PDA

View Full Version : Micro 3D Printer print... Not good!! :(



Paul_Boland
07-02-2015, 06:29 PM
Hi Folks.

I got my Micro 3D Printer a few weeks ago and have been playing around with it. First thing I printed was a filament holder at medium quality and it turned out great! Today I created my first piece in Lightwave to be printed. I used the STL exporter plugin mentioned in a thread on these forums from the ChromeCow website. The model loaded into the Micro 3D software fine. I set it up and set it for BEST print quality!! Let's just say I am not impressed with the end result...

I'm wondering how good that STL exporter really is? I tried loaded the STL file into an online STL viewer and the results were not great. I can't contact M3D about this at the moment as their website is down, but I posted the pic to their Facebook page. I'm not sure if it's the printer at fault, or if Lightwave exported a bad STL file.

Anyone got any thoughts on this please?

m.d.
07-02-2015, 07:37 PM
Nothing wrong with the STL exporter...been printing with it for a year, some pretty intricate and precise parts

Most issues with printing come down to speed, temperature and acceleration....
It takes a fair amount of trial and error to get these down....not sure what filament you are using but you may want to slow it way down and see the results you get....maybe by as much as half the speed

Also just my opinion but do yourself a favor and look into simplify3d.....by far the best and simplest 3d printing software out there. Honestly makes a huge difference over manufacturers stuff....check that your machine is supported first

What material are you using....what temps, and what is your printing speed set to, I might be able to help you dial it in
Each different filament requires different speed and temps, for instance nylon to have to print at less then half the working speed for ABS or PLA

m.d.
07-02-2015, 07:39 PM
And remember the STL is just the shape...
It's your printing software that decides the tool path speeds and temps, that's were simplify3d is way way ahead of the factory stuff

Paul_Boland
07-02-2015, 07:48 PM
Hi M.D.

I checked out Simplify3D, it doesn't work with the Micro 3D Printer. I'm printing with PLA, light blue. I think the temperature is set to 210 (c? f?). I can slow down the print speed I think.

Sensei
07-02-2015, 07:52 PM
You're using plentiful of n-gons (show wireframe)...
You're using probably not connected geometry.
Try making it single connected mesh..
You're using obsolete LightWave version (10.1).
Dunno how it would work for us with new version..

Paul_Boland
07-02-2015, 08:02 PM
I think perhaps there is too much geometry there for Lightwave and the printer to work with. Even through I have Boolean it all together, and tripled the polygons, when I zoom in I'm seeing rogue polygons inside the model. Perhaps this is confusing the printer??? Also, the STL exporter is throwing up an error message saying invalid argument number or something like that when I try to export the entire tripled model. I'm guessing the text is just too detailed for it all to handle.

As for working with Lightwave 10, can't afford to upgrade so have to work with what I have.

m.d.
07-02-2015, 09:07 PM
Looks like it will support it, probably no built in confuse though
Edit "no built in presets",....stupid autocorrect:)
http://micro3dfans.com/viewtopic.php?f=3&t=282

Cura is also a very good open source alternative.

210 is pretty good, but if you have lots of oozing or dependent on the PLA 190 can work as well....it's so dependent on brand etc
Mesh lab should handle STL exports if LW 10 ain't cutting it....mesh errors can complicate things but the better slicers can handle it

JoePoe
07-02-2015, 11:30 PM
I think perhaps there is too much geometry there for Lightwave and the printer to work with. Even through I have Boolean it all together, and tripled the polygons, when I zoom in I'm seeing rogue polygons inside the model. Perhaps this is confusing the printer??? Also, the STL exporter is throwing up an error message saying invalid argument number or something like that when I try to export the entire tripled model. I'm guessing the text is just too detailed for it all to handle.

Uh, too much geometry for lightwave/printer to handle?? Really? Occam's razor suggests you don't have to look much further than your rogue geometry and error message. Care to post the model for inspection ?

Paul_Boland
07-03-2015, 06:33 AM
Uh, too much geometry for lightwave/printer to handle?? Really? Occam's razor suggests you don't have to look much further than your rogue geometry and error message. Care to post the model for inspection ?

I am becoming more and more convinced that this is the real problem, that there is bad geometry in there somewhere. I am going to try and see if I can sort it out. In the meantime, I have no problem with sharing the model, find it attached to this post. This model has all the elements on separate layers before they are Boolean together. If you want to try and sort it for me, that would be hugely appreciated, but please don't put yourself out on my behalf.

Thanks.

Sensei
07-03-2015, 06:57 AM
After fast review... All sins mentioned already in post #5 plus geometry inside of geometry, and flying in air geometry (2nd layer)...

JoePoe
07-03-2015, 07:54 AM
Okay, sooooo ...... Paul.

I was hoping to see the model "as printed" to identify problems including anything that may have happened after the final Boolean (but before triple.... very hard to work with after that step).

But we've got a bunch of stuff to talk about with the layered elements.

Layer 3 is the only one that looks okay to me at this point.

Layer 1: the two halves are not joined.

Layer 2: You're missing the vertical face on the front of the floor. But also, the vertical front face of the side walls aren't joined to the floor. You can see it's one big edge.... should be two. 128804. I mean one edge split in two sections vertically. Now there are two edges (one long and one shorter) right next to (on top of) each other.
A lot of the elements in this layer are crisscrossing and intersecting each other. Some print software can handle that.... some can't. Did you boolean all of this stuff together? Do you even need the floor of this piece?? Looks like it's underneath the main shield level.

Layer 3: :thumbsup:

Layer 4: you've got 5780 2-point polys. After getting rid of those Unify Polys got rid of 12 more polys. And there are some interior poly walls.

So all of that needs to be addressed before the final Boolean. Which by itself can cause some issues. Remember to merge afterwards and check for left over "rogue" geometry ;).

Edit: Aw heck....went ahead and fixed it up. Just a couple minutes.
Layers 1-4 corrected. Layer 6 is all Booleaned together. Layer 7 all tripled.

Edit 2: Almost forgot to post the thing :D. BTW, you've got some real delicate geometry in the D of Dungeoneer. Hopefully that's what puts the Micro in Micro3D. If this doesn't work then it's your speed and temp ;).

Edit 3: Had to change one last trouble spot. See post below for .lwo.

Paul_Boland
07-03-2015, 09:56 AM
WOW!! Thanks for all this help, much appreciated!! I was STUNNED to find the two parts of the shield in layer 1 were not unified! I have fixed this and it is now one solid object. I have removed the backing bit on the portcullis door and also separated the frame and the portcullis into individual layers. I have removed the two point polygons from the KnightTrek Productions logo and unified the polygons, removing another twelve. I am attaching the new model file to this post if you care to check it out. Thank you all again for the help, much appreciated!!

JoePoe
07-03-2015, 11:00 AM
Okay, so I haven't looked at your revised object yet.... but I'm posting my fixed up version.

There was one extra tricky spot. The bottom right sharp corner on the "u" of Dungeoneer was crossed over itself.

Layer 1-4 = separate elements (Layer 2 all booleaned together & layer 4 reconstituted the tripled text back into quads and ngons).
Layer 6 = Booleaned together.
Layer 7 = tripled and ready to go.

------

Looked at your revisions:
So far so good.
Layer one is fine.
Layer two is fine.
layer 3... all those tubes need to be booleaned together.
layer 4.... fix the u
layer 5.... remove internal polys in Rs Os and D.

spherical
07-03-2015, 03:09 PM
Ignoring the bad model for the moment, as JoePoe is on that case, the issue I see with the print looks like the Prime/Suck values for the De-string function are not set correctly. This causes ooze and stringing that accumulates on the nozzle and eventually gets dragged off. These values differ from material to material and temperature to temperature.

210C for PLA is okay. Might try going down as far as 195C. Be aware that feed rate (velocity of the head across the surface) has a relationship to temperature, because they have a relationship to flow rate. If you increase the feed rate, the temperature must go up. You're forcing goo through a tiny orifice. The faster the head moves, the more goo per second is required. The more goo per second, the lower viscosity it needs to be in order to apply the required amount through that tiny hole. Inverse is true as well. Slower means thicker or you get ooze and too much material deposition.

Be wary of models found on Thingiverse. The majority are junk and need to be fixed before slicing. As you are learning, models need to be clean and watertight. No internal cruft. No holes. One solid surface or the slicer will get confused as to what is the inside and what is the outside.

If your printer can accept standard G-code, get a copy of KISSlicer. It has a strict mesh analysis function that will balk when it sees cruft in a model and tell you what and where it is.

Attached is an instruction sheet that I put together to calibrate KISSlicer to printers (or the other way around). The concepts in it should be able to be applied to any slicer.
128814

BigHache
07-03-2015, 09:19 PM
On the point of models being "water tight" or manifold, layer 5 brings up problems. In Edge Mode bring up the Statistics panel. All edges should join to exactly 2 polygons. No more, no less.

128817

The type knockouts generate an internal poly at the cut line when you extrude. Always gotta remember to delete them.


128815

Here you have internal polys below the shield surface. I've been able to get away with this on Shapeways, but my slicing software has never allowed me to have this. I haven't necessarily HAD to boolean everything, but I do need to at least move the surfaces to where they are exactly touching. Boolean/unifiying polys is a better practice, but internal polys should generate bad results.

ianr
07-04-2015, 09:34 AM
Yep big hache right,
There STL open source checker programs out there>
BUT Water tight... means, weld, weld, weld, Paul!
Standard LW 3d Typefaces where hassle.
' you gotta check yer boat fer leaks before U float!'
BTW what size is that word 'productions' meant to be,
cos in millimetres terms/point size terms can a definition
be reached that small with your rig?


p.s. Wish i had a Metal printer?

Paul_Boland
07-04-2015, 01:18 PM
WOW!! Thanks very much for all the feedback!! I have a lot to take on board here. I'm hoping to make another attempt at printing the shield soon, I will keep you posted.

JoePoe, thanks for a updated shield model, very much appreciate you taking the time to do this for me.

Spherical, thanks for the info about KISSlicer. Mesh analysis is what I want really so I'm going to check this out. I'm not sure if my printer will work with the software, but if it can look at the mesh and highlight problems, I'm all for it!!

BigHache, I read somewhere that having a mesh inside a mesh was ok for printers but I'm guessing it's not??? Boolean stuff together can be a nightmare and this model particularly so with all that lettering. I don't know how to fix the edges problem though...

Ianr, I've welded my two half of the shield together into one solid object. Checking the boat for holes is not an easy task though, I'm hope KISSlicer's mesh analysis will help me with that.

I'll keep you all posted on progress...

Paul_Boland
07-04-2015, 01:39 PM
JoePoe, I can't open your model. When I try, Lightwave prompts me for an extension called NodeHandler??? I'm guessing this is because you're working with a higher version of Lightwave than I am, I'm using 10.

Dexter2999
07-04-2015, 02:27 PM
Try telling it to skip the node handler.

JoePoe
07-04-2015, 02:28 PM
Just say No when prompted and keep going (it may ask more than once, and yes I'm in v 11.6). Should still open. Nothing special going on.... just polygons.

Also, when using the STL exporter it automatically calls up Mesh Repair which gives some good feedback on the viability of the model... and tries to fix things for you. That's how I found the issue with the "u" (my eyes aren't THAT good ;)).
But once identified I like to fix the issues myself (and in this case, Mesh Repair didn't do a good job anyway). Not sure if Mesh repair was in v 10 though.
At any rate, it passes with flying colors now..... but test away.

Paul_Boland
07-04-2015, 09:29 PM
Ok, I got it to load. I see the Layer 7 all Boolean and ready to go. It's 4:25a.m. here now so I will try and print this tomorrow. I hugely appreciate all the help everyone here has given me.

- - - Updated - - -

Right, I got it to load :). I can see layer 7 all booleaned together and ready for printing. I'll do it tomorrow as it's all hours of the morning here now. I'll keep you posted...

spherical
07-04-2015, 10:58 PM
Just be sure to copy/paste the polys from that one layer into a separate object, then convert to STL. If you don't, there will be duplicate geometry; most of it overlapping polys. STLs don't understand layers, so everything gets put into the one model. Overlapping polys will wig out any slicer, and the difference between layer 7 and layer 6 is significant.

BTW, there are a number of unnecessary polys on the underside surface. You can weld the ones on the loop just inside the perimeter to the ones adjacent to them on the perimeter. Then weld the one just above center to the one in the center. Then weld the ones on the straight upper side to their corresponding points on the side where the first break happens. Then identify and delete the 1-point and 2-point polys left over from the reduction operation. That will get you a model that looks like the attached:
128821

Don't know about your printer or slicer, but KISSlicer needs the model sitting on the origin in +Y; not in the middle of it. The latter places the sliced file into the bed. It needs to be sitting on top of it or you get a nozzle crash when it goes to print the first layer. Most printers will prevent this happening, but you only get the top half of the model; if it tries to print it at all.

Below is a screen grab of the KS Mesh Error Display when it loaded your original model, after my having put all geometry onto one layer:
128823

Paul_Boland
07-05-2015, 08:50 AM
Thanks for the additional info. I've set up the new layer 7 tripled model in the Micro 3D Printer and the new print has begun... KISSlicer doesn't report any issues with it so I'm keeping my fingers crossed. Print time is over three hours so bear with me and I'll let you know how it goes...

BigHache
07-05-2015, 09:04 AM
I read somewhere that having a mesh inside a mesh was ok for printers but I'm guessing it's not??? Boolean stuff together can be a nightmare and this model particularly so with all that lettering. I don't know how to fix the edges problem though...

It really depends on the software. My experience is the software in home printers cannot understand it. There may be occasions where it's OK, but best practice is to not have overlapping.

I totally understand understand about Boolean complicated geometry. I've spent quite the time doing this. If you don't already have it, this will change your life: 3rd Powers Boolean Tool (http://www.3rdpowers.com/index_store.html). I bought this plug-in specifically for 3D printing. Doesn't necessarily make your geometry perfect, but I've found it to get me a heck of a lot closer to the end goal and save lots of time.

Paul_Boland
07-05-2015, 10:06 AM
40% complete printed and continuing. Gunk strands building up around the mesh! Not happy!! Ggrrr...! I'll keep you posted...

128828

Paul_Boland
07-05-2015, 10:26 AM
58%, it's starting to print the letters. Meanwhile the shield surface is a mess!!

128830

JoePoe
07-05-2015, 10:31 AM
Uh oh.
It's scaled up from the lwo right? That object was only 100mm in the biggest axis, and some of the lettering elements are only .12mm wide. The object looks bigger in the bed.... or is it a really small bed?

Spherical had a good point too..... did you rest the object on the ground?

Other than that, (if I do say so myself ;)) we can rule out the model (or not..... :stumped:). Time to focus in on all that other technical stuff.

Paul_Boland
07-05-2015, 12:54 PM
Uh oh.
It's scaled up from the lwo right? That object was only 100mm in the biggest axis, and some of the lettering elements are only .12mm wide. The object looks bigger in the bed.... or is it a really small bed?

Spherical had a good point too..... did you rest the object on the ground?

Other than that, (if I do say so myself ;)) we can rule out the model (or not..... :stumped:). Time to focus in on all that other technical stuff.

Ok here it is, printed. Yes, the Micro 3D Printer is tiny so I made the model to fit the print bed at 100mm and to be fair, the printer did print at that size. But I'm not overly happy... While KnightTrek and Dungeoneer came out ok, only the P in Productions is printed. The frame of the portcullis is there, but the portcullis itself is not. And the whole surface of the shield is covered with crust! Yuck! And this was printed on BEST PRINT QUALITY!! Bah!! Yes, the model is resting on the print bed. I agree, it's not a problem with the model now, it's an issue with the printer. Going to show this to M3D and see what they have to say. In the meantime, I'm printing it again now, same settings, but with my silver filament, to see if it looks any better.

128831 128832

- - - Updated - - -

Another pic.

128833

Paul_Boland
07-05-2015, 01:55 PM
BEST PRINT QUALITY! Nasty...!!

128834

spherical
07-05-2015, 04:27 PM
That has nothing at all to do with the quality setting. I see whole, round, uniform diameter strings in this view. This says that the flow rate is not large enough for the layer thickness chosen. IOW, the new layer is not being extruded onto the previous layer with enough "interference". An extrusion cross section should be round-sided and mostly flat. If printing at 100% fill, each path should completely meld with the previous one. There should be no division between them and it should push well upon the layer below.

Did you run the calibration procedure I posted? This is essential. Trying new prints over again without doing this is a waste of time and materials. There are a few things that can cause issues like this and getting to a good Thin Wall and Solid Cube Calibration can lead you to what is actually wrong.

This could be a mechanical issue with the transport mechanism. A few layers print OK and then fail. Points to belt tension or stepper motor skipping; where the layer increment is as expected for a while and then jumps to a thickness that is too high. When that happens, the extrusion isn't squished into the layer below and has no adhesion.

If the stringing happens after the flat section is completed, this points to bad de-string settings. The flow needs to stop and a small retract is performed to pull the extrusion back into the nozzle just a bit (called: "Suck") to prevent ooze just dribbling out of it onto the print while it travels across empty space to the next part. Just before the next part is reached, a short, high speed positive flow (called: Prime) is done to re-fill and pressurize the nozzle to the condition is was prior to the Suck. The letters create a LOT of these jumps.

If your printer uses a Bowden push/pull cable method of feeding filament to the extruder, where the stepper motor is not right on the nozzle but at the other end of a long tube, the Prime/Suck values must be set extremely high. There is inherent slop in a Bowden cable and hysteresis alone will just ignore small back and forth moves, thereby essentially not doing any de-string at all.

Therefore, I would strongly suggest NOT using this complex a part as your first project. Start simple and get to know your hardware and how to set it first.

The industry would like everyone to believe that this is as easy as falling off a log. It isn't. I pity the librarians out there that have to nursemaid 3D printers that have been thrust upon them. May get to the point where it is "push a button and walk away easy" but it's not now.

spherical
07-05-2015, 04:45 PM
OK, just checked the M3D site and, while I have yet to determine the filament delivery system type, I see something that is typical. They say that their software "automatically sets the temperature for the filament loaded into it". Bull pucky. This is probably true for the filament that they sell but they also state that any filament may be used. Filament from different manufacturers has different physical properties. Heck, even filament of a different color from the same manufacturer requires different settings; sometimes radically different, to obtain good results. Sometimes, I've had to adjust when using the same color from the same manufacturer. Many things can change from reel to reel. Melt temp can be off by 10C-20C, diameter can be off by 5%. Both of these will require changes in the settings to bring the extrusion back into the sweet spot. I'll do some more research, but this was a big red flag.

Paul_Boland
07-05-2015, 05:00 PM
WOW!! Thanks for such a technical reply, Spherical. I will work through it tomorrow, too late now. I promise to do your full calibration system but I'm not sure if it will work with the Micro, the printer software interface is child-friendly (as in there is very little user input, it's really just clicking a button and go!).

Here is my finished print with the silver filament. It looks better than the first but I want a print on par with the ones M3D use to promote and sell the printer!

128846 128847

spherical
07-05-2015, 08:07 PM
They say that the printer is compatible with other slicers. This means that KS will work and the printer will read its G-code. That means that KS will be able to fine tune its settings, based upon the results you obtain from the calibration, plus any minor adjustments necessary for a particular material/speed/object, and the printer should follow those settings. If, on the other hand, there is no way to eliminate their one-size-fits-all-and-you-don't-need-to-know solution, then it really isn't compatible with open source slicers as they claim. Their Quality setting is nothing more than choosing the layer thickness and, along with that, a default set of adjustments to feed rate, flow rate and temperature. This is not enough to be able to print everything without issue. Too many variables.

So, the Best Quality setting is just the thinnest layer possible, so that the side surface resolution becomes finer than the lesser quality settings are. Their smallest layer thickness is 0.05mm. That is asking a lot of an FDM printer that forces hot goo through a hole. The nozzle is prone to clogging and, hence the drive mechanism prone to stripping the filament, when the area into which it has go force goo is that small. Also, it had better slow the feed rate down A LOT when printing that thin a layer. You have to think about the physics involved. Just like the speed/temperature dynamic, the available area into which material is to be forced creates serious resistance that must be factored in.

I can get ours down to 0.1mm-0.075mm but the experience level required to get all of the factors into alignment and dance well with each other is very high.

My suggestion is to choose a lower quality setting, say 0.25mm or 0.5mm and see how things go. Then get fancy.

spherical
07-05-2015, 08:41 PM
I promise to do your full calibration system but I'm not sure if it will work with the Micro, the printer software interface is child-friendly (as in there is very little user input, it's really just clicking a button and go!).

Yes, figured that. But, they also say that there's and Expert Mode that allows use of other slicers and filament. That's what you'd use.


It looks better than the first but I want a print on par with the ones M3D use to promote and sell the printer!

Not to be flippant or anything but, good luck with that. You have to know that the parts are chosen for the best presentation. After all, you don't really know if they are demonstrating using the default settings or not. I'm sure you'll eventually get there but it will take knowing a large portion of what the designers know.


While KnightTrek and Dungeoneer came out ok, only the P in Productions is printed.

I'll bet that the other letters didn't even generate a path, due to size. If a feature is smaller than the Extrusion Width x 2, so as to create a full Loop, it would require Crowning to print a single Extrusion Width, and I don't imagine that their slicer knows about Crowning; it's a KS feature used to fill in where a complete Loop misses.

EDIT: Checked with KS. Indeed, those letters are too small. KS generated only Crowning paths (single width, no complete Loop) until I dropped the Extrusion Width down below the diameter of the 0.35mm nozzle. Once I got to 0.30mm, the P would generate a tenuous, but complete, Loop. The other letters remained as single paths. That's not good for many reasons. Usually, Extrusion Width is at or slightly above Nozzle Diameter. With it less than Nozzle Diameter, you don't get any squish and, hence, weak or no adhesion; which looks like the issue with your prints.

gordonrobb
07-06-2015, 03:20 AM
Do your supplier not offer sample models, and have you tried printing one of them?

ianr
07-06-2015, 05:06 AM
Sorry to hear ,i told u so,
But i did say the word 'production'
was too small.

I think that you have just perfect
a makers model preferably with typeface
on it...That will be a restart to tweaking
your machine.

Then incrementaly decrease the model
size until you find the text break size.

With that solved, resolve that size into
your text on your own lwo shield model.

In other words calibrate the machine
To your needs then ammend the lwo
model to the caibrated machine.

ps whenyou can get 3rdpwers Boo's
they all work back to rev9.6.

Paul_Boland
07-06-2015, 01:03 PM
Hi Folks.

Thanks for the continued detailed feedback.

Spherical:
I had a look at the printer software and I see no way to use a different slicer software (is that what 3D printing software is called? ;) ). There are advanced options but they are very basic such as calibrating the level of the print bed and saying if you want the raft foundation used or not. I don't see any way to direct it to take input form another print source but I will look at KISSlicer later and see if it can talk to the printer. I have not yet worked through your detailed calibration instructions yet either and I'll tell you why... Like you suggested, people on the M3D forum told me to print at medium quality to get a thicker print layer for better results. I did so today and this is the result...

128856

The shield came off the print bed while I was in having dinner and a mess got printed! Suffice to say I have not stopped trying to print this and will wait to hear back from M3D themselves on the matter. I've also noticed on the silver shield print, the shield surface is very thin! If I put my fingers behind it, I can see them through the shield. I have a copy of your calibration instruction here on my computer. I will do them, I promise, just bear with me.

As for the size of the lettering in Productions, I agree, it's tiny! Perhaps way too small to be printed. It did a decent enough job on the other letters through.

Gordonrobb:
No models came with the printer itself but M3D do have a filament holder on their website which I downloaded and printed. It printed the arms great but the support bar was a bit rough. Still, it's working, it's attached to my printer and holding the spool, though I think the arms need to be longer as the spool keeps getting caught on the edge of the printer.

128857

Ianr:
Yep, I tend to address. But what is bugging me is that even the flat shield surface is crusty!! Ggrrr...!!!

I'll keep you posted, folks! Thanks for all the help, really do appreciate it!!!

spherical
07-06-2015, 04:48 PM
I had a look at the printer software and I see no way to use a different slicer software (is that what 3D printing software is called? ;)

Yes. There must be an Expert Mode somewhere, as their Features List talks about it.


I don't see any way to direct it to take input form another print source but I will look at KISSlicer later and see if it can talk to the printer.

KS doesn't "talk to the printer", as in directly controlling it. It generates G-code files that the printer firmware reads and follows the coordinates written in it. It has different firmware choices, so as to generate the type of control needed for a particular style machine. Unsure what choice would be made for a Micro. That'll take some research. Choose the wrong firmware and it all goes wonky, as the machine doesn't match what it is being fed. IOW, a Delta printer uses different commands from a Cartesian and a 5D is different from that. Currently, there are four choices:

Bits from Bytes
5D Relative E
5D - Absolute E
5D - Absolute E no reset (makerbot)


The shield came off the print bed while I was in having dinner and a mess got printed!

Various methods have been devised to increase adhesion to the bed:

A single weight sheet of glass with a thin coat of Aqua Net hairspray (don't know if they have that brand over there) works wonders.
A very watery slurry of ABS dissolved in acetone (I use ABS cement, which is already dissolved and add more acetone to thin it) wiped upon the glass as only a film.
Blue Painter's Tape covering the bed.
Kapton film covering the bed.
Heated bed, perhaps combined with one of the above.



I've also noticed on the silver shield print, the shield surface is very thin! If I put my fingers behind it, I can see them through the shield.

Unclear what you mean, here. Does the overall height of the print match the model? Is the print solid inside or is it filled with Sparse Infill?


But what is bugging me is that even the flat shield surface is crusty!!

Possible causes of "crust" like that comes from nozzle clogging, then it prints air for a while, the extrusion finally begins to flow again, material accumulates on the nozzle and then deposits a stringy blob on the part when the blob gets big enough to touch the surface again. Nozzle distance to part (the area into which the material is to be flowed) is a critical factor. Hence, the increased difficulty in printing very thin layers. The amount deposited must constantly match the amount being supplied or the nozzle will clog.

Paul_Boland
07-12-2015, 12:33 PM
Hi Folks.

This came off my Micro 3D Printer just now...

128911 128912 128913 128914 128915

Awesome!!

I was in contact with M3D and they had me go through three different printer drivers. They told me to print on medium quality with a higher fill rate (I set it to medium too). I covered the print bed with Scotch Blue Painter Tape and the model stuck to it great! The corners still did lift during printing but it took a bit of prying to get it totally free at the end. As stated on this thread, the letters in "Productions" are too small to print. I might try to enlarge them and see what happens, or I might just go without. Either way, I love this print, it looks great! Not 100% perfect, but a hundred times better than previous prints :)!!

Thanks to everyone here for their help and input, very much appreciated!!

Paul_Boland
08-03-2015, 05:30 PM
Hi Folks.

I'm wondering if anyone can help me out again... I have been working on my Dungeoneer shield again and I have increased the size of the word, Productions, for printing. However no matter what I do I can't get Lightwave to balloon it together for me :(. When I try adjusting the size of the word in the ballooned model supplied to me earlier in this tread, my STL exporter won't work, it tosses out an error message. I have Lightwave 10 and it doesn't export STL natively so I have to use the Chromecow STL exporter, located here:
http://www.chromecow.com/

Here is the error message I get:
129120

I am attaching the new shield model. If anyone can give me some pointers on how to successfully get it ballooned together, that would be great!

spherical
08-03-2015, 06:37 PM
I assume you meant Booleaned? Bool Op worked fine here. Did each individually. No errors, though three sets of two points needed manual welding. Did you get any errors during that operation? Before trying the convert to STL, did you save a separate file of only the Booled model with no other layers?

I note that you still have all of those unnecessary polys in there on the front and back. The front can be one poly, the back can be one poly, eight polys creating the ridge going across the top can be reduced to four. Boundaries crossing the letters is where the issues were with the points needing welding.

Try the attached in your STL converter and see if there's any difference.
129124
Also attached is an STL. It was loaded into our slicer and the mesh checked out OK. Remove the trailing ".lwo" extension as the admins haven't figured out that we sometimes work in 3D printing and milling, so need to upload .stl files from time to time. :D
129125

A quick search turned up another plugin that may work as expected. It's by Chilton and found on Liberty3D:
http://www.liberty3d.com/2013/03/free-lscript-export-to-stl-2-0-for-lightwave-9-x-through-11-5/
Works in 9.x through 11.5, is an LScript, so 32-bit and 64-bit compatible,

Paul_Boland
08-04-2015, 06:24 AM
Spherical, thank you very much for the help, really do appreciate it. The new STL file loads into KISSlicer without any errors being reported so I'm going to try and print it now and see how it goes...

In regards to Booleaned (yes, that's what I meant! Oops!!), when I tried it, no matter what order I did it in, Lightwave 10 hung on me, just froze up and I had to manually close it down.

I'll let you know how the new print goes... Thanks again!!

BigHache
08-04-2015, 02:41 PM
Hi Paul,

If I could offer some (hopefully) helpful advice. For the gate/portcullis, if you need more height for definition you could try something like this:

129129

Originally those cross members were floating which will cause issues when printing with filament, especially at a small scale.

spherical
08-04-2015, 03:18 PM
They're Booleaned in, now, so part of the one volume. I agree that the rounded top profile adds not only height but nice detail that differs from the surrounding doorway; if indeed the bars are intended to be round, not make from riveted flat stock. Depends upon the era of the dungeon, I guess. :D

Paul_Boland
08-05-2015, 08:06 AM
Failed print. It was working fine till it got to the finishing off the top layers which for some reason got printed misaligned which has ruined the print. Email sent to M3D tech support (again!).

129136129137129138

zapper1998
08-05-2015, 08:23 AM
The reason it does not print correctly is it is not a single object.
I ran it thru Simplyfi3D slicer and had errors.

You need to make it a single closed object.

Mike

Paul_Boland
08-05-2015, 12:27 PM
The reason it does not print correctly is it is not a single object.
I ran it thru Simplyfi3D slicer and had errors.

You need to make it a single closed object.

Mike

Hi Mike.

Before I tried to print the new STL file from Spherical above, I opened it in KISSlicer and it didn't report any problems at all. Can you tell me what Simply3D said and how this object is not a solid closed object so I can correct it? Thanks.

spherical
08-05-2015, 03:38 PM
I'd wager he's looking at the LWO, not the STL I made; which is one object, one layer and one volume.

The print looks like a hardware issue. Something in the transport mechanism is slipping. If it were the G-code, you'd see it in the Paths display.

EDIT: Ok. Sliced my STL and ran the G-code through an independent Sim. Paths look perfect.

zapper1998
08-06-2015, 07:11 AM
Hi Mike.

Before I tried to print the new STL file from Spherical above, I opened it in KISSlicer and it didn't report any problems at all. Can you tell me what Simply3D said and how this object is not a solid closed object so I can correct it? Thanks.

With my printer. MB 2X
I make the object, Solid, One closed water tight.
When sliced in a slicer from an stl file it looks solid but it is not.
Believe me they print different.

Will load a screen shot of the preview when I get home.

spherical
08-06-2015, 05:36 PM
Then there must be some cruft in your model: duplicate edges, overlapping edges, edges on only one poly, edges on more than two polys, unwelded points, internal polys that are not part of the "outside", duplicate polys, overlapping polys that Unify won't find, etc.

Trust me, that STL file is one object, one volume, is water tight and exhibits a complete set of correct paths in a separate simulator that reads the G-code file generated after slicing. The bad print is a hardware/firmware issue; not the model.

Paul_Boland
08-11-2015, 01:09 PM
With no feedback from M3D tech support, I decided to print the shield again today just to see if the previous print was a fluke and went bad. This is the result... It looks great!! The printer is not going to win any prizes for detail, but I'm happy :).

129243

spherical
08-11-2015, 02:58 PM
Yep. Glad you got a good result. This does point to a physical hardware issue (check for slippage) or perhaps a spike in the line power that screwed the mainboard's brain and it wigged out for the remainder of the print. Is the printer protected by at least a surge suppressor?

Is this sliced with KS or the software that comes with the printer? If the former, you might want to experiment with Partial Infill and set a Skin Thickness equal to the number of Loops x Extrusion Width. This will print a pattern that you select (straight, octagonal, rounded) at a fill percentage you select between the solid fill "skins" on top and bottom and takes a lot less time and filament.

Paul_Boland
08-13-2015, 10:43 AM
Yes, I have the printer plugged into a surge protector socket. This was printed with the default software that ships with the printer so I'm wondering if the previous trip was just a fluke that resulted in a bad print. I'm printing it again for a friend, hopefully I'll get another great result.

Appreciate all the help here!! Thanks, Folks!!

spherical
08-13-2015, 05:39 PM
OK. Gotta get the terminology clear, so I know what you're talking about. There's Slice and there's Print. The two do not have to be combined. Many printers do have a slicer in the printing software that drives the mechanism, so it "looks" to be all one operation. Even then, it isn't. Other printers do not have a slicer as part of the driving software. Some of the ones that do, have the ability to accept code generated by another slicer and then use the printer software to only drive the mechanism. In those that do not have the ability to read externally generated code, the users have developed their own workarounds, as the bundled slicer usually isn't all that good. It's the bare minimum, in order to give the appearance of a one-stop-shop/pushbutton "ease". Printer manufacturers are, quite naturally, printer-centric. They look at the FDM concept from that viewpoint. Writers of slicing software do not; instead looking deep into the deconstruction/reconstruction of the shape in the most efficient way possible and then providing adjustments to minimize undesired effects that the printer will introduce.

Slicing is a very complex operation. One look though the KS tabs and you get the idea that it is far from "one size fits all". There are a blizzard of settings, many of which interact with each other, so that you may obtain the best slice and, therefore, the best print.

ConjureBunny
08-13-2015, 05:43 PM
As for working with Lightwave 10, can't afford to upgrade so have to work with what I have.

Hi Paul,

Please try this plugin. I built it specifically to address issues in other STL implementations I'd tried, and to fix an issue I'd had trying to make a valid STL that worked with some of the online shops.

So far, I've received no complaints about problems, so I think the problems I was having, that are fixed, were the same ones others had, too. And it should work fine with LightWave 10.

It's also free.

http://www.liberty3d.com/2013/03/freebie-export-stl/

Edit: I just read the rest of the forum posts and it looks like you're doing just fine :)

-Chilton

Paul_Boland
08-18-2015, 01:32 PM
Thanks, Folks, for the info and the new STL exporter, appreciated.

Just an update... Had more failed prints. Practically gave up on the getting the shield printed successfully again (only had one successful print so far). Finally, today, M3D wrote back to me and recommended I use the new driver they had just released. I installed it and got the shield printing again. The old printer driver took five hours to print. The new driver said six hours. And sure enough, the print head was moving slower. Well... It's now five and a half hours and it's only at 39% complete!! This is insane!!

So I went to the M3D forums (which for some reason are not public and you have to be a M3D printer owner to see them!?!). I posted about my issue with the slow new driver and asked were others having the same problem? Then I looked at the forum to see what other topics were there and I found one about problems with layers shifting... I read it...

Turns out LOADS OF PEOPLE are having this issue! Prints are being ruined left, right and centre by lots of folks. M3D have issued a reply via their tech support to some people (not me) telling them to perform surgery on their printer. You have to open up the pulley system casing and trim off parts of the pulleys!! It sounds bloody insanely dangerous to me because if you mess it up you ruin your printer!!

Lots of people are losing faith in the machine and I'm with them. I'll leave this print finish and see how it looks, it will take about another six to eight hours by the looks of thing! But for a Kickstarter campaign that asked for US$50,000 and got US$4,000,000!!, I'm not impressed with the end product. I can't really see me doing much with this printer, at least not the plans I had to make stuff for my games and hobbies and stuff. I'll get what fun I can from it, but when it finally gives up the ghost on me, I won't be buying another Micro 3D Printer, I'll look for something with a better reputation and reliability.

spherical
08-18-2015, 04:46 PM
Well, at least we now know that it really has been a hardware problem. What are they asking you to do to the pulleys?

Just as a benchmark, I sliced the final STL to be printed on our 3D Touch with 0.25mm Layer Thickness, 2 Loops, 1mm skin, 33.3% Infill and it came out at 2 hours, 42 minutes.

Paul_Boland
08-18-2015, 07:02 PM
My print is still going!! Don't know what the hell it's doing :(... As for what M3D has told some of people to do, here is what I copied from the forum:

I've made a quick tutorial that may fix the shifting. This fix is for Y-axis shifting...I don't know if that is the issue you're having but shifting on the X axis is a lot less common.

In the factory we've began to cut the ends of the X Rack, which is the toothed rod the print head moves on.

1.One side of the X Rods will be glued to the sliders, one side will not. Hold the slider in one hand, and the X Rack in the other.

2.While holding the slider and X Rack, gently pull apart from each other. Please be careful as to not bend the Z Rods, only a small gap is needed.

3. Rotate the slider so the pegs point upwards. The X Axis is now free from the sliders.

4.Trim piece of plastic sticking out from X Rack. Use wire cutters or diagonal cutters for a clean cut.

5.Install X Axis in reverse order.

spherical
08-18-2015, 07:16 PM
I'm sure it makes more sense when seeing the parts they're referring to. What puzzles me is, this has to do with modifying the X axis, when it is the Y axis that is missing steps that shift the head to an incorrect position. ?

jwiede
08-18-2015, 08:02 PM
Is the printer belt-and-cog driven (with teeth), or pulley-and-cable driven (without teeth/cogs)? Pulley-and-cable/line driven 3DP systems are well-known for having accuracy and reliability issues related to slippage and expansion of the cables/lines versus the pulleys. Basically all printer accuracy in such printers winds up dependent on the cable or line being very taut, not stretching, and not slipping on pulleys/drive. Historically, most pulley-and-cable/line printers are lower-cost versus competitors, but also require much more maintenance, and suffer from ongoing reliability/durability issues.

Paul_Boland
08-19-2015, 07:08 AM
The printer has cogs and belt with teeth. Here is a pic from the maintenance post above that they asked some people to perform.

129346