Makerbot Replicator Bowden Tube experiment

[delsydsoftware]

Since the 3-point adjustment mod I did worked so well, I figured I'd do something a little crazier. I decided to convert the Replicator to a Bowden tube setup.

 I made a custom bracket to mount the extruder stepper motors to the back of the Replicator, and attached some Teflon tubing. I'm still designing parts to hold the tubing in place on the print head, so for now the tube is just wedged in place with a tiny screwdriver. That works surprisingly well, actually. I also added a Lin Engineering stepper motor damper to the extruder, which seems to really clean up blobbing caused by backlash. I'm using the Minimalistic MK8 mod to feed the filament directly into the teflon bowden tube, which leads straight into the print head.

The guardian squirrel was printed at 50 microns layer height, with print-o-matic set for 60mm/s and 70mm/s extrusion speed. With no motors attached, the print head is really fast. I was able to do some test prints at 150 mm/s at about the same quality as an 80mm/s print with the original setup. Echoing is greatly reduced as well. The machine is also amazingly quiet now. You can occasionally hear the damper knock when the stepper motor makes a drastic direction change, but that's really not a big deal. The y axis on my machine has always been the loudest, and it quieted right down. I'm planning on adding a damper to it as well, to see if I can reduce the noise a bit further.

The only catch right now is that I need to design a bracket for the right extruder that hits the X limit switch. I've been tapping the switch by hand when the print head resets. I had overlooked that little detail, because I assumed that Makerbot wouldn't have used the side of the stepper to tap the limit switch. Bad assumption! There is nothing like hitting the print button and hearing your belts grinding.  :(

I'm also thinking about using a different set of heatsinks for the print head. Only a small portion of the heat sink is making contact, so the rest is basically wasted surface area. I doubt that the top of the heatsink is really doing much work. I was thinking of using some passive ram heatsinks instead, the type you see on Corsair Dominators.

Here's a vid of the extruder/damper in action. I'm going to work on the rest of the bracket designs when I get home:

http://www.flickr.com/photos/delsydsoftware/8054221343/in/photostream

Total cost so far is $13 for tubing and $25 per damper. I used spare screws from the MK8 build to attach the bracket to the motor, so those expenses were already covered. So, basically $63 total for a great increase in speed and high-res print quality.

[JohnA.]

Looks great.  I'm curious if it'll require some tuning of retraction for some prints with numerous tall / thin items...  There's always talk of that 'spring' action of having all that filament post motor.  

JohnA.

[Dan Newman]

The other direction we need to take extruders is *faster* feed rates.
Right now with the Jetty Firmware, we can run the bot's quite fast
(> 300 mm/s), but the 1.75mm filament extruders cannot handle those
speeds. A 3mm filament extruder likely could as you only need to
ingest filament at one third the rate (1 / 2.93).

Sure momentum is an issue at those rates and thus having a less
massive extruder is very helpful. But you can compensate for a massive
extruder by have small max accel/decels and tighter jerk control.
Doing that slows down the print and for small prints may mean never
hitting the top/target feed rates. For large prints, there's
plenty of "take-off" room in which to hit top speeds.

Dan


On 4 Oct 2012 , at 11:09 AM, delsydsoftware wrote:

>
>
> <http://farm9.staticflickr.com/8182/8054156782_5c25417a6a_b.jpg>
>
> <http://farm9.staticflickr.com/8310/8054155728_7cd5b65a09_b.jpg>
>
> <http://farm9.staticflickr.com/8178/8054151859_bfd10d2a5d_b.jpg>

> --
> You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
> To view this discussion on the web visit https://groups.google.com/d/msg/makerbot/-/vuCY4hzHVmsJ.
> To post to this group, send email to make...@googlegroups.com.
> To unsubscribe from this group, send email to makerbot+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/makerbot?hl=en.
>

[ddurant]

> So, basically $63 total for a great increase in speed and high-res print quality.
 

Uh...

 

I hope Mark appreciates that I'm trying to restrain myself here..

[delsydsoftware]

Well, considering that I could only hit 60 micron layers at 35mm/s before, I'd call that a nice increase. I know other printers are faster and better. I don't own one of those.  :)

[Luis E. Rodriguez]

It was awesome, always has been. That is why I bought one for the Science Center. You heard me.

http://blog.ultimaker.com/2012/09/24/ultimaker-3d-print-farm/martijnelserman-ultimaker-printfarm/

Luis E. Rodriguez

Are you a Maker?

On Thu, Oct 4, 2012 at 2:17 PM, Mark Cohen <markc...@gmail.com> wrote:

And because of that they print fast. Vibrate less.
They also have nice cutouts in the top of each case that meshes with
the legs of another and so can be stacked.
I saw this at Makerfaire.

On Thu, Oct 4, 2012 at 3:15 PM, Mark Cohen <markc...@gmail.com> wrote:
> I certainly do. LOL
> I'll say it.
> Ultimaker has been doing the Bowden in their machines for quite some time.

>> --
>> You received this message because you are subscribed to the Google Groups
>> "MakerBot Operators" group.
>> To view this discussion on the web visit

>> https://groups.google.com/d/msg/makerbot/-/l4JE57wtio0J.

>>
>> To post to this group, send email to make...@googlegroups.com.
>> To unsubscribe from this group, send email to
>> makerbot+u...@googlegroups.com.
>> For more options, visit this group at
>> http://groups.google.com/group/makerbot?hl=en.

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

[Dan Newman]

> I certainly do. LOL
> I'll say it.
> Ultimaker has been doing the Bowden in their machines for quite some time.

And there's tradeoffs: less mass which allows for higher print speeds without
loss of quality. OTOH, there's a different set of challenges as witness the
many threads about bowden jams (multiple causes, some hopefully fixed with the
UM v2 hotend?) and "Stringless prints…" [how to achieve them] (that one is
a thread of ddurant's from August ;).

As with any extruder design there's tradeoffs. However, it's nice to see
more experiments with Bowden extruders since they have a lot of promise.
(And they also might need some more attention on the fw side of things.)

Dan

[Aaron Double]

Hey Luis,

Guess you decided on which toy to play with.

I want to hear your impressions when it comes in.

Aaron

[Luis E. Rodriguez]

Being assembled now, during "free time."

Luis E. Rodriguez

Are you a Maker?

[Aaron Double]

I just ordered a B9 creator, when it comes in in December, I'll give my opinion.

Aaron

[ddurant]

> OTOH, there's a different set of challenges as witness the
> many threads about bowden jams (multiple causes, some hopefully fixed with the
> UM v2 hotend?)

 

Hopefully the v2 hot end will fix most of those.. I've had it happen once in about 18 months and that was because I left the machine sitting while it was holding +200C. Think I forgot I'd set the temp and went to watch cartoons or something. The new hotend is said to be much better at temperature isolation and should prevent a lot of that from happening. Jams from crappy PLA that gets stuck in the tube are a lot harder to address.. Not sure they can do much about that.

 

> and "Stringless prints…" [how to achieve them] (that one is a thread of ddurant's from August ;).
 

Yes.. An upcoming blog post, once I find some time to do some experiments then type it up. I'll let others that were at MF:NYC talk here about the quality of UM prints, if they want, and how good they have gotten at getting rid of strings. Except to say that I was carefuly to talk to Erik, Martijn & Siert (the UM founders) about the prints they were displaying and all of them told me there was 0 post-processing. Those prints were all straight off the machines..

 

Also, WTH?? I didn't spot Aaron or Luis (and didn't know to look for Mark) and MF.. Were you guys really there???

[ddurant]

Nope. I think I met her at MF:NYC 2010 but didn't talk to her or Tony last weekend..

 
On Thursday, October 4, 2012 9:21:53 PM UTC-4, Mark Cohen wrote:

I think i was in front of you if that was you talking to Tony Busers wife at the 3d printer area. You may be the guy she said hi Dave to.

--

You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

To view this discussion on the web visit https://groups.google.com/d/msg/makerbot/-/BPgC9Le_7kYJ.

[Laird Popkin]

Very cool. Looks like a nice improvement in performance. Is there any 'lag' between the stepper movement and the movement of the plastic into/out of the extruder hot end? I guess other bowden users (e.g. Ultimaker) have figured that out, but I'm curious about how it's working for you. 

 

The X limit switch is in a reasonable place - the extruder motor is in a fixed place in MBI's design. But yeah, if you move it you'd have to put a bracket or something across the gap.

For holding the tubing in place, I use the upgraded bracket http://www.thingiverse.com/thing:26880, which uses paint gun fixtures which hold the tubing quite firmly, to the point where it's impossible to pull out unless you press down on the retainer clip. The original clips are OK, but the paint gun clips are much firmer, and since they're solid metal, chromed, they look nice, too. Surprisingly cheap.

Are you going to share your designs? I'm inspired to give this a try, once you do.

So what's the fastest that you've been able to run for printing? I saw an Ultimaker last weekend dialed up to 400 mm/s (printing "stretch bracelets") and it was amazing. Admittedly it wasn't printing anything detailed ("stretchy bracelets") but man, was it fast!

[delsydsoftware]

I haven't noticed any extrusion lag so far. Also, I can tell that the damper is making a good difference with retraction. I have tested with ReplicatorG and Makerware, and ReplicatorG prints are much cleaner than they were before. Makerware prints are about the same. I'd love to put dampers on all of the steppers, but I think I could only fit one on the y axis without serious modifications to the existing parts.

I reset the X and Y Max Acceleration on the machine to 3000mm/s (from 1500 in firmware 6.0) and it's still pretty smooth and relatively quiet. Last night, I printed out a slug trap in PLA ( http://www.thingiverse.com/thing:23467 ) at 130% scale at 200mm/s. It looks pretty much identical to a previous slug trap I printed at 80mm/s, so I think I could probably crank the speed up a bit more before quality got too bad. I don't know if I could even approach 400mm/s without some serious tweaking of values. I have been impressed with the increase in vertical resolution. I was able to do a print at 40 microns last night. I think I may be able to achieve better than that with tweaked settings in RepG.

Once I get the designs worked out, I'm going to post them on thingiverse. I'm still tweaking the parts right now. Thanks for the link to the connectors. It's amazing how much pressure those lines are under. I had also looked at some compression fittings from McMaster-Carr. The tubes are 5mm in diameter, which is probably a bit small for most fittings.

[lassikin]

hi,

are the dampers something you add to the stepper motor and where are they available from? also, is the bowden cable something special or for which use is the cable normally sold for, guiding brake cables and such?

nice mod.

-lassi

[Jetty]

Just a heads up, the acceleration and feed rates you think you're
getting
you may not be due to limitations within the firmware / and or the
print being small.

Also, the xml is another limiter to speed.

Also, past 150mm/s, the increased pressure in extruder chamber
generally
means the MK8 strips the filament.

It's cool what you're doing, if it ends up being workable it should go
a long way to
reducing the vibration issues in the HBP caused by the mass up top.

[Luis E. Rodriguez]

This may well help the stripping, if it occurs: http://www.thingiverse.com/thing:15718 or maybe the extruder on the Rostock: http://reprap.org/wiki/Rostock

It's obvious it will work since the Ultimaker has been doing it for years. You could always try their extruder setup as well. 

Nice project!

Luis E. Rodriguez

Are you a Maker?

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

[Jetty]

> It's obvious it will work since the Ultimaker has been doing it for years.
> You could always try their extruder setup as well.

Just because you tell the firmware to print at 400mm/s, doesn't mean
it can or will. Same goes for the extruder.

The Arduino Mega is only a 16MHz 8 bit processor, that's a serious
lack of
horse power for fast speeds.

Additionally small prints don't give the printer enough time to reach
full speed on all moves,
and extruders that run for a few seconds at top speeds, don't
continue to run as the pressure builds / equalizes in the head. Also
there's
thermal mass to consider which if not enough causes extrusion to pulse
on output.

When testing the Replicator to find out the speed it's capable of, the
acid test is long
stretches at the top speed. I.E. A rectangle that takes up the
longest axis of the
build platform. That way the extruder gets a chance to run for a
while at top speed. Then combine
it with a stop watch, to see if that speed is being reached.

[Zak Kus]

Hate to be the only ignorant jackass in a room, but what is a bowden tube and why is it cool exactly?

[Aaron Double]

An example of a bowden cable is the shifter cables on your bicycle.

It's cool because it allows you to get rid of the mass of the motor from the moving parts of your gantry and that allows you to go faster!

Aaron

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

To view this discussion on the web visit https://groups.google.com/d/msg/makerbot/-/ygMNC644enIJ.

[delsydsoftware]

Thanks for the heads up. The prints are definitely faster, but it's not a huge difference. With the slug trap, which is 5 inches across and 3-4 inches tall, I could print the base in about 2.5 hours hours using the default accelerated settings in RepG before. The test I did last night printed in close to 2 hours. The infill is definitely faster, but perimeter speeds didn't seem to be massively faster. It could be because the perimeter of the slug trap is curved, so it never gets close to full speed. I need to do some tests with prints that have more long, straight areas.

i'd like to do some more thorough testing. That will probably have to wait until I get a reliable way to keep the teflon tubes locked in place. I hope I can fix that over the weekend.

[delsydsoftware]

The damper is just a tiny weight suspended in a flexible rubber grommet. The damper I'm using is made by Lin Engineering. The video at this link explains it better than I can:

http://www.linengineering.com/line/contents/stepmotors/Nema17_Damper.aspx

I found out about these dampers when I was searching the web for ideas on making the stepper motors in the Replicator quieter. People add these to their CNC machines to reduce vibration, and it has the added benefit of making motor movements a bit more precise. The only vendor I've found so far is Newark:

I'm using the rear-mounted version of the damper, which is out of stock because I bought all 3 of them :)

Here's the front-mounted version. I think the only difference between the two is that the rear-mounted version is a bit longer.

http://www.newark.com/lin-engineering/057-00009/inertia-damper-front-mount/dp/79R1398

[Doogiekr]

I did a little searching and there are some DIY alternatives out there as well, including one that uses caster wheels or roller blade wheels

 

http://www.solsylva.com/cnc/dampers.shtml

 

I would imagine since wheels can be used, then it would also be possible for us to design / 3D print something that would work well

 

It may take a bit of creativity, but this group is full of creative people =)

[delsydsoftware]

Just as an aside, before I tried the bowden cable idea, I toyed around with the idea of using a flexible shaft cable (similar to the type of cable that runs from the engine of a weed eater to the spinning head) to externally power the extruders. The stepper motors would be in the same position that I have mine in now, but with flexible shafts attached. The print head would have a mount for the shaft and an adapted version of the mk8 minimalistic mod. From the tests I did, it would probably work in a printer designed specifically for that sort of setup.

The flex shafts can't bend beyond a certain point, so you would need to have some clearance at the back of the printer to make sure that the cables didn't get kinked. The Replicator doesn't have enough room behind the print head for the cables to stick out, unfortunately. If you could design a geared system which allowed the flex shaft to come in from the top, it might just work. But then, you also have the mass of the flex cable to consider.

 It seems like it would have been a nice tradeoff----reduced weight with the advantage of the filament being forced into the extruder from a short distance. The only real issue would be hysteresis in the flex shaft. I have considered trying this idea out with the X-axis stepper on the replicator, though. I think that might be more workable, and would further reduce the mass of the print mechanism

[Charlie Prevost]

I may be showing my ignorance here, but with a Bowden setup, are the fans required? i thought they were to prevent the stepper motors overheating (thus increasing their lifetime) Would it be possible to remove them also and make your extruder even lighter?

[hellphish]

Most bowden setups are for PLA, for which a fan greatly helps print quality.

To view this discussion on the web visit https://groups.google.com/d/msg/makerbot/-/to5FxZ7KX74J.

[delsydsoftware]

Sorry for the lack of updates on this. I had to have a wisdom tooth removed, so that slowed things down a bit. My impression so far is that the bowden tube experiment does a lot more for noise and detail level than it does for speed. But, I am making some great progress with print quality at a small scale. I think the reduced weight on the print carriage is reducing some sag. I can successfully print at 20 microns (with print-o-matic set to 30mm/s print speed and 45mm/s feed rate). I tried a couple tests at 10 microns and 8 microns. 10 microns shows some promise---it looks like I just need to get the feed rate worked out to prevent blobbing. 8 microns was a little too greedy. I think Skeinforge is having a hard time at this resolution as well, because a significant portion of the perimeter of the squirrel's head was missing.

So, in order from left to right is 8 microns, 10 microns, and 20 microns. The 8 micron print took a little over 4 hours and was 1800 layers. The 10 micron print took an hour and a half, and the 20 micron print took 45 minutes.

Any suggestions on cleaning these prints up? I suspect that 8 microns is beyond what the printer can cope with, but it looks like 10 is definitely possible. there is a huge blob on the 10 micron print. I'm not sure how that happened, because I've been starting these prints before I pass out. Wisdom teeth suck.  :)

On Thursday, 4 October 2012 14:09:39 UTC-4, delsydsoftware wrote:

Since the 3-point adjustment mod I did worked so well, I figured I'd do something a little crazier. I decided to convert the Replicator to a Bowden tube setup.

 I made a custom bracket to mount the extruder stepper motors to the back of the Replicator, and attached some Teflon tubing. I'm still designing parts to hold the tubing in place on the print head, so for now the tube is just wedged in place with a tiny screwdriver. That works surprisingly well, actually. I also added a Lin Engineering stepper motor damper to the extruder, which seems to really clean up blobbing caused by backlash. I'm using the Minimalistic MK8 mod to feed the filament directly into the teflon bowden tube, which leads straight into the print head.

The guardian squirrel was printed at 50 microns layer height, with print-o-matic set for 60mm/s and 70mm/s extrusion speed. With no motors attached, the print head is really fast. I was able to do some test prints at 150 mm/s at about the same quality as an 80mm/s print with the original setup. Echoing is greatly reduced as well. The machine is also amazingly quiet now. You can occasionally hear the damper knock when the stepper motor makes a drastic direction change, but that's really not a big deal. The y axis on my machine has always been the loudest, and it quieted right down. I'm planning on adding a damper to it as well, to see if I can reduce the noise a bit further.

The only catch right now is that I need to design a bracket for the right extruder that hits the X limit switch. I've been tapping the switch by hand when the print head resets. I had overlooked that little detail, because I assumed that Makerbot wouldn't have used the side of the stepper to tap the limit switch. Bad assumption! There is nothing like hitting the print button and hearing your belts grinding.  :(

I'm also thinking about using a different set of heatsinks for the print head. Only a small portion of the heat sink is making contact, so the rest is basically wasted surface area. I doubt that the top of the heatsink is really doing much work. I was thinking of using some passive ram heatsinks instead, the type you see on Corsair Dominators.

Here's a vid of the extruder/damper in action. I'm going to work on the rest of the bracket designs when I get home:

http://www.flickr.com/photos/delsydsoftware/8054221343/in/photostream

Total cost so far is $13 for tubing and $25 per damper. I used spare screws from the MK8 build to attach the bracket to the motor, so those expenses were already covered. So, basically $63 total for a great increase in speed and high-res print quality.

[delsydsoftware]

I tried this, and it's a bad idea. The heatsinks are actually serving 3 purposes:

1. Keeping the motors from overheating
2. Keeping the ABS parts of the print head from heating up too much
3. Keeping heat from soaking into the upper part of the extruder

In just a few minutes of test prints, I started getting filament jams. The top of the extruder was getting hot enough to make the filament expand and lock into place. I really don't recommend it. The fans and heatsink really don't weigh much more, anyway.

[Clinton Hoines]

Any more updates? Will you be sharing the mount etc on this design?

Looks like it works great, really like the less weight on the carriage idea . Also fees up room for a filament fan :)

[delsydsoftware]

I need to rethink my extruder design. The bowden tube worked great with Firmware 5.5 and 6.0. But, it can't extrude fast enough for Sailfish. With the reduced carriage mass, the print head can accelerate faster than than extruder can extrude under Sailfish.  Also, I'm not sure that the MK8 Minimalistic mod produces enough pressure on the filament to prevent skipping. Small prints were great, but larger prints would invariably have issues when the filament would get stripped. It looks like most bowden extruders on thingiverse have strong springs and are fully enclosed.  I also want to try a narrower bowden tube. The tube I used has an interior diameter of 3mm and an exterior diameter of 5mm. I want to see if a tube with an interior diameter of 2mm would help with extrusion pressure, since there is less room for the filament to move around inside the tube.

[Charlie Prevost]

regarding layer resolution. I think I read somewhere recently that for ABS, resolution becomes limited due to the fact that it isn't a smooth 'liquid' as such. At small scales, little lumps are present that would cause problems with extrusion (be like trying to extrude lumps of stewed apple, rather than a nice pasty plastic, as is my understanding).

Additionally, have you tried using sailfish with the bowden? I know that it made massive difference to my print surface quality with a stock rep1 hardware. I have managed to print successfully at 50 microns (haven't tried smaller yet, but with some slicing adjustments, am confident that I could get to maybe 10 microns)
I would be very interested to see if you get any improvements with a bowden. I feel that it might mainly be speed related at fine resolutions, as you could potentially increase the maximum acceleration rates while not having to have too high a feedrate.

I'll be following with interest :)

[ddurant]

Did you also turn up the temperature? As you print faster and faster the plastic spends less and less time in the hot end. If you don't add heat as you increase the feed rate, the plastic doesn't have enough time to melt and you end up trying to push a solid(ish) mass through the nozzle - a sure reciepe for stripping filament.

 

> I also want to try a narrower bowden tube.

 

This is extra important for things like retraction. If there's slop in the system, it effectively acts as a pause when you change extruder directions.. If the extruder is only going in one direction (no reversals), it's less of an issue.

 

You're also still pushing the Y motor around so if you're really trying to make things out, you should probably have different acceleration settings on X than you have on Y.

[delsydsoftware]

I had the temperature set to 240C, and it was still having issues. The only way I could correct the gaps was by setting the Restart Extra Distance setting in Dimension. That wasn't a perfect solution, though. It worked great on large models, but when printing small details, it would extrude way too much abs at the beginning of the layer. I finally used a stopwatch to time a regular extruder vs my bowden extruder, and the bowden extruder was about 20% slower. I'm not sure if it's because of slop in the system or just an inefficient extruder setup.

[delsydsoftware]

I have been able to print at 10 microns with Sailfish and the regular Replicator extruder just by using Print-o-matic (with the MK8 minimalistic mod). You just have to have an extremely level HBP to do it. I've done several prints at that resolution with no issues. Unfortunately, the issues I experienced with the bowden setup were under Sailfish. But, it's not the firmware's fault. I think my Bowden setup was just screwy.  I'm going to try one of the bowden designs on thingiverse and see if that improves the situation.

Sailfish makes some of this moot, though. There was a huge difference in print quality, speed, and noise using the bowden setup with firmware 5.5 and firmware 6.0.  Sailfish is so good at acceleration that there is only a small difference between the standard print head and the bowden head when it comes to noise and print quality. The only real difference is that the bowden print head definitely accelerated faster, and there was less vibration.

[Dan Newman]

On 22 Oct 2012 , at 8:25 AM, delsydsoftware wrote:

> I have been able to print at 10 microns with Sailfish and the regular
> Replicator extruder just by using Print-o-matic (with the MK8 minimalistic
> mod). You just have to have an extremely level HBP to do it. I've done
> several prints at that resolution with no issues.

At such small layer heights, set deprime=0 in the Sailfish firmware.
You should get better results. There's sooo little plastic being
extruded, that the firmware deprime doesn't buy you much. (I reserve
the right to change my opinion in the future when I again have time
to experiment.)

> Unfortunately, the issues
> I experienced with the bowden setup were under Sailfish. But, it's not the
> firmware's fault. I think my Bowden setup was just screwy. I'm going to
> try one of the bowden designs on thingiverse and see if that improves the
> situation.

Be warned: SF 50 + RepG 39 likes to generate copious amounts of

disable A axis

soon followed by

enable A axis

It does that before travel moves (and maybe layer changes as well). I
suspect that this is intended to help prevent smears with dual extrusion,
but that's conjecture. You should have been seeing this
as well with MBI's 5.5 and 6.0 firmware, so this shouldn't be a new
issue. However, I have to wonder if it is part of the issue.

So, this may be an issue with a Bowden extruder: it's turned out to be
an issue with 3mm extruders which have much higher pressure in the melt
chamber compared to 1.75mm extruders. We're working on a long term solution
for Sailfish -- an extruder hold option. In the short term,
you can hit SF's replace.csv and have it replace M103 commands with
";M103" to comment them out. That impacts software deprime which
uses the "extruder stepper motor disable" commands as an aid to
detecting a true travel-only move. But this is just a short term
work around you can attempt.

Dan

[RocketSled]

On Sunday, October 21, 2012 11:02:29 PM UTC-4, delsydsoftware wrote:

With the reduced carriage mass, the print head can accelerate faster than than extruder can extrude under Sailfish.  

Not so.  The control system is open loop, the controller doesn't "know" how fast the print head is moving.  It commands the steppers to step at a given rate and simply assumes there is enough motor torque to move the print head at that rate. 

The stepper motor isn't a free running motor like in an electric fan.  Quite the opposite.  It's a special kind of motor designed to move in discrete increments, or steps.  Power up a fan, it accelerates to a speed that's a function of the voltage, motor "winding" and the load.  Power up a Stepper, it moves a fraction of it's total rotation and stops (typically, 200 steps/rev). 

On a CNC, acceleration is simply the speed at which the stepping rate changes.  And it's controlled by the profile generated by the controller FW (Sailfish).  The FW can't see that the print head is lighter and so, accelerate it harder.  Whether the print head weighs 1 gram or 1 kilogram, Sailfish says "move this many steps at this rate" and simply assumes the stepper motors have the power to make it happen.

If the motor can't "make it happen", you lose steps and the positioning of the print head relative to the build surface drifts and your print comes out looking like spaghetti. 

If Sailfish was already accelerating the print head too hard, meaning it wasn't reaching it's target speed during a given move, lightening the print head would matter. But for this to be true, you'd have been missing "steps" before you made the print head change, nothing would have printed.

Now, Sailfish might be good to 240mm/s with the mass of the stock print head.  The motors don't have enough torque to reliably accelerate to faster speeds. You lighten the print head, you might be able to push that quite a bit higher.  Telling the controller to achieve a higher speed requires it accelerate harder, which it can do with a lighter print head.  But keeping all settings the same, assuming you're printing correctly with the stock print head, lightening the print head will not result in any change in print speed.