle-hotend-material combination (22mm3/s). Unfortunately, the issue with only using one nozzle-hotend-material combination to infer the maximum capability of an extruder is that it gives an incomplete picture as the force/speed curve for each set up varies.
For example, by changing from PLA to PETG you would see a higher limit. Similarly, if you were to change to a 1.4mm High Flow nozzle you would see a higher limit again. Please check out our Shorts to see a video of Roto printing with a 0.6mm High Flow nozzle in PETG at 31mm3/s and a 1.4mm High Flow nozzle in PETG at 49mm3/s.
If it suits your schedule, we’d be up for coming on The Meltzone Podcast to discuss your other points/design choices further 😊.
Apologies it took us so long to reply, TH-cam wasn't liking the links we had included.
Thanks,
Team E3D
Speed demo - th-cam.com/users/shortsTvgeMpdz6JM
Oooch! I hope this comes to fruition! Really interesting stuff :)
if you watch the video it does not relavant with flow, its simply you have a tiny motor that couldn t push as much as compaired to others(in the video) at 24V or lower.
stop your marketing excuses. the point is it's much worse than the competition at 2-3 times the price.
just change the motor, it's not hard nor more expensive. makes no sense.
Would love to see a full review of the Smart Orbiter v3! Glad to see it in this testing.
I can still find no reviews of the SO3, just an unboxing and a teardown. Looking forward to one
A guy named NightHawkInLight had a somewhat recent video on how you can test multiple parameters that influence each other and test them with only a few tests. I think it would be worth a check so that you can be able to do testing of more variable while not sacrificing time
That method is called ortogonal array: en.wikipedia.org/wiki/Orthogonal_array
NightHawk was talking about Taguchi arrays, as Tom is an engineer I'm sure he knows about them.
Indeed, clever evolution of concurrent testing with smart filtering of the results. Not sure how that could be accomplished given the tests can't be performed concurrently.
Design of Experiments (DoE) is the methodology he’d want to use. We call testing each and every point full factorial. With multiple variables you can quickly to an untenable amount of points to run.
processing data in spreadsheets is also pretty slow... this is where data science and the use of a python library (Pandas) comes in handy. Thomas thinks that's a lot of data, but it's definitely small as far as data science subsets go
Don't think of this as a waste of filament, Tom... Its science!
Or pay a visit to Stefan of CNC Kitchen to get it all recycled.
I believe that Thomas is working on that recycle system as well. Hope it goes better than his MP-CNC where they got mad at him for making an improvement.
and better than Klipper where their minions attacked him for not being a mouthpiece for them.@@WreckDiver99
@@WreckDiver99 The guy that made the MP-CNC is apparently way crazy than just that incident. Unsurprising to me 😂
@@SiFireHasSpeedYikes...I actually stopped looking at the MP-CNC after Thomas' issues, not just with the creator, but the unit itself. The creator side of it sealed the deal on avoiding that thing.
Appreciate your honesty, Tom. Many would chicken out to tell a company bad news and I also appreciate E3Ds way to respond. Kind of shocking that the Smart Orbiter, as a one man show development, can not only compete but is superior in some ways
I’m looking forward to E3D’s response to this. Keep us posted!!
This is typical of E3D. They release their overpriced f-ed up products and later on they fix them. That is why i don't buy their products and don't trust them. They are almost like Apple.
Tom has kindly pinned our reply, see above :)
Love the follow up Video ! If you have experiments with a lot of variables consider design of Experiments (statistische Versuchsplanung) and statistical Result testing to qunatify the results. It is a little intimitading at first but really pays of in these complex analysis.
I think nighthawkinlight has a video about this. It was very fascinating to watch and learn
Getting software to analyse such experiments is pretty price though
@@Max-tj7bp You can easily analyse it with a small self-written script, even in excel if you set up your equations by hand. The mathematics behind is quite basics, if you stay at the classical fractional factorial design.
@@Max-tj7bp for commercial yes, but you also save a lot of money as work time is also extensive. there are also Python libs for free 👍
Agree with @Kaliumcyanidful on this. I was using design of experiments/DOE/DOX back in the 90s to optimise injection moulding settings. I'd rather implement it in a dedicated stats package like Minitab than Excel, or use a relevant library in R. There are also dedicated DOE packages too, which are great, but pricey, unless you can get sponsorship from them.
PS Taguchi designs are interesting but in my opinion, you need to be a stats ninja to run them with confidence.
Hey Tom, thanks for taking the time and investigating the Roto Revo again, great video!
I would love to see more of the smart orbiter v3.0 - it looks really promising
Absolute awesome insight! Never thought that all this factors might have such a big impact on extrusion.
Also: it's a joy to see someone use excel in a proper way than just for basic calculations
google sheets*
You confused Excel for a spreadsheet. Can do most things in all 3 main spreadsheet programs. Ms office. Google sheets or libre office
This is really interesting. I would not have expect the smart orbitor 3 actually went this well. I would really like to see a video on reviewing the smart orbitor 3 in the future.
Good deep(er) dive with the comparsion to the Smart Orbiter 3 because I was having trouble justifying the way higher cost of the Revo Roto. It's very nice but couldn't understand the cost difference.
thanks for the follow up. Would love to see a dedicated video on the Smart Orbiter v3 - maybe combined with a Hermit Crab v2 CAN board would be the dream, as I'm eying with that combo :)
you should send the test poops to Stefan for turning into recycled filament
Take the hot end out of the test by setting a constant force on the filament. Fewer parameters to change, no filament to waste. A 'reverse extruder' on the output to adjust the back pressure.
Thought about that, too! I'd still need to run filament through the extruder though, as the force produced at standstill is quite different to the one at speed, and I'd probably also need to use fresh material each run as the extruder teeth chew up the filament quite a bit. On the upside, at least I could still re-use the chewed-on filament for simple prints in another machine.
A magnet field is caused by current not the voltage. What a driver does is open transistors to apply the voltage which results into a current raising over time with the factor e ^(t/tau) where tau=r/L. tau is the time constant. After x % of a PWM period the transistors are switched off and the energy of the magnet field getting converted into current which is sinking with the factor 1-e^(t/tau). If the PWM frequency is too high or tau too large, the driver will never be able to control the current to Imotor_max. If the voltage is too low or r too high the system will never be able to get to Imotor_max either.
Hence the motor might the wrong type with a too high inductivity or coil resistance, since system parameters like the 24V can't be changed.
I don’t know what the Trinamic drivers are using internally, but there are numerous ways to implement a constant current design. They are probably using a digital averaging design for RMS current, but they could also be using a D/A voltage reference for a true set current. The problem with the Roto motor is that it has too many windings, thus it requires a higher voltage to reach the required current.
@@cnc-maker They are using PWM with a H-Bridge current measurement measuring every cycle, which is then eventually averaged with a low pass. But this isn't interesting, because the current can only be achieved when the impedance of the coils isn't to large, which is here clearly the case. My guess is E3D had another motor in the lab than in production. It's a QM problem. I had a similar problem with hydraulic valves, actuated by a magnet field of a coil. Usually the PWM has a frequency of 150Hz but an engineer gave us a cheap card with 30kHz and the range of the valve was 0%-30% only.
This video illustrates how much integrity and objectivity Tom has. Thanks Tom
Woow, lots of analysis. Great Job. At one point, I was a bit lost, you needed to put a lot in the explanations. But at the end of the trending is really clear and your conclusions seem really interesting. Thank you Thomas. Have a nice day.
Was looking to "Shrink" my extruder footprint on my CR10 and then also include the Beacon Probe. I am heavily invested in Revo nozzles over all my printers so Roto looked interesting.
In my case, I don't need above 6mm/s so Roto would be a good fit. But.... Then came the price and this was unfortunately the deciding factor.
I am rather waiting for Orbiter 3 with the Revo integration to see what the pricing then looks like.
Once again, your videos and time taken into the details are very much appreciated.
For experiments with many variables like this, you can use fractional factorial design to cut down the number of tests, which is basically the "official" version of what you did with the graphs in excel
320 bucks? That's what patents do! And it should be called out.
As always, I appreciate your honesty and transparency. It's great to be able to see what these extruders are/aren't capable of.
Thomas, consider using dark backgrounds for the graphs :D
Love the chalk board, use it every video from now on.
The “issue” with the Roto is the 17mm nema14 stepper, the orbiter uses a 20mm wich can handle higher current
I wonder if they tested as well as you have here. Excellent work.
You are missing induction aka Henry, or for coils mH(enrys). It's a key to how sudden a stepper can move on a change, but also how much it can hold as a contra.
I would relay want to see a 48V test.
In a bygone era (3 years ago) such considerations (very high speed and flow) were a non-consideration. For example I have put a Sherpa Mini on my old QQS-Pro. It prints at 80mm/s and the moons NEMA14 with 5:1 gearing works fine, so yeah this Revo Roto would have been a lovely project for that machine. Yeah retraction speeds are limited, but it works. Some things I have also discovered over the years; always set your TMC current to the RMS value of the max Stepper current ( multiply by 0.707) as it will inavrably push less than this in general use, and I see no need for safety margins that people sometimes recommend; temperatures of the motors are fine and have been running for years now.
Also some major things about extruders; get your acceleration correct! Turn off Stealthchop and also turn off interpoleration; these settings alone will return your extruder to actually sounding like and extruder, if you're unsure. All these things alone can reduce accuracy of stepper movements; the TMC data sheets will actually tell you this, as well as some Klipper documentation out there. It's interesting as all this information is out there and written down in some shape or form, but not easy to pick out and collate in general (as in everything in electronics). As such well done for bringing some of this information with actual testing to the masses!
interesting that the revo motor is basically at Vmax the entire time, wasn't expecting that outcome, and you're right, quite surprising they didn't work with LDO on it, I wonder if Jason and Co will make a swap in LDO motor for the roto
is vmax volume, velocity or voltage?
They said they were working on a "different" motor version almost as soon as it was released because of the max temperature of the stock motor.
So maybe an opportunity there to work with a different manufacturer.
@@marc_frank voltage
The real shocker Here is that he actually got a roll of das Filament B-Grade.
The surprise boxes are gone in Like 3 Seconds.
I know a guy 😉
the stepper on the hermera is a 24v stepper therefor its internal resistance is low. on those other ones there are for much higher voltage so there internal resistance is lower.
if you want a stepper hat can push until gears breaking use lower voltage steppers and diver that can handle higher current. there is no problem over driving motors and if they get to hot, there are heat sink collars you can buy for it.
the thing is if you want a fast direct drive system you need a lighter head but most of the weight of the head is the motor, but if you want a lighter motor you need a smaller one. but the problem is if you make the motor smaller and want the power you have only 2 option, first is less internal resistance for more current but most stepper driver are limited in current and cant push any harder. therefore you need a stepper with higher voltage to lower the current that is required for the same power but your system uses 24v and you using a 48v stepper because it weight less you also have less power. this may no big deal for a ender 3 or other 50mm/s printer but machines that can 200mm/s this limits you.
bigger stepper are less a problem because the stepper driver sets the limit with its max current. for example you have normally a 24v stepper with 1A and your driver can 1,5A the stepper uses its 24v on full power but if you buy a 48v stepper with 1A it will use also the 24v but because of its internal resistance it will only use 0.5A. the Current is mainly the reason for the size of a motor. so bigger stepper can handle more current. but if you buy a 12v stepper with 1A and use it on a 24v system the motor uses now 24v and up to 2A but you are limited to 1.5A so in the end you overdrive the stepper a bit.
I would be quite interested to see a 48 or 60V setup, just to see if that improves performance on the Roto. You might be able to get a setup where you can pass the data lines from a 12 or 24 volt board to a stepper driver, but have the driver powered at the higher voltage from a separate power supply.
.9 stepper is the problem. Everybody thinks more precision, better. But this is an extruder we are talking about here, giving up half the power was a mistake.
Thanks for the detailed investigation and sharing your results.
Not only pushing force, but stealthchops causes so many headaches! I can only see the point on slow workhorse printers where it's used for skip detection and stuff like that
Thomas, I would like to commend your approach and digging in entirely for what appeared to be an obvious limit of flow in your previous video. Here you answer what the bottleneck is, and it's wonderful information for any printer enthousiast.
To me this exposes a product that seems to fall in the category "because we can".
I honestly don't like the response from E3D that they posted here, it does nothing to address the potential issues, and there are plenty:
- Stepper motor needs more than 24V when running fast due to back-EMF from the inductance aspects of a stepper motor (coils don't discharge instantaneously)
- Load on CPU is also way higher, as the Revo Roto extruder has 2682 steps per mm. This is such a big increase from a regular setup that this has to be investigated to find the implications when you are running Marlin on a relatively weak board (Arm M0+ for instance on a BTT SKR MINI E3 V3)
- Stepper motor should run really really hot. How hot? Nobody knows.
- Driver might also run hot because it has a lot of work to do. How hot? Nobody knows. Maybe you even need to turn off interpolation to alleviate some limits. But again, nobody knows.
These limits that could all have become relevant, and all for just being able to have very high torque from a very weak motor in a very light package. A light package which will only shine when....you don't go slow, but start cranking the speeds. This product missed the mark. Could have gone 1.8 instead of 0.9, without any downsides, could have stayed with a Nema17, but then go full pancake and still have a pretty lightweight extruder.
The Orbiter 3 now has a Revo edition. Also there's an addon tooboard, the SO3 toolboard that works great,(Extra $50).
I really don't like the direction E3D is going. Closed Revo ecosystem, very expensive, less performance without super expensive high flow nozzles and heaters, etc.
Personally I think that they have so much passion to tinker that they spend a lot of money on it.
As they spend a lot of money on it they need to get it back somewhere...
Is it sustainable? Only time will tell.
I’m not sure I agree.
The closed ecosystem is a bummer, but they are partnering with other manufacturers for products that otherwise wouldn’t exist (cht, biqu, sprite upgrade?).
The Revo hotends aren’t that expensive considering what they are and what they come with, though I do wish the nozzles were more affordable.
They are consistently putting out great products and standing behind their products when they make mistakes (Revo heater core).
Obxidian was a LONG wait, but you should check out the blog post about the process if you haven’t. It makes sense things can go wrong when pushing a new process into production at a mass scale.
When we compare the level of refinement we’re seeing in the current extruders, hotends, and nozzles to the original V6 era designs, I really can’t see how the value isn’t backing up the price.
As I said, I do wish the ecosystem was open, but based on my AliExpress suggestions, they aren’t going full Slice on those listings and there are tons of knock off nozzles and accessories available.
The REVO idea of uniting the nozzle and heartbreak is great at eliminating jams at the nozzle-heartbreak interface.
However their flow rates, even “high flow rate”, have lagged behind many competitors. It’s telling when the printers that push the limits do not use any E3D product.
Also the Obxidian fiasco was terrible. Marketing getting ahead of production by 2-3 years.
Tbh, not convinced by the Roto.
However I do love the Revo system and do wish Prusa used them on the MK4 but also understand Prusas want for backwards compatibility
The concept of REVO is nice but very expensive. I even invested in a diamond nozzle.... Not worth the money if you consider mediocre performance on a "normal" brass nozzle level. 😂
Wait, doesn’t LDO recommend to disable stealthchop for the orbiter?
Great tests! I hope you're sending all the filament to Recycling Fabrik, so it get's a second life 🙂
Hot DAMN this is a great video, Tom.
This would be an excellent use case for multivariate testing.
It seems like were getting at the borders of E3ds abilities to design, source, and spec out to manufacturers smaller powerful systems. It may not be a physics limit, but an organizational one and or systemic in the ability to find cheaply producable components to keep driving size and weight down. A lot (most?) 3d componentry isnt developed as consumer 3d printing technology initially,. We're getting refinements of tech and designs that found other uses.
For example an old M3D printer I had used motors commonly used to move the louvres on mini-split AC units for their geared extruder steppers. The more bespoke the solution is required to get these specialized designs going the higher the cost and harder it is going to be to find a manufacturer willing to make the relatively small runs that E3d can drum up.
E3D seem to be pushed towards making more high-end parts that offer specific benefits. The mid- and low-end market is quite saturated with products that all easily pass the "good enough" threshold and are often produced with maximum cost optimization in design, materials and labor. E3D, being a UK company, simply can't compete there.
@@MadeWithLayers So then how can they explain the smartorbiter v3.0? It's 1/3 the price, and the only downside is that is doesn't use Revo, which might not even be a downside. E3D has charged a premium for their products, and it has been worth it, but they need to take a look at what they are offering for the money they are asking.
Would be interesting to see how Induction heating can influence flow rate compared to various resistive heaters. Plasmics offers Induction heaters at 24 and 48 volts.
this is extremely well done
A screening test would have likely helped you focus on fewer variables.
DoE for the win.
Do you publish your raw data?
There’s a Revo version of the smart orbiter V3 now as well even less reason to get the Roto
Wow. Im so glad you found that out not me. I had a tatty E3D titan Aero replica on my baby Kingroon and was looking at the Roto as its lightweight and fast but now i see its not, As luck would have it i picked up a Creality Efit extruder instead (for £30 brand new) Nothing flashy but imho, a seriously great bit of kit (considering its make) Ive not had any problems pushing it hard for 48hrs while printing the super long bedslinger parts im building for one piece RC plane wings. The quality the little KP3S is banging out now has also shocked me going on the amount of bad rap the thing gets elsewhere. Could you test one of these Efit extruders? See how it compairs the the Hemera as the New bedslingers going to be needing an extruder soon.
I really wanna make a high voltage extrusion system now. I prefer having excessive power and tuning down than knowing there's a limit. Plus having a 60 volt system is just hilarious
A bit of work making sure insulation is up to scratch. You can get away with a lot running 24 V, but 60 V is going to find ways to evade the paths you want unless you are more careful.
Use resettable fuses.
PLEASE PLEASE PLEASE. do a full review of the LTO smart orbiter 3. I believe it has inbuilt tool board with accelerometers and USB connectivity? Are the hardened gears standard or an optional extra. What is the hot end max temp???? Please do a review. 🎉
Well, this is tempting to invert motors on the LDO and the E3D...
great video
You didn't even finish it
@@teitgenengineering I saw it on the front page when it was 6 seconds old and thought this would be the funniest thing to post
@@jessestridge5708 why?
@@teitgenengineering primarily because of this conversation
@@jessestridge5708 😂
Flow rate tests aren't a good way to test extruders, instead they are good for testing nozzles and heater blocks.
The flaw is that higher force extruders may not really achieve a higher flow compared to what the Revo hot sides are actually rated for -- I think its more likely the extruders are just forcing colder, harder plastic through the nozzle. The Revo nozzles are pretty well known at this point at producing specific volumetric maximums in different materials, which are well-documented on E3D's website. Those flow ratings are not a function of the extruder, but rather the material and how fast that material can be heated.
These days, pretty much every extruder can produce far more force on the filament and hotend than is necessary, especially at small nozzle sizes like the 0.6mm that was tested. This can be proven with intuition -- you only need light finger pressure to constantly advance filament into your standard 0.4mm hot setup! The *real* job of the extruder (outside of exotics / flexibles) is to advance and retract at very specific speeds/steps. The ultimate test data to compare extruders would be something like: 1) does it meet the target volumetric rating of the hot side in use? 2) measure precision 3) measure flexibles feeding 4) can it dial in proper amount of gear backlash? 5) long term gear-wear 6) motor temperatures 7) conditions where steps start to skip, such as current/temp/backpressure
TLDR: The nozzle and heater block are almost always the problem with flow limitations, not the extruder. Need different tests for extruder performance.
PSA: The orbiter is horrible to take apart! To access the filament path you need to take out all screws (which are all different lengths and have different diameters and you need two different allen keys), and once you have it taken apart, the planetary gears fall out as well as the ball bearing for the filament sensor. Reassembling is even more annoying...
Good analysis !
Love all of this
Interesting, so how does stealthchop compare on motion in 3d printers. Ive seen my printer skip with stealthchop on at 20k accel but never with it off. Would be interesting to see data on that too.
It might be interesting, but if this is a printer anywhere near a person, you wouldn't want to run with stealthchop disabled on the kinematic system motors. If you want higher acceleration, get motors that can support those accelerations at the speeds you are moving.
Great video. Thank you. Did you check, if it would be possible to change the motor?
The motor is not user replaceable.
Would have been nice to see how the Revo roto would perform with the orbiter v3 motor...
Using higher voltage and lower current means the copper windings can be lighter (size of copper is dictated by current). This makes for a lighter motor
Not quite - you also need more turns on the stator when using less current to achieve the same magnetic flux, so the material savings effectively get cancelled out.
The magnetic field strength(magnetic force) is directly proportional to the current but the current is proportional the the square of the weight of the copper wire. @@MadeWithLayers
As a very quick test, you should measure the DC resistance of the steppers. The Revo Roto would probably have a higher reading. Of course an inductance measurement would but better but most people tend to have DC multi meters and the measurements could possibly add to the overall conclusion you have arrived at.
The resistance is actually listed online, at 7.5 Ohm for the Revo Roto (but no inductance, which I'd assume would be rather high as well): e3d-online.com/pages/help-centre-roto
@@MadeWithLayers do you have way to measure inductance of the stepper?
Would Taguchi arrays help narrow down the affecting variables with a lot fewer tests?
I'm not familiar with Taguchi methods, but I'm sure there are quite a few DoE optimizations that I could take. Still challenging to present such multi-dimensional data though.
What is the purpose of this? You should be limiting extrusion speed for volumetric flow rate which will happen way before you start skipping steps.
Tom, can you comment on how you (and perhaps Stephan) arrive at the extrusion forces you use? I know that if I applied 5 to 10 kilograms of force to a 1.75mm diameter PLA filament in a 0.4mm nozzle heated to 210C, I would tear through a spool of filament in a matter of minutes, if not seconds. Can you clarify, please?
You're pretty far off with the numbers you "know" - there's no way a hotend could even melt that amount of filament in "seconds". Look at the speeds these extruders are achieving, they're around 10-20mm/s, using a 0.6mm HF nozzle. Assuming the 12mm/s I found for the Revo Roto, a full 1kg spool of PLA would take almost 8 hours to extrude. Not an unrealistic time to output that amount of filament tbh.
Manches erinnert mich an ein altes Vidoe von mir zum Thema Stepper Driver 😊
I was going to be putting revo roto's on a 4 tool, toolchanger system. Looks like I will be using a Smartorbiter v3.0 now. I cannot justify spending triple the price for a package that is barely smaller and lighter, and performs worse. The only saving grace of the revo roto, is the revo hotside, but it's not worth $200+ as well as the more expensive and more limited nozzle selection.
Actually a quantum tunneling extruder would easily be the next big thing in 3D printing
it's a shame they're using custom gears on the motor, if they used 10 or 9 teeth gear on it, we could change it for a different motor
@MadeWithLayers with all that poop, maybe it’s time for a diy filament recycling build?
I stick with my hemera until they fix this. Excellent info for sure.
Well if the time it took for Hemera XS is any metric (which was obvious need from even the days of Hermes), you'll have to wait a long time for fixed Roto
Чувак спасибо за тесты )
thanks i was about to buy the roto but given the number you gave i will be limited as i inteded to print at 140mm/sec with 0.6mm nozzle
Good afternoon,
Assuming standard flow geometry for a 0.6mm nozzle (so a layer height of 0.3mm and a track width of 0.72mm), 140mm/s xy speed would convert to 30.24mm3/s. Here’s a video of Roto doing just that (31mm3/s) with a 0.6mm Revo High Flow nozzle - th-cam.com/users/shortsTvgeMpdz6JM.
Thanks,
Team E3D
Did you ever do any followup reviews on this?
no follow-up but a previous video th-cam.com/video/X5UpN0QaxGY/w-d-xo.html
Could it be bad quality wiring from board to motor? If wiring resistance is comparable to motor internal resistance it would explain
Awh heck hadn't realized analog devices aquired trinamic. Guess the analog tax is gonna apply to any future chips they release.
It was all the way at the end, but it's definitely one of the most important parts: they're using a 0.9 degree stepper. In testing, a 0.9 degree stepper has about HALF the torque of a 1.8 degree stepper which pretty much all other extruders use. A simple stepper swap could likely massively improve the performance of the Revo Roto.
I really like E3D and respect what they've done for the 3D printing space, but they don't seem to recognize that faster and high flow printing is becoming more and more the norm.
Afternoon,
The design philosophy behind Roto was to have all the extrusion force you need in as lightweight a package as possible. Changing to a larger or more powerful stepper motor wouldn’t enable you to print any faster. Roto can already achieve the maximum volumetric flow requirements for every nozzle in the Revo ecosystem. If you increase flow beyond the rate that a given nozzle-hotend combination can melt it, you’ll just be pushing harder and cold extruding.
The spike in force requirement in this situation is because the filament is not fully molten. Printing like this will lead to poor quality prints. If you were to put a bigger nozzle (with a higher maximum flow rate) onto the Roto you would be able to achieve higher extrusion speeds.
Thanks,
Team E3D
I like Bondtech's LGX Lite. It's Swedish... made by the Volvo folks.
swap the motors?
I love e3d!
next step is just to integrate the stepper motor as part of the extruder
Mixed feeling:
- Roto is ill-dimensioned for current 24V ecosystem standard
- Roto is "future proofed" for the more desirable 48V ecosystem which is (very) slowly being introduced on the market (I am personally already transitioning to 48V using the last duet control boards)
- by the time 48V ecosystem becomes a standard, the non-electrical side of the Roto might be outdated.
--> maybe E3D should propose a 24V version and a 48V version of the Roto. The 24V might be a bit heavier but engineering = compromising, isn't it.
Hey Raoul,
You don’t need to run 48V, 24V is more than sufficient to achieve the maximum volumetric flow requirements for every nozzle in the Revo ecosystem! Please see our pinned comment for some more information.
Thanks,
E3D
Interesting Test, but still I think that the Roto shines on other characteristics. Not everybody needs maximum extrusion torque. Being much lighter, with a center of gravity that is much more in the center are harder to quantify but more important for most users motion systems IMHO.
The weight is definitely an advantage over E3Ds earlier offerings, but the weight between it and the Smart Orbiter should be relatively comparable in a full setup. The Roto weighs 146 Grams + 8g Nozzle (listed) + 20g Heater Cartridge (estimation) = 174g. The Smart Orbiter v3 is listed at 175g for the whole system (as it includes everything). Center of Mass should also be mostly similar between the two.
Weight and Center of Mass are only important when you are printing at high accelerations. You don't gain anything from running high accelerations if the extruder can't keep up. Especially as Revo emphasizes the ability to quickly change to the best Nozzle Diameter for each print, maxing out at ~100mm/s with a .6mm high flow nozzle is a weird choice. It is pretty likely that someone running this on a high performance printer would regularly run against the limits of this extruder. Choosing a slightly bigger and heavier motor would have been a more sensible option.
@@scratchfisch6425 Thanks for the corrections, might be comparable in weight after all...
But I have to disagree with the latter. High wheight and bad center of gravity becomes a dealbreaker at high speeds, but it still affects each and every print significantly also at lower speeds. Having less energy to jerk around every corner was and always will be better.
Lower weight allows you to print faster which this can’t do. So what’s the point?
@@rbaile508 Yes it does also that, but it also improves the quality at every speed slightly, and rreduces general wear an tear on every motion system obviously.
@@perchte I agree to a point. Running something like a Hemera at 388g will cause artifacts at much lower accelerations than a lighter extruder. Extremes are definitely bad. But basically all "current gen" extruders (BMG, Sherpa Mini, Orbiter, Revo Roto etc.) weigh less than 175g. At that weight, the remaining differences in weight/center of mass will not have any noticeable impact on print quality at reasonable accelerations (keep in mind that input shaping drastically increased maximum accelerations without noticeable ghosting)
Extruders need to have the right balance of weight and power, and the roto definitely misses the mark here.
Nice video :)
Can't you turn off stealthchop?
To me it sounds like someone miscalculated the winding count & wire gauge requirements for the Roto. From what you’ve presented, there are too many windings in the Roto. The higher the winding count, the higher the voltage requirement. They should be using fewer windings and a larger gauge wire, which would then work with a 24v system.
I think it's more to do with step angle. 0.9Deg needs twice as many steps as 1.8Deg, which then require higher voltage to push the current
@@Papinak2 step angle has very little, if anything at all, to do with this. This was proven when he dropped the multiplier from 16x to 8x. The fact that you can’t hit Imax of the motor at 24v is the issue.
@@cnc-maker it has a lot to do with it - he reduced microstepping multiplier, which removes bottleneck on the MCU side, but it doesn't reduce frequency through motor winding, which is large contributor to inductive resistance - double the frequency reduces current to half. So motor that needs half as many steps to complete full rotation would be able to run at roughly double the speed, everything else being equal.
@@Papinak2 we'll have to agree to disagree.
@@cnc-maker what do you disagree with? That more steps results in higher frequency and lower current?
I don't disagree that different winding could increase the speed, but at the same time, it decreases torque - it can be compensated by higher current thanks to thicker wire, but I'm not sure how well it scales.
1:28 Point at *Analog Devices*, since they bought Trinamic.
Yes, I was going to make that joke, but decided against it in the end. Trinamic are still sort of their own thing and the brand will most likely stick around for quite some time.
if those two extruder really have the same motor size and current is the only factor, you could swap motors to prove the point
They have their gears integral to the motor, not sure if they would match up. The motor from the Orbiter should improve things on the Roto, as it's not a 0.9° unit, but the Roto is still much more geared down.
@@MadeWithLayers seriously if the motor is geared down it means it does much more rews than it wouldn't do if it was runing as direct drive . so only reason to put 0.9 degree motor in such a thing is to make it more silent and that doesn't increase the torque but makes mcu work harder for no other benefits
if anything it would be much better to go with ~50-100 steps/rev motor to give it some torque and accuracy would be increased by the gear ratio anyway
the faster you want to print the lower the accuracy regardless of any motor specs because of fluid dynamics
We really are going to need FPGA-based motion and motor controls, and servo extruders, aren't we ...
So um I got a question what if I just slap on like a giant nema like 420 and just like print with that would that be enough force to like pick up the printer?!?! 😂
That one extruder cost more than three of my 3D printer..(it was on sale)
As an electronic engineer I'm completely confused. According to the documentation from E3D the resistance per phase (coil) is 7.5 ohms. At 24V you'd be passing 3.2 amps through each phase (short circuit). So to not exceed the recommended 0.6A at that resistance you'd only need 4.5v. I'm not sure what the whole speech you gave about needing a higher supply voltage to drive it better.
Happy to be corrected. Also happy to get out my oscilloscope and measure the voltage waveform across an extruder stepper coil.
Context of voltage is that driving the stepper motor. Almost seems like a stepper motor has mismatched specification for the application. 24V power supplies are most common.
Note this is just the extruder stepper motor, not the driver motors for controlling print head movement. It's like the extruder is targeted for a specific custom printer architecture.
@@AerialWaviator what!? I have no idea what the point you're trying to make is. It's just a stepper motor attached to an extruder. It has certain characteristics. Phase resistance and max current are just two of the dimensions that vary for stepper motors. Stepper motors have a max torque output based upon those characteristics. Supply voltage has little to do with it, unless the driver doesn't have enough voltage to drive the desired max current.
As a mechatronics engineer, I'm happy to help. At standstill, the static coil resistance is the only thing determining the voltage requirement, but as the motor spins up, it also starts producing significant back EMF en.wikipedia.org/wiki/Counter-electromotive_force, essentially acting as a generator working against the driver. This is where the relatively fixed RPM limit on Revo Roto seems to stem from, as at this point the generated voltage is starting to approach the supply voltage, essentially disabling the driver from pushing any current through it.
@@MadeWithLayers Clapping back to an "as an electronic engineer" with an "as a mechatronics engineer" is stone cold though.
Yay!
Now we know it's the extruder that is flawed. =)
BTW: 0.9° extruder motor? E3D: SERIOUSLY?!? O_O
You kept referring to “stealthchop drivers”, but stealthchop2 is just a mode/capability *of* the driver-it can be changed in firmware or on the board. It might be better to use a more modern mcu/firmware (RRF or Klipper) to power that rig so you can test the same drivers under different modes and switch back between them easily.
Stealthchop reduces peak torque and, in turn, may reduce positional accuracy (especially when combined with interpolation). Its purpose is to reduce noise during linear movement, so I can’t see any reason why an extruder should ever be set to Stealthchop mode. It is never moving at a noise-producing speed for any appreciable length of time, so all it’s doing is just reducing available torque.
Price wise is e3d taking itself out of business.. i never owned a product of them and with the looks of it i never will!
Are you not over complicating things, and simply running up to the limit of how fast the stepper can spin? Are they the same motor? Can they be swapped?
The planetary gear boxes make it hard to swap motors because they are made with an expected teeth count, LDO has many motors with different teeth counts because of this and other extruders. The current ones they carry are: 6, 8, 9 and 10. At least those are the ones I remember.
@@MatrixRay19 yep, 6 is custom for annex engineering and their sherpa extruders, 8 is standard, 9 is custom for galileo 2 and a longer spur gear than usual, 10 is again standard. All have a module of 2mm and the 8, 9 and 10 teeth steppers have press fit gears, the 6 teeth one has the gear integrated onto the shaft due to size constraints. There are additional variations, for example the stepper that ships with the lgx lite has a steel gear, so you can use it with the heavy duty lgx gears made entirely from steel while standard is a brass gear.
To answer your question at the very end of the video, I don't have a 3D printer.
Lol
Looks like a DOE. Nice
0.9 motor for a extruder is most useless thing i've ever heard...
Especially on an extruder with 10:1 gear ratio.
First thing I do on all my printers is to disable stealth chop. Klipper even recommends doing it.
So the roto is an engineering failure. The makers didn't consider all variables and didn't do enogh testing. This thing is flawed and pretty useless as a drop in replacement since 99% of all hobby 3D printers run 24V or 12V.
That's some Stepper Fu right there...