Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s  

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Tim

Bob - what acceleration and max feed are you using with the E395 command?  Or is that short hand for E60 F395?  But question remains for acceleration.

You are correct - those are Fxxx speeds, not Exxx! Only extruder moves at start up firmware defaults. 

Argh. Curse the inability to edit posts. Now I'm responsible for misinformation and anybody reading from top to bottom is going to spew out a LOT of filament needlessly.

Team Prusa, PLEASE RESTORE OUR ABILITY TO EDIT PRIOR POSTS. This isn't a forum for political debates in which we need to worry about ninja edits. This is a technical support forum in which the ability to accurately convey complex technical information is essential!

Also - at what temp did you run PID calibration?

Ah, good point. 210C. Another data point that might tie into testing.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Sembazuru

Hey, bobstro, I looked at your MuppetLabs just now about calibrating new filament. I've been thinking that I should add figuring out volumetric flow based on your topics above.

I'm starting to compile enough notes to add a new page on the topic. The "illusion of speed" topic is already overly long, so a linked page is likely. I do updates on the fly as I have time, so it'll probably be a few days.

But my disagreement is with your order (though, the reasons for my disagreement may be based on new information discovered here that you may not have had last time you edited your page). Because it seems extruder temperature (and probably nozzle, but that just muddies the water for the moment) affects the max volumentric speed, I would find the max volumetric speed after doing a temperature tower and I've decided on a temperature that has the best compromise for surface finish, overhangs, stringing, bridging, etc.

That's a good point, to be sure. Until now, I've been a bit like an ancient blacksmith, gathering bits of mythology, superstition and a bit of wisdom to work my craft. Previously, I had been limited to simply confirming the results with ad-hoc observation and testing ("ah, clicks stop when I keep MVS at this level..."). I've never been happy about simple citing 11.5mm^3/s without supporting data, so I'm glad to have a more solid basis. Hopefully we can put together enough information to make it easy for others to do their own testing. Limiting testing at the speed I'll actually be printing at just makes sense. Good stuff, thanks!

While it's been a challenge for us Prusa users using E3D hotends, the poor souls using a lot of the Chinese budget printers are in a very sorry state. They stick with magical speeds and settings for the most part, and nobody really has a clue what their hotend capacities are. Candles into darkness and all that!

Up until now, based on your pre-experimenting thoughts (I'm thinking that "advise" is too strong a word), I was assuming that your 11.5mm³/s wasn't conservative enough so I'd been just throwing 11mm³/s in to be even more conservative.

Sounds like you've done some project estimation work! When getting started with a new filament, I've often done the same just to get started.

Granted, I rarely print something large enough that I even approach the max volumetric speed on my stock 0.4mm nozzle, especially since I usually use gyroid infill which isn't really the fastest infill available... Though, not knowing how to find the max volumetric speed has been one of the reasons why I have yet to try using flexibles yet (even though I have a spool in my collection...).

Definitely this! I like the idea of doing a bit more rigorous testing to get to that initial recommended MVS setting for a new filament. This is one significant value-add that filament profiles for Slic3rPE/PrusaSlicer has. A community library of tested settings would be very helpful.

I do have a question (that you might not have an answer for): What would you say is the range of amount of filament used for dialing in extrusion? How many 6cm extrusions do you end up having to do, average and maximum? A dozen? More? Less?

I'd say 8-10 based on the 2 runs. It takes about 6 60mm "shots" to bracket the speeds that trigger the clicks at each temp. Once I get to a safe upper end, I run 2 more to verify it doesn't cause clicking intermittently. This is per-temperature. Now that I've done it a couple of times, I'd expect it to be a 5-10 minute job for a couple of temps. I end up with a little spaghetti ball of filament that I feel guilty about, but I've made worse on accident.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch

[...] you asked for an analysis, so here is the physics. 🙂 And thanks for the measured data, that kind of research is most valuable.

Hey, this is where I welcome the people that can explain why we're getting these results. I appreciate your contributions. In particular, please point out where I'm doing something wrong, or at least something that needs more testing. Extra points for pointing out how!

[...] Even steel is 20x more thermally conductive than PLA and so it is always the filament that is the limiting factor in the melting process and the nozzle material doesn't make much difference. That's why the E3D Volcano, the Supervolcano and the Raise3D Matchless nozzles exist - they try to increase the contact area between the nozzle and the plastic to pump heat into the plastic faster.

The length of the hotend is a major factor. I had read up on the nickel-plated copper hotend from E3D but decided to pass with the understanding it wouldn't help much, and that the longer hotends on the Volcano and Super Volcano would make a much bigger difference. Now I'd really like to be able to test anyhow!

[...] In your experiments, you're basically measuring what is the limit of the nozzle → filament heat transfer. With increasing speed, the temperature of the extruded plastic decreases until it is below the melting point and the extrusion stops. Increasing the temperature of the nozzle increases the heat flux into the filament thanks to a bigger temperature gradient, but the extruded plastic temperature will be the same in all the maximum extrusion rate situations you measured, somewhere around 180 °C.  This would result in very poor layer adhesion and possibly other adverse effects.

So using a higher temp just gets the filament to melt temps faster, allowing more filament to be melted and pushed through in a given amount of time, correct? That makes sense. In researching some of this, I found reprap discussion back from 2013 or so discussing pre-warming filament and other tricks. Of course, this was likely before we had a wealth of hotends to choose from.

As for the filled filaments, I limit my carbon-nylon printing to 3.6

Things really started clicking (pardon the pun) when I realized the MVS for PETG was set to 8. I have a few filaments that I use 5mm^3/s for, and Colorfabb XT-CF20 only works well for me at 1mm^3/s. Now to do some more testing of these problem children!

Your findings support your original statement that a slicer limit of 11.5 mm³ would be a more reasonable default for PLA. With that, there is enough headroom for the nozzle to heat the plastic to a reasonably consistent temperature.

I had encountered 11.5mm^3/s in several E3D V6 discussions as a realistic value, so I don't claim any credit for coming up with it, only validating results using it. But yes, I am after a "safe recommendation" and not a tuning parameter for performance. As we approach summer, I'm expecting all of the extruder click and heat creep posts to increase, and I'd like to convey useful information for new users.

Fortunately, in real prints, the extrusion rate is varying considerably and it's enough for the average to be under the sustained rate limit. Acceleration and speed constraints for particular features tend to limit the print most of the time. That's why the 15 mm³ default can still work - as an infrequent stop gap to prevent the printer stalling on the quick parts of the print.

I think one of the traps of 3D printing is that we want to test everything, but we're impatient. We do a lot of testing and calibration on small parts, then are surprised when larger multi-hour prints fail far into a print. These numbers and limitations really come into play with larger prints where acceleration, jerk and extrusion lengths don't constrain speeds.

Good stuff, thanks for your thoughts!

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Taking a simple example of a 0.45 mm by 0.2 mm layer at 200 mm/s extrusion... the math says the most it can take is 18 mm3/s.  Add acceleration and deceleration and it's something less.  Change the layer to 0.15, and we only need 13.5 mm3/s maximum.

Change to a larger nozzle, like 0.6 mm, and the scope of the issue changes accordingly. 

Still thinking over my extrude test ... getting closer to a sound methodology.

ps: I wonder if we shouldn't also consider heat loss due to the print fan running....

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Tim

[...] ps: I wonder if we shouldn't also consider heat loss due to the print fan running....

If nothing else, we may wind up with a good data point for testing alternate fan designs.

Here are my max volumetric rate calculation spreadsheet results for 8mm^3/s and 11.5mm^3/s. I was driven to this madness by Simplify 3D's frustrating "speeds as percentage of default speed" settings. Blue are layer heights below 25% of nozzle size, red are layer heights above 80% of nozzle size. Speeds exceeding printer safe maximum speeds are italicized.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch

Posted by: Sembazuru

Because it seems extruder temperature (and probably nozzle, but that just muddies the water for the moment) affects the max volumentric speed, I would find the max volumetric speed after doing a temperature tower and I've decided on a temperature that has the best compromise for surface finish, overhangs, stringing, bridging, etc.

The converse is even more true: Speed changes the temperature of the extruded plastic. (See my other reply.) So when you select a particular temperature, any increase in speed will change the print quality. While there is enough headroom, the change is negligible, but as you start to approach the speed limit, the temperature of the extruded plastic will go down quickly.

So it's about finding a good combination, rather than tuning one and then the other, or vice versa, unfortunately.

I do agree with this, but the temperature tower I use is small enough that I doubt I really start reaching really fast extrusion rates. I haven't actually checked, so I will. Ok... This is funny. I just sliced the temperature tower that I usually use and set the filament MVR to 15, the actual printed MVR is 11.5, just a small triangle on one internal layer of each tower block:

My hope in dialing in a reasonable (doesn't even need to be optimal for speed) MVR for each filament is so I can use that setting in the Filament Settings Profile to throttle (if needed) the speed settings in the Print Settings Profile so I only have to change one profile when I change filaments. This may be a fools errand, but it is my current goal.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Sembazuru

[...] I do agree with this, but the temperature tower I use is small enough that I doubt I really start reaching really fast extrusion rates.

I haven't actually checked, so I will. Ok... This is funny. I just sliced the temperature tower that I usually use and set the filament MVR to 15, the actual printed MVR is 11.5, just a small triangle on one internal layer of each tower block:

My hope in dialing in a reasonable (doesn't even need to be optimal for speed) MVR for each filament is so I can use that setting in the Filament Settings Profile to throttle (if needed) the speed settings in the Print Settings Profile so I only have to change one profile when I change filaments. This may be a fools errand, but it is my current goal.

I think the idea of doing a quick test after doing the temp tower makes a lot of sense. No purpose is served by testing speeds you won't print at. Select a temp for good bridging and stringing results, then do the MVS tests to determine what your safe max. MVS is for that filament. It sounds like a bit of work, but is actually quicker than some of the "try a setting" prints I think. 

I like to set MVS under Filament Settings for filament-specific limits, but I do keep my "maximum speed limit" set under Print Settings as a 2nd level. If I get to the point where I've got actual tests per-filament, I'll be more comfortable setting it only under Print Settings.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

I don't see a reason to lower max volumetric speed. It fits to the overall (pretty high in my opinion) default speed settings. 

Most of the prints I've done so far were smaller size (<200g) and complex enough, so the speed got reduced anyway. But I'm reducing the speeds manually because 200mm/s is way too loud and prone to fail for some filament brands. I think hotend might get into trouble with 15mm/s flow rate especially with HTPLA, without the sock and cold ambient temp. But on daily base and most likely in the enclosure with warmer ambient temp and good PLA (wide temp range) this settings will never bother you.  

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: nikolai.r

[...] But on daily base and most likely in the enclosure with warmer ambient temp and good PLA (wide temp range) this settings will never bother you.  

I'm going to respectfully disagree on this. During the hot summer months, I responded to many posts in which people were describing "heat creep" and "extruder click" problems that were resolved by reducing MVS. Not every case, of course, but several. The high MVS default coupled with the aggressive 200mm/s speeds used for infill can really cause problems on larger prints where the nozzle actually starts moving at specified speeds. Alarming clicks are one thing, but dislodged and ruined prints, nozzle wear and extruder jams are significant problems that really frustrate new (and old) users. IMO better to provide safe defaults and let those who have figured things out do the tweaking to use higher values.

Note that the use of an enclosure can sigificantly exacerbate these problems as the air-cooled E3D V6 hotend is only rated to 40C.

I've started compiling my notes into this page.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch

PLA thermal conductivity is 0.13 W/m·K, steel is around 25 W/m·K, brass is 109 W/m·K, aluminum is 205 W/m·K, copper is 385 W/m·K.

Even steel is 20x more thermally conductive than PLA

I missed a 0. Steel is 200x more thermally conductive than PLA.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: bobstro

The length of the hotend is a major factor. I had read up on the nickel-plated copper hotend from E3D but decided to pass with the understanding it wouldn't help much, and that the longer hotends on the Volcano and Super Volcano would make a much bigger difference. Now I'd really like to be able to test anyhow!

It's not just the length. Of course the length helps, because the filament stays in contact with the hotend longer as it goes down. But the Matchless nozzle solves it differently, without changing the hot end, by having a more complicated internal structure that touches the filament with a bigger surface, with good results, too. Matchless nozzle:

I have an own design for a fast melting nozzle in mind, I just need to figure out where to make it to be able to test it ...

So using a higher temp just gets the filament to melt temps faster, allowing more filament to be melted and pushed through in a given amount of time, correct? That makes sense. In researching some of this, I found reprap discussion back from 2013 or so discussing pre-warming filament and other tricks. Of course, this was likely before we had a wealth of hotends to choose from.

YES, precisely. A higher temperature means a higher gradient means a higher heat flux means heat is entering the filament faster despite its low thermal conductivity. Melting quicker. Not necessarily getting hotter if there isn't enough time in the hotend for that.

Things really started clicking (pardon the pun) when I realized the MVS for PETG was set to 8. I have a few filaments that I use 5mm^3/s for, and Colorfabb XT-CF20 only works well for me at 1mm^3/s. Now to do some more testing of these problem children!

Cool. When we're at it, I probably should test with some of the nylons using your test protocol.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch

It's not just the length. Of course the length helps, because the filament stays in contact with the hotend longer as it goes down. But the Matchless nozzle solves it differently, without changing the hot end, by having a more complicated internal structure that touches the filament with a bigger surface, with good results, too. Matchless nozzle:

Is this Matchless nozzle something widely available and compatible with the Prusa printers? A quick search turns up some for 2.85mm setups, but none for 1.75mm and specifically for the E3D V6. For this purpose, that may be in the "interesting but not useful" category.

For my purposes at least, I'm going to stick with the assumption that we're running a stock Prusa with an E3D V6 hotend and perhaps a variety of nozzles for that hotend. It would be interesting to see other hotends added to the results, but I have no way of testing those at present.

I have an own design for a fast melting nozzle in mind, I just need to figure out where to make it to be able to test it ...

It would be interesting to see others.

YES, precisely. A higher temperature means a higher gradient means a higher heat flux means heat is entering the filament faster despite its low thermal conductivity. Melting quicker. Not necessarily getting hotter if there isn't enough time in the hotend for that.

Based on the discussion here, I'm thinking testing a new filament over a range of values isn't necessary, though perhaps still interesting. If the user prints a temperature test tower prior, all they need to do is test at the temperature they intend to print the new filament at. The inner workings of the hotend are interesting technical details, but not necessarily something somebody out to do a print cares about at the time. Again, I'm going to assume a stock E3D V6 for now. Anybody running a Volcano or other exotic customized setup can add some great data!

Cool. When we're at it, I probably should test with some of the nylons using your test protocol.

I'm going to try with some not-so-nice PLA to get a sense for the range of values. I also have PETG, a few filled PLAs and Taulman Bridge that I can test with. Other nylons, PC etc. would be very interesting to see. My assumptions now are:

• PLA can be used to narrow down the volumetric rate the hotend can work at. Unless there are common filaments out there that flow more readily than PLA, assume unfilled plain PLA results will provide the "best case" for hotend capacity. I'll try some white PLA. I don't have any natural/uncolored. This is the value I enter as MVS on the Print Settings tab.
• Other materials will impose their own volumetric rate limits. This is the value I enter as MVS on the Filament Settings tab.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: bobstro

 

Is this Matchless nozzle something widely available and compatible with the Prusa printers? 

I have an own design for a fast melting nozzle in mind, I just need to figure out where to make it to be able to test it ...

It would be interesting to see others.

Let me model it and share. I think I figured a company nearby that could make me one ...

Based on the discussion here, I'm thinking testing a new filament over a range of values isn't necessary, though perhaps still interesting. If the user prints a temperature test tower prior, all they need to do is test at the temperature they intend to print the new filament at. The inner workings of the hotend are interesting technical details, but not necessarily something somebody out to do a print cares about at the time. Again, I'm going to assume a stock E3D V6 for now. Anybody running a Volcano or other exotic customized setup can add some great data!

That makes sense.

Cool. When we're at it, I probably should test with some of the nylons using your test protocol.

I'm going to try with some not-so-nice PLA to get a sense for the range of values. I also have PETG, a few filled PLAs and Taulman Bridge that I can test with. Other nylons, PC etc. would be very interesting to see. My assumptions now are:

• PLA can be used to narrow down the volumetric rate the hotend can work at. Unless there are common filaments out there that flow more readily than PLA, assume unfilled plain PLA results will provide the "best case" for hotend capacity. I'll try some white PLA. I don't have any natural/uncolored. This is the value I enter as MVS on the Print Settings tab.

I've found colored PLA/PHA mix (ColorFabb) flows even more readily than PLA alone. Plus there are some PLAs specifically made for high speed printing with lower melting points (TreeD Gonzales). But I think a common cheap PLA is a reasonable benchmark.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch
Yes, 3D Solex sells them individually or as a Prusa Speed Doubler set.

Fascinating. I have to say I've spent a fair amount of time scouring the web for V6-compatible nozzles and never encountered these. It looks like a very interesting concept. Do you have experience with them? I'm tempted to pull the trigger on an order for 0.40, 0.60 and 0.80mm sizes, but really should do a bit of research first. Thanks for the link!

I've found colored PLA/PHA mix (ColorFabb) flows even more readily than PLA alone. Plus there are some PLAs specifically made for high speed printing with lower melting points (TreeD Gonzales). But I think a common cheap PLA is a reasonable benchmark.

Inland is a good economy choice for the US, but I'm not sure about world-wide availability. ColorFabb is great stuff, but a bit pricey to ask someone to buy just for testing. Is eSun readily available world-wide? 

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Tim

ps: I wonder if we shouldn't also consider heat loss due to the print fan running....

As long as the reported nozzle temperature doesn't fluctuate (meaning a properly calibrated PID), the fan should not have any impact on the ability to extrude. 

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: bobstro

[...]

Is it possible you've got it set in another location? You can set MVS under Print Settings->Speed->Autospeed (advanced) and under Filament Settings->Advanced->Print speed override. That said, MVS is still only likely to start rate limiting on either larger prints with smaller nozzles, or when using larger nozzles. With a 0.80mm nozzle, you can start hitting MVS limits at 20mm/s or lower for example.

I didn't notice the MVS setting in the Print Settings. Seems I had that at 15mm³/s. (Leftover from when I customized the Prusa settings for "magic number" back in early alpha versions of Slic3r PE 1.42. The more recent Quick and Quality 0.2mm layer height profiles provided by Prusa have this set a zero.) I set this to zero and pushed the filament MVS to 150 (i.e. something arbitrarily large) and the temperature tower that I used slices to a max of 20.something mm³/s. I'm not sure if it would actually reach that during print given accelerations and linear advance (K value) settings.

Not that this matters at the moment. Just closing the loop on correcting an error I made.

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Voj - this is right up your alley:  find the error in this:

G1 X200 Y0 E6.652 F12000 ; extrudes 6.652 mm of filament in a trace 200 mm x 0.2 mm x 0.4 mm (16 mm3)

The theory: print 200 mm at 0.2 mm x 0.4 mm, requires 16 cubic mm material. Or 6.652 lineal mm of filament.

D = nominal filament diameter
V/mm = filament volume per mm

mm = length of extrusion
h = layer height
w = extrude width
a = area (cross section)
v = volume (x*y*a*mm)
v/s = volume per second (based on movement feedrate)
Fs = feedrate (per second)
F   = feedrate (per minute)

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Vojtěch

Posted by: Tim

ps: I wonder if we shouldn't also consider heat loss due to the print fan running....

As long as the reported nozzle temperature doesn't fluctuate (meaning a properly calibrated PID), the fan should not have any impact on the ability to extrude. 

Not so sure - the fan is moving air across the heater block and and removing calories from the system. At 40w, there are only so many calories.  Unless there is some other limit on heat exchange that makes 40w unimportant (i.e., 30w is enough to exceed some thermal exchange limit).

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: bobstro

• At 200C E420 was safe (~7.00mm/s * 2.405 = ~16.83mm^3/s)

I tested on my printer. Super cheap Gembird PLA Natural, 200 °C, Stock Prusa E3D v6,  Bondtech BMG 1:3 geared extruder, 0.9° LDO pancake at 0.96A, E3D Nozzle X 0.4mm. Result: F570 was safe (9.5 mm/s · π ·  (1.75 mm/2)² = ~22.8 mm³/s). The extruded plastic wasn't very liquid at that rate and extrusion wasn't smooth. No clicking, though. Smooth extrusion at F550 (~22.0 mm³/s) .

RE: Slic3r PE: Please reduce default values for "max volumetric speed" from 15 to 11.5 mm³/s

Posted by: Tim
Not so sure - the fan is moving air across the heater block and and removing calories from the system. At 40w, there are only so many calories.  Unless there is some other limit on heat exchange that makes 40w unimportant (i.e., 30w is enough to exceed some thermal exchange limit).

FWIW - I am testing with the part cooling fan off, or rather not turned on. I do have a silicon sock on my hotend.

Of course the hotend fan is running. I wouldn't expect it to matter, but ambient temps may affect things in strange and mysterious ways. I'm still that ancient blacksmith hammering away at this thing without fully understanding it.