Pancake stepper extruder

What's the smallest motor possible for an extruder?

Looking at the E3D titan, some people are claiming 18Ncm nema 17 motors, which weigh around 140g. Would this provide enough push on a 1.75mm filament? I'm dubious that this will do it, but my only experience is with my direct drive extruder, running a pretty beefy 59ncm motor. I'm pretty sure this is too much, but I've not had any problems so far apart from the weight.

Anyone had any experience with running these small pancake steppers on an E3D titan?

... the size depends on possible speed and gearbox-ratio - best/smallest would be a small DC-motor with encoder ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I an running an E3D Titan with a pancake motor and it is working great. You need to take into account that it has a 3:1 gearing so the torque it can apply is greater than the motor alone. You get the added torque at the cost of speed. But for an extruder, speed is not a major factor.

Quote
Origamib
What's the smallest motor possible for an extruder?

Looking at the E3D titan, some people are claiming 18Ncm nema 17 motors, which weigh around 140g. Would this provide enough push on a 1.75mm filament? I'm dubious that this will do it, but my only experience is with my direct drive extruder, running a pretty beefy 59ncm motor. I'm pretty sure this is too much, but I've not had any problems so far apart from the weight.

Anyone had any experience with running these small pancake steppers on an E3D titan?

I run a NEMA11 with a 5:1 gearbox [www.omc-stepperonline.com], it has done 1.75mm filament just fine. Coming up I have some of the Sanyo 11.6mm NEMA17's (70g each) that I'm going to run with 4:1 gearing as they have greater torque out of the box than the 11. I don't expect any issue.

-edit to clarify, I don't run the geared stepper into the titan, but the results are still relevant

Edited 1 time(s). Last edit at 06/08/2016 12:07PM by Koko76.

Seems that 2-3kg of force is enough to push the 1.75mm filament through the regular 0.4mm nozzle. I have a bowden setup with a similar large NEMA17 with Mk8 gear running at half its potential (i.e. 1A instead of 2A) to keep it cool. That's the reason for the current attempt to do tiny DC motor driven extruders: [forums.reprap.org]

Quote
Koko76

Quote
Origamib
What's the smallest motor possible for an extruder?

Looking at the E3D titan, some people are claiming 18Ncm nema 17 motors, which weigh around 140g. Would this provide enough push on a 1.75mm filament? I'm dubious that this will do it, but my only experience is with my direct drive extruder, running a pretty beefy 59ncm motor. I'm pretty sure this is too much, but I've not had any problems so far apart from the weight.

Anyone had any experience with running these small pancake steppers on an E3D titan?

I run a NEMA11 with a 5:1 gearbox [www.omc-stepperonline.com], it has done 1.75mm filament just fine. Coming up I have some of the Sanyo 11.6mm NEMA17's (70g each) that I'm going to run with 4:1 gearing as they have greater torque out of the box than the 11. I don't expect any issue.

-edit to clarify, I don't run the geared stepper into the titan, but the results are still relevant

That's pretty good to know actually! What is the weight of your extruder+motor with that set up? Using the smallest nema 17 motor on omc-stepperonline and a titan extruder, 200g is possible (although I'm told it is actually around 170-180g as their website is wrong for weights. This gives 13Ncm torque, although more with the titan extruder (possibly 30Ncm? not sure on the efficiency of the titans gearing)

I have found their website is wrong about a lot of things, the motor/gearbox I listed weighs almost double the listing (close to 200g). All up my motor/gearbox extruder block (this serves as frame for all that stuff, tension mechanism, and water cooler for the hot end) as well as the heater block itself is right around 300g. That was for a quick test of some concepts, so I didn't look to shave every gram. With the 70g pancake I think I can get to under 200 for everything pretty easy.

20mm nema 17 direct to an E3D titan, 0.3mm nozzle 1.75mm filament - no problems at all. Total extruder mass 200g,

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

This is with a nema 14 pancake 7Ncm with a 1:4.3 reduction
[www.youtube.com]

I will not advise this setup, because is skips steps. But with the nema 15 12nm pancake is works perfect

---

P3steel DXL, with Due/RADDS/Raps128 dual Wade's extruder

Would it be worth looking at smaller nema motors, say nema 14, and then printing a plate to convert it to an E3D titan? It's hard to say if this would be worthwhile, as its hard to find exact weights of the motors online. I suppose it would be better to use a geared nema 14, but this means relying on possible inefficient gearing (But how good is the titan I wonder?) and its harder to mount the hotend, so would require extra weight for a printed extruder/mount

[www.pololu.com] Is the 70g nema17. They have a 14 pancake as well. There is also a thicker 17 pancake there.

In my experience, you need about 3:1 gearing to get enough extrusion force using a 20Ncm motor and a RepRapPro hobbed insert; or to put it another way, with a standard 1.8deg motor the product of full steps/mm and torque in Ncm should be more than 500. From this you can work out what gearing you need for less powerful motors.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
In my experience, you need about 3:1 gearing to get enough extrusion force using a 20Ncm motor and a RepRapPro hobbed insert; or to put it another way, with a standard 1.8deg motor the product of full steps/mm and torque in Ncm should be more than 500. From this you can work out what gearing you need for less powerful motors.

The OMC Stepperonline motor I linked above has 6Ncm of torque listed, and I've run it for some time now at 5:1 with no issues, that would be roughly half of what you are saying. I do use 1.75mm filament and I don't print crazy fast. I'm also using a .3 nozzle. Not trying to say you are wrong in your assessment, but this setup has given me some great prints.

@Koko76
6Ncm x 5 x 200steps/mm=600
so it fits Daves rule of thumb?!

Quote
o_lampe
@Koko76
6Ncm x 5 x 200steps/mm=600
so it fits Daves rule of thumb?!

Yes, but 20 Ncmx3 is 60Ncm and he is saying he needs that, which is double he number in the rule of thumb. My question is why if he rule of thumb is half of what he said is required.

1. The steps/mm in the formula takes account of the gearing and the hobbed insert diameter. So don't multiply the motor torque by the gearing ratio, just multiply it by the steps/mm.
.
2. I am not saying that you can't extrude with less torque than my rule suggests, but in my experience you need that amount for reliable extrusion in a range of situations and with as range of filaments. My rule also allows about 25% margin, because you want a little spare torque and you don't want to run the motor at full current. You can increase temperature to allow extrusion with lower torque, but that generally increases stringing when printing PLA.

Some examples:

Direct drive extruder with 100 microsteps/mm @ x16 microstepping and 60Ncm motor: 100/16 * 60 = 375. Below 500 so marginal.
Ormerod extruder with 420 steps/mm @ x16 microstepping and 22Ncm motor: 420/16 * 20 = 525. Above 500 so OK.

Edited 1 time(s). Last edit at 06/10/2016 09:56AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Ok, so I have approximately 600 steps per mm, 16 micro steps. That's 37.5x6Ncm or 225 by your formula. I have done many 10+ hour prints with abs and pla of a variety of manufacturers/temps etc. I would characterize it as reliable. I'm trying to understand why you need or want more than 2 times that number.

So 270steps/mm 1/16 stepping 13Ncm = 219.

However no problems printing with abs and pla. Sure there's not much headroom I'd accept that and the motor runs warm.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
Koko76
Ok, so I have approximately 600 steps per mm, 16 micro steps. That's 37.5x6Ncm or 225 by your formula. I have done many 10+ hour prints with abs and pla of a variety of manufacturers/temps etc. I would characterize it as reliable. I'm trying to understand why you need or want more than 2 times that number.

Maybe it's because I run a delta that does PLA infill at 100mm/sec using a 0.4mm nozzle and I don't like to compromise on extruder temperature. Pick a figure that works for you.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
Koko76
Ok, so I have approximately 600 steps per mm, 16 micro steps. That's 37.5x6Ncm or 225 by your formula. I have done many 10+ hour prints with abs and pla of a variety of manufacturers/temps etc. I would characterize it as reliable. I'm trying to understand why you need or want more than 2 times that number.

Maybe it's because I run a delta that does PLA infill at 100mm/sec using a 0.4mm nozzle and I don't like to compromise on extruder temperature. Pick a figure that works for you.

And I run a smaller nozzle, so higher back pressure. Pla at 190c. Fastest I'm going in infill is 80 with a layer height of .2 (or less). I still don't see the doubling of the number being needed. I'm genuinely curious as to why we have different results, telling me "pick a number that works" isn't awful scientific. Furthermore if I saw that someone else was getting better performance I would be curious to know why.

Like many other 3D printing related things, there are too many variables with no meaningful measures on most of them. There's no formula that works for all extruder/motor/hot end/nozzle geometry/filament material combos. The best you can do is err on the side of too much torque, etc. If you push the limits of performance expect to have failures.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Why not locate the motor somewhere else?

Connect the motor via flexible shafting (drive cable like they use in car speedometers) and use a worm gear to drive the hobbed gear in the extruder. That would allow for both a small motor and very light extruder. The one unknown for me is just how much lash is inherent in flexible drive cable.

Well this is not a new idea since the flex3drive product from Mutley3D does exactly this and has been around quite a while now. I've got one, and whilst there is some lash/backlash in the flexdrive itself its geared 40:1 with a worm gear so this has far less effect on the drive gear in the extruder than at a lower gear ratio.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

The flex3drive sounds great, I don't think I'd ever go for the speedometer cabling though as it's not a particularly tested method. The simplicity of direct drive beats weight savings in my opinion, and hopefully even lighter extruders will continue to come out that rival Bowden / flex3drive.

An update to this thread though, the 20mm Stepper works great! I've used abs, PLA and ninja flex through it so far with no problems at all. Any withdrawals I had about this set up are all gone. Next I'll be trying even smaller motors to see what is possible.

Yeah I wonder whether we've become accustomed to using a 40mm nema 17 for extruders as it was easier to just order 4xmotors and that type are the most common and least costly and 3mm needed higher torque.

Modest speed machines i3's etc were not capable of high speeds due to other limitations so lightweight print heads were a secondary consideration. With delta/corexy speed is not secondary and I reject the argument people make about quality being all that counts, it's important but surely the aim is high speed and high quality. Now lightweight print heads are highly desirable.

I like the efforts going on to make a really light direct extruder but at 250g a titan and 20mm nema 17 is reasonably priced, performs superbly and is reliable. The cycloidal drive extruder looks good but reliability looks difficult. 100g, reliable and less than $100 and now you're talking.

Edited 1 time(s). Last edit at 08/13/2016 04:40AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

dc42 - Interesting rule of thumb. If I apply that to my Wade extruder 747 steps/mm 9:47

required motor torque = 500 /(747/16) = 10.7Ncm

which could be easily be handled with a NEMA 14 or NEMA 17 pancake ~120-140gr

So if for nothing else than to reduce overshoot, I should be replacing my 40Ncm 285gr NEMA 17 on my existing Wade with a NEMA 17 pancake.

It's not a common problem for a wade extruder gear to wear out or break. So the Wade could be put on a diet. Gears also printed in nylon. Lower ratio. Lower gear module. (I print functional servo gears down to 0.5 module).

Seems like the advantage of the Titan is that it is small, light weight, and off the shelf....but where's the fun in that? I think we're all getting lazy and drinking the Cool-Aid. I've got this 3d Printer thingy sitting in the corner of my basement...

Has anybody seen a pancake stepper, 3d printed gear, direct drive/non-bowden extruder for a core XY like a d-Bot?

Quote

Has anybody seen a pancake stepper, 3d printed gear, direct drive/non-bowden extruder for a core XY like a d-Bot?

What's the difference in terms of "fun" to print another guys design instead of buying a Titan extruder?
Actually I ordered a Titan for my D-Bot and hope to get rid off the Bowden tube. The idea is to mount it with the clear plastic cover towards the hotend-carrier.

Hmm this got me thinking, I think my extruder is on 16th steps, its working fine but maybe if I changed it to full or 1/4 it might be a little better.

It will effectively have more torque on microstepped positions and about the same on full step positions. It'll need few step pulses to run, and if your extruder is (highly) geared or you're running 0.9 degree motors it wont really matter in terms of resolution. I don't think I'd do that on a mk8 it would be at 25 steps/mm!

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
o_lampe
What's the difference in terms of "fun" to print another guys design instead of buying a Titan extruder?
Actually I ordered a Titan for my D-Bot and hope to get rid off the Bowden tube. The idea is to mount it with the clear plastic cover towards the hotend-carrier.

What is your experience with the Bowden? I'm still building my "D-bot" and struggling with the ideal of using a Bowden feed. On the other hand, I am targeting a sustained chamber temp of 60C+ which creates challenges cooling an extruder stepper. I just checked and pretty much the only thing my machine has in common with D-bot, is that it is a roughly a cube.

But the D-bot design was a great starting point and enabled endless quantities of "FUN". Many kilos of proto parts. Did you consider mounting the Titan above the X carriage and using a short PTFE to the E3D? Or is there another way to not lose X-Y travel?

I love what E3D has done for our way of life, but I really struggle hitting the "submit order button" on a genuine. Mfg cost is about $14 + their tooling amortization and engineering which they are totally entitled to....but fair market price should be ~$40. I do plan to pickup a genuine for reference if something shows up for Black Friday. The Titan will probably find its way into my machine as a stop gap, just to get it printing.

My big takeaway is the std Wade drive mechanism is over powered (from dc42's rule of thumb) and overly robust, which should create opportunities for a bucket of FUN. I don't think we've hit the limit on reliable 3D Printed gears.

(see next post - NSTT)

Edited 1 time(s). Last edit at 11/19/2017 11:21PM by to-the-nth.