Extruder force/gear ratio choices

I'm looking at extruder designs and am confused: the available drive strengths are all over the map. I mean, the classic Wade extruder geneology use gear ratios of about 4:1 with a final 8mm bolt. But my Prusa i3 MK2 has it connected directly to the (apparently ungeared) NEMA 17 stepper with a pulley that is only slightly smaller than 8mm. And it works great, of course.

How much drive force is really required? Are all those big Wade gears really just needless, or is there another trick at play on the Prusa? Has anyone done any real world measurements of how much drive force is required for different plastics at different speeds and temperatures?

I think it has to do with some of the larger 3mm filament. I believe it takes a lot of force and torque for 3mm PLA. Most of this is negligible with 1.75mm filament, but different filaments do take more or less torque than others. I imagine, people just use the geared extruders to either use smaller motors (NEMA17 pancake steppers or NEMA14 motors) to save weight.

That's a good point. 3mm is going to produce ~3x less pressure in the melt chamber for the same drive force (ignoring friction), and the Wade extruder was originally designed in the 3mm era (where I was looking at printers, but not building any). But the Prusa claims to work with 3mm filament too, and again does it with essentially the same motor and 4x less torque from the gearing.

I'm just struck by the mismatch and the fact that the "best practices" designs don't seem to agree. If I were to design my own extruder, how should I gear it? I'm hoping the answer leads to someone's blog post where they checked out PLA and ABS extrusion rates and determined the point where the motor started missing steps. I could work backward from there to a torque.

It would be interesting to talk to the folks at E3D and figure out why they designed the Titan for 3:1 ratio. I know they had said something in terms of being capable of having enough force and being able to keep up with feedrates for the Volcano nozzles.

I couldn't get 3mm to extrude reliably with an un-geared NEMA 17, so I switched to a Wade style with I think 4.9 to 1 ratio. That works with NEMA 14 steppers, but they're running flat out to do so.

I use a figure of about 2 kg force as the "must have", and above 4 kg as the "never actually need that much force".

So you can juggle motor torque, gear ratio, and total weight to get a good compromise. Note that motors are usually specced at currents that would end up too hot for my taste, so I derate the specced torque by 50% to get the extruder motor to run cool, and still have enough push.

Might help you to get in the ballpark.

btw I've been looking for 6 months to get the correct sized stepper and gear ratio, to put into an MK8 style extruder assembly. I ended up wanting about a 3:1 - 5:1 planetary gear ratio, and a thinner style NEMA 17 motor. I was about to buy one of those , when I came across the Titan -- and instantly bought one. I'm very happy with it.

Their 3:1 gear ratio is perfect, the gears are a lot lighter and thinner than a planetary gearbox, and it's an extremely well designed piece of kit. The filament path through the extruder is very well thought out. You pull the lever and push the filament through. It just goes straight through, no poking around at all. I haven't tried a long print with flexibles, but I think it should work really well.

Just for reference, the numbers above are for 1.75mm filament. I had one of the direct drive systems with a 40mm NEMA17, but I found that the force just wasn't quite enough to print fast. I used to be a believer in direct drive extruders, but now I think that some form of gearing is a good thing.

The moment you throw the Remote Direct Drive idea into the mix it gets even more complex.
For instance, we use a 30:1 reduction, mainly due to the fact that we use a worm gear.

We measured pushing forces of around 5 to 6 kilo's with a normal Nema17 stepper. I am now using a little Nema 14 stepper with the amps turned down low.
As for speed, we feel we are in the right speed range and there are others that go up to a 40:1 ratio, so there still is some leeway.

To me, the fact that so many extruders use different solutions and different torque ratios means that we are nowhere near optimising to the perfect settings.
For a long time I have wanted to build a load cell into the extruder to see what the actual pressures are, while printing, but never got round to it. Now that I am working integrating a piezo sensor into the extruder, for bed calibration, I might be able to find a way to include a load cell. The piezo is not really ideal for measuring pressure, only for pressure changes.

Edited 1 time(s). Last edit at 02/23/2017 04:25AM by Lykle.

---

Lykle
________________________________________________

Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]

My take on it having used titan 3:1 (current) for 1.75, mk8 ungeared for 1.75, printed extruder 5:1 planetary for 1.75, flex3drive remote drive 40:1 for 1.75mm and wades (11:45) for 3mm is that a little gearing helps both to provide enough torque (necessary for 3mm) and to smooth out the pulsing that occurs with ungeared extruders at low flow rates, especially with 1.8deg motors. It also allows for cooler motors, or smaller motors and less current and heat generated.

The other argument is about retraction speed. Here I disagree with the received wisdom that retraction has to be fast to work. It doesn't. My flex3drive (40:1) can only retract as I currently have it setup at 11mm/s, but I don't get blobs or strings. However if I start turning the speeds up, then the proportion of the time spent retracting and unretracting is longer. So fast retractions (using low or no gearing) are important with faster print speeds so you can keep the speed up but not so you can actually retract the filament. Of course I'm only retracting it 2mm so at 11mm/s it still only takes around 0.4sec to retract/unretract.

So far the best extruder I've used has been the titan, I'd love to try the flexion but I don't yet print flexibles and it's got a hefty price tag.

Edited 1 time(s). Last edit at 02/23/2017 06:21AM by DjDemonD.

---

Simon Khoury

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

Quote
nebbian
Their 3:1 gear ratio is perfect, the gears are a lot lighter and thinner than a planetary gearbox, and it's an extremely well designed piece of kit. The filament path through the extruder is very well thought out. You pull the lever and push the filament through. It just goes straight through, no poking around at all. I haven't tried a long print with flexibles, but I think it should work really well.

I've spent way more time than I really want to admit trying 1.75 flexibles on the Titan extruder. It does extrude flexibles better than most extruder designs, but it doesn't really extrude enough flexible filament properly (ie: extruding 100mm, won't give you 100mm and trying to dial in the tension will result in the filament being stuck to the drive gear and being pulled out of a very tiny hole).

After I realized this, I built a CoreXY with the Flexion extruder with the intent of only printing the NinjaFlex 1.75 flexibles on this machine. It's an ungeared setup, but due to the design of it's idler and drive gear it doesn't stick and get pulled out of the path. It's also sort of wierd. Its not an extruder you calibrate. They have a defined steps/mm to use, and anything higher than this causes jamming. But comparing a print from the Flexion extruder vs the Titan extruder is the same as looking at a properly extruded part vs an underextruded part. I have been extremely happy with my Flexion purchase and in my case I bought their E3D Retrofit kit so I could keep using my E3D V6 hotends with silicon socks. I really didn't think the price was too bad for the extruder section, but the support via phone/email from Flexion was well above and beyond what you would expect.

Quote
DjDemonD
The other argument is about retraction speed. Here I disagree with the received wisdom that retraction has to be fast to work. It doesn't. My flex3drive (40:1) can only retract as I currently have it setup at 11mm/s, but I don't get blobs or strings. However if I start turning the speeds up, then the proportion of the time spent retracting and unretracting is longer. So fast retractions (using low or no gearing) are important with faster print speeds so you can keep the speed up but not so you can actually retract the filament. Of course I'm only retracting it 2mm so at 11mm/s it still only takes around 0.4sec to retract/unretract.

Michael Hackney has a great post about retraction and how fast is often quite wrong, I've been using 5mm/sec 0.9mm retractions quite successfully with the Nimble, as you've seen in the bridge test I did. for those that haven't here is a link to it: Walters bridge test, printed on Zesty Nimble

Edited 2 time(s). Last edit at 02/23/2017 03:27PM by briangilbert.

I have written a wiki page on extruder drive selection at [duet3d.com].

Regarding retraction speed, it's critical to use a high retraction speed if you use a Bowden extruder that needs a lot of retraction.This is because you need to relieve the pressure in the nozzle quickly to avoid continued extrusion after the end of the extruding move. I can believe that it's much less critical if you only need to retract a small amount.

---

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
it's critical to use a high retraction speed if you use a Bowden extruder that needs a lot of retraction.This is because you need to relieve the pressure in the nozzle quickly to avoid continued extrusion after the end of the extruding move.

Yes, this totally makes sense, and I used to, but now that I'm using the Nimble and effectively direct drive not a concern thankfully.

Edited 1 time(s). Last edit at 02/23/2017 05:25PM by briangilbert.