3D Printing Upgrade to High Flow Hotends (goal = high-speed, large nozzle, vase mode for large nose cones)

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

BigMacDaddy

Well-Known Member
TRF Supporter
Joined
Apr 30, 2021
Messages
2,121
Reaction score
3,360
Location
Northern NJ
I just purchased a Prusa Mk3s clone kit that comes with a crappy hotend (so needs to be replaced). I have also been pushing the limit of my stock Mk3s to try to get faster prints and realized I need to go up to larger diameter nozzles and larger layer heights.

I am wondering if anyone has upgraded their hotends to do something similar and what was your experience?

As I was looking into the Dragon High-Flow ends and similar popular options I came across the Rapido Hotend from Trianglelabs. It is rated at 75mm^3/sec and costs about the same amount as other high-flow hotends ($100 or so). Not too many reviews for this but a few that seem to suggest the flow rate on these is really high and quality is judged to be pretty good.
https://trianglelab.net/products/rapido-hotend?VariantsId=10030
I also came across the CHC Pro which looks like a good deal (even with an upgraded Heatbreak it is still around $50). So wondering what folks think of that unit (just from specs / pics I suppose).
https://trianglelab.net/products/chc-pro-hotend?VariantsId=10272
 
I run an E3D V6 with a copper block (50W heater) and a 0.5mm nozzle-X. I also use a bond-tech dual-drive geared extruder, but that's another discussion.

My dad runs a slice engineering mosquito on at least one of the printers for the high school robotics team.

A E3D volcano may be a good 1st step into vase mode w/large nozzles.

Look up CNC Kitchen and Thomas Sanladerer on youtube, they both has some great videos on the current latest and greatest nozzle tech (and all other things 3d printing). In the last year or so there has been some interesting new tech and hacks to get better melt quality.
 
I have never been able to get high flow to equal higher speed. The drop in quality makes it now worth my time.

I did change to 0.6mm nozzle with no change in quality.
 
Copper block, long nozzle, 0.6 or higher nozzle. Have printed with 0.9. As I print ABS mainly. Use latest version of Cura and import the "filament Friday fast" settings to Cura. Copper block makes biggest difference. Filament Friday next largest. Search on You Tube for settings. Importing them may require hand editing the printer name in the settings to make it work for your printer. These are simple things you can do to a standard printer. Or you can spend the money and go for a full change of head.
Good luck on your journey.
 
I have never been able to get high flow to equal higher speed. The drop in quality makes it now worth my time.

I did change to 0.6mm nozzle with no change in quality.
I saw your post from a while back about the TriangleLabs Dragon hotends. Glad that you had good results with the Trianglelabs products.

The latest products they have put out seem to be the Rapido and/or the CHC (Ceramic Heating Core) versions that I am looking at. The CHC Pro and the Rapido have 145w heaters while a couple of the other CHC models have 45w heaters. I cannot get flow rates from the CHC models (they responded: "I did not specifically check the flow rate" which seems odd for a hotend advertised as high flow) but a few folks have suggested that the CHC pro should be similar to the Rapido. I will see what I can find out and post my experience. I may just pull the trigger on a few of these hotends since it is not such a big risk at $30 or so.
 
I saw your post from a while back about the TriangleLabs Dragon hotends. Glad that you had good results with the Trianglelabs products.

The latest products they have put out seem to be the Rapido and/or the CHC (Ceramic Heating Core) versions that I am looking at. The CHC Pro and the Rapido have 145w heaters while a couple of the other CHC models have 45w heaters. I cannot get flow rates from the CHC models (they responded: "I did not specifically check the flow rate" which seems odd for a hotend advertised as high flow) but a few folks have suggested that the CHC pro should be similar to the Rapido. I will see what I can find out and post my experience. I may just pull the trigger on a few of these hotends since it is not such a big risk at $30 or so.
Bear in mind that not all materials respond to heating in the same way. ABS doesnt really fully melt like PLA. its more like an extrudable hot slush. So different printing materials will have different results due to their properties. The increase in speed you can achieve for one might be greater than another material.
 
Bear in mind that not all materials respond to heating in the same way. ABS doesnt really fully melt like PLA. its more like an extrudable hot slush. So different printing materials will have different results due to their properties. The increase in speed you can achieve for one might be greater than another material.
Good point -- and from what I read some of the lower temperature filaments like PLA actually have the biggest issues with the higher-flow nozzles due to heat creep and retraction issues.

I am using ABS - goal is not to go hotter, just maintain the steady high temperature necessary to print more volume of filament per second.
 
FWIW - I totally appreciate the typical trade-off of speed vs. quality but have been having some decent luck with 120mm print speeds with simple geometry.

For example, here is the airfoil ring-tail for the Coleoptere printing at 120mm/sec. I am printing it with 3 walls with a .4mm nozzle, .4mm wide .2mm thick lines. It still takes 1hr 10min to print but at least it is not 3+ hours. Would love to be able to do this with a single wall in vase mode. Some other higher-speed prints have not been as smooth as these but I am trying to experiment and push my printers.

Cooling is probably the biggest limiting factor now on these simple prints. Not as bad when I print 12 cones at once but this might be printing slower than 120mm simply because it is required to have 5 seconds per layer for layer to cool. I am going to need to upgrade fan setup on the new hotter hotends.

 
Last edited:
Looks like my MK3s clone is going to get here on Sept 2nd instead of Nov 5th like was originally estimated. I guess I will go ahead and order a couple of these CHC hotends to upgrade that unit as well as my Ender 6. The Ender needs a bunch of tweaking to get reliability up.

I went with:
 
Last edited:
The owner of hotends.com (J-Head hot ends) was in my cidery this weekend. He was looking at the printed sample trays i made, and I showed off the Talos Terrier I had sitting in my trunk. (Didn't get to fly - very sad).

I've moved my CR10S to 0.6mm for single wall vase-mode nosecones and tube sleeves (easier to finish than the homemade paper tubes they go over) and have been very pleased compared to 0.4mm. I have not pushed the speed, so the stock heater is still fine.

I abandoned stainless nozzles - I don't think SS conducts heat well enough. The above suggestion of a copper block makes some sense to me - Though I recently watched a heater system eat itself as the copped plates oxidized and then wedged the alu-bronze pieces apart. But that was at 600C.
 
The owner of hotends.com (J-Head hot ends) was in my cidery this weekend. He was looking at the printed sample trays i made, and I showed off the Talos Terrier I had sitting in my trunk. (Didn't get to fly - very sad).

I've moved my CR10S to 0.6mm for single wall vase-mode nosecones and tube sleeves (easier to finish than the homemade paper tubes they go over) and have been very pleased compared to 0.4mm. I have not pushed the speed, so the stock heater is still fine.

I abandoned stainless nozzles - I don't think SS conducts heat well enough. The above suggestion of a copper block makes some sense to me - Though I recently watched a heater system eat itself as the copped plates oxidized and then wedged the alu-bronze pieces apart. But that was at 600C.

That is cool!

How wide are you extruding? I have been reading that you can extrude up to 200% your nozzle width pretty reliably.
 
I'm telling Cura 0.6mm. I need to revisit the tuning - I'm pretty sure I'm getting a wall thickness greater than intended.
Check the weight. Cut 50 mm of your filament and weigh it. Manually extrude 50 mm and weigh it. This will tell you if you are under or over extruding. At speed you'll extrude slightly less. At the limit of your print speed it will significantly reduce resulting in underextrusion and tearing in the finished print. The trick for best speed is to find what works for your material type. The copper block I'm using is nickle plated.
Cura doesn't care what nozzle you actually have fitted. It just works on what you tell it you have fitted for the recommendations it makes for layer height etc.
Ive printed a 0.4 nozzle at 0.8 and a 0.25 at 0.4. It depends on the quality you are prepared to accept. ABS can easily be finished up with sanding and acetone. Other materials cannot.
YMMV
Norm
 
Check the weight. Cut 50 mm of your filament and weigh it. Manually extrude 50 mm and weigh it. This will tell you if you are under or over extruding. At speed you'll extrude slightly less. At the limit of your print speed it will significantly reduce resulting in underextrusion and tearing in the finished print. The trick for best speed is to find what works for your material type. The copper block I'm using is nickle plated.
Cura doesn't care what nozzle you actually have fitted. It just works on what you tell it you have fitted for the recommendations it makes for layer height etc.
Ive printed a 0.4 nozzle at 0.8 and a 0.25 at 0.4. It depends on the quality you are prepared to accept. ABS can easily be finished up with sanding and acetone. Other materials cannot.
YMMV
Norm

I have a particular part that I print for calibration - including wall thickness. And I have the appropriate micrometer. I don't recall redoing my calibration process after switching from 0.4 to 0.6mm, though - so I'm sure it's my bad.

Doing it by weight's an interesting thought. I've got an Ohaus scout handy that will do .1g resolution. Without taking out and dusting off the analytical balance.

I print almost entirely PETG. I find it easier to file than to sand.
 
I have a particular part that I print for calibration - including wall thickness. And I have the appropriate micrometer. I don't recall redoing my calibration process after switching from 0.4 to 0.6mm, though - so I'm sure it's my bad.

Doing it by weight's an interesting thought. I've got an Ohaus scout handy that will do .1g resolution. Without taking out and dusting off the analytical balance.

I print almost entirely PETG. I find it easier to file than to sand.
You can get a set of 0-50 gm scales at 0.001 accuracy for about $10 delivered.
I use the weight method as a starting point for initial calibration. Once thats ok print a test cube.
This also gives you an easy way to check you're not getting drive slippage on filament or blocked nozzle creeping up on you.
 
I bought the E3D Revo Hemera head for my MK3S. I used the Hemera Odyssey mounting https://github.com/tetra3dprint/HemeraOdyssey.

The head is all metal construction. It has two gears to wind the filament through. Great heat control. It has exchangeable rapid change nozzles, including a 0.8. I sort of consider it hobbyist 1.0 in terms of hot end tenchology and quality. Finally, something at the hobbyest level that approaches professional grade in terms of quality and features. Prints are stunning.
 
You can get a set of 0-50 gm scales at 0.001 accuracy for about $10 delivered.
I use the weight method as a starting point for initial calibration. Once thats ok print a test cube.
This also gives you an easy way to check you're not getting drive slippage on filament or blocked nozzle creeping up on you.
I saw a video where they created G-code to extrude fixed weight amounts of filament into small piles (like soft-serve ice cream) so that you could calculate maximum volumetric flow at different temperatures. I cannot find it again but will share if I find it.
 
I saw a video where they created G-code to extrude fixed weight amounts of filament into small piles (like soft-serve ice cream) so that you could calculate maximum volumetric flow at different temperatures. I cannot find it again but will share if I find it.
Was it the volcano link I posted above? I don't know if they went into the g-code creation but it was all about dumping piles of goo to measure maximum volumetric flow for various nozzle configurations.
 
I accidentally started a new thread but moving this over here...

I have been running my new MK3s+ clone with Triangle Labs CHC Pro hotend for the last couple of weeks on and off and overall quite happy with things.

I had tested things with the original .4mm nozzle and some pretty high-speed (80-120mm/sec) two-wall shells (no supports and 0% infill) and all worked very well (Benchy for comparison). One layer shift in this print that was not too bad and could be sanded / filled out (wish I had packed grease in the bearings prior to assembly) but no major issues.

1663621501652.jpeg


I just swapped out for a 1mm nozzle with the goal of printing 1.5mm wide extrusion x .5mm layer height x 40mm/sec (like to go faster) = 17+ mm^3/sec volumetric rate single wall in Vase mode. First couple of tests with larger nozzle were failures. I am pretty happy with the quality of the print before it failed and volumetric rate seems fine (I will likely sand these and paint w/ filler primer). Layers 'seemed' to be stuck together well and print was strong but when I took it off the bed 1st layer stuck and started unwinding the other layers (so poor layer adhesion I guess).

1663622200699.jpeg


1663622218123.jpeg


1663622280013.jpeg



I set the slow down at 5 seconds which has worked for me with smaller nozzles. Obviously a larger extrusion holds quite a bit more heat. If I slow down more it kinda defeats the purpose which is to get large nosecones in short times. Setup has dual fans and they are set at 100% for more or less the entire print.

1663622556812.png

I was able to get this to print successfully by raising the temperature to 225 and then to 250 in 2nd half of print (fan speed is 100%). Adhesion is better -- cannot really see a difference in the lower vs. upper half. I also slowed down to 40mm/sec and 5 sec slow down (I think my earlier print had a 2 second slowdown). This is 1.5mm width, .5mm layer height, single wall vase mode.

1663802498499.jpeg

This printed in 38min. Normally if I was using a .4mm nozzle with 3 walls and .2mm layer height this would take 2hr 22min at 80mm/sec exterior walls and 120mm/sec interior walls. That being said, not sure it is worth the trade-offs -- design as well as quality / detail.

I would normally print with ABS but this is not in an enclosure so have been using PLA on this printer for testing purposes.
 
Last edited:
I've solid/infill printed tips with a shoulder and glued them into the vase. I either cut the vase print short, or design it to end short up front.
 
I've solid/infill printed tips with a shoulder and glued them into the vase. I either cut the vase print short, or design it to end short up front.

Thanks for confirming that -- it was what I was thinking to do if I stuck with this method (shoulder and tip of nosecone printed separately and main cone printed vase mode).
 
Good job.
I can only manage 41 mins with 2 walls 0.8 nozzle printing 0.85 wide. 80mm/sec print speed.. But I've not actually printed it yet. Having a couple of issues that started when the heated bed thermistor decided to reach end of life without any notice.........
ABS and PLA are very different materials. PLA melts properly and sticks to the previous layer. ABS extrudes as a hot slush that if you are lucky sticks to the previous layer before shrinking so much it rips itself apart.... :)
Have you tried the "filament friday" fast profiles? Gut says you should be able to save another 5 mins on that without too much tweaking. Depends on how important speed is to you.
 
I design all my printed NCs to have the last ~10mm be solid. The first NC I printed had some deformation at the tip so I added a minimum layer time parameter to my slicer and have had no issues since.
 
Back
Top