External perimeter speed?

DocTrucker

Morning all,

How much do you wind back your external perimeter speed compared to the standard perimeter and infill settings?

Just another canvas to see if my current direction makes sense.

Thanks.

fcwilt

Hi,

I use Simplify3D with the default setting (80% as I recall) but I'm making functional parts so appearance is secondary.

When it comes to settings like that I have simply been printing test cubes, etc and trying different values.

Frederick

droftarts

Cura drops the external perimeter to 50%. The reason for dropping the speed is to improve the finish, and the reason the finish at full speed may not be as good is down to the quality of the printer! If you've got a solid printer, well-tuned extrusion, speeds that don't exhibit ringing for the frame etc, there probably isn't much reason to slow it down.

Ian

Phaedrux

It varies for me. Either it's the same as inner perimeter or if the model has a lot of overhangs or would benefit from higher quality I'll drop it. Some of the filaments I use don't like to change speed very often, so print speed is held constant. Polyalchemy elixir for instance.

The newer versions of Cura have the experimental feature to detect overhangs based on a configured angle and reduce the print speed and increase fan speed for those sections. This has been working very well for me.

Generally though I try to keep print speeds in the 50-80mm/s range and use layer height and extrusion width to maximize the extruder throughput.

DocTrucker

Thanks for the figures.

I'm trying to get up to the 10mm3/sec on the hotend (E3D V6, 0.4mm) and found While I could make parts at that throughput the outer perimeters needed to be around 6.4mm3/sec to get the shine on the outer surface. Push much faster and the shine rapidly dulls.

I'm assuming this shiny to dull transition is linked to the linear rate at which the polymer flows through the nozzle, rather than volumetric flow rate. So taking my figure of about 6.4mm3/sec results in an average a linear exit speed of 50.93mm/sec. In retrospect obvious but if you scale the track width and layer height with increases in nozzle size you end up targeting the same print speed.

Nozzle (mm dia.)	Track Width (mm)	Layer Height (mm)	Vol. Throughput (mm3/sec)	Print Speed (mm/sec)
0.20	0.24	0.125	1.6	53.3
0.40	0.48	0.250	6.4	53.3
0.50	0.60	0.313	10.0	53.3
0.80	0.96	0.500	25.6	53.3

It's not so much the shine I'm after, more a consistent finish. If the extrusion speed is above the shiny speed then things become shiny if the slicer slows it down for small perimeters, or short layers making those sections shiny and the rest dull.

Interestingly even with the following throughputs:

Extrusion Geometry	Vol. Throughput (mm3/sec)
external perimeters	6.40
perimeters	9.59
infill	9.59
solid infill	9.59
top infill	7.25
support material	6.40

...the average throughput relating to total build time is way down at 2-3mm3/sec due to losses from accelerations and travel moves.

Edit: I'm already running hot. 230C on 3DFilaprint's Premium PLA.

sinned6915

i just stumbled upon this.. those theoretical max numbers of 10 or 15mm^3/sec arejust that. printing speed does not mean anything unless you can melt enough plastic to push it through. you figured that out. the faster you can melt it, the faster you can go.

Dyze design had some interesting blog posts about this too. September I think of alst year.

doctrucker- have you tried kiss slicer and its wizards? there is a flow wizard to computer that. it varied flow across a range you specify.
http://wiki.snolabs.com/index.php?title=Tuning_Wizards#Flow_Rate_.5Bmm.5E3.2Fs.5D

on my machine, 24V, E3D V6 knock-off bowden system, with 0.4mm MicroSwiss nozzle, PLA+ or PETG or PC+, I max out at 6.4 mm^3/sec. Trying a volcano upped it some but it I was not happy with the overall performance.

i dont believe the surface finish has to do with flow rate, it has more to do with temps and cooling than anything.

try varying your temps 5-10 degrees with same flow rate and see how it looks.

on your first graph, instead of holding speed constant, hold volume constant and adjust speed instead. i think you will find the surface finish will change there too at the same temps.

DocTrucker

@sinned6915 thanks for the links. I will look through them.

Half the confusion is we don't control polymer melt temperature, we control the temperature of one heat zone. Increasing flow rate will decrease the time for the polymer to reach temperature and hence (if there is no temperature stabilisation in the melt zone) a lower melt temp.

You are definitely right to mention temperature as a contributing factor. The hotter you get the faster you can go before the extrusion dulls, that's why I am runnig so hot at the moment.

The crux of the difficulty in investigating the problem is there is no real way thay we can check exit temperature of the polymer with what is available to us at the moment.

DocTrucker

@Phaedrux said in External perimeter speed?:

Some of the filaments I use don't like to change speed very often, so print speed is held constant. Polyalchemy elixir for instance.

What are the symptoms of a polymer that doesn't like to change flow rate?

OBELIKS

I use PrusaSlicer, so usually it maxes out at 15mm3/s on PLA. There was some discussion that 11mm3/s is more reasonable, but that usually kicks in only on infill where small filament underrun is not critical. If I try Volcano nozzle I can go up to 25mm3/s.
And for perimeter speeds I went with 60mm/s for inner perimeters, 30mm/s for outer (visible) and 20mm/s for small. Infill is at 80mm/s, but it can be upped to 100mm/s for non solid infill.
As far as I can see S3D is tricky to set for bigger nozzles, because of missing volumetric speed cap. Not sure about Cura.

DocTrucker

I'm using Slic3r. There is a volumetric flow rate option but I like to be able to have a little more control over the process than giving it all up to the slicer, at least until I understand more fully the decisions it is making, and can agree or disagree with it! It does also have a flow rate levelling feature, but that really stuffed up a few revisions ago under certain circumstances leading to intermittent layers being processed really slowly.

I will investigate other slicers at some point but I am not comfortable with aligning too closely with a slicer that is closely tied to a specific manufacturer rather than seeking their drive from a community or independent leader. I haven't looked for slicers for some time.

I'm assuming the really high print speeds that are occasionally mentioned (150mm/sec+) only really come into play when you run with much lower layer thicknesses, or running a long heat zone hot end (e.g. volcano) with a comparatively small nozzle - something that I want to try for TPU once I'm happy that I understand my test bed well enough!

OBELIKS

You can still control, the volumetric speed is only the maximum, so yo do not make a mess. That way you v can set infill speed to let's say 200mm/s and it will still be OK

DocTrucker

@OBELIKS yeah I appreciate that thanks.

deckingman

@DocTrucker said in External perimeter speed?:

.............................. I'm assuming the really high print speeds that are occasionally mentioned (150mm/sec+) only really come into play when you run with much lower layer thicknesses, or running a long heat zone hot end (e.g. volcano) with a comparatively small nozzle - something that I want to try for TPU once I'm happy that I understand my test bed well enough!

I'm a bit tired of saying this but if you want to print fast, best of all is to use a mixing or multi input single output hot end such as a Diamond. For two reasons. Firstly you have multiple melt chambers so vastly increased surface area through which the heat can be transferred to the filament. Secondly, each filament passes through each melt chamber at fraction of the speed of a single melt chamber/single filament design. So for example, with a 3 input hot end, with mixing ratio set to 33.3% each individual filament takes 3 times as long to pass through it's respective melt chamber. So greater time in contact with a greater surface area = greater throughput.

With a 5 input diamond and mixing ratio set to 20% for each input, using a 0.5mm nozzle and 0.3mm layer height @ 195 deg C, I managed 240 mm/sec before any signs of under extrusion. Which equates to a flow rate of around 36mm^3/sec. Bumping the temperature up to 210 deg C allowed me to print at 300mm/sec which equates to a volume flow rate of 45mm^3/sec. And with a great big 0.9mm nozzle and 0.6mm layer height, I managed 100mm/sec at a flow rate of 54mm^3/sec. All with a single 40 watt heater cartridge I might add.

Write up here https://somei3deas.wordpress.com/2018/10/14/real-3d-printing-at-high-speeds-and-even-higher-melt-rates-with-a-large-nozzle/ .

Accompanying video here https://www.youtube.com/watch?v=rUV5IZxfAxU&feature=youtu.be

droftarts

@deckingman said in External perimeter speed?:

I'm a bit tired of saying this but if you want to print fast, best of all is to use a mixing or multi input single output hot end such as a Diamond.

True, but then you're flinging a lot of weight around, which causes other problems... but you know that, and have fixed it, already!

Ian

DocTrucker

@deckingman said in External perimeter speed?:

I'm a bit tired of saying this but if you want to print fast, best of all is to use a mixing or multi input single output hot end such as a Diamond.

So in other words yes:

...or running a long heat zone hot end (e.g. volcano)

This thread is mainly trying to establish if the behaviours I am seeing on my machine are typical or if I still had some serious issues to overcome. For a 12V machine running on a v0.6 board I think this machine is approaching the limit on balance in returns to effort spent, and so I am happy with it.

I do intend to fire up the three way again at some point but I've concerns about swinging the weight of a direct extruder from the 8mm smooth rods at this stage which comes before my interest in colours as I have an interest in a TPU application. The rods are bouncy and I would like to swap to a v-slot extrusion before doing much more heavy head stuff.

Edit: My comment on ultimate speed was merely a curiosity. My main interest was main vs external/perimeter speeds when you are trying to get a consistent part finish.

deckingman

@DocTrucker I take your point - you are unlikely to be able to fully exploit high melt rates unless you have a sturdier frame. But the Diamond isn't a direct extruder - you still have Bowden tubes. And to get the same size melt zone you'd need a Volcano that is 3 times as long as a std V6. I've never compared the weight of a (super) volcano with a Diamond but I guess the Diamond would still weigh more because of the multiple heat sinks. But then again. one long melt chamber still has the filament speed disadvantage compare to multiple (parallel) melt chambers. And of course, you need multiple extrudes (in my case 5) which probably cost more than many people spend on an entire printer.

But surface finish is a function of both speed and temperature IMO. If you run high temperature to get high flow rate, then you really don't want to slow it down too much on perimeters. TBH, with those high speed tests I did at up to 300mm/sec, I didn't slow down the perimeters at all. The surface finish wasn't all that shiny but not too bad at all.

deckingman

@droftarts said in External perimeter speed?:

True, but then you're flinging a lot of weight around, which causes other problems... but you know that, and have fixed it, already!

Ian

What problems?

DocTrucker

@deckingman said in External perimeter speed?:

...Diamond isn't a direct extruder....

Direct extruder is a requirement of the TPU, I'll be running that into an E3D hotend on these machines.

@deckingman said in External perimeter speed?:

But surface finish is a function of both speed and temperature IMO. If you run high temperature to get high flow rate, then you really don't want to slow it down too much on perimeters. ...

Generally agree with you a few posts before yours External perimeter speed?. But definable need to slow a reasonable amount on my system. This is likely to be largely do the larger time the the polymer stays molten in your system. I suspect the polymer is there long enough in yours to reach a stable temperature, and hense why you have said in the past you are able to run (and indeed the 3 way did before I broke it down) at lower temperatures.

In a crude sense in my system the melt is out of control due to the very short residence time for the melt and the polymer melt temperature being largely effected by changes in process rates and hot end temp. At 230C if the head stays still it will leak a significant amount of polymer in a short space of time. Without the high temp the process rates need to be slowed down significantly to avoid patchy parts.

Essentially I'm talking myself into trialling a volcano soon with a small diameter nozzle!

Edit: Requirement for direct extruder comes from trial work I did on Duet 2 Ethernet core XY with direct mounted genuine BMG. It was utter crap before I used direct.

deckingman

@DocTrucker I think we must have been talking at cross purposes. You mentioned something about being concerned about swinging the weight of a direct extruder and I assumed that you were referring to the diamond - hence my mentioning the fact that it isn't.

And yes, filament is funny stuff. It's generally a poor heat conductor so one can get the situation where it is fully molten on the outside but only partially molten on the inside. Which is why I think that time in the melt zone is as important if not more so, as the surface area of that melt zone.

DocTrucker

@deckingman I meant to add a comment saying I think we are talking at cross purposes. To many things going on at once here!

Edit: I have said many times on here polymer is weird!