Temperature or filament width?

[Chris Fastie]

I have printed many dozens of this part with SainSmart PLA on a Replicator 1 Dual. Today I received my first spool of Toy Builders PLA (they have nice natural colors and now the price is the same). The first print fell apart in my hands as I removed it from the build plate.

This has never happened before, and these parts have to support cameras flying on kites (http://publiclab.org/notes/cfastie/03-27-2014/the-redstone-rig-is-ready). So I guessed that Toy Builders PLA needs either:

• higher temperature (to bond the layers better) or 
  
• a smaller feedDiameter (the filament is narrower than SainSmart).
  

So instead of using MakerWare's default PLA profile with:

• feedDiameter  1.77
  
• extruderTemp0  205
  
• feedstockMultiplier  0.93
• layerHeight  0.2
• numberOfShells  2

I made a profile with:

• feedDiameter  1.74
  
• extruderTemp0  210
  
• feedstockMultiplier  0.93
• layerHeight  0.2
• numberOfShells  2

This seemed to print well, but the (different) piece I printed next also broke immediately (this piece has never broken before).

I have printed these parts with more than a dozen spools of nine different colors of SainSmart PLA and never changed feedDiameter or feedstockMultiplier. So what is the best parameter to change next to make strong parts with Toy Builders PLA? 

Chris

[Dan Newman]

On 24/10/2014, 9:54 PM, Chris Fastie wrote:
> I have printed many dozens of this part with SainSmart PLA on a Replicator
> 1 Dual. Today I received my first spool of Toy Builders PLA (they have nice
> natural colors and now the price is the same). The first print fell apart
> in my hands as I removed it from the build plate.
>

> <https://lh6.googleusercontent.com/-o-kd97ryUVc/VEslZGGek0I/AAAAAAAA-_o/oSxi-A9Hs68/s1600/KAPtery-542-25.jpg>

>
> This has never happened before, and these parts have to support cameras
> flying on kites (
> http://publiclab.org/notes/cfastie/03-27-2014/the-redstone-rig-is-ready).
> So I guessed that Toy Builders PLA needs either:
>
>

> - higher temperature (to bond the layers better) or
> - a smaller feedDiameter (the filament is narrower than SainSmart).

>
> So instead of using MakerWare's default PLA profile with:
>

> - feedDiameter *1.77*
> - extruderTemp0 *205*
> - feedstockMultiplier 0.93
> - layerHeight 0.2
> - numberOfShells 2

>
> I made a profile with:
>
>

> - feedDiameter *1.74*
> - extruderTemp0 *210*
> - feedstockMultiplier 0.93
> - layerHeight 0.2
> - numberOfShells 2

>
> This seemed to print well, but the (different) piece I printed next also
> broke immediately (this piece has never broken before).
>
> I have printed these parts with more than a dozen spools of nine different
> colors of SainSmart PLA and never changed feedDiameter or
> feedstockMultiplier. So what is the best parameter to change next to make
> strong parts with Toy Builders PLA?

First, you should always measure your filament diameter with calipers and then
use that diameter when slicing. Doing otherwise, is asking for problems regardless
of what filament you use. When doing even modest size prints, having too low or
too high of a diameter will lead to under or overextrusion, either of which will
lead to print failures.

Second, the default feedstockMultiplier is bogus. MBI decided to just average
the value for ABS (0.85) and PLA (1.0) and use that as the default. You should
be using a value closer to 1.0. But further, you should calibrate your slicer
to your extruder and then stick to that value. (Mind you, some folks will
calibrate for each spool of filament. Me, I'm happy to just calibrate for PLA
and then again for ABS.) See,

http://www.sailfishfirmware.com/doc/tuning-slicer-calibration.html

for a discussion on calibrating your slicer to your extruder. The part of the
discussion which will interest you the most begins in Section 5.1.4.

Dan

[mfitz73]

the default extruder temp for my rep2 using pla is 230C
why is it lower for you?

[Ryan Carlyle]

Makerbot PLA is a bit higher-melting than most other PLAs. That's how they get away with recommending 230C. It's widely considered silly though, and they dropped the default PLA temp for the 5th gen line. Toybuilder PLA works great for me at 210-215C.

One common cause for weak layer adhesion with PLA is excess cooling directed at the nozzle -- you do need a brief moment where the filament is hot so the layers stick properly.

[Chris Fastie]

Dan -- I did try to measure the filament diameter, but I paid less than $100 for my caliper, so with a "resolution" of 0.01 mm, I can't really tell the difference between 1.74 and 1.76 mm. But by taking repeated measurements at exactly the same place and getting a feel for the range (and then repeating at other places), I gathered that the Toy Builders spool was about 1.74 and the last spool of SainSmart was about 1.76 mm. That is what prompted me to change feedDiameter from 1.77 to 1.74. Neither of the Toy Builders prints looked like underextrusion -- there is plenty of plastic flowing and no "see-through" layers. The parts look great. But I guess I should print a calibration cube.

mfitz73 -- Sorry for the confusion, I typically use the default profile but change 230 to 205 and might also change layer height, number of shells, and speeds in MakerWare. It's high time I started using custom profiles more regularly.

Ryan -- I have printed with SainSmart in the same room in all four seasons with extruder temperature from 200C to 225C. and never had a problem with layer adhesion. I don't have a fan to cool the filament. It seems odd that all the variability of many spools of filament, varying room temperature, and different extruder temperature has never produced an easily breakable part. Toy Builders PLA must be substantially different from SainSmart, so some careful testing is called for. Which means I get to try really hard to break my printed parts.

Thanks for your help.

[Dan Newman]

On 25/10/2014, 6:53 AM, 'mfitz73' via Makerbot Users wrote:
> the default extruder temp for my rep2 using pla is 230C
> why is it lower for you?

As Ryan pointed out, MBI's PLA when they released the Rep 2 worked okay at 230C.
Mind you, lots and lots of people had serious issues with PLA jamming with the Rep 2.
In a large part that was because they initially used a filament drive system which worked
adequately for ABS but poorly for much harder PLA.

But, there's possibly another reason behind that 230C. With the Rep 2, MBI started down
the path of trying to hide complexity from people. Rather than have users tell their
slicer, "I'm using PLA" or "I'm using ABS", MBI decided to just use settings which
would work okay for both: a single feedstockMultiplier, a single set of temperatures,
filament diameter, etc. They made things appear simpler to new users by reducing the
number of choices presented to users. Presenting a single temperature likely fell
victim to that, even if it wasn't the ideal temperature for either plastic. (For ABS
I use at a minimum 235C and more often than not 240 or even 245C, depending upon how
fast I expect the extruder to be generating output.)

As an aside, reducing the choices presented to users was arguably at odds with
their concurrent decision to focus on the professional market rather, putting
less attention in to the home market and abandoning the DIY market (their roots).

As a further aside, when MBI came up with that 230C recommendation in 2012, they were
using Village Plastics as their filament supplier. When VP got bought by 3DS in 2013,
MBI stopped using VP and went to a different filament supplier. So, it may no longer
be the case that MBI's PLA will work "okay" at 230C and that may be another reason
why, for the Gen 5s they dropped the temp. The company which bought VP, 3DS, is
Stratasys' #1 competitor. And Stratasys -- SSYS -- owns MakerBot.

Dan

[TobyCWood]

I have had spools from toybuilderlabs that sometimes needed a bit more heat. In particular the black... 

Each spool from every vendor varies in all aspects. Vendors that claim virtual reliability and repeatability are lying.

[teamca...@gmail.com]

I personally won't buy toybuilders pla anymore. I've bought 3 spools in the last year, and have had problems with each one. For some reason toybuilders stuff air prints way too easy (and yes, I do measure filament, calibrate extrusion width, install sailfish, etc). All three spools were super brittle once I used up a third of the filament. The stuff would just snap to pieces when coming off the spool and I ended up throwing the rest of each spool in the trash. And it's not a humidity thing, after my first spool I setup a dry storage box with rechargeable dessicant. It didn't make a difference. I've had really good luck with sainsmart. Their filament has actually been the closest to 1.75mm than any other brand I've tried. The only drawback with sainsmart is it usually prints stringy for me, regardless of temp or retraction settings

[Jetguy]

Let me give you a tip here. If you really had problems with Toybuilderlabs filament, you should have contacted them and they would resolve the issue. That's WHY I buy from them.

Part 2 is that they are selling Esun plastic and one reason to buy from them is they check every roll in a shipment for damage. Not every vendor checks for shipping damage.

I'm saying that because there are now quite a few other vendors selling Esun plastic and if you really don't like it, then fine, make sure you don't accidently buy more Esun plastic from another vendor.

The flipside is that have seen the exact stringing problems from many other brands of filament (Deltamaker) is one of the worst.

It simply comes down to plastic formulation. If your environment is abusive, pure PLA seems to degrade (hint, eSun and thus TBL is about the purest PLA I know of) and thus plastic with extra additives might not get brittle, but then it also might string like hell.

Again, I nobody should throw out a roll of plastic- especially not from Toybuilderlabs. If you have a problem, I'm sure they will try to make it right.

[teamca...@gmail.com]

Thanks for the feedback Jetguy, I've always had a lot of respect for your knowledge in all the forums that Ive seen you in, I need to go educate myself on Esun plastic. I keep my printer and filament in a spare bedroom in the back of the house, and when the filament isnt on the printer I have it in a big plastic storage bin with a large Eva-dry desicant pack in it. I do go through my spools fairly slowly, especially if its a nice color, so maybe I'm just not using it fast enough if it truly degrades? What would you consider an abusive envorinment? I'm tempted to try again because I need to buy another spool of white pla for some lithophanes I make. And even though I've had bad spools, I keep seeing nothing but good reviews, plus your personal endorsement so I'm willing to possibly give it another try. I couldve just ended up with spools from a bad batch. I did find other people online with similiar problems. My problem was that the PLA was so brittle that it would snap just from the small amount of flex needed to straighten out enough to enter the filament tube on my replicator 2. It was like in these videos:

https://www.youtube.com/watch?v=3gqnKMeafjo   (at about 1:10 into the vid)

https://www.youtube.com/watch?v=jreOtnRui6E

Any chance you have any input on what makes pla go so brittle?

On Sunday, October 26, 2014 2:19:43 PM UTC-7, Jetguy wrote:

Let me give you a tip here. If you really had problems with Toybuilderlabs filament, you should have contacted them and they would resolve the issue. That's WHY I buy from them.

Part 2 is that they are selling Esun plastic and one reason to buy from them is they check every roll in a shipment for damage. Not every vendor checks for shipping damage.

I'm saying that because there are now quite a few other vendors selling Esun plastic and if you really don't like it, then fine, make sure you don't accidently buy more Esun plastic from another vendor.

The flipside is that have seen the exact stringing problems from many other brands of filament (Deltamaker) is one of the worst.

It simply comes down to plastic formulation. If your environment is abusive, pure PLA seems to degrade (hint, eSun and thus TBL is about the purest PLA I know of) and thus plastic with extra additives might not get brittle, but then it also might string like hell.

Again, I nobody should throw out a roll of plastic- especially not from Toybuilderlabs. If you have a problem, I'm sure they will try to make it right.

On Sunday, October 26, 2014 4:59:08 PM UTC-4, teamca...@gmail.com wrote:

[Jetguy]

I always assumed it was moisture. I live in a relatively high humidity region and run AC constantly but I still find if I leave a spool loaded for longer than a day, the filament in the guide tube and not on the roll will get brittle and snap. But every PLA filament I have seems to exhibit the same failure. I have some genuine MakerBot filament, Toybuilderlabs, Deltmaker, Flash Forge, and several other off the wall brands.

Translucent filament seems to be the most sensitive and others are clearly less prone, but all have done it at some point. By abusive, I mean high humidity, maybe temperature changes, and most important is how much flexing the filament takes between coming off the spool before it enters the extruder. And small spools force a sharper bend coming off on the last bit from a spool. That's one reason I like TBL, they use MakerBot OEM sized spools.

Again, I find that the first say 6-12 inches of any spool can be brittle, especially if you ever loaded it previously and then unloaded and put the spool away. I wish I had the magic answer why, but it's just something I deal with. I always cut a new fresh end and try to flex it a little to see if it is brittle.

I know you are trying to store it right and I'll tell you, I'm horrible and never store mine in anything. They are just lying around. So that's why it's hard for me to understand how others have such problems and honestly, other than the first say 6-12 inches from a spool that was loaded and in a guide tube previously, I haven't had a lot of brittle problems. I must have 20-30 rolls of Toybuilderlabs filament.

[teamca...@gmail.com]

sorry about the slight diversion, back to the original topic though. Chris, to me that pic "almost" looks like an abs print that split due to warping, I recieved ABS by accident once (dont remember who the supplier was, it was through amazon) and I didnt realize it at first. The plastic was actually mislabled. I figured it out using the acetone test.(if you're not familiar, thats where you take a small piece of the filament and see if it melts in acetone, if it does, its ABS, if it doesnt its PLA)

On Friday, October 24, 2014 9:54:11 PM UTC-7, Chris Fastie wrote:

[Chris Fastie]

I have printed five more parts with the spool of Toy Builders brown PLA with five different combinations of feedDiameter and feedstockMultiplier. I also raised the extruderTemp0 from 210°C to 215°C. Here are photos of a top surface of an early and a late print in the series:

The later print looks better because there are fewer gaps between beads of PLA. But I am surprised that my big changes in parameters had such a subtle effect on the result. Maybe I am not modifying the profile file correctly (with ProfTweak), so I have attached the miracle.json file (renamed) for a late print in the series (MakerBot Rep 1, MakerWare, Makerbot firmware). This should be MakerWare's default PLA profile with a few parameters changed. 

The later prints are much stronger than the earlier ones, but not as strong as many dozens of prints with SainSmart PLA. The parts break between layers and not so much between shells, although crisscrossing beads generally prevent breaks in that dimension. In spite of the lingering gaps between beads, it looks like there is plenty of PLA being extruded. Here is a photo of a break in a part (that I made with some effort) showing a cross section of a two shelled outer wall. There is hardly any airspace between the layers or the shells (feedDiameter=1.70, feedstockMultiplier=0.90).

So I guess I will raise the temperature to 220°C and see if I get better adhesion among layers. For reference, the parameters I have used for dozens of strong prints with SainSmart PLA are 

• feedDiameter  1.77
  
• extruderTemp0  205
  
• feedstockMultiplier  0.93

Does anybody have a MakerWare profile for Toy Builders PLA for a MakerBot Replicator 1 Dual that produces strong parts? Or at least what do you use for feedDiameter, extruderTemp0, and feedstockMultiplier?

@teamcarlisle2 -- The PLA can't be ABS, because it smells like PLA and it did not curl off the blue tape with the build plate at 45°C!

Thanks,

Chris

[Ryan Carlyle]

I don't think you're getting the calibration process here. Changing variables around at random to try to find a combination that works is going to cause nothing but pain.

You need to MEASURE average filament width with a good set of calipers. Then you print a 100% infill 20x20x10mm calibration box, with roughly the same speed/layer/temp settings you intend to use. Then look at the cal box. If the top is bulging up, INCREASE feedstockMultiplier by about 0.05. If there are gaps or the top is concave, DECREASE feedstockMultiplier by about 0.05. Repeat until it looks right.

Once you get feedstockMultiplier tuned once, that value will always work for that one type of filament run through that specific extruder. Re-calibrate for different materials or if you change extruder hardware.

[Chris Fastie]

I increased the extruder temperature to 220°C and the parts are now strong. So I think the weak parts did not have much to do with filament diameter. That's a really big temperature difference between two PLA filaments.

Ryan, I think you are right that I am not quite getting the calibration process. So I made six calibration blocks with feedDiameter 1.73 and feedstockMultiplier from 0.80 to 1.05. The one with a feedstockMultiplier of 1.00 looked best, and it was very instructive to examine these blocks. However, I wonder about the process:

1. Digital calipers for $20 to $300 generally display 0.01 mm increments and claim accuracy of 0.02 to 0.04 mm. So unless you paid a lot for your caliper, reliably determining that one filament is 1.76 mm and another is 1.73 mm may not be possible. Are we fooling ourselves about measuring filament?
   
2. "Repeat until it looks right" is not exactly a repeatable calibration procedure. None of my blocks was actually concave or convex on top, but with some it was obvious that the nozzle was extruding more plastic than it needed to. Once you train yourself to recognize this, you don't need to print blocks. So are the calibration blocks more of a training device than a calibration procedure?
   
3. Are feedDiameter and feedstockMultiplier the only parameters that influence how much plastic gets extruded? What other parameters should you pay attention to (roofThickness, floorThickness, layerHeightMaximum)?
   

Thanks for your help,

Chris

[Ryan Carlyle]

You're right that cheap ($20) digital calipers are probably not accurate enough for this. Personally I prefer a good analog dial caliper. But if you can get +/-0.01mm precision (repeatability) then it will work fine. You don't need a super-fancy NIST-traceable caliper to do that. Accuracy is not critical because the multiplier fudges out most diameter offset error. You just need repeatable measurements.

Even high-quality filament generally has +/-0.02mm diameter variation along the filament length used on a large print, and often some ovality, so there isn't much point in measuring a lot more precisely than your filament's variance. You just need to be pretty close. It is obviously better to use high-quality tools and high-quality filament though.

If you think in terms of volumes, there is a HUGE difference between 1.70mm and 1.75mm and 1.80mm. 0.05mm diameter error is about 5-6% volume error, which is going to be clearly visible. But it's actually WAY worse than that, if you overextrude on multiple layers. 100% fill layers (including floors/roofs) are extremely sensitive to over-extrusion because the volume error accumulates on each layer. If you're 5% over on each layer, after 20 layers, you have an entire layer's worth of excess material, and your print looks like crap, and your nozzle/print gap is completely blocked. 

The reason why 100% infill calibration cubes work so much better than eyeballing a typical print is because they are 100% infill. That makes volume error extremely visible. Normal prints have non-solid infill layers and not a ton of roof/floor layers, and thus any excess just squeezes to the sides and is very difficult to detect. 

Measured filament diameter and calibrated volume multiplier should be the only parameters you need. Although your slicer may or may not have additional overlap fudges. For example Makerware allows you to set the overlap between infill and shells. My personal opinion is that you should calibrate volumes with zero additional shell/shell overlap or shell/infill overlap, but most people are happy with the defaults. 

[John B]

Ryan

I have been following this with interest. Everything you say makes perfect sense and over the past year or so that I have been printing with a Rep 2 I have done the calibration cube a number of times. Now for me I can not see or measure any appreciable difference in the cube. Last night I printed one with PLA @ 100% and a diameter of 1.74. And then another with the same filament but with the setting changed to 1.85 on the slicer. Both cubes are the same, should I not see a big difference with this large a change or am I missing something?

[Ryan Carlyle]

Under-extrusion is much less visible -- it mainly makes your parts weaker. So when you raise the diameter input, the slicer will extrude less filament, and it may not be all that noticeable. Basically you're getting air gaps and reduced packing density which aren't visible from the outside.