Thanks for all the input on the Highflow's wavyness! Couple points to note: - 100mm at 10mm/s makes for only a 10-second extrude move - way too quick for PID to affect anything, especially not at the frequency it's showing up. - The force ripples align well with the start of the extrude move, making it unlikely (but not impossible) that off-center extruder gears would be the culprit. It's pretty unlikely that the starting gear positions would be so similar between runs - Issue 602 is a potential culprit for the E3D Hemera, too, but the wavyness that creates is of much higher frequency, with about one wave per extruder gear tooth (0.5mm). I'm seeing one wave every 20mm or so - The load cell did see significant creep over time with everything slowly heating up (which is why in the automated test runs, it calibrates zero before every extrude move), but again, this is happening over a much longer time scale I think overall @MirageC's explanation seems the most plausible, with a heating / cooling cycle creating a resonant effect between filament flow and the (limited) elasticity in the coupling between the hotend and the extruder.
Like I stated above, I really enjoy these more technical videos. Any chance of doing a collaboration with Stefan? You two together would be amazing, the natural extension of the pod cast.
Some napkin math tells me it has something to do with the filament drive gear. the napkin math: I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz). The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps) So we have: 397 microsteps/mm 16 microsteps/step 200 steps/rotation 10 mm/s This means at 10mm/s, the stepper shaft (gearbox input) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!! since the Hemera has a 3.32:1 gear ratio, this means the output of the the gearbox (aka the hobbed gear) is spinning at 1.24/3.32 = 0.374 Hz. 0.374 x 2 = 0.748 which is prety close the the ripple frequency. This could be a coincidence, but it seems the period of the revolution of the filament drive gear corresponds to the period of the observed ripple. As for why it lines up. Perhaps the filament preloads the drive gear by pushing it against the shaft? idk
@@mayukh124 What about the teeth on drive gears a/ displacing some of the filament from an even volume when its undamaged to an uneven volume along its length? b/ causing the filament to speed up /slow down as the teeth dig in ..... like a multi flat sided wheel on a car would cause the speed to be gerky
@@doogssmee9742 my math was off as I had mixed up the input and output shaft. So in fact it seems there is a correlation between the filament drive gear and the ripple. Maybe the ripple lines up with every trial because the filament preloads the drive gear by pushing it against the shaft?
Thanks for taking the time to review Revo High Flow Tom. Apologies for the 404's, looks like you screen-grabbed whilst we were in the process of undergoing some web maintenance, these should now all be fixed. Team E3D
Oscillations from the PID heating control loop perhaps. Like you say, so many paramters could have been measured but where do you stop, But would be curious to log the heater drive and see if it correlates to the oscillations you see.
I had already seen all the videos on the new revo hf, but I knew your approach would be something different. Having tried testing nozzle combinations before, I really appreciate the time that went into this. Cheers!
❤ Awesome job Thomas! I love that machined automated rig! That being said, this is the kick i needed to speed up the return of my hotend olympics series! As for your observations with regards to the initial pressure creep before extrusion run, it is caused by a solidification of the filament at the junction of the heatsink and heat zone. And the pressure oscillation when reaching medium pressure levels during the run is due to the cold filament core obstructing the path, then melting and obstructing again and repeat. Enough talk... I am getting back to work! 😅
Some napkin math tells me the pressure oscillation has something to do with the stepper and/or it's connection to the gearbox. the napkin math: I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz). The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps) So we have: 397 microsteps/mm 16 microsteps/step 200 steps/rotation 10 mm/s This means at 10mm/s, the hobbed (output) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!! since the Hemera has a 3.32:1 gear ratio, this means the input the the gearbox (aka the stepper shaft) is spinning at 1.24/3.32 = 0.374 Hz. 0.374 x 2 = 0.748 which is prety close the the ripple frequency. This could be a coincidence, but it seems the period of the revolution of the stepper shaft corresponds to the period of the observed ripple.
@Mayukh Nath Napkins should be kept for what they are intended to ;) 6 periods in 80mm @ 10mm/s is like saying 6 occurrences over 8 seconds. This means 1.33 Hz. And the Gear Ratio is already factored in the 397 steps per mm @ 16 micro-steps. ;)
@MirageC Good point I should leave my napkins math-free haha. Are you sure the 6 periods in 8 seconds is 1.33 Hz? Hz is /s. Anyways, as for the 397 steps/mm from E3D, I completely agree that it factors in the gear ratio. I was wrong and mixed up the input and output shaft. However, that means the @10mm/s the stepper shaft is spinning at 1.24 Hz. And that means the output of the gearbox (aka the filament drive gear is in fact spinning at 0.374 Hz. So, I was wrong about any correlation between the stepper motor shaft and the ripple; I should have said the filament drive gear and ripple appear to have a correlation .
You should try see if the wavey-ness correlates with the circumference of one of the drive gears; Could indicate an extruder issue. Also, I'd guess this is more prone to clogs.
Clogs happen because heat travel up the tube into feeder area, and cheap filament that expand a lot under heat. I had lots of clogs with Creality Spider ver. 1, upgrade to high speed ver 3, no more clog on same cheap filament at same temperatures and print speeds.
@@Soravia Woodfill filament or filaments with fibres in it are more likely to clog, even without heat creep. Smaller channels are more likely to get blocked. Hence almost no one prints woodfill with a 0.2 nozzle.
It may have been mentioned before but steppers do not have constant torque, it will vary with position, especially if the driver is micro-stepping. That would explain the 'waviness in the force'. Basically, the stepper can push harder when the angle between the magnetic poles is ideal. If this is the case, the wave frequency would correlate to (a factor of?) the number of steps per revolution, which is usually 200.
I have a volcano style bondtech CHT .6mm Nozzle. If i push it really hard, especially with high speed PLAs it does make the same sounds extremely wet filament makes. Sounding almost like a drop of water in an oiled pan. It only does that at the beginning of a high flow extrusion move and only at the beginning of the print. I always thought it was some PID thing, where the sudden in rush of Filament absorbs a lot of energy, the heater kicks in and overshoots a little and that causes the filament to crackle. Never thought about till you mentioned high flow noises :)
The problem I have with revo again and again is cost. You buy one RevoHF vs a rapido and a CHT and it’s similar but the second you add a second or even a replacement nozzle it becomes a bloodbath. Given the entire point of Revo is supposed to be changing nozzles, why is everyone not giving this part much attention? It’s a huge premium for a tiny time saver which isn’t really all that much better in any way.
I totally understand companies wanting to make things proprietary, open is ofcourse great for the consumer and having more choice. But, having more choice, ensures E3D to make less money. And in the end, that's what keeps it alive. I had a colleague who developed a fancy camera, but closed it incredibly well as China would immediately jump on it and make it for 10x less. In this day and age, you can't compete almost and my prediction is, things are going to be more and more closed source. I wouldn't mind to buy E3D nozzles, but only if the support is there and good, because that is what I would be paying extra for. ps. love the in depth analysis!
I think people skip over a really big benefit to the revo system that not that many other companies are doing. It eliminates a connection point in the heat zone. I don't know how many times I've had issues from filament either finding its way between the nozzle and heat break or residual filament getting stuck in the crack when I change to another filament type causing other issues. I could see the revo increasing reliability.
You should see if the wavy graph correlates with the circumference of the extrusion gear (or specifically half of it). If those teeth are eccentric by as little as 0.01-0.1mm for whatever reason, the extrusion speed will be wavy like this. Was explored rather deeply by @MirageC in the video "will this new extruder gear solve the issue".
Really interested to see whether this load cell tech gets better integrated into printers, (I know the MK4 has it, and its really the only "complete" integration I know of at this point, in a consumer machine). In addition to bed probing, I can see automated flow rate calibration, as well as automated pressure advance calibration coming, and that is really interesting, because flow rate has been traditionally something that was mostly guessed at by hobbyists. I know right now I mostly just pick a number in prusaslicer for the filament profile that seems okay, and I know that other people dont even do that. It would be especially interesting to see things like layer adhesion numbers graphed by flow rate for a certain setup. It would be orders of magnitude more testing, but layer adhesion by flow rate by fan speed would be really cool. I dont think I've seen someone directly measure how "melted" the filament is either. For example on a 1mm plain nozzle, how solid is the core of the extrudate at various flow rates? A CHT nozzle? How does that affect things like shrinkage and curling? It would be very interesting to be able to optimize for the "most perfect" print, instead of just the fastest print, especially since the fastest print is most likely complete unusable garbage.
Load cells... hah! I worked in the plastics equipment industry 30 years ago... we used load cells for "gravimetric blending" - that is, we weighed a hopper to calibrate the flow of pellets against measured rpm of the feed augers. Load cells are great until they fail, of course... in our case, IIRC, they were aluminum blocks with a sensor wire providing the variable resistance to the ADC of our control boards. They always required some sort of calibration, but I haven't looked at what type of load cells they are using for auto-bed tramming and other 3D printing applications.
@@seljd but does the firmware actually get the load cell value? And is extrusion force part of that? Or is it just a glorified switch, and the load cell is mounted in such a way that it can only measure bed to nozzle force? Thats what I meant by complete integration.
@@seljd the vyper is a strain gauge, not a load cell (which are similar in what they do but 2 different things to my knowledge, happy to be proven wrong though)
When I saw the loadcell on the MK4 I thought the same thing.. I also hope they include filament jam and strip detection, both should be possible too, but after that, yeah.. A calibration routine to work out pressure advance would be a really impressive feature to see as, unlike input shaping(That they can calibrate from the factory), that can change with fulfillment.
The Revo system, while neat in concept, is just not for me at this time. I'm so invested in the V6 system with my various nozzles (including a couple CHT nozzles) I'm just not tempted. For those invested in Revo, it's nice that they have a higher-flow nozzle. Super cool testing setup, by the way.
Awesome video! Major props to putting in the obvious effort here and showing the first real results. Though I think you missed an opportunity to create a nice little #short - "Unsolved Mysteries of the Revo HF! Why is this pulsing???" From the thumbnail, the most interesting part was buried.
My Titan clone was pushing about 14mm/s3 (PLA 220C) thru my microswiss hotend when it started deflecting the mount and making me nervous. That rhythm seems to show itself on the surface of my prints, but it's more likely a stepper kinematics issue since the spacing is about 2mm. Nice work analyzing that mountain of data.
I had my fair share with high flow hotends, namely volcano (ptfe as well as steel heatbreak), rapido hf, rapido uhf as well as a dragon hf. My experience was that with a higher maximum flow you get less flow artifacts like i call them. A typical flow artifact would be the transition between high and low flow sections like the windows of a benchy. As soon as the bow ends and only windows are printed you can see a visual difference on the surface. There is also a clear and very linear corelation between melt zone length and melting performance, around 0,6 to 0,7 mm3 per mm of melt zone length, number vary depending on material used and structure of the hotend, mostly where the thermistor sits in corelation to nozzle and heater. Sadly i haven't had a chance to try a cht nozzle, but may try the bimetallic cht one as im not wanting to switch nozzles for the one or two glass fiber or carbon fiber prints a month and steel nozzles are just a bad compromise in this case. Back to the high flow hotends. I dont see more stringing compared to standard flow ones, maybe its luck with retraction settings or so, but i use the same values regardless what filament i use. Personal favourite is of course the rapido. The low thermal mass in combination with high power heater and thermistor sitting very close to heater and filament make for a very quick reacting system, basically negating the need for a good pid tune. The only fluctuations mainsail shows is background noise since its the same variation as when the hotend is cold. The hotend i like the least is surprisingly the dragon hf. Not for clogging, totally fine even at 30 degrees ambient with 30mm fan, but rather for the position of the heatsink and melt zone extender. On a mini stealthburner it used 70% of the 40w heater just to stay at a constant temperature, no fans or flow involved. Thats because the hotend fan blowing onto the extender. Without fan its down to 30%. Designed a shroud to block airflow at the lower portion.
Hypothesis. Your chats show a 100mm extubation isn't enough to desaturte the heat in the nozzle as the force is still increasing at a steady rate at the 100mm mark. I'd like so see the heater power logs ploted against the force. This would also show if the heater is responsible for the waves if frequency matches albeit out of phase. This could be a method to get a better heater pid tune. I'd also like to have seen some sample weighing to calibrate maximum acceptable extinction force. Keep up the good work.
The slow pulses in the flow "pressure" may be due to the temperature fluctuating as it cannot keep up with the amount of heat transferring into the plastic. It could be the PID loop in the temp controller not being tuned for a high heat transfer and it falls behind, then catches up as the error grows, causing the plastic viscosity to fluctuate. Maybe monitoring temp during the test would catch that. Cheers,
So do the waves coincide with the rotation angle of the extruder's stepper motor? i.e.: is it something related to the spindle angle such as a non-circular gear changing the pushing force as its radius slightly changes?
I could not wait for this test. Then I'd got inpatient and bought one and tested it. Sorry! But now I saw it. Great job! ...as usually. I found only one down side of the HF nozzle. I printed (vase mode) with a very viscous rubber like flex material with a 1.2mm HF nozzle (no regular option for this size) and there was constant under-extrusion and also delamination at some places. Though the 0.8 mm regular nozzle had absolutely no issue with the same material at the same flow rate and temperature.
It seems like we need temperatur avance. If you have a heater block with a small thermal mass, you can quickly turn the temperature up, just before high extrusion.
Its possible the wavy data could be the current feedback loop between the stepper driver and the stepper it self. Bouncing off the upper limit of its setting, as it regulates the current to the motor. That loop could also doing a dance with the heater / temp sensor loop trying to maintain temp as so much heat is pulled away with all that flow.. An oscilloscope would come in handy to graph the power signatures to see their phase relation to the wavy data your seeing.. Graphing the heater "On Time" vs wavy oscillation would allow you to see phase relation with your setup as is.
you need a belt bed under that to get rid of the extruded material as it might influence the results when and if the material build up high enough to push against the nozzle
Maybe that rhythmic pattern is moisture in your filament? The moisture is creating a bubble out of steam, the bubble is collapsing and causing a consistent pattern. If you think about it the bubbles should be forming at the same spot within the hotend at the same flow rate. So it would be easy to image it would be rhythmic like that. Just a theory. If you agree, or disagree, please share your thoughts.
I'm wondering if the waviness is that spool rocking on that holder. I've found that holding filament in a way that allows it to rock as it's pulled has some surprising impacts on print quality.
+1 Back when I got an Anet A8 and started doing some mods, I was surprised how upgrading the spool holder (the stock holder was just a acrylic frame and a metal rod) improved the printing quality! I was really not expecting that mod to do anything other than making the holder look a bit neater
I wonder if that Revo design has them locked into 0.8 being the highest they can go? I have great experiences with the Slice Magnum+ using 1.0, 1.2 and 1.8mm nozzles. I like E3D but they don't seem to support anything bigger like Slice does.
The oscillating(waviness) you see is eccentric on the extruder gear. It is the way it is mounted to the shaft. If there is just a slight but of play the grubscrew sets the gear off center
The bouncing of your data that you mentioned is 99% likely to be the extruder. The gear may not be 100% concentric or the gear meshing might not be at its ideal lash so the force transfer vector probably isn't perfectly in line. The extruder is the cause of a lot of artifacts.
The pulsing *might* be an artifact from the extruder gear - the gears are not perfectly circular and dig into the filament to some degree, resulting in the filament to be not only slightly mis-formed but also coming in with a tiny speed differential: the effective shape of the gears is a pair of n-gons pushing, and as a result, there is a slight speed differential.
My revo micro (from time of preorder) with a full set of nozzles got only a few minutes installed on my voron. Swapped back to a high flow hotend with cht-clone nozzle. Was not the best Investment, but now i want to try revo high flow just the price is holding me back.
I imagine once the parents age, someone will (if not e3d before this happens) will make a revo nozzle that you can thread a standard nozzle into. So a cartridge for your favorite nozzle. Want a revo with a tungsten carbide nozzle? There ya go. Ruby? Got it. Cht? Go get one. If that happened, e3d could still sell the cartridges
I need to upgrade to a high flow brain to keep up with this. Can you please test the effects of the newer high flow filaments, as this is something we can all use immediately.
Some napkin math tells me it has something to do with the filament drive gear. the napkin math: I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz). The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps) So we have: 397 microsteps/mm 16 microsteps/step 200 steps/rotation 10 mm/s This means at 10mm/s, the stepper shaft (gearbox input) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!! since the Hemera has a 3.32:1 gear ratio, this means the output of the the gearbox (aka the hobbed gear) is spinning at 1.24/3.32 = 0.374 Hz. 0.374 x 2 = 0.748 which is prety close the the ripple frequency. This could be a coincidence, but it seems the period of the revolution of the filament drive gear corresponds to the period of the observed ripple. As for why it lines up. Perhaps the filament preloads the drive gear by pushing it against the shaft? idk
I'm curious if there is any mixing inside both the Revo HF and CHT nozzles. Might be interesting to see how those multi-color filaments (MatterHackers Quantum and similar) come out. Distinct colors or some color mixing?
I think Stefan did some testing on this once, iirc since you're only getting nice laminar flow inside the CHT geometry, coextruded filament do not mix in these nozzles.
I with respect to maximum pushing force that the extruder can exert before it starts damaging the filament: I'm curious how Prusa's Nextruder stacks up. Since it has two idlers and a much longer contact area.
@16:20 I believe what we're seeing is a thermal bounce caused by the smaller mass of the Revo design and the HF design pulling a lot of heat faster than the thermistor is close enough to read. As mentioned, more thermal mass can compensate for this. I've been thinking there could be productive thermal mapping that can be calculated for and preheat rather than react to thermal conditions. This would also fix inconsistent melting on complex parts. Which is a problem i had to deal with last night when printing an ITX case. Larger segments showed under heating and smaller sections showed proper melting, but i didn't want to up to temps more as it would cause malformation on the smaller parts. Sure!! I could print at 120mms like a noob in 2016, and it'll look great!! But i like my prints done in the same day 😢 *I'm using a 40 watt heater in a volcano hotend with hybrid copper spacer and hardened 0.5mm nozzle. Copper anti-seize on the spacer and nozzle 👌 would be IMO best to make a CHT spacer for volcano, so you can use any V6 nozzle 🥰 *Side note, I'm thinking of making my own but not sure about the best way. Might end up having to chop one up and thread it all the way to the tip where I would cut it down till the opening has the 1.75mm to mate to a V6
One problem not mentioned with this type of design is in keeping the flow at a fairly even amount through the different parts of the nozzle. I could imagine some foreign matter partially blocking one internal hole and this restriction causing the local flow to diminish while increasing rate at the other holes. The temperature then would also be hotter at the semi-blocked one. Just curious, is the revo shaft slightly tapered into the heater core to cause a snug fit?
High Thomas, I'm a great fan of your channel, congrats for an amazing content! I have a question around the high-flow nozzle which I've emailed E3D with no response. How does it interact with the improvement of velocity given by the input shaping technology? I have a Prusa MK3S+ which I'm going to upgrade to 3.5 (which provides it with Klipper and input shaping capabilities). My questions are: a) Does the increase in speed given by input shaping add up to the increase in speed of a high-flow nozzle (v.s the regular), or do the speeds sort of cancel each other out? I'm going to buy some Oxbidian nozzles to be able to print abrasives, and I'm unsure of buying the regular steel Oxbidian or the high-flow version Oxbidian, as there isn't much difference in price. b) What are the downsides of the high-flow nozzle? I understand it's more difficult to clean with an acupunture needle if there are cloggs, given the shape of the bores, anything else that comes to mind? Many thanks! Pedro
Some more investigations regarding the initial spike would be nice. Especially how repeatable it is. E.g. does it only exist after long times of no extrusion or could it lead to extrusion inconsistencies if it also occurs during quick changes in feed rates in general (which can happen a lot during printing). This is especially interesting because it the spike is much more pronounced with the new nozzles compared to the old V6.
Appreciate the detailed analysis. Was at RMRRF for 2 days, literally opposite the nice E3D folks and had no idea that they had these on display. Lesson learned, say hello and be curious.
cant believe I had not subscribed in the past, have watched lots and lots of your very well thought out tests, tips, trials and tribulations. Again you have not disappointed. I guess the whole point of this type of nozzle is ... speed of print? My Burning question is.... what about quality of print at the high flow and print speeds. If the quality and accuracy of the end product degrades...... what is the point?
That spike at the start might be a airgap or stretch between the cold and molten filament. Your fluiddynamics-wavyness seems to have the same frequency as the loops when extruding, can you try one without a floor the filament can stop against?
I dont know if i agree with your distinction between the Volcano and the CHT style nozzles. They both are just increading the surface area the material is exposed to. Yes, a Volcano is going to have more of a buffer but the flow rates have been tested.
The core difference is that CHT ("core heating technology") heats the filament from the inside out as well, while for a regular nozzle the biggest thermal resistance will always be the filament itself as the heat tries to make it to the center of the filament.
Just a question as I go through boxes after a move. Can the original Anet A8 be brought back to life in 2023? Of course there would need to be some updates that can be done to strengthen the frame and other new features available now that were not out back then. Just a thought.....
I wonder if it's possible in the HF if the waviness in The Force is something to do with the filament sucking heat out of the hotend? The comparisons between the HF and the Volcano are probably the most interesting, no?
I’m not any genius but for the waviness in the force, could it be related to the pull out drag of the filament. Would be interesting to see what the reading Returned could be with just a meter of unrestrained filament. ***Just a wild thought ***Cheers!
I have two questions that apply to both this nozzle and the volcano. What is retraction like? It seems to me that if you flow all that plastic around something inside the nozzle, that when you pull back a little, it won't be as effective as a straight bore nozzle. Is there any way to clean the inside of these nozzles? On a straight bore nozzle, you can do a cold pull. But if the filament is melted around the inside of the nozzle, you won't be able to pull it out.
Thanks for the test. I am very much interested in the sample rate used on the load cell and if you checked for variance in sample rate timing as the the data seems kind of coarse for some parts of further data analysis?
don't you think, the drop at the beginning is caused by a drop in temperature due to the change in push of colder material? Just until the PID compensates.
wavyness in the data points seems familiar to me. are you by any chance powering hot end and extruder motor from the same power suply? in my homemade printer i noticed that heater was clamping power rails and stepper motor had momentary less voltage to work with and that produced rythmical artefacts on my prints that only showed up at higher speeds. but they didn't show up when i was building the printer , so i tracked it down to single power suply, adding separate power suply for the heater made motors work smoother and wavy artefacts disapeared
don't you think, the waviness is also a function of PID loop actively heating/not heating? hotter and colder temperature probably even more for less perfect PID values.
looks like a hx711 loadcell, wondering if you use it at 10 or 80hz. But if you want more details in the reading (samples per second), try Adafruit NAU7802. It can read up to 320hz (think the loadcell reader in prusa new extruder use a hx717 with 320hz, but could not find any simple board with it). Maybe you could them see the waves also in more detail. Or the start ramp up, what maybe only caused by the static friction of the filament against the walls. Also wondering if the v6 force was against a full metal, or some with PTFE tube, could also explain maybe the overall less pressure for the V6, with less metal in contact to the filament.
I'm reading the HX711 at 10Hz, which I think gives me enough data to even see some of the intricacies. Extrusion is still a pretty messy process, so I'm not sure if more temporal resolution would make for more information 😅 Hotends were all regular all-metal ones.
Could the "waviness" of the extrusion force be due to the spur gears on the Hemera being slightly eccentric instead of concentric on the shafts? What I mean by that is that the set screws (or grub screws I think is the non-American term) pushes the gear tangentially against the shaft, causing the gears to be slightly eccentric. I have seen that exact situation happen on a robotic system I've worked on. There was a tolerance stackup for the ID on the spur gear and the OD of its mating shaft that eccentuated the problem to the point that it would bind every 360 degrees in the location of the set screws. Edit: it looks like other people have basically said the same thing I've said.
I don't have an issue with patents as long as they aren't used to lock away a system. There is no reason E3D could have a reasonable cost license to allow for 3rd party nozzles.
I'd be surprised if they weren't already utilising this. It's common practice for interference-fit parts, using the thermal expansion of those parts to get such a fit as you describe.
No disrespect to you Tom but you need to design a printer for the market for sale your good in design and making a good product .I'd buy one with no doubt that it would be a great product
Thanks for all the input on the Highflow's wavyness! Couple points to note:
- 100mm at 10mm/s makes for only a 10-second extrude move - way too quick for PID to affect anything, especially not at the frequency it's showing up.
- The force ripples align well with the start of the extrude move, making it unlikely (but not impossible) that off-center extruder gears would be the culprit. It's pretty unlikely that the starting gear positions would be so similar between runs
- Issue 602 is a potential culprit for the E3D Hemera, too, but the wavyness that creates is of much higher frequency, with about one wave per extruder gear tooth (0.5mm). I'm seeing one wave every 20mm or so
- The load cell did see significant creep over time with everything slowly heating up (which is why in the automated test runs, it calibrates zero before every extrude move), but again, this is happening over a much longer time scale
I think overall @MirageC's explanation seems the most plausible, with a heating / cooling cycle creating a resonant effect between filament flow and the (limited) elasticity in the coupling between the hotend and the extruder.
Like I stated above, I really enjoy these more technical videos. Any chance of doing a collaboration with Stefan? You two together would be amazing, the natural extension of the pod cast.
Some napkin math tells me it has something to do with the filament drive gear.
the napkin math:
I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz).
The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps)
So we have:
397 microsteps/mm
16 microsteps/step
200 steps/rotation
10 mm/s
This means at 10mm/s, the stepper shaft (gearbox input) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s
This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!!
since the Hemera has a 3.32:1 gear ratio, this means the output of the the gearbox (aka the hobbed gear) is spinning at 1.24/3.32 = 0.374 Hz.
0.374 x 2 = 0.748 which is prety close the the ripple frequency.
This could be a coincidence, but it seems the period of the revolution of the filament drive gear corresponds to the period of the observed ripple.
As for why it lines up. Perhaps the filament preloads the drive gear by pushing it against the shaft? idk
@@mayukh124 What about the teeth on drive gears
a/ displacing some of the filament from an even volume when its undamaged to an uneven volume along its length?
b/ causing the filament to speed up /slow down as the teeth dig in ..... like a multi flat sided wheel on a car would cause the speed to be gerky
@@doogssmee9742 my math was off as I had mixed up the input and output shaft. So in fact it seems there is a correlation between the filament drive gear and the ripple.
Maybe the ripple lines up with every trial because the filament preloads the drive gear by pushing it against the shaft?
periodic thing at 17:00 could be the hotend PID, as you push filament it cools the extruder and the PID increases the duty cycle.
Thanks for taking the time to review Revo High Flow Tom. Apologies for the 404's, looks like you screen-grabbed whilst we were in the process of undergoing some web maintenance, these should now all be fixed. Team E3D
so when is Revo Volcano coming out?
Any plans for a Revo high flow obsidian nozzle for abrasives?
May i ask is there a 2.85 mil one also planned?
Oscillations from the PID heating control loop perhaps. Like you say, so many paramters could have been measured but where do you stop, But would be curious to log the heater drive and see if it correlates to the oscillations you see.
I'd like to see the 30w vs 60w, to see if that has an impact on the behavior.
Just came to say exactly that.
I had already seen all the videos on the new revo hf, but I knew your approach would be something different. Having tried testing nozzle combinations before, I really appreciate the time that went into this. Cheers!
❤ Awesome job Thomas! I love that machined automated rig! That being said, this is the kick i needed to speed up the return of my hotend olympics series!
As for your observations with regards to the initial pressure creep before extrusion run, it is caused by a solidification of the filament at the junction of the heatsink and heat zone.
And the pressure oscillation when reaching medium pressure levels during the run is due to the cold filament core obstructing the path, then melting and obstructing again and repeat.
Enough talk... I am getting back to work! 😅
I loved hotend olympics, can't wait for the next episode!
Some napkin math tells me the pressure oscillation has something to do with the stepper and/or it's connection to the gearbox.
the napkin math:
I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz).
The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps)
So we have:
397 microsteps/mm
16 microsteps/step
200 steps/rotation
10 mm/s
This means at 10mm/s, the hobbed (output) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s
This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!!
since the Hemera has a 3.32:1 gear ratio, this means the input the the gearbox (aka the stepper shaft) is spinning at 1.24/3.32 = 0.374 Hz.
0.374 x 2 = 0.748 which is prety close the the ripple frequency.
This could be a coincidence, but it seems the period of the revolution of the stepper shaft corresponds to the period of the observed ripple.
@Mayukh Nath Napkins should be kept for what they are intended to ;)
6 periods in 80mm @ 10mm/s is like saying 6 occurrences over 8 seconds. This means 1.33 Hz.
And the Gear Ratio is already factored in the 397 steps per mm @ 16 micro-steps.
;)
@MirageC
Good point I should leave my napkins math-free haha.
Are you sure the 6 periods in 8 seconds is 1.33 Hz? Hz is /s.
Anyways, as for the 397 steps/mm from E3D, I completely agree that it factors in the gear ratio.
I was wrong and mixed up the input and output shaft.
However, that means the @10mm/s the stepper shaft is spinning at 1.24 Hz. And that means the output of the gearbox (aka the filament drive gear is in fact spinning at 0.374 Hz.
So, I was wrong about any correlation between the stepper motor shaft and the ripple; I should have said the filament drive gear and ripple appear to have a correlation .
@@mayukh124 mouahah!!! so I guess your Napkins are better than my excel for the Hz calculation! You have the point for this one :)
You should try see if the wavey-ness correlates with the circumference of one of the drive gears; Could indicate an extruder issue. Also, I'd guess this is more prone to clogs.
Probably a result of the set screw on one of the gears.
Clogs happen because heat travel up the tube into feeder area, and cheap filament that expand a lot under heat. I had lots of clogs with Creality Spider ver. 1, upgrade to high speed ver 3, no more clog on same cheap filament at same temperatures and print speeds.
@@Soravia Woodfill filament or filaments with fibres in it are more likely to clog, even without heat creep. Smaller channels are more likely to get blocked. Hence almost no one prints woodfill with a 0.2 nozzle.
@@elminz that's a benefit of Revo. It's easy to swap to a 0.6mm or something for woodfill.
It may have been mentioned before but steppers do not have constant torque, it will vary with position, especially if the driver is micro-stepping. That would explain the 'waviness in the force'. Basically, the stepper can push harder when the angle between the magnetic poles is ideal. If this is the case, the wave frequency would correlate to (a factor of?) the number of steps per revolution, which is usually 200.
I have a volcano style bondtech CHT .6mm Nozzle. If i push it really hard, especially with high speed PLAs it does make the same sounds extremely wet filament makes. Sounding almost like a drop of water in an oiled pan. It only does that at the beginning of a high flow extrusion move and only at the beginning of the print. I always thought it was some PID thing, where the sudden in rush of Filament absorbs a lot of energy, the heater kicks in and overshoots a little and that causes the filament to crackle. Never thought about till you mentioned high flow noises :)
If you clog a nozzle like this your pretty screwed. For that amount of money, No way.
That's part of the game we all play printing.
The problem I have with revo again and again is cost. You buy one RevoHF vs a rapido and a CHT and it’s similar but the second you add a second or even a replacement nozzle it becomes a bloodbath. Given the entire point of Revo is supposed to be changing nozzles, why is everyone not giving this part much attention?
It’s a huge premium for a tiny time saver which isn’t really all that much better in any way.
I totally understand companies wanting to make things proprietary, open is ofcourse great for the consumer and having more choice. But, having more choice, ensures E3D to make less money. And in the end, that's what keeps it alive. I had a colleague who developed a fancy camera, but closed it incredibly well as China would immediately jump on it and make it for 10x less. In this day and age, you can't compete almost and my prediction is, things are going to be more and more closed source. I wouldn't mind to buy E3D nozzles, but only if the support is there and good, because that is what I would be paying extra for.
ps. love the in depth analysis!
Thank you for covering the high flow options. I’m just switching over to Revo and it really helped me decide which nozzles I want to start out with.
I think people skip over a really big benefit to the revo system that not that many other companies are doing. It eliminates a connection point in the heat zone. I don't know how many times I've had issues from filament either finding its way between the nozzle and heat break or residual filament getting stuck in the crack when I change to another filament type causing other issues. I could see the revo increasing reliability.
You should see if the wavy graph correlates with the circumference of the extrusion gear (or specifically half of it). If those teeth are eccentric by as little as 0.01-0.1mm for whatever reason, the extrusion speed will be wavy like this. Was explored rather deeply by @MirageC in the video "will this new extruder gear solve the issue".
Really interested to see whether this load cell tech gets better integrated into printers, (I know the MK4 has it, and its really the only "complete" integration I know of at this point, in a consumer machine). In addition to bed probing, I can see automated flow rate calibration, as well as automated pressure advance calibration coming, and that is really interesting, because flow rate has been traditionally something that was mostly guessed at by hobbyists. I know right now I mostly just pick a number in prusaslicer for the filament profile that seems okay, and I know that other people dont even do that.
It would be especially interesting to see things like layer adhesion numbers graphed by flow rate for a certain setup. It would be orders of magnitude more testing, but layer adhesion by flow rate by fan speed would be really cool. I dont think I've seen someone directly measure how "melted" the filament is either. For example on a 1mm plain nozzle, how solid is the core of the extrudate at various flow rates? A CHT nozzle? How does that affect things like shrinkage and curling? It would be very interesting to be able to optimize for the "most perfect" print, instead of just the fastest print, especially since the fastest print is most likely complete unusable garbage.
Load cells... hah! I worked in the plastics equipment industry 30 years ago... we used load cells for "gravimetric blending" - that is, we weighed a hopper to calibrate the flow of pellets against measured rpm of the feed augers. Load cells are great until they fail, of course... in our case, IIRC, they were aluminum blocks with a sensor wire providing the variable resistance to the ADC of our control boards. They always required some sort of calibration, but I haven't looked at what type of load cells they are using for auto-bed tramming and other 3D printing applications.
Creality CR6SE and Anycubic Vyper have them since 2020
@@seljd but does the firmware actually get the load cell value? And is extrusion force part of that? Or is it just a glorified switch, and the load cell is mounted in such a way that it can only measure bed to nozzle force?
Thats what I meant by complete integration.
@@seljd the vyper is a strain gauge, not a load cell (which are similar in what they do but 2 different things to my knowledge, happy to be proven wrong though)
When I saw the loadcell on the MK4 I thought the same thing.. I also hope they include filament jam and strip detection, both should be possible too, but after that, yeah.. A calibration routine to work out pressure advance would be a really impressive feature to see as, unlike input shaping(That they can calibrate from the factory), that can change with fulfillment.
The Revo system, while neat in concept, is just not for me at this time. I'm so invested in the V6 system with my various nozzles (including a couple CHT nozzles) I'm just not tempted.
For those invested in Revo, it's nice that they have a higher-flow nozzle.
Super cool testing setup, by the way.
Awesome video! Major props to putting in the obvious effort here and showing the first real results.
Though I think you missed an opportunity to create a nice little #short - "Unsolved Mysteries of the Revo HF! Why is this pulsing???"
From the thumbnail, the most interesting part was buried.
Actually, you could name it "I detected a Disturbance in the Force"!
How's that for honest clickbait? 🙂
My Titan clone was pushing about 14mm/s3 (PLA 220C) thru my microswiss hotend when it started deflecting the mount and making me nervous.
That rhythm seems to show itself on the surface of my prints, but it's more likely a stepper kinematics issue since the spacing is about 2mm.
Nice work analyzing that mountain of data.
5:41 there is one like that, it's a V6 compatible that uses a copper insert with 3 bores.
And no CNC kitchen it's not infringing on the CHT crap.
I had my fair share with high flow hotends, namely volcano (ptfe as well as steel heatbreak), rapido hf, rapido uhf as well as a dragon hf. My experience was that with a higher maximum flow you get less flow artifacts like i call them. A typical flow artifact would be the transition between high and low flow sections like the windows of a benchy. As soon as the bow ends and only windows are printed you can see a visual difference on the surface.
There is also a clear and very linear corelation between melt zone length and melting performance, around 0,6 to 0,7 mm3 per mm of melt zone length, number vary depending on material used and structure of the hotend, mostly where the thermistor sits in corelation to nozzle and heater. Sadly i haven't had a chance to try a cht nozzle, but may try the bimetallic cht one as im not wanting to switch nozzles for the one or two glass fiber or carbon fiber prints a month and steel nozzles are just a bad compromise in this case.
Back to the high flow hotends. I dont see more stringing compared to standard flow ones, maybe its luck with retraction settings or so, but i use the same values regardless what filament i use.
Personal favourite is of course the rapido. The low thermal mass in combination with high power heater and thermistor sitting very close to heater and filament make for a very quick reacting system, basically negating the need for a good pid tune. The only fluctuations mainsail shows is background noise since its the same variation as when the hotend is cold.
The hotend i like the least is surprisingly the dragon hf. Not for clogging, totally fine even at 30 degrees ambient with 30mm fan, but rather for the position of the heatsink and melt zone extender. On a mini stealthburner it used 70% of the 40w heater just to stay at a constant temperature, no fans or flow involved. Thats because the hotend fan blowing onto the extender. Without fan its down to 30%. Designed a shroud to block airflow at the lower portion.
Hypothesis. Your chats show a 100mm extubation isn't enough to desaturte the heat in the nozzle as the force is still increasing at a steady rate at the 100mm mark.
I'd like so see the heater power logs ploted against the force. This would also show if the heater is responsible for the waves if frequency matches albeit out of phase. This could be a method to get a better heater pid tune.
I'd also like to have seen some sample weighing to calibrate maximum acceptable extinction force.
Keep up the good work.
The slow pulses in the flow "pressure" may be due to the temperature fluctuating as it cannot keep up with the amount of heat transferring into the plastic. It could be the PID loop in the temp controller not being tuned for a high heat transfer and it falls behind, then catches up as the error grows, causing the plastic viscosity to fluctuate.
Maybe monitoring temp during the test would catch that.
Cheers,
Could also be slight misalignment from the dual gear extruder setup.
So do the waves coincide with the rotation angle of the extruder's stepper motor? i.e.: is it something related to the spindle angle such as a non-circular gear changing the pushing force as its radius slightly changes?
I'd say it's the heater kicking in
I could not wait for this test. Then I'd got inpatient and bought one and tested it. Sorry! But now I saw it. Great job! ...as usually.
I found only one down side of the HF nozzle. I printed (vase mode) with a very viscous rubber like flex material with a 1.2mm HF nozzle (no regular option for this size) and there was constant under-extrusion and also delamination at some places. Though the 0.8 mm regular nozzle had absolutely no issue with the same material at the same flow rate and temperature.
Still waiting on my obsidian nozzle, maybe these will come out before 2030
2035 at the least, with the next change of royal clown in Britain
They're already out. ObXidian too, in stock at E3D and other resellers in various sizes
Thomas showing many of his skills. I am always impressed. Not to mention the hawaii shirts. Thank you, GOOD JOB :)
It seems like we need temperatur avance. If you have a heater block with a small thermal mass, you can quickly turn the temperature up, just before high extrusion.
Its possible the wavy data could be the current feedback loop between the stepper driver and the stepper it self. Bouncing off the upper limit of its setting, as it regulates the current to the motor. That loop could also doing a dance with the heater / temp sensor loop trying to maintain temp as so much heat is pulled away with all that flow.. An oscilloscope would come in handy to graph the power signatures to see their phase relation to the wavy data your seeing.. Graphing the heater "On Time" vs wavy oscillation would allow you to see phase relation with your setup as is.
I sense a wavyness in the force...
you need a belt bed under that to get rid of the extruded material as it might influence the results when and if the material build up high enough to push against the nozzle
Maybe that rhythmic pattern is moisture in your filament?
The moisture is creating a bubble out of steam, the bubble is collapsing and causing a consistent pattern.
If you think about it the bubbles should be forming at the same spot within the hotend at the same flow rate. So it would be easy to image it would be rhythmic like that.
Just a theory. If you agree, or disagree, please share your thoughts.
I'm wondering if the waviness is that spool rocking on that holder. I've found that holding filament in a way that allows it to rock as it's pulled has some surprising impacts on print quality.
+1 Back when I got an Anet A8 and started doing some mods, I was surprised how upgrading the spool holder (the stock holder was just a acrylic frame and a metal rod) improved the printing quality! I was really not expecting that mod to do anything other than making the holder look a bit neater
This is exactly what I was thinking. it could cause both an increase and decrease in the load cell measurement, as the spool is pulled, then recoils.
I wonder if that Revo design has them locked into 0.8 being the highest they can go? I have great experiences with the Slice Magnum+ using 1.0, 1.2 and 1.8mm nozzles. I like E3D but they don't seem to support anything bigger like Slice does.
There's Revo nozzles up to 1.4mm...
The oscillating(waviness) you see is eccentric on the extruder gear. It is the way it is mounted to the shaft. If there is just a slight but of play the grubscrew sets the gear off center
The bouncing of your data that you mentioned is 99% likely to be the extruder. The gear may not be 100% concentric or the gear meshing might not be at its ideal lash so the force transfer vector probably isn't perfectly in line. The extruder is the cause of a lot of artifacts.
should be tested with something like Prusa nextruder to see if that's the problem
Wavyness could be torque variance from the stepper motor or from the gears and the contact radius.
The pulsing *might* be an artifact from the extruder gear - the gears are not perfectly circular and dig into the filament to some degree, resulting in the filament to be not only slightly mis-formed but also coming in with a tiny speed differential: the effective shape of the gears is a pair of n-gons pushing, and as a result, there is a slight speed differential.
Wavyness could be down to a lack of perfect concentricity in the extruder?
My revo micro (from time of preorder) with a full set of nozzles got only a few minutes installed on my voron. Swapped back to a high flow hotend with cht-clone nozzle. Was not the best Investment, but now i want to try revo high flow just the price is holding me back.
Really enjoyed this video, I’m looking forward to more data from your automated test setup!
THanks Tom for another great video. Love to see more with the test extruder rig.
I imagine once the parents age, someone will (if not e3d before this happens) will make a revo nozzle that you can thread a standard nozzle into. So a cartridge for your favorite nozzle. Want a revo with a tungsten carbide nozzle? There ya go. Ruby? Got it. Cht? Go get one. If that happened, e3d could still sell the cartridges
id bet thats the heater on off cycling.
great information
I need to upgrade to a high flow brain to keep up with this. Can you please test the effects of the newer high flow filaments, as this is something we can all use immediately.
Impressive rig
Like your approach with the loadcell.
Thanks for sharing your experience with All of us 👍😃
Some napkin math tells me it has something to do with the filament drive gear.
the napkin math:
I am estimating the ripple frequency to be about 0.75 Hz (6 periods in 80mm @ 10mm/s -> 0.75 Hz).
The recommended steps/mm value from E3D is 397 steps / mm (at 16x microsteps)
So we have:
397 microsteps/mm
16 microsteps/step
200 steps/rotation
10 mm/s
This means at 10mm/s, the stepper shaft (gearbox input) gear will spin at (397/16)/200 * 10 = 1.24 Hz ) or 1.24 rotations/s
This doesn't seem to correlate to the 0.75 Hz of the ripplel; HOWEVER KEEP READING!!!!!!!!!!
since the Hemera has a 3.32:1 gear ratio, this means the output of the the gearbox (aka the hobbed gear) is spinning at 1.24/3.32 = 0.374 Hz.
0.374 x 2 = 0.748 which is prety close the the ripple frequency.
This could be a coincidence, but it seems the period of the revolution of the filament drive gear corresponds to the period of the observed ripple.
As for why it lines up. Perhaps the filament preloads the drive gear by pushing it against the shaft? idk
I'm curious if there is any mixing inside both the Revo HF and CHT nozzles. Might be interesting to see how those multi-color filaments (MatterHackers Quantum and similar) come out. Distinct colors or some color mixing?
I think Stefan did some testing on this once, iirc since you're only getting nice laminar flow inside the CHT geometry, coextruded filament do not mix in these nozzles.
I with respect to maximum pushing force that the extruder can exert before it starts damaging the filament: I'm curious how Prusa's Nextruder stacks up. Since it has two idlers and a much longer contact area.
The wave in the force could be related to the PID on the temperature controller.
@16:20 I believe what we're seeing is a thermal bounce caused by the smaller mass of the Revo design and the HF design pulling a lot of heat faster than the thermistor is close enough to read.
As mentioned, more thermal mass can compensate for this.
I've been thinking there could be productive thermal mapping that can be calculated for and preheat rather than react to thermal conditions. This would also fix inconsistent melting on complex parts. Which is a problem i had to deal with last night when printing an ITX case. Larger segments showed under heating and smaller sections showed proper melting, but i didn't want to up to temps more as it would cause malformation on the smaller parts. Sure!! I could print at 120mms like a noob in 2016, and it'll look great!! But i like my prints done in the same day 😢
*I'm using a 40 watt heater in a volcano hotend with hybrid copper spacer and hardened 0.5mm nozzle. Copper anti-seize on the spacer and nozzle 👌 would be IMO best to make a CHT spacer for volcano, so you can use any V6 nozzle 🥰
*Side note, I'm thinking of making my own but not sure about the best way. Might end up having to chop one up and thread it all the way to the tip where I would cut it down till the opening has the 1.75mm to mate to a V6
Could the rhythmic flow changes be caused by the PID heating algorithm?
Like with many things, I probably won't be getting one of these, but it is interesting to see how they work
One problem not mentioned with this type of design is in keeping the flow at a fairly even amount through the different parts of the nozzle. I could imagine some foreign matter partially blocking one internal hole and this restriction causing the local flow to diminish while increasing rate at the other holes. The temperature then would also be hotter at the semi-blocked one.
Just curious, is the revo shaft slightly tapered into the heater core to cause a snug fit?
High Thomas, I'm a great fan of your channel, congrats for an amazing content! I have a question around the high-flow nozzle which I've emailed E3D with no response. How does it interact with the improvement of velocity given by the input shaping technology? I have a Prusa MK3S+ which I'm going to upgrade to 3.5 (which provides it with Klipper and input shaping capabilities). My questions are:
a) Does the increase in speed given by input shaping add up to the increase in speed of a high-flow nozzle (v.s the regular), or do the speeds sort of cancel each other out? I'm going to buy some Oxbidian nozzles to be able to print abrasives, and I'm unsure of buying the regular steel Oxbidian or the high-flow version Oxbidian, as there isn't much difference in price.
b) What are the downsides of the high-flow nozzle? I understand it's more difficult to clean with an acupunture needle if there are cloggs, given the shape of the bores, anything else that comes to mind?
Many thanks!
Pedro
Is the wavy effect maybe from the gears and/or the stepper motor where it's not consistent and exactly equal/even pressure over time?
Some more investigations regarding the initial spike would be nice. Especially how repeatable it is. E.g. does it only exist after long times of no extrusion or could it lead to extrusion inconsistencies if it also occurs during quick changes in feed rates in general (which can happen a lot during printing). This is especially interesting because it the spike is much more pronounced with the new nozzles compared to the old V6.
I would be curious if the oscillations are caused by the dual drive gears meshing
thanks for the video!
I always hear it's recommended to not wear gloves on rotating machines
the wavieness is the heater turning on and off probably
Appreciate the detailed analysis. Was at RMRRF for 2 days, literally opposite the nice E3D folks and had no idea that they had these on display. Lesson learned, say hello and be curious.
For sure Interesting data! Thanks!
Zu meiner Lehrzeit schauten schon einige Extrusionswerkzeuge ähnlich aus vom Materialfluss her
cant believe I had not subscribed in the past, have watched lots and lots of your very well thought out tests, tips, trials and tribulations. Again you have not disappointed. I guess the whole point of this type of nozzle is ... speed of print?
My Burning question is.... what about quality of print at the high flow and print speeds. If the quality and accuracy of the end product degrades...... what is the point?
14:10 could be the the nozzle minimum diameter forcing the filament in the cause of this back pressure?
That spike at the start might be a airgap or stretch between the cold and molten filament. Your fluiddynamics-wavyness seems to have the same frequency as the loops when extruding, can you try one without a floor the filament can stop against?
some haunted video artifacts at 14:21 :O
I dont know if i agree with your distinction between the Volcano and the CHT style nozzles. They both are just increading the surface area the material is exposed to. Yes, a Volcano is going to have more of a buffer but the flow rates have been tested.
The core difference is that CHT ("core heating technology") heats the filament from the inside out as well, while for a regular nozzle the biggest thermal resistance will always be the filament itself as the heat tries to make it to the center of the filament.
That wave you were seeing could actually be the sine wave for the motors.
3 channels ? I thought it was 4 channels inside the RevoHF... I wonder if the Revo hotends would perform better or the same with a V6 heating block ?
Just a question as I go through boxes after a move. Can the original Anet A8 be brought back to life in 2023? Of course there would need to be some updates that can be done to strengthen the frame and other new features available now that were not out back then. Just a thought.....
I wonder if it's possible in the HF if the waviness in The Force is something to do with the filament sucking heat out of the hotend? The comparisons between the HF and the Volcano are probably the most interesting, no?
I’m not any genius but for the waviness in the force, could it be related to the pull out drag of the filament. Would be interesting to see what the reading Returned could be with just a meter of unrestrained filament. ***Just a wild thought ***Cheers!
Would be interesting to see other materials and nozzle diameters being tested.
I have two questions that apply to both this nozzle and the volcano.
What is retraction like? It seems to me that if you flow all that plastic around something inside the nozzle, that when you pull back a little, it won't be as effective as a straight bore nozzle.
Is there any way to clean the inside of these nozzles? On a straight bore nozzle, you can do a cold pull. But if the filament is melted around the inside of the nozzle, you won't be able to pull it out.
Thanks for the test. I am very much interested in the sample rate used on the load cell and if you checked for variance in sample rate timing as the the data seems kind of coarse for some parts of further data analysis?
don't you think, the drop at the beginning is caused by a drop in temperature due to the change in push of colder material? Just until the PID compensates.
Im guessing that you could create very fancy internal structures with SLM 3D printing, I'm wondering what you could do with that
Oooooh, but the surface roughness!
@@MadeWithLayers I've guess... Idk, I'm wondering how much would affect it... Maybe this is a job for your sponsor PCBway
could you do a test with "chunkier" filament, i can imagine CHT nozzle and alike would get clogged more easily?
wavyness in the data points seems familiar to me. are you by any chance powering hot end and extruder motor from the same power suply?
in my homemade printer i noticed that heater was clamping power rails and stepper motor had momentary less voltage to work with and that produced rythmical artefacts on my prints that only showed up at higher speeds. but they didn't show up when i was building the printer , so i tracked it down to single power suply, adding separate power suply for the heater made motors work smoother and wavy artefacts disapeared
I purchased a revo voron for my 0.2. I wish it came standard with the high flow nozzle. It would have made things slightly cheaper.
Somebody should make aerogel insulating socks for nozzles.
don't you think, the waviness is also a function of PID loop actively heating/not heating? hotter and colder temperature probably even more for less perfect PID values.
What about unclogging those with a accupucture needle, though - 3+ times the work and a lot of luck? ;-)
In stock next year!!
looks like a hx711 loadcell, wondering if you use it at 10 or 80hz. But if you want more details in the reading (samples per second), try Adafruit NAU7802. It can read up to 320hz (think the loadcell reader in prusa new extruder use a hx717 with 320hz, but could not find any simple board with it). Maybe you could them see the waves also in more detail. Or the start ramp up, what maybe only caused by the static friction of the filament against the walls.
Also wondering if the v6 force was against a full metal, or some with PTFE tube, could also explain maybe the overall less pressure for the V6, with less metal in contact to the filament.
I'm reading the HX711 at 10Hz, which I think gives me enough data to even see some of the intricacies. Extrusion is still a pretty messy process, so I'm not sure if more temporal resolution would make for more information 😅
Hotends were all regular all-metal ones.
Are the waves maybe the reason for what was once called 602 error (i think) wavy outer layer pattern?
The wavyness is way too low-frequency for it to be 602
s that rythmic pulsation caused by hysteresis in the thermal management of the hot end?
what if you need to do a cold pull, how would you do it with these nozzles?
Could the "waviness" of the extrusion force be due to the spur gears on the Hemera being slightly eccentric instead of concentric on the shafts? What I mean by that is that the set screws (or grub screws I think is the non-American term) pushes the gear tangentially against the shaft, causing the gears to be slightly eccentric. I have seen that exact situation happen on a robotic system I've worked on. There was a tolerance stackup for the ID on the spur gear and the OD of its mating shaft that eccentuated the problem to the point that it would bind every 360 degrees in the location of the set screws.
Edit: it looks like other people have basically said the same thing I've said.
Wavyness might be from a bent extruder/gear, making the gear move excentric, slightly increasing and decreasing speed.
Just noticed a small mistake @ 11:20. It does top out with 15mm^3/s, because it's limited in the settings done by Prusa.
I don't have an issue with patents as long as they aren't used to lock away a system. There is no reason E3D could have a reasonable cost license to allow for 3rd party nozzles.
At this time e3d should Just hire the German RUclipsrs for scientefic Testing. ..😂
SL_ST 3xEXTRUDER designed to be compatible with ordinary V6 nozzels;)
But, in theory, three times more productive.
Maybe it 's time to build a StealthBurner for Voron ?
The waviness might be caused by the heating element as the strain cells are sensitive to heat.
If they froze the tips and warmed the main body, they could get a tighter fit and it'd be a perma-bond.
I'd be surprised if they weren't already utilising this. It's common practice for interference-fit parts, using the thermal expansion of those parts to get such a fit as you describe.
@@lio1234234 What I heard from another youtuber's video is that it's "just press fit", I think.
Can you do a comparison of the normal revo vs say mk6 CHT?
Isn’t the cht technology a 3DSolex patent that bond tech license?
No disrespect to you Tom but you need to design a printer for the market for sale your good in design and making a good product .I'd buy one with no doubt that it would be a great product
The default heater core is 40W not 30W as you said at 22:13
THe Revo heater's power drops as it gets to temperature, so its effective power while printing is much lower than the rated 40W.
What was with the artefacts whenever you were showing B roll of the display? Were you trying AI video again?
No AI, real numbers, just in slow motion to be able to read them better.