Why non-Bowden?

[Matt Keveney]

I've become curious about CoreXY printers.  Seems that the configuration offers the same speed advantages as a delta (low mass moving parts) without the complex geometry.

But I see a number of CoreXY designs that use a conventional (non-bowden) extruder, like SmartRapCore.  By moving that heavy stepper around, it seems to me that these designers are missing the point. 

Is there some other advantage to coreXY that _I_ am missing?  ...or are these designers merely intimidated by the Bowden mechanism?

-Matt

[Jetguy]

Bowden extruders have a large distance between the motor controlling the filament and the hot end. Filament compresses, tube stretches, connections are also changing distance under pressure and retraction. Bottom line is Bowden is not equal to a direct drive short distance extruder.

Core XY moves the X motor (or whatever axis you decide it is) off the moving gantry. This reduces both a flexing separate wire harness and point of failure, along with a reduction in total moving mass. This means that in turn, we don't need to reduce mass and can use a direct drive extruder for finer pressure and volume control of filament feed resulting in higher quality printing.

Trust, me, we FULLY understand Bowden. It's not the advantage you are claiming it is.

Many of us are not going Core XY to set some speed record. It's to get the best quality print at a reasonably fast speed with long life and limited fuss.

[Jetguy]

In other words, your assumption is Bowden=Direct drive in extruding quality.

You second assumption is reduction in moving mass trumps everything else.

This is the world you seem to be coming from:

Delta arguably is the lowest moving mass in many cases. As such, Bowden is regular design implementation because the mechanism does not respond well to increased moving mass. Since a lot of folks are not using direct drive extruders, or using light geared extruders that simply are not the same thing, the theory is Bowden is "fine".

The Cartesian world has been using direct drive extruders because of lower parts count, lower overall cost, easy implementation, and the mechanics can handle it.

Core XY is seen as a refinement in the Cartesian realm.

While Bowden is often discussed, hundreds and hundreds of times it's been tried in Cartesian world, but never takes a foothold because the general consensus is that direct drive offers finer filament control, retraction, pressure, and just flat out less problems and more reliable.

Basically, Cartesian direct drive + the refinement of instead of focusing on trying to reduce the mass of an extruder, we reduce the moving mass by moving the one motor off the gantry giving us a net reduction in moving mass.

[Ryan Carlyle]

In my opinion, there are two equally-viable strategies for overall design of a "high performance, low cost" 3d printer:

• Minimum moving mass bowden designs with high jerk, z-hop, and super travel speeds... basically your goal is to use rapid motion capabilities to mitigate blobbing and stringing caused by the "bowden lag" in the extruder
• Direct-drive designs with lower acceleration and jerk... the precise extruder volume control means you don't have to move all that fast during travel and through corners to avoid blobbing/stringing, and less aggressive acceleration tuning makes up for the heavier moving mass.

There's a size component to this -- a Mini Kossel effector literally doesn't have enough space for any normal-sized direct drive extruder, so you're forced to use bowden. Short bowden tubes perform well. Whereas in a large Cartesian or CoreXY printer, the bowden tube length becomes unreasonable for good volume control, and you have all the space in the world to mount an extruder on top of the gantry. 

Then, consider the relative mass of the effector compared to the mass of the drivetrain. A Mini Kossel's moving parts are incredibly light (and thus don't need to be strong/stiff), and a NEMA 17 stepper would add an enormous amount to the moving mass and drivetrain flex. Whereas the linear hardware in a long X-bridge is likely to be considerably heavier than the extruder stepper. Adding the stepper mass to a 300mm or 500mm metal rod/rail linear stage isn't making that big of a difference.

So, long and short of it is, small and nimble printers tend to be good candidates for bowden extruders, while medium and large printers experience more downsides from bowden tube length and incur relatively lower penalties from the direct drive mass. MOST people building CoreXY printers are in the 200mm+ size range. 

[Eric Pavey]

I started my 12x12x21" core xy with a bowden, running a .9-1.2mm volcano nozzle. What a nightmare: like mentioned above, the long distance from stepper to extruder made the thing drool like crazy based on (what I presume is) back pressure forming. When I switched to direct drive, quality immediately improved. Plus, I was maxing out at 60mm/sec for those nozzles (low speed, but huge extruded volume), which isn't taking advantage of the theoretical speed increases the Bowden can give you.

[Raphael Remma]

I just finished my smartrapcore alu-like with 22x27 bed. Just realised cant get enough from Bowden. Path too long.
Then just bought a Chimera extruder yesterday. Believing Bowden would be the best.
Do you guys, think is too much place two steppers would be really bad?(since the printer is a bit fragile)

If is the case, I got two 3,5cm steppers lying around. Could they handle using gears?

Cheers guys. Appreciate all the stuff you did for the group. I've been learning a lot here.

Raph

[Raphael Remma]

This is what I've got so far.

[Ryan Carlyle]

Two steppers seems like it'll be a lot of weight on a gantry with rods that long. What's the diameter and length of your X and Y rods?

If you do use two steppers, I'd recommend short-stack motors like the Moon's 17-HDN series motors Makerbot uses in the Replicator 1/2/2x. 

[Raphael Remma]

8mm rods. 

They still holding well (using ball bearings instead of linear bearings on X,Y).

Most backlash comes from the Z screws. Tiny bit of rigging.

Printing at 80 mms. Wouldn't mind to reduce that a bit. Built that scavenging a Prusa and it's been A LOT better.

Cheers,

Raph

[Ryan Carlyle]

8mm rods with CoreXY and one extruder stepper I'd run up to 250-300mm span, with two steppers I'd only go to 200-250mm span. Longer than that and you'll get a lot of ringing. But you can always drop the jerk and accel to compensate. Speed itself isn't such a big issue except on complex organic curves where the acceleration code doesn't work very well. 

[Alex Borro]

1) 3M Production Printer at 80mm/s and 400mm/s of travel (this is the first worldwide 3M product manufactured by 3D Printer):
https://www.youtube.com/watch?v=Kit-3rx95RM

(600mm of Bowen tube, 300mm x 300mm useful bed area)

2) Same model, printing honeycomb infill at 100mm/s and XY Jerk at 80mm/s
https://www.youtube.com/watch?v=LH71MEjywxU

3) Same model, printing some parts at 100mm/s and 400mm/s travel:
https://www.youtube.com/watch?v=0LiVASCSSEI

This last video has just minor issues on first layer due bad bed leveling, but I let the print go on because the issue will not affect the parts..

Well tuned bowden, great quality and speed.

--
You received this message because you are subscribed to the Google Groups "H-bot and CoreXY 3d printers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to h-bot-and-corexy-3d-...@googlegroups.com.
Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

--

"Não é o mais forte da espécie que sobrevive, nem o mais inteligente. É aquele que se adapta melhor as mudanças"  ( Charles Darwin )

Alex Borro

[William Cook]

Ryan Carlyle, thank you for so clearly explaining these trade-offs. What is your current preferred direct-drive extruder design (and combined with which hotend) ? For that matter, what do you favor in a Bowden setup? I've been through a number of designs on the bowden side, and I have one that works, but I have little experience with direct-drive and in either case I'm guessing you have more experience than I do with both.

[whosawhatsis]

Some people have started doing dual extruders with one direct and one bowden, and I think it's a really good solution. This enables some useful combinations (flexible and stiff material, with the flexible in the direct, soluble support with the support in the bowden, etc.) while keeping the effector mass close to that of a single direct extruder. The one thing I was hoping to add was the ability to disconnect the direct extruder when it's not needed, to get a bowden-weight effector.

[Raphael Remma]

Nice. Thanks guys, for clarify. And thanks Matt, for bring up a super relevant question.

Raph

[Ryan Carlyle]

William, I still like the Replicator 2 style ungeared short stack stepper and ~10.5mm hobbed drive gear, with a metal spring-arm idler or David Headrick's 3in1. Performs just fine with "typical" hot end hardware, without having so much torque that it's likely to strip the filament grip rather than harmlessly stalling. That then goes extremely well with the Replicator 2 style cooling bar and threaded thermal barrier that lets you print any material reliably with bare-minimum total Z thickness. (Groovemounts with cylindrical heatsinks like E3Ds are disgusting wastes of Z height, and I don't like PTFE tubes.)

There are a number of other good options, I just get good enough results with that kind of setup (in my R2x and CoreXY Cube) that I don't have any meaningful reason to try other stuff.

For Bowden drives, I use my B'Struder with a small 5.2:1 geared stepper. Works great. I have one running with a hybrid Rep2 style cooling bar system, and one with an E3Dv6. Either is fine. The Rep2 style setup is a lot more compact but doesnt have a ton of service history so I can't promise it's as reliable. (Hasn't given me any issues with PLA or PETG.)

[Matt Keveney]

Thanks all, for the interesting discussion.  

I've recently completed a Cherry PI IIIs, my first Bowden. The pneumatic fittings are easy to work with, and I don't get stringing issues.  It's early days, and I'm certainly no expert, but the Bowden system has been the least of my problems; I just assumed that it was pretty well sorted out by now... I guess not.

Ryan Carlisle mentions this point:

> ...Whereas the linear hardware in a long X-bridge is likely to be considerably heavier than the extruder stepper. 

> Adding the stepper mass to a 300mm or 500mm metal rod/rail linear stage isn't making that big of a difference.

Agreed, but that merely argues for making the linear hardware in the X gantry lighter, which ought to be doable, if we dispense with the heavy stepper.  Perhaps aluminum extrusion with Delrin wheels here?  Any other thoughts?

-Matt

[Matt Keveney]

Ryan Carlisle mentions...

That's "Ryan Carlyle."  

Apologies Ryan.

[Ryan Carlyle]

Maximizing stiffness-to-mass ratio is desirable for any printer, but it's not the only viable design strategy. Other design drivers can take precedence. For example, some downsides to V-wheels:

• They take up a lot of space.
• They don't resist torsion around the long axis of the extrusion very well. That side-loads the radial bearings and puts bending moments on the carriage/wheel axles.
• Such motion stages rely on the accuracy of the extrusion, which may or may not be suitable over long spans. (Check the max acceptable tolerances on OpenBeam sometime -- it can be "in spec" and yet completely unacceptable for use as a 3D printer rail.) 
• What if you have a nick on your extrusion from shipping damage? Oops, can't use it any more, you'll get a ridge in the print where the wheels roll over the nicked surface.
  
• Delrin doesn't work too well in heated build chambers due to creep issues. In fact, even in non-heated environments, Delrin takes a compression-set if tightened and left stationary for an extended period, which will put flat spots on the wheels. 
• Pre-loading requires the use of fiddly eccentrics or tightening mechanisms built into the carriage, which is more mechanical complexity.

A high-quality steel linear rail is a frickin' dream to work with compared to V-wheels. Durable, compact, precise, heat-resistant, "install and forget" kind of hardware. Yeah, it weighs more, but who cares? There's more to 3d printer engineering than stiffness/weight ratios. 

Main point is though, a good direct-drive extruder delivers superior extrusion volume control. See Jetguy's posts. Not everybody cares, that's fine, you CAN get good results with Bowden drives, but any reasonably-long Bowden tube will introduce print flaws and additional tuning requirements. 

[whosawhatsis]

Btw, I once ran a print with a test extrusion loose in the bottom area of a printer running on v-slot so that as the platform ran back and forth, a wheel (the lone wheel on a three-wheel setup) would run over it (it would get between the wheel and the extrusion. This was on a bukito prototype, where the platform was made of laser-cut acrylic, which would flex enough to roll over it (I think the final metal version would either push it out of the way or skip before rolling over it). Because of this, there's a print of a tardis floating around with a strange deviation along the X axis where the platform shifted out of its linear path each time it crossed the plastic.

--

[Uglybob]

I thought v-slot was awesome... now my corexy dreams have been thrown in to chaos. I dont like this game any more. =\

[Raphael Remma]

Me too. Now I'm scared.

--
You received this message because you are subscribed to a topic in the Google Groups "H-bot and CoreXY 3d printers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/h-bot-and-corexy-3d-printers/LQDx0WxRa8Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to h-bot-and-corexy-3d-...@googlegroups.com.

Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

--

Raphael Rema

Senior 3D Generalist

Mob.: 07586539171

[Ryan Carlyle]

Hey, don't let me scare you off V-wheels, you can make great printers with V-wheel drivetrains. A good setup is smooth-running and light-weight. Some of my favorite printer designs (like the Fusion 3 F306 and F400 CoreXY) use V-wheels. I'm just trying to point out that there's no one "best design" for linear stages. Everything is a tradeoff. 

To unsubscribe from this group and all its topics, send an email to h-bot-and-corexy-3d-printers+unsub...@googlegroups.com.

Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

[Raphael Remma]

I am running small ball bearings on a steel rods. wobbles a bit but I'm getting there. At least is A LOT better than LM8UU (had that before).

To unsubscribe from this group and all its topics, send an email to h-bot-and-corexy-3d-...@googlegroups.com.

Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

--

Raphael Rema

Senior 3D Generalist

Mob.: 07586539171

--
You received this message because you are subscribed to a topic in the Google Groups "H-bot and CoreXY 3d printers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/h-bot-and-corexy-3d-printers/LQDx0WxRa8Y/unsubscribe.

To unsubscribe from this group and all its topics, send an email to h-bot-and-corexy-3d-...@googlegroups.com.

Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

[whosawhatsis]

On Saturday, April 9, 2016 at 20:43, Ryan Carlyle wrote:

William, I still like the Replicator 2 style ungeared short stack stepper and ~10.5mm hobbed drive gear, with a metal spring-arm idler or David Headrick's 3in1. Performs just fine with "typical" hot end hardware, without having so much torque that it's likely to strip the filament grip rather than harmlessly stalling. 

There are reasons that you might want more torque. For instance, E3D has started using an operation they're calling "ram-purging" to get plastic out of the hot zone when switching extruders. What they do is move to a dump zone, then try to extrude much faster than the thermal transfer to the filament can maintain so that they end up with cold filament in the hot zone (with a lot of back-pressure and die swell in the process). They then immediately retract that cold filament all the way out of the hot zone and let it cool while the other extruder is active. This eliminates ooze because there's no plastic in the other nozzle, but it doesn't work if your extruder isn't strong enough to fight that excessive back-pressure. This is on their new Titan extruder, and they're using a motor that would probably be sufficient without gearing, then adding a 3:1 reduction (it's basically a wade's extruder in a much smaller package).

[Ryan Carlyle]

I think E3D has been ram-purging with regular non-geared direct drive extruders too. Couldn't tell you if it works as well, but they first revealed the process on the dualstrusion BigBox, and it uses a very vanilla NEMA 17 ungeared spring-arm setup. 

[whosawhatsis]

Ah. They were talking about it at MRRF with their Titan extruders on the bigbox. I haven't been following the bigbox stuff that closely because I'm not terribly impressed with the mechanical design. I do like some of the components though, like the platform heater with heater power concentrated around the edges (I can't believe it's taken this long for someone to finally make one of those, I spent years telling Roy from Panucatt that he should do that).

--

You received this message because you are subscribed to the Google Groups "H-bot and CoreXY 3d printers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to h-bot-and-corexy-3d-...@googlegroups.com.

[WZ9V]

Here is another take on keeping the effector end "light"  http://mutley3d.com/Buy-Flex3Drive/ 

[Raphael Remma]

Wondering if a dremel flexible shaft would work...

[Ryan Carlyle]

Dremel shaft is only really good for one-direction motion... has issues with retraction. I have heard the Mutley extruder uses a model boat propeller shaft. 

[Matt Keveney]

Yes, and I remember this one too:

   http://www.thingiverse.com/thing:15132

Sounded intriguing at the time, but I never heard anything more about it.  I suppose the coreXY concept could be extended (CoreXYE?)

-Matt

[Ryan Carlyle]

Yep, it's been proposed, but you need custom firmware to multiplex the signals, and it's just not worth the hassle. 

A couple ways it can be done: https://groups.google.com/forum/#!searchin/3dp-ideas/corexyab/3dp-ideas/Xv-tC_bf708/7YAdByltg0sJ

Note that the CoreXE concept shown there has nothing to do with CoreXY, and they shouldn't have called it that. 

[Raphael Remma]

Cannot open this link, Ryan.

[Alex Borro]

I think the main problem in this "shafted extruder" is the shaft elasticity. Way more than the filament spring effect.

2016-04-13 16:33 GMT-03:00 Raphael Remma <djr...@gmail.com>:

Cannot open this link, Ryan.

--
You received this message because you are subscribed to the Google Groups "H-bot and CoreXY 3d printers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to h-bot-and-corexy-3d-...@googlegroups.com.
Visit this group at https://groups.google.com/group/h-bot-and-corexy-3d-printers.

[Ryan Carlyle]

This one work? https://groups.google.com/d/msg/3dp-ideas/Xv-tC_bf708/7YAdByltg0sJ

[Ryan Carlyle]

On Wednesday, April 13, 2016 at 3:41:39 PM UTC-5, Alex Borro wrote:

I think the main problem in this "shafted extruder" is the shaft elasticity. Way more than the filament spring effect.

Yes, Mutley uses a 40:1 worm gear at the extruder end so that the shaft flex and backlash is reduced 40x at the drive hob. You lose some retraction speed, but it basically eliminates the flex issue.

[Raphael Remma]

The link worked. That gave me a "windows blue screen" on my head.
Looks complex. Did you test that?

[Ryan Carlyle]

On Wednesday, April 13, 2016 at 4:18:38 PM UTC-5, Raphael Remma wrote:

The link worked. That gave me a "windows blue screen" on my head.
Looks complex. Did you test that?

Nah, too complicated and not enough benefit over a more traditional bowden drive. Basically it isn't worth the effort required for detailed design. 

[Chris (crispy1)]

I do like some of the components though, like the platform heater with heater power concentrated around the edges (I can't believe it's taken this long for someone to finally make one of those, I spent years telling Roy from Panucatt that he should do that).

Every Fusion3 ever built has a bed heater like this.  We've been doing it since 2013 :P 

[Ryan Carlyle]

E3D does a lot of not-particularly-original stuff that their user base hasn't seen before and therefore assumes is new.

Now that I think about it, that's kind of the plague of hobbyist 3D printing in general, people not knowing what "state of the art" is.

[whosawhatsis]

I stand corrected. What I should say is that I'm surprised that I haven't heard about anyone actually doing this (and I've looked, but it's not obvious from looking at the Fusion3 site) until now.

On Saturday, April 16, 2016 at 08:42, Chris (crispy1) wrote:

I do like some of the components though, like the platform heater with heater power concentrated around the edges (I can't believe it's taken this long for someone to finally make one of those, I spent years telling Roy from Panucatt that he should do that).

Every Fusion3 ever built has a bed heater like this.  We've been doing it since 2013 :P 

[whosawhatsis]

That and the fact that bad ideas that look superficially like good ones (the canonical example is putting a bearing at the end of a Z screw) spread faster than the reasons that they're actually bad ideas.

On Saturday, April 16, 2016 at 10:20, Ryan Carlyle wrote:

E3D does a lot of not-particularly-original stuff that their user base hasn't seen before and therefore assumes is new.

Now that I think about it, that's kind of the plague of hobbyist 3D printing in general, people not knowing what "state of the art" is.

[Raphael Remma]

True. Suffered for the first few months on my first Mendel.