Ultimaker:Too hard for general users to achive high quality?

[polyoptics 0]

Hi all,

I am a new user wondering if the kit and printer are too complex and problematic for me.

I realize this is an impossible question, but I would like to ask your advice on the subject.

I have been following 3d printers for many years. I am a professional 3d artist and have been waiting for a time to jump into 3d printing when cost/quality were at a level that I think is right for my needs.

This type of quality would be fantastic (of course): :shock:

http://www.hive76.org/insane-3d-printin ... -the-micro

Reading on the forums here and on the google group, I see professional engineers trying to tackle problems and I feel like this is way, way, over my head. Also, though there are occasional posts saying that the build is very straight forward and getting it working is not difficult.

My best hope is that I could achieve this quality, or close, with careful building, calibration, a lot of test prints, and a little luck... along with some good tips from these forums.

My worst fear is that the printer becomes a source of frustration and unpredictability as it requires a tinkerer's mind and love of building machines and constant attention to keep it working. :oops:

Is there any advice that I could get on this or some stories of personal experience that you might be willing to share?

Thanks much,

Brian

[Daid 287]

I'm a professional engineer. In software that is. So I do have a bit of a tinkerer mind, but I rather have working mechanical parts then that I tinker with them. My printer needs to work, when I want. It's my tool, not my tinkering toy.

My print from today. With SkeinPyPy's default settings (only layer size to 0.1 instead of 0.2)

I didn't change anything in my machine, it's all stock hardware. It's not as good as some of the highest details Ultimaker prints, but I have never tried anything below 0.1mm, which already looks very good.

Hardware

It's simple, once it works, you can get this quality quit simple. I had some tinkering problems in the beginning, but that had to do with me not following the assembly instructions properly. After that I only added belt tensioners, and oiled it a few days ago. It's quite an amazing machine. Oh, and you will replace the stock fan shroud, but that's almost part of assembly instructions these days

The real tinkers can be found building RepRap machines, they spend more time fixing/upgrading machines then actual printing.

Software

The software options are a bit chaotic at the moment. There is ReplicatorG from Ultimaker, NetFabb (not free), Slic3r, Skeinforge, PrintRun. And then I'm not even mentioning some rarer options.

However, I'm trying to solve all that with SkeinPyPy. Especially for possible users like you. You don't want to tinker with 10.000 software options and configuration settings. You just want to print. SkeinPyPy allows everyone to do that. You don't have to take my word for it. Other people have been saying "I've been struggling for 3 months, and suddenly now I'm getting good prints thanks to SkeinPyPy". It's a project not aimed at making the best possible prints. It's a project aimed at making the best user experience.

[ddurant 0]

This type of quality would be fantastic (of course): :shock:

http://www.hive76.org/insane-3d-printin ... -the-micro

 

Some of the pix here will make you drool, too..

http://davedurant.wordpress.com/2011/10 ... of-prints/

Paul Candler is the one that did the headlining pictures. I'm not sure what he does for a living but I do remember him saying that the Ultimaker is his first printer and he'd only had it a couple months before doing that print.

I think one of the big problems new people have is that they hear Ultimaker can do crazy fast speeds and really high resolution so they try to start with BOTH of those, which turns out to mostly just make messes. There are always trade-offs in speed vs quality..

If you take your time and have realistic expectations then yes, you can do stuff like this.

[polyoptics 0]

Thank you both for your great replies!

In my research I read that I should definitely start with the stock firmware to get things going before I experiment with other types, so I would do that. For my purpose though, it would seem to me I will need to go the netfab/marlin route in order to get prints as smooth as I need them. Is it difficult to use the this software combination, I know I can load templates, but I don't really know much more than that.

I have currently been using shapeways which is very expensive and takes 3+weeks to receive a printed piece, but of course the quality is incredible. I know I probably won't be able to get that result from a printer that costs a fraction of what they are using, but if I can get close to some of those other guys prints (probably .06mm or lower) then I think it will be close enough.

Daid: Thanks for the encouragement. If I can get as far as you have gotten, I would be very happy indeed! It may not be enough for what I need, but it would put a big smile on my face.

Ddurant: Yes those images are amazing, I found the thread where he was posting his initial results and he seemed like a very advanced user, a modest fellow based on that post you read. I could be mistaken, but didn't he add a heatsink amoungst other things?

From my work in computer animation I am very aware of the quality vs time trade offs, so that should be ok.

I would love to hear some more first hand experiences if anyone wants to share!

[Daid 287]

Daid: Thanks for the encouragement. If I can get as far as you have gotten, I would be very happy indeed! It may not be enough for what I need, but it would put a big smile on my face.

You will.

I have a very big time constrain on my prints. I need to watch my printer while it's printing, so I only have a few hours for each print (my longest print was 6 hours). I have a cat who loves to craw into my printer. So I need to keep him out during prints. I cannot let it run over-night.

I'm also a bit inpatient The above dragon took 30 minutes for the chest (which is 100% filled as the glow-in-the-dark works better if it's fully filled) and 45minutes for the dragon itself. Both printed on average speed 80mm/s and 0.1mm layers.

The ultra-high quality Yoda (in gold) took 13 hours to print (if I remember correctly). Which is still a lot quicker then waiting for shapeways

[ddurant 0]

Ddurant: Yes those images are amazing, I found the thread where he was posting his initial results and he seemed like a very advanced user, a modest fellow based on that post you read. I could be mistaken, but didn't he add a heatsink amoungst other things?

He's certainly figured things out but I think he did those things while still pretty much a newbie.

For the heatsink, the original Ultimaker kits came with smaller motors. These worked fine but could get really hot on long prints. A number of people put heatsinks and/or fans on them to keep the temperature down - I put fans on mine, not because it made the machine better but because I was sick of touching the motors by mistake. Not hot enough to actually damage skin but I sure knew it when I touched one.

Ultimaker started including bigger motors around October or so, which run far cooler. I upgrade my machine to use these for about US$20/motor and now, even after like 5 hours of continuous printing, the motors are just over room temperature.. So.. Worthwhile investment for batch-1 machines, already fixed on new machines.

[polyoptics 0]

Is it just your cat and anticipation that keeps you watching the prints, or is there something you need to do while it prints or watch out for?

I have noticed that almost all the prints I see do not have overhangs/undercuts, is there any reason for this? Have you had any experience with models that need a lot of support structures?

I will be trying to print this specifically (shown here is the shapeways print):

 

I wonder if the radar piece could be printed as one part, or if I would need to split it up and manage the model to make the print possible, or model my own support structures, which I am not sure how to do with some of these thin curved surfaces with protruding details. -- that is one thing I liked about Shapeways. :roll:

Eventually I would like to try and print the big white guys shown here (shapeways print):

 

 

As you can see around the folded arms there are tons of undercuts, gaps and ..areas that would need supports, like the elbow?

Would these fit into a 'difficult to print' category?

[polyoptics 0]

Ddurant: I could be mistaken, but didn't he add a heatsink amoungst other things?

He's certainly figured things out but I think he did those things while still pretty much a newbie.

For the heatsink, the original Ultimaker kits came with smaller motors. These worked fine but could get really hot on long prints. A number of people put heatsinks and/or fans on them to keep the temperature down - I put fans on mine, not because it made the machine better but because I was sick of touching the motors by mistake. Not hot enough to actually damage skin but I sure knew it when I touched one.

Ultimaker started including bigger motors around October or so, which run far cooler. I upgrade my machine to use these for about US$20/motor and now, even after like 5 hours of continuous printing, the motors are just over room temperature.. So.. Worthwhile investment for batch-1 machines, already fixed on new machines.

Ahh I see! Thanks for the clarification -- I believe he also added the LCD screen and SD card reader, I may have read that that kit also comes with an improved way to read the temperature? Also, he mentions that in some of his blue prints he had stripes of black material as a result of sealing something around the extruder area (where he used black plastic), is that part of the regular build or an extra procedure? --Excuse my ignorance here please Do you think that these modifications (which I could *probably* do) are essential for printing at very high quality levels?

The stock kit comes with a fan mounted on the extruder (doesn't it?) for which I can print the fan shroud modification as Daid has mentioned, and the belt tensioners. Are there any other essential mods for high quality printing? I have read about feeder tube clips on both ends and modifications to the main.. (ugh i don't know the terminology yet, sorry) pipe that the material comes through to avoid backups and blockage etc...

I'm not sure how much of this will be needed, and I fear some of those fixes (although not mandatory) may become necessary if I encounter a problem, and are out of scope as someone without access (and skills) to proper tools and a workshop.

That is some messy post... sorry for that! Do you get what I mean though?

[Daid 287]

The LCD and SD reader are optional. Ultimaker is working on "productizing" it I think. Atleast, I've seen a few of those at the Ultimaker HQ. Same for heated platforms.

Stripes can happen if you switch color, and material is left in the head. If you look at my print above, you'll notice the bottom of the chest is black. This was because there was still some black on my printer head, and the glow-in-the-dark is slightly translucent.

The "sealing" he did is optional, normally the printer head is sealed with some Teflon tape against leaks. It's part of the assembly instructions.

The default fan mount is a bit... problematic. It has a tendency to disassemble itself. If you super-glue it it will stay in 1 piece. Other fan mounts provide different types of cooling. I have one with a very small nozzle, because I print small things I want localized better cooling.

You don't need access to a workshop. I did everything I do at home, with my tools being a box-knife, a screw driver and a hammer. It's one of the goals of the ultimaker to be easy to assemble and maintain. It's even shipped with the proper screwdriver you need for the type of screws it uses.

My experience with support structures is limited. I've used automated support from SF, with limited succes. My first print was a mess and the support was impossible to remove. But the dragon shown also had automated support for the hand on top of the chest (two separate prints, joined later). Removing support is quite a task. But my limited experience doesn't help in this case. Maybe some tweaks will make it easier to remove.

For:

 

I would print it as multiple parts, glued together. The "beast" can be 1 part with support (impossible to print the underside of the belly otherwise). I would print the radars as 2 different parts, the arms as different parts and finally the center piece as another part.

Any angle more then 60 deg slope is a problem for the Ultimaker. (with horizontal being 90 deg) However, there is a concept we call "bridging" shown in these tests:

http://www.thingiverse.com/thing:12925

which make it possible to print pure horizontal lines from a tower to a tower.

But, yes, these are in the "difficult to print" category. (awesome paint job btw on the radar beast) With support you get stringing problems. To explain stringing problems, see this horrible case of strings:

http://daid.eu/~daid/IMG_20110927_000944.jpeg

(Lower temperature greatly improved the amount of strings, and thickness of them, but the effect is still there)

[polyoptics 0]

Thanks Daid for your response! I have been rereading it and doing research since your post. You've been extremely helpful!!

It seems that support structures and models with undercuts/overhangs aren't really a hot topic, and I fear that although things are moving at a really high pace (regarding software/hardware) these issues aren't at the top of the list for development. I could be wrong on that, of course.

I wish I could find some information on anyone (just one person!) who is using the Ultimaker in the same way that intend on using it - to print character models, which will consist almost 100% with complex overhangs and protruding details. Which sort of leads me to believe that there must be reason for that.

Still a huge part of me just wants to take the plunge and see what happens!! I wonder what printers are on the horizon and what Ultimaker2.0 will be capable of... ?

[Daid 287]

Maybe this photo will turn you off:

http://daid.eu/~daid/IMG_20111212_194956.jpeg

It was printed with support, at 0.2mm layers. Cleanup didn't work very well, and the end result was very ugly. It was my first experiment with support. (The green dragon was the 2nd).

But how does it look before support removal? (picked this from google)

http://cdn.shopify.com/s/files/1/0030/7 ... 1286806837

 

[polyoptics 0]

Hah Yes those are scary photos. It would seem, as you mentioned, that printing in pieces is really the way to go for this sort of stuff!

Any pics of complex objects that have been assembled from multiple parts to avoid this? Your dragon hand was a good example and turned out really well by the way!

[Daid 287]

 

http://daid.eu/~daid/IMG_20111007_010749.jpeg

This tank was assembled from 4 parts (2 sides, 1 center, 1 top), standard 1 euro coin added for size.

Same for this one:

http://daid.eu/~daid/IMG_20111008_005844.jpeg

Both put together with superglue.

Note that these where made with thicker layers, and before I managed to sort out the software, and thus my extrusion amount was off (to little plastic). I should print the first tank again, it will look a lot better now I think. (It already looks quite good, black is very hard to take pictures from, as it shines like crazy)

Another photo you might like to see is:

http://daid.eu/~daid/IMG_20120125_211716.jpeg

 

[polyoptics 0]

Hey they turned out pretty good!!

I have been really amazed these past few weeks seeing how far things have come just in the last couple of years. I think for my purposes, its clear that I am expecting a little too much from the machine at this time, but its **SO** close. I imagine in the next year or 2, I will take the plunge. Until then I will continue to keep a close watch here.

Thanks so much Daid, and all the early adopters, for paving the way!

[Daid 287]

New print of the tank top:

http://daid.eu/~daid/IMG_20120327_ ... small.jpeg

Compared to the old print:

http://daid.eu/~daid/IMG_20111007_010722.jpeg

You can see some stringing between the barrels. The main differences between these prints are caused by a new firmware (Marlin) and easier to configure software (Cura)

[polyoptics 0]

That's a huge difference there, awesome!!

At this point I feel pretty confident in the print quality, although the software looks a little challenging, it seems the hardware is up to the task, and then some.

For the past few days I have been obsessing over support structures. I had never considered that as a factor before, but it seems now that this is more a limiting issue than the resolution. I havn't had much luck, except for some old posts from 2010 where some people are using PLA as a dissolvable support via hard chemicals. I have seen the printed moving gear images from flicker using this technique with PVA (the guys name escapes me),but I am pretty sure he is not using a makerbot or ultimaker.. TOM? And that he may even be modeling the support structure himself and merging stl files.

This has me looking at the new makerbot replicator, although there is no proof yet, just talk, it seems like a dual extruder and PVA could be an amazing asset for creating supports from safely dissolvable material. I realize that most any printer can be moded with an extra extruder, but being a not so tech savy person, it's a major plus to be able to buy a pre assembled kit with dual extruders.

The stopping point seems to be the software. It seems to be so new (dual extruding pla, and pva support material) that there are no existing tests and therefore only speculation about what software may or may not support this feature.

It seems at the best of times, support structure has taken a back seat to other issues, perhaps this new-ish technique with dual extruders will push things along a little faster?

That could be a long wait... meanwhile, affordable liquid resin printing ( looks incredible) is slowly starting to pop up here and there in prototype videos and images.

Thoughts?

[joergen 1]

yes, I agree with you, the hardware is pretty good, and IMO it is mostly the software that is holding it back, but it has made such great progress over the last 6 months, that there is hope.

http://ultimachine.com/catalog/print-ma ... va/pva-3mm

just stocked up on PVA, which is water soluble, and netfabb can print the support material with it. quite possibly, the other slicer can do it as well, I just haven't seen the options for it yet.

[polyoptics 0]

Are you using a printer with dual extruders?

Also, some info on Slic3r and Replicator here:

http://forums.reprap.org/read.php?263,1 ... 86,quote=1

If you ever get to test this, please post your results!

[Daid 287]

quite possibly, the other slicer can do it as well, I just haven't seen the options for it yet.

It's "doable" in SF. But it's not a simple "switch you turn on/off". You'll have to make a "support_start.gcode" and "support_end.gcode" which switch to the support extruder.

As for the support question. The top of my small tank, see those barrels? They are floating. During printing they had 3 small support blocks under it which I cut off after printing. Automatic support works, but sometimes manual support in the model is better.

[polyoptics 0]

Thanks Daid -- I also found this larger example, which is promising:

http://www.thingiverse.com/thing:3856

Regarding Replicator and Ultimaker -- I get the feeling after reading a lot about them both, that Ultimaker is easier to get higher details with. As far as the Replicator goes, it does have dual extruders which may (or may not) prove to be extremely valuable, however it seems that getting down to, and below, 0.1 mm is a tougher task?

Personally, the idea of a prebuilt kit, although not really in the spirit of true DIY, is very appealing to someone like me. I can see why MBI is going that route, for now.

As for the Replicator being the latest printer with the most stuff... I can only imagine that we will see a new Ultimaker kit at some point in the near-ish future. Interestingly, the ultimaker forums seem to be quite well developed and active VS the makerbot forums.

Both printers use different extruders (correct?)

Daid, how do you feel about the Replicator and Ultimaker -- I know they both have pros and cons .. - I am curious what your thoughts are regarding these two machines and how they approach 3d printing in different ways. Its possible this topic hasn't been beaten to death yet, right?

[Daid 287]

Regarding Replicator and Ultimaker -- I get the feeling after reading a lot about them both, that Ultimaker is easier to get higher details with. As far as the Replicator goes, it does have dual extruders which may (or may not) prove to be extremely valuable, however it seems that getting down to, and below, 0.1 mm is a tougher task?

The Ultimaker sort of has room for dual extrusion. There is room in the printer head for a 2nd hot-end. But I'm not sure if anyone has build one of these dual extruder Ultimakers yet.

 

Daid, how do you feel about the Replicator and Ultimaker -- I know they both have pros and cons .. - I am curious what your thoughts are regarding these two machines and how they approach 3d printing in different ways. Its possible this topic hasn't been beaten to death yet, right?

Nope, topic hasn't been beaten to death. But I am biased, let me warn you about that. I've also never seen the replicator except on pictures. And I have an Ultimaker, which works great. I've also visited the Ultimaker HQ twice already. So I will like the Ultimaker better. But I'll try to put them side by side.

The Ultimaker has a 205x205x220 build area, for a 350x350x3?0 machine. The Replicator is advertised as printing BIG objects, and only has a 225 x145 x150 build area for a 320 x 467 x 381 machine. That's less build volume for a bigger machine.

The Ultimaker uses 3mm filament. Replicator 1.75mm. The 1.75mm filament is more expensive.

The Ultimaker prints at speeds between 50mm/s and 150mm/s. The Replicator is advertised with 40mm/s.

The Ultimaker does not have a heat bed option. The replicator does. Making printing with ABS much better suited for the Replicator.

€1,194 vs $1,749, making the Replicator slightly more expensive. But add shipping and you'll find that it depends where you are what the most expensive option is. Replicator also comes with an LCD, which isn't buyable for the Ultimaker yet.

IMHO, the Replicator is a step back, not forward from the Ultimaker, even tough the Replicator was released later. The Replicator is simply a larger Thing-O-Matic with a 2nd extruder option slabbed on there. The Ultimaker is designed from scratch as the best possible machine they could think off.

For the dual extrusion. Like I said, the Ultimaker has room for it. But hasn't "made" it yet. I don't fancy dual extrusion that much. (I DO fancy chocolate extrusion, but that's another thing) It's also expensive with the Replicator (+250$, that's half the price of a PrintrBot!) Also, oddly enough, their dual extrusion print examples are blue and green. Why would that be? Hint: You don't see traces of blue in green that well, or the other way around.

The stringing problem is huge for dual extrusion. And if we find someone who finds time to experiment with a lot of settings to find the best way to avoid strings then we might be able to solve the stringing problem. (there are like 20 settings in Skeinforge designed to help against stringing) Which would make dual extrusion a more interesting option.

As for the whole support structure question. The larger the object, the larger the details, the easier it will be to remove the support. If you are willing to share the "radar beast" model files with me, then I can try to print it with support and photo it for you. If you want. I won't share the model files with anyone else. I think it would make a cool print. (The tank shown earlier is also a "unsharable" model)

[polyoptics 0]

Thanks Daid,

I realize you are biased, I think its a good way to get some critical feedback on the Replicator.

Regarding the points you mentioned, that I am unsure of:

1)- Filament size -- Cost aside, are there any differences with the thinner filament size? Why do you think MBI went this route? I found this but I am not sure what implications that may or may not have on quality:

>>1.75 requires less torque, so you can turn the current down on the stepper.

>>1.75 requires more steps for the same flow, but in reverse this gives better flow control especially with a small nozzle.

2) Print speed -- is not an issue for me, unless faster prints can impact the quality in a positive way, which seems counter intuitive, but I may have read something about faster speeds creating thinner threads of plastic and higher quality ??

3) heated printer bed -- I do like this, even for PLA it seems to help - on the Ultimaker with PLA wouldn't a HPB help with quality, even slightly?

4) Ultimaker uses the boden extruder, not sure what the Replicator uses, but there are pros and cons for each correct?

TOM vs the Replicator:

1) - The TOM uses the mk7 extruder, and the replicator the mk8, I have no idea what the difference is there, do you? Here is what I found, but it still leaves me vexed:

> > * What is the difference between the MK8 and MK7 extruders?

>

> Mk8 is just a Mk7 with an angle on the nose of the extruder. We wanted

> an angled nose on the Mk7, but some manufacturers tolerances made it

> hard to get right.

2) - Also, the TOM has the moving platform and the Replicator uses the 'ultimaker style' moving extruder, which should show improvement in quality, correct?

From researching it seems the old fashioned moving build platform on the TOM, is enough reason to make the Replicator a 'better' printer .. perhaps?

I see what you mean about there not being a huge difference between the TOM and the Replicator though.

---

As for my radarFox model. Yes I would absolutely love for you to test some thing on it, if you don't mind!!! It's a model that I would never want to go public (at this time) but I feel very confident that I can trust you with that. Maybe you can PM me your email address and I can send over the file tonight. Would you prefer the file as STL?

Thanks Daid for taking the time to chat about this with me, I am seriously ignorant when it comes to the details. I really appreciate it and I hope I am not trying your patients too much! :oops:

[Daid 287]

You can email me at daid303@gmail.com, it's no secret ;-) I won't share them at all. And I have the tools to "fix" any problems I would have with them.

The difference between a TOM/Replicator extruder and the Bowden extruder is the location of the "extruder drive" with the TOM/Replicator it sits on the moving head, directly controlling the extrusion. With the Ultimaker it hangs on the side of the machine. Which causes more flex in the filament drive, and a pressure buildup. I think the TOM/Replicator design is better against strings, while the Ultimaker design is better for speed and move quality (less mass to move around it higher accuracy in the movement)

[polyoptics 0]

Interesting, I suppose there is no perfect solution yet! -- I sent along the radarFox file, let me know if you get a chance to play, and thanks so much, again, for all your help!!

On a side note, I wonder when we can expect Ultimaker2.0 ??

[ddurant 0]

quite possibly, the other slicer can do it as well, I just haven't seen the options for it yet.

It's "doable" in SF. But it's not a simple "switch you turn on/off". You'll have to make a "support_start.gcode" and "support_end.gcode" which switch to the support extruder. ...

If you use the same material for support as you do for the rest of the object, SF already does this. The trick is telling it to print support structures at a lower density and enabling it is just a couple easy(ish) settings..

It's pretty dated but I wrote about this a couple years ago. This was with a crappy makerbot DC extruder - what we have today is a lot more flexible..

edit: and, if we do move to dual extruders on Ultimaker, the setup isn't really that much harder. Somebody makes good support_start/_end.gcode files then everybody else just plug 'em in. It's really not a big deal.