Slowing down to debug

The one, the only, the Atremis. Forum for the monster metal machine from SeeMeCNC
Post Reply
Posts: 180
Joined: Wed Apr 19, 2017 7:18 pm

Slowing down to debug

Post by wepollock » Fri Mar 09, 2018 12:29 pm

I just wanted to point out an experience I just had playing around.
In my parts bin I had a spare titan extruder. Just to play around I swapped in the titan extruder. I had to print a 6mm spacer because the shaft on the stepper motor is not milled down most of the stepper shaft. Then I had to adjust config.g

;M569 P3 S1 ; Drive 3 goes forwards (E0) original
M569 P3 S0 ; Drive 3 goes forwards (E0) titan

Then I had to calibrate the extruder;

;M92 E182.0:182.0 ; Set extruder steps/mm original
M92 E810.0:810.0 ; titan extruder steps/mm yielding 100mm extrusion

The net result was a "string bomb" on the prints using a simplify3d profile without any adjustments.

Simplify 3d outlines all the settings you are using in comments.

; extruderRetractionSpeed,6000
; extruderUseCoasting,0
; extruderCoastingDistance,0.5
; extruderUseWipe,1
; extruderWipeDistance,1
; primaryExtruder,0
; layerHeight,0.25
; topSolidLayers,3

I am asking the extruder to move 450% faster with the geared extruder.
I decided to tune very low on on extruderRetractionSpeed and then work up

This is an excerpt from a print job (although the printer prints out fluidly, each gcode step is distinct and sequential)

G1 X1.100 Y0.597 E0.3891
; infill
G1 X0.499 Y-0.499 F4000
G1 X0.000 Y0.000 E0.0524 F4200
G1 X-1.499 Y-1.499 E0.1573
G1 E-3.2000 F500 ; here is the extruder moving at 500 speed..
G1 Z2.515 F1200
; layer 5, Z = 1.715
; inner perimeter
G1 X49.500 Y3.500 F4000
G1 Z1.715 F1200

Guess what, at the lower retraction speed setting there is NO stringing whatsoever.
I can now tune upward from 500.
and around 1300 (in the geared extruder) I should hit the same revolutions as the ungeared extruder.

Perhaps this is where 24v might be an advantage; because it will allow more energy potential to fastly-push into the stepper motor vs 12v.

Anyway, slowing down (and working upwards) seems to be a great way to debug.

Posts: 437
Joined: Mon Mar 07, 2016 10:17 am

Re: Slowing down to debug

Post by dc42 » Sat Mar 10, 2018 10:17 am

It doesn't help that the motor in your Titan is 0.9deg, whereas a 1.8deg motor is adequate for a geared extruder. You might like to plug your figures into the motor speed/voltage calculator at

Posts: 180
Joined: Wed Apr 19, 2017 7:18 pm

Re: Slowing down to debug

Post by wepollock » Sat Mar 10, 2018 7:39 pm

DC thanks.. that is a great set of software tools...
I am really going to enjoy that link.

I am trying the geared extruder even though I realize that there are some advantages to direct drive. I mentioned the same to a cnc staff member who confirmed.

With direct extrusion my guess is that your speed can be higher and your distance can be higher; perhaps you might get less stringing risk with a direct drive)

With the geared extruder;
I am at retraction of 3mm retraction speed of 700 and overall print speed of 4200 (which is higher than where I started).
I may want to try a bigger zhop.
Because of the torque in 3:1 you might be able to use less current or a less beefy stepper.

Some very small advantages of the titan;
The printer is even more quiet - if that is belivable
And there is a tensioning screw on the titan.
There is more torque

What changes would actually be advantageous
I am trying the geared extruder simply because i have a spare.
12v vs 24v
1.7mm vs 3mm
Geared vs Non Geared extruder
Enclosed vs Open
Stock vs Aftermarket (btw I don't think the E3d hot ends will actually improve the stock printer)
Software vs Hardware (I do think that lots of things can be done in software with the stepper drivers and firmware; I prefer a hardware approach first and software approach second. )
I did change the bowden tube to a capricorn tube - mostly because I had some around; and the original bowden had some friction.

On another interesting note is a SCARA
IMHO I am much more impressed with a delta than a cartesian (owning a flashforge and a printerbot)
A fat delta would be great; but with a SCARA you can remove two towers.

PS.. today I embed a metal object in a print using duet pause; it worked like a charm.

User avatar
Posts: 1557
Joined: Wed Sep 23, 2015 2:55 pm

Re: Slowing down to debug

Post by Xenocrates » Sat Mar 10, 2018 10:41 pm

24V is usually better, because you can use smaller wires and less current (and also, it can drive steppers faster with a given inductance), reducing the moving mass (and copper usage) for a given heating capacity.
1.75mm filament is going to be better almost universally, because of it's higher resolution in an extruder (IE, the extruder can move a fractional degree, and with a given drive wheel size, moves a given distance of filament, which contains a given volume proportional to the diameter), and the better surface area to volume ratio. For instance, 3mm (2.85 more properly in most instances) filament has a volume of 6.38 mm^3/mm, and an area of 8.95mm^2/mm, or a ratio of 1.402mm^2/mm^3. 1.75mm filament has instead, an area of 5.5mm^2/mm, and a volume of 2.41mm^3/mm, and a ratio of 2.28mm^2/mm^3. As we can't change the thermal conductivity of a nozzle or plastic within a given class, in order to get the best coupling of the measured temperature to actual, as well as to allow the fastest extrusion speeds, we want that ratio, which effectively tells us the area getting hot divided by the material to heat, to be as high as possible. It also means a more homogeneous temperature in the melt zone, which is generally good.
A geared extruder is going to give more torque and resolution, but a lower speed with a given motor. Now, if the speed has to go up, the driving voltage could be increased as well, but this is generally harder to change in an assembled printer. If you mostly don't challenge the speed capabilities, a geared extruder can be a good choice. However, the maximum torque is set by the material after a certain point. Beyond that, the filament will strip, instead of the motor skipping. So the biggest gains are in resolution, but the motor can deliver that just as well, and usually with fewer moving parts and a more compact design.
An enclosure helps control the printing environment, and makes the machine safer to work around. My Rostock lives in a workshop with little actual heating, so the enclosure is necessary to actually print in the Michigan winters. It also keeps sawdust and other contaminants mostly off the printer (admittedly, it isn't perfect. I had to clean the E3D heatsink recently). It also helps keep the temperature up to reduce ABS warping and cracking. On the other hand, it also traps heat that might be better vented if I were printing PLA in the summer. It also constrains my choice of fans somewhat severely, since I regularly print to near the edge of the print area. I've got another printer that's a dedicated PLA machine, and it has an enclosure I'm glad about, since it lives in the house with the cats and occasional small children, but it has a pair of 80mm fans to exchange the air inside regularly, and is much larger than the build area. Done right, there's no reason not to have an enclosure other than cost.

I wouldn't pick SCARA for 3D printing as a geometry. It's intentionally got play in the Z-axis, and while fast, thanks to the lever effect, you can potentially have a very variable resolution in the work area, even beyond what a delta has. The pathing is also neat, but also kinda a bitch, because there are some locations that you need to be in either a right or a left handed configuration to hit, and swapping between the two makes for a very non-trivial acceleration profile if you try to do it while also following the path at a given speed. As it scales up, you end up needing closed loop control of the motors, as the second one will throw around the first section of the arm, and again, the sag will be partially dependent on position, as the effective lever of the arm on a given joint changes as it moves, making for a very interesting bed map. That all said, Nicolas is a brilliant guy, and while I would not pick the Helios as a design to emulate given my experience with industrial SCARA machines (largely ancient Adept robots), if he can make it work, and I bet he can with enough effort, the largest quality of life issue would be the weirdly shaped workspace.
Rostock Max V2, Duet .8.5, PT100 enabled E3D V6 and volcano, Raymond style enclosure
Automation Technology 60W laser cutter/engraver
1m X-carve router

Sic Transit Gloria Mundi

Posts: 437
Joined: Mon Mar 07, 2016 10:17 am

Re: Slowing down to debug

Post by dc42 » Sun Mar 11, 2018 5:27 am

I favour 1.75mm filament and extruders with about 3:1 gearing. I've always used 1.8deg motors to drive them, not the 0.9deg motors that E3D suggests for the Titan; however it looks like with 12V power and 0.9deg motors, around 50mm/sec should be achieveable, which is probably sufficient.

I blogged about my experience with a cheap SCARA 3D printer at ... d-printer/ and the following two posts. SCARA is interesting for its large build area, but its cantilevered nature and propensity to backlash make it in other respects less suitable for 3D printers. We were originally going to build a Helios, but Nicholas appears to have stopped working on the design.

Post Reply

Return to “Artemis”