Reasonable expectation for print speed

334 views
Skip to first unread message

Joergen Geerds

unread,
Mar 18, 2012, 12:40:46 PM3/18/12
to ulti...@googlegroups.com
Thanks to Skeinpypy and slic3r, I have finally found the upper limit for extrusion on my Ultimaker: 15.6mm^3/sec

with some experimentation, and some bigger prototypes, I was printing at 125mm/s and 0.2mm layer height and 200C real temp and 0.5mm nozzle... since my infill becomes incomplete at that speed, I'm adjusting the extrusion with the M221 S125 in pronterface to get decent infill. pushing it to M221 S150 has the filament slipping, which means I have reached the max pressure setting for my setup. other extruder setups will allow for more pressure, but my setup is still somewhat in the original design spec (other than some bowden reinforcement at the printhead, similar to the owen bowden clamp, which UM should include by default into each pack when it's shipped).

So, in theory, finding the upper limit for your UM shouldn't be so hard anymore, and should give you an idea what your UM can do and can't do. it is very dependent on the temp setting, since you can push more PLA through the nozzle by raising the temp by 5C, but that increases your chance of more stringing, and finding the right balance between how many strings you can tolerate takes a couple of prints.

example:
200C and 0.5mm nozzle and 0.2mm layer height:
extrusion: 0.2mm x 0.5mm = 0.1mm^2 (simplified math, I know it's not a rectangle)
printing at 125mm/s: 0.1mm^2 x 125mm/s = 12.5mm^3/sec
adjusting extrusion to 125% before slipping at 150%: 12.5mm^3/s x 1.25 = 15.625mm^3/sec (or about 2.6mm filament per sec)

once you have that number, you can do your own math to see what is reasonable on your UM... and maybe it can help to explain why your filament is slipping on your UM (maybe because you are pushing your UM too hard, after you rules out all other factors).

(max extrusion per sec) / (layer height x nozzle diameter) = max print speed

For example, I should be able to safely print at 250-300mm/sec at 0.1mm layer height, but only 90mm/sec for 0.35mm layer height.

and please keep in mind that a 0.4mm nozzle can push less PLA per sec than a 0.5mm nozzle. and this also doesn't take cooling into consideration, another issue/problem all together.

Daid: it would be great to enter that number into the prefs of the next version of SPP, and calculate the color code warnings based on that. right now, the speed setting warns that 150mm/sec is unreasonable on a UM, but as we can see above, it depends on the volume/sec extruded.

maybe SPP could get 2 modes: normal and expert, showing 2 interfaces, depending what the user chooses. the expert interface would continue in the direction SPP is going wight now anyway (more settings, more control), while the "normal" interface would take a default UM setting into account (a 0.4mm nozzle has only 64% the opening of my 0.5mm nozzle, so a standard UM should be able to push 10mm^3/s), and maybe offer 3-5 presets, while showing the associated numbers. this would be similar to Paul's netfabb profiles, but taking the volumetric concept to the right/next level: 1. fast & rough (0.2mm height, 0.4mm nozzle, 125mm/sec) 2. nice&easy (0.15mm height, 0.4mm nozzle, 100mm/s) etc etc. maybe a UI with sliders would be interesting: i.e. one ranges from zero to the max mm^3/sec etc.

this would allow new users to get decent prints out of the box (together with updated instructions on the wiki pages), while still letting the users dive into the finer details of 3D print settings, since each print/object has it's own requirement, and this might help speeding up the learning process.

even better would be some awesome ninja math, and taking the viscosity of the molten PLA at certain temperatures into consideration, if you raise the temp by 5C, you get x% more max extrusion...

thoughts and comments are welcome.

Jelle Boomstra

unread,
Mar 18, 2012, 4:02:29 PM3/18/12
to ulti...@googlegroups.com
Awesome Joergen!

Maybe we can derive some more easy to remember benchmarks? The order of magnitude for grams/hour would be about 100 iirc, with your calculations I get 15.6 *1.25 = 19.5 mg/s = 70.2 g/h. That looks very much in the ballpark, as the 100g/h referred to a 0.8 nozzle.
For a 0.4 nozzle, that would mean around 45g/hour.
In meters filament, that would be...
0.5 nozzle:
15.6mm^3/sec/ pi*1.445^2 = 2.38 mm/sec filament into the extruder,  143mm/min or 8.56 m/hour.
0.4 nozzle:
9.98mm^3/sec = 1.52 mm/sec, 91.3 mm/min or 5.47m/h

off course, the two decimals in these numbers are not justified at all, way to little precision to warrant that...





--
You received this message because you are subscribed to the Google
Groups "Ultimaker" group.
To post to this group, send email to ulti...@googlegroups.com
To unsubscribe from this group, send email to
ultimaker+...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/ultimaker?hl=en
 
For information about the Ultimaker, visit: http://Ultimaker.com



--
Met vriendelijke groeten / with kind regards,
Jelle Boomstra

ProtoSpace | FabLab Utrecht

Make your vision come to life!

www.protospace.nl


Joergen Geerds

unread,
Mar 18, 2012, 5:46:53 PM3/18/12
to ulti...@googlegroups.com
Maybe it could be part of the initial calibration routine (although I think we should be able to derive a base value that is good for all stock UM's).
after the stepper calibration (moving 100mm through the stepper without extrusion), we could ask the user to heat up to a middle of the road temp (i.e. 210C), and push 20mm filament through the nozzle at a given speed (steps/sec or mm/sec), starting with i.e. 1mm/s, and ask the user "did you feel the filament slipping yet, Yes/No, if no, continue increasing the speed by 20%, and ask again.
That way you can derive a upper save value for the extrusion rate without having to do too much math, plus we can give the user a direct feedback, like, "congrats, your UM can print xx mm^3/s, or xx gr/h"... in an more advanced calibration mode, we can do the same steps, but ramping the temp, and get an even smarter slicer out of it.

Taylor Alexander

unread,
Mar 18, 2012, 6:09:08 PM3/18/12
to ulti...@googlegroups.com

That would be great, and would also help us test and verify improved extruder designs like the Greg's and MK6 based one I'm testing.

--

Daid

unread,
Mar 18, 2012, 6:13:05 PM3/18/12
to ulti...@googlegroups.com
"did you feel the filament slipping yet" that's pretty hard to observe for new users.

But I like the idea of a warning on setting the settings so that the extrusion rate will be to high for an average machine. (All warnings and errors can be ignored in SkeinPyPy, so you are always free to experiment beyond that)

Joergen Geerds

unread,
Mar 18, 2012, 6:24:09 PM3/18/12
to ulti...@googlegroups.com
On Sunday, March 18, 2012 6:13:05 PM UTC-4, Daid wrote:
"did you feel the filament slipping yet" that's pretty hard to observe for new users.
I will make a video tutorial about it when time comes.
 
But I like the idea of a warning on setting the settings so that the extrusion rate will be to high for an average machine. (All warnings and errors can be ignored in SkeinPyPy, so you are always free to experiment beyond that)
great that you are on board with this.

here is another idea:
instead of overstuffing spp with too many setup and calibration functions, why not create a third app that handles ALL the calibration things, and updates the FW, from setting up the PiD, to measuring the temperature (to get decent TC calibration), extruder stepper calibration, to the afore mentioned max extrusion setting. all those values can then be shared with the prefs inside SPP.

I don't want to start any feature creep in your dev, but maybe this is something to consider for V2

Joergen Geerds

unread,
Mar 18, 2012, 6:49:30 PM3/18/12
to ulti...@googlegroups.com
On Sunday, March 18, 2012 4:02:29 PM UTC-4, Jelle Boomstra wrote:
Maybe we can derive some more easy to remember benchmarks? The order of magnitude for grams/hour would be about 100 iirc, with your calculations I get 15.6 *1.25 = 19.5 mg/s = 70.2 g/h. That looks very much in the ballpark, as the 100g/h referred to a 0.8 nozzle.
I personally prefer standard units that are used elsewhere in the software: we talk about mm/sec everywhere, so it is difficult for a user to grasp what mm/min or m/h is in this context... on the other hand, the software could just give all 3 numbers as result, just to be nice to the whet-ware sitting in front of the computer.
I also thought Joris did push almost 1000 gr/h, or was he using an even bigger nozzle than 0.8mm?

And I agree, it would make more sense to put the "your print is using xxx gr PLA or ABS" into the foreground, since it's super easy to put a coil of filament on a scale to see if it's enough for the a, while I would leave it to a local BDSM club to measure how many meters of filament are left on a springy PLA coil from hell.

Taylor Alexander

unread,
Mar 18, 2012, 6:55:33 PM3/18/12
to ulti...@googlegroups.com

while I would leave it to a local BDSM club to measure how many meters of filament are left on a springy PLA coil from hell.

As one who frequents such clubs, I can tell you that only the harshest Dom would force their submissive to do such a task.

Taylor Alexander

unread,
Mar 18, 2012, 7:01:02 PM3/18/12
to ulti...@googlegroups.com
Of course I'm joking. Most Doms would enjoy putting their sub through such torture, haha. Thanks for the suggestion. ;)

Boman33

unread,
Mar 18, 2012, 7:06:20 PM3/18/12
to ulti...@googlegroups.com

I like your practical and scientific testing to come up with real performance numbers.  It is the first test that I have been planning to do once I get my UM up and running. 

If the test is run in air it will remove all the variables relating to layer thickness and movement speed.

We are down to:  mm^3/second vs. Material- Temperature- Nozzle size.

A second trickier test would be to plot the extrusion rate vs. layer height vs. print speed.  Presumably the backpressure will also be dependent on the print speed.  Moving very fast there is less back pressure than a slowly moving head.

Bertho

  ======================================

Joergen Geerds

unread,
Mar 18, 2012, 7:46:34 PM3/18/12
to ulti...@googlegroups.com
On Sunday, March 18, 2012 6:09:08 PM UTC-4, Taylor wrote:

That would be great, and would also help us test and verify improved extruder designs like the Greg's and MK6 based one I'm testing.

It would also help us trouble shooting some of the typical beginner problems.
If the standard UM has indeed a 10mm^3/sec limit, and a new user is trying to print a 0.3mm layer with a 0.4mm nozzle, the max extrusion speed is around 80mm/s (ignoring the fact that a .3x.4 extrusion is not a rectangle), going faster than that might be the signs of slipping extruders we've seen.

Joergen Geerds

unread,
Mar 18, 2012, 7:57:33 PM3/18/12
to ulti...@googlegroups.com
On Sunday, March 18, 2012 7:06:20 PM UTC-4, Bertho wrote:

If the test is run in air it will remove all the variables relating to layer thickness and movement speed.

the "extrude 20mm at x.xmm/s" test would be into mid air of course 
 

A second trickier test would be to plot the extrusion rate vs. layer height vs. print speed.  

Possibly, but one can take this calibration thing too serious... remember, the test is to find the upper limit, or the point of failure for a given UM. there is so much to print-speed vs layer height that depends on experience, practice, and most of all the model you are printing.
 

Presumably the backpressure will also be dependent on the print speed.  Moving very fast there is less back pressure than a slowly moving head.

it's the other way around, the faster you go, the more filament pressure you get, since that was the reason for my filament to start slipping at 125mm/s, 0.5mmx0.2mm and 150% extrusion.
slow printing is easy, even a makerbot can do that :-)

mastory

unread,
Mar 18, 2012, 8:20:14 PM3/18/12
to Ultimaker
> And I agree, it would make more sense to put the "your print is using xxx
> gr PLA or ABS" into the foreground, since it's super easy to put a coil of
> filament on a scale to see if it's enough for the a, while I would leave it
> to a local BDSM club to measure how many meters of filament are left on a
> springy PLA coil from hell.

I was just thinking about this today as my only current spool is
dwindling and I am starting for the first time to wonder if I have
enough media to do the job. I guess you could get an idea from the
modeling software, but it complicated by the fill percentage.

Boman33

unread,
Mar 18, 2012, 8:30:36 PM3/18/12
to ulti...@googlegroups.com

From: Joergen Geerds   Sent: Sunday, March 18, 2012 19:58
--à snip

Presumably the backpressure will also be dependent on the print speed.  Moving very fast there is less back pressure than a slowly moving head.

it's the other way around, the faster you go, the more filament pressure you get, since that was the reason for my filament to start slipping at 125mm/s, 0.5mmx0.2mm and 150% extrusion.

slow printing is easy, even a makerbot can do that :-)

-----------------------------------------------

Either I do not understand something or I did not explain my thoughts properly.  (Please remember I have never fired up my UM yet……)

 

For a fixed, unchanging extrusion rate of nnn mm^/sec and if we then start to print a thin layer, I assume there will be back pressure since the space between the nozzle tip and previous layer will impede the flow.  An extreme endpoint would be a stationary head and a puddle would form.  If the head is moving extremely fast there would almost be a pulling effect as the plastic is ejected from the nozzle.  The practical case is presumably in-between these extremes.

 

I do agree and understand that for proper printing the extrusion rate needs to go up proportionally to printing speed to get the required amount of material deposited.

Bertho

Joergen Geerds

unread,
Mar 18, 2012, 11:21:42 PM3/18/12
to ulti...@googlegroups.com
On Sunday, March 18, 2012 8:30:36 PM UTC-4, Bertho wrote:

Either I do not understand something or I did not explain my thoughts properly.  (Please remember I have never fired up my UM yet……)

For a fixed, unchanging extrusion rate of nnn mm^/sec and if we then start to print a thin layer, I assume there will be back pressure since the space between the nozzle tip and previous layer will impede the flow.  An extreme endpoint would be a stationary head and a puddle would form.  If the head is moving extremely fast there would almost be a pulling effect as the plastic is ejected from the nozzle.  The practical case is presumably in-between these extremes.

 Having played with my UM for some time now, and having thought about some aspects of it, 3D extrusion is a 3 dimensional thing, meaning the pressure after the nozzle will squeeze the plastic to the left and the right also (this is also the edge that forms the surface later). together with the residual heat of the nozzle tip, the plastic has many ways to go, and the pressure build-up isn't that great, unless you consider the first layer, where the pressure buildup is real. the complex pressure buildup that is actually the real problem is the fluid dynamics inside the nozzle, where you have PLA in a wild mix of temp gradients flowing. but the overall idea is to not have crazy amounts of pressure build up, since that could cause backflow (molten PLA wandering UP the brasstube), and having the extruder grind away the filament is in some ways a safety feature (maybe, hopefully, dunno), since the next safety layer would be the stepper missing steps, and that is difficult, since they are awfully strong (at least on my UM, but maybe I got overly careful since I had so many problems last year).
but once you get your UM, you will see :-)

Daid

unread,
Mar 19, 2012, 6:50:19 AM3/19/12
to ulti...@googlegroups.com
I've added a warning to SkeinPyPy, based on the 10mm/s max extrusion on a 0.4mm nozzle. If you have a bigger nozzle configured it accounts for it.
It's a warning so it can be ignored. It also reports the current "print speed volume" configured if you go over the max. So you can experiment with higher values.
Reply all
Reply to author
Forward
0 new messages