This topic contains 43 replies, has 11 voices, and was last updated by 
Martin Sloth Hansen 4 months, 1 week ago.
-
Posts
-
We reduced speeds significantly in previous releases of AutoMaker to concentrate on quality, specifically the nozzle control which is unique to Robox. In the aim to make Robox easy to use and reliable many settings are tuned down to ensure flaws are not introduced or reliability reduced.
To take part in this experiment you will need to have set safe mode to OFF.
The simple way to increase speed is within your Filament profile, increase the
Feed Rate Multiplier
This is typically set to 1.0 but a setting of 3 or 4 is acceptable, it will do as it says, Multiply the Rate of everything….not ideal but very simple. If you are changing the speeds elsewhere please keep this setting close to it’s default, normally around 1 to prevent confusion.
More control is offered in the Print Profile under the Speed tab (looks like a car speedometer), you can alter the movement speeds in a lot of detail.
Click for bigness.
Edit these values with care, increments of 5mm/s would be acceptible, print the same simple parts each time to try and find the best setting.
Increasing the movement speeds will effect quality but this can be tuned with another settings which is in the firmware but accessible via a gcode command.
Enter the following in the gcode box and hit enter or press the Send GCode button. The default values for AM 102.00 are displayed in the image above
M503
You will see a lot of information, look for
Acceleration: M201 S12
This M201 command is setting the acceleration to a speed (S) of 12. If you have increased your movement speeds you may find the movements are jerky and violent. Reducing this M201 setting to S6 will slow down the changes in direction but should allow some higher speeds without sacrificing quality.
To alter the setting in the GCode entry box type this and press enter after each line
M201 S6
M500
M500 saves the setting.
So after typing M503 again the setting will look like this
Acceleration: M201 S6
Please note that the settings you make to the firmware will be overwritten with a firmware upgrade. You can copy the M503 result text from the gcode console and save it for reference. Each software version has a matching firmware version, if you reload the same software version the firmware will not be changed.
-
This topic was modified 1 year, 10 months ago by
Pete. Reason: altered "Feed Rate Multiplier" mistake
For official support please visit www.cel-robox.com/support/ and create a ticket@pete This is great. Exactly what is needed to bring the community together.
A few weeks ago I put together a new profile based on Normal but with UM2 speeds. The acceleration gcode could be what I was looking for to get this to work.
I will test it in the morning and let you know the results.
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.comOk so I jumped right in, drafted a simple STL and went for a fast version of the CEL Normal profile using figures based on what the “Ulti”/fine on UM2 does.
Attached is my profile and a comparison of the standard CEL “Normal” print compared to my “UM2” profile print. The time improvement was 27 mins to 21 mins but I expected more. Then I realised why it wasn’t more, the two columns were not affected because I believe they are controlled by the areas greyed out such as “Small perimeters” (I am sure there is a good reason they are not editable).
Also note that I have not yet changed the acceleration settings in the gcode.
So as I am restricted with the speed changes in the profile I am now going to work on the Feed Rate Multiplier. I will report back later.
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.com@simonhoult It’s quite easy to check if you are being restricted by any un-editable settings by comparing the Gcodes.
You can find the G codes in My documents/CEL Robox/printjobs
You should now be greeted by a long list of folders - if you sort by time and date newest first, you will be able to select your prints with the most recent being at the top.
If you work out which print job folder is for which print, you can then retrieve the G code from the folder and use a txt comparison program such as winmerge to see if the Gcode differs. Checking the .roboxgcode is probably best as this will account for any changes made by the post processor.
If you can dig these two G codes out I’d be really interested to take a quick look at them!
S
Well that little test was exhilarating and the Robox enjoyed stretching its legs.
As you can see on the attached I increased the feed rate multiplier and compared the standard 100% with 200%, 300% and 400%.
I found it very interesting that the top layer finish got better when the speed increased! Also I found it strange that slowing the acceleration by GCode actually made the prints worse.
However increasing the speed with the feed rate multiplier is not ideal as it speeds up everything including the first layer which I think screwed up some of the prints as it was too fast.
@pete I would like to work on this in more detail. Is there any way to edit the greyed out boxes in AM so I can adjust different elements of the print profiles?
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.com@pete Also would I be right is saying that changing the Feed rate multiplier to 2.00 would literally double the speeds shown in the print profile?
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.com@simonhoult You wrote << Also I found it strange that slowing the acceleration by GCode actually made the prints worse. >>
This is due to the inertia of the plastic material flowing out from the nozzle. You can’t control the extrusion rate so quickly and precisely to compensate for the movements of the Head that has to continuously change speed. If the speed was constant then you would get a nice, constant flow deposited on the previous layers, but as speed (see it separately, X and Y) cannot be constant due to the need to change direction (especially on corners), you cannot get constant flow out of the nozzle.
The best acceleration, quality-wise, is the highest allowed by the hardware without incurring in missing steps. By reducing the acceleration, if the extruder rate is properly compensated (taking X/Y acceleration into account) you’ll get thicker extrusion during decelerations, and thinner extrusion during accelerations anyway due to the inertia of the pressure build up in the chamber. You can’t avoid this. This because the change of pressure in the nozzle chamber can’t cope up with the change in speed of the Head. It’s inherently a slower-reacting component.
That’s why a lighter Head is always a good thing, it lets you use higher X/Y accelerations without incurring in missing motor steps.
I am quite excited by my results. I have marked all my prints underside so for a bit of fun jumbled them up and put them in order of quality to my eyes. I did this 3 times and got the same result each time….
1) 200%
2) UM2 Profile (not complete)
3) 100% (standard Normal)
4) 300%
5) 400%
6) 400% M201 S6
7) 400% M201 S3Obviously this is simple model but I am optimistic.
-
This reply was modified 2 years ago by
Simon.
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.com@siymon The printing of the columns were not changed by my profile. I could see it by eye, the speed wasn’t effected.
If I was able to access the greyed out boxes in AM profiles I feel I could put together a pretty fast profile whilst retaining good finish.
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.comI will try to rephrase it to make it more clear: imagine you want a nice, constant flow out of the nozzle, that is what in the end gives you quality.
To make things simple, let’s build 4 walls, like a square (as seen from top).
Now, you start from the bottom left, then you move that nozzle to the right, and when you arrive to the extreme right, you gotta turn the direction up.
If you have infinite acceleration, then no problems (approximately, as the round cross section of the nozzle causes some too, but they’re negligible in this case).
But now imagine your Head has a lot of inertia, and if you stepped with infinite acceleration, you’d get missing steps. Of course this must be avoided, else the print will be completely ruined.
So you set a certain, maximum allowed acceleration. For ease, let’s imagine this acceleration is really low.
So now, re-imagining the same scenario of before, you start from the bottom left corner, with speed 0, and you must accelerate to your target speed, while moving on the right. Let’s say that in 1/4 of the length you manage to reach top speed (respecting the acceleration limits), now you have to be alert that 1/4 of the length of the segment before the rightmost limit, you must begin to slow down, respecting the (de)celeration limits you got imposed. Then you’ll turn the nozzle in the up direction and in 1/4 of the segment length you’ll increase speed again to your target (profile) speed.What happened in that corner? The pressure build up in the nozzle chamber is quite slow, so the plastic has kept on flowing at the same rate although the Head was slowing down (to respect the positive/negative acceleration limits). This has caused over-extrusion.
If the system is not compensated, there will be over-extrusion even when the corner is gone and the speed is getting built up again. I believe this is the case, from your pictures on the right.
This is the reason why you get better extrusion consistence (especially on corners) when using slower speeds, and/or higher accelerations. As I already wrote elsewhere, if you are printing an object with sharp corners, you should avoid high speeds (this not only for the problem we’re discussing, but also for X/Y resonances), but if you’re printing a “nicely round” object, you can increase printing speed safely.
With a compensated extruder software module, you could take this problem into account, and realize that you have extruded too much material, because the Head hasn’t changed speed instantly. So you gotta extrude a bit less, after the corner.
With a pre-compensated extruder software module, you could take into account everything and predict when you gotta decrease pressure in the nozzle chamber, so that the pressure (and the corresponding flow, which is proportional) starts to be reduced just when the Head is starting to slow down, and the two things go together, thus you get constant flow/extrusion_rate.
This will be a little tricky, but it’s possible. I have done that but I have my own “slicer” (getting 3D models not from STL files though, but from Voxel ones).
@omega64 Thanks for the reply. I do understand the theory (I dabble in Aerospace Engineering) but I wrongly assumed that AM, with the help of the needle valves and accurate extruder would compensate for this.
I suspect although it isn’t automatically compensated something must be happening in the existing setup to vary the flow during acceleration/deceleration already?
So reducing the acceleration in the GCode was never going to have a good effect, apart from making sure Robox didn’t shake itself to pieces? Actually it wasn’t as violent as I expected, although even more noisy than usual.
STEELMANS 3D PRINT.
3D Hub: www.3dhubs.com/service/steelmans
Website: www.steelmans3d.com@simonhoult You are right: reducing acceleration will reduce vibration, shake, etc.. especially at high speeds, but it has a detrimental effect on quality.
Aerospace Engineering? Now that’s a cool thing. There must be a reason why they call it “rocket science”. 😉
-
This topic was modified 1 year, 10 months ago by
You must be logged in to reply to this topic.
