Bad prints out of brand New Form 3


#266

thank you


#267

Just had to try one more shot with different lighting:

Robert


#268

@monk2002uk, what does the bottom look like?


#269

Sorry to interrupt the tank discussion; I’ve got some good news. @larsenstephen sent me his ship bridge model and I printed it in Black on my Form 2 and [early 2019] Form 3. I’d say the results look great on both.

Print settings:

Form 2 Form 3
Resin Black V4 Black V4
Layer Height 50 μm 50 μm
Preform 3.4.5 3.4.5
Firmware 1.19.15-220 1.6.2-576
Tank LT V1
Layers 903 918
Estimated Volume 8.07mL 8.44mL
Estimated Time 5h 11m 4h 39m
Actual Time 5h 49m* 5h 13m

*My Form 2 has a slow heater and the time for that machine includes approximately 1 hour heating time before the build platform lowered and its print began.

His original supports were used for the Form 2 print. They were imported (including touchpoint sizes) to generate supports for the Form 3.

Both parts underwent 15 minutes in the Form Wash (since my alcohol is getting a little old), then were hand rinsed with fresh 99% IPA from a squeeze bottle and spot dried with canned air. They sat overnight, then were cured for 40 minutes at 60°C.

The surface details like the piping and junction boxes are crisp on both, and the portholes look round. The window corners are comparably square (maybe just ever so slightly more rounded on the Form 3, but it’s barely discernable).

Front view:

(Added) Front view #2: The Form 3 did a better job between the fire control boxes. Reviewing his photos, I think @larsenstephen’s Form 2 in Grey does a better job than my Form 2 in Black on delineating this particular feature.

Front windows closeup:

Starboard side view:

Port side closeup:

Here’s an identical angle on the Form 3 with a slightly different focus:

I did find one thin strand of bridging on a port side girder on the Form 3, which wasn’t present on the Form 2. It didn’t recur on the starboard side. Click to zoom in:

Top closeup:

Feel free to point out anything I missed in these photos, or ask questions if you want to examine any other features in detail. Sharper eyes than mine may spot other, subtle differences.


Form 3 vs. Form 2 Faceoff!
Form3 - let's do a proper comparison with Form2 - how good is "LFS"?
What resin for highly detailed small figures
#270

That is a dramatically better print by the Form 3 in black resin than what the Form 3 is now printing in standard gray V4. I’m very excited to see smooth surfaces (no warping), defined detail, and accurately-shaped openings (circular portholes) in black at 50 microns. I could sell that model.

Let’s hope FL can tweak the Form 3 to work well with gray V4 as well or better than black.

:grinning:


#271

Why can’t you make them and sell them in black?


#272

Good question. These results are intriguing.

Short answer: 1) the Form 2s I have print best in gray. 2) customers are accustomed to and prefer gray models for several reasons, some of which are described below.

Long answer: although a color change would appear to be an effortless change to make with no new costs incurred, actually it’s not an easy change at all.

Generally, products need to be printed in a resin consistent with how they appear in the store’s catalog. And all printers ought to be printing models with the same resin in the same color, just as they appear in the already established catalog. (I would be printing with a mix of Form 2s and Form 3s).

The catalog is huge, composed of over 1400 available models. Most of them, but not all, have been printed. Those that have been printed are printed in gray V4. Those models have been photographed and those photos appear in the catalog. Customers require the models they buy to be consistent with how they appear in the catalog. Rebuilding a catalog that large to include product photos in a new color would require a very costly and lengthy transition, potentially confusing customers and harming sales.

Customers have become accustomed to gray and like it. Customers prefer a material color which helps them with painting. Gray is a very good material color for that purpose.

There are some production efficiency reasons too, a bit tedious to describe here.

For those reasons and some others, I would prefer to stay with gray.

Hopefully, FL will get the Form 3 printing well in gray at 50 microns. These black resin results tend to indicate that the Form 3 is not fundamentally flawed. The Form 3 can achieve excellent quality as evidenced by the photos seen above. These black resin results give me high hopes that FL will be able to achieve Form 2 print quality in gray V4 resin, too.


#273

That’s really encouraging. Now if I could actually buy any black resin, and a new tank. Sadly, the Formlabs store isn’t working, and my third party vendor only has the V1 tank. But I suppose that’s for a different thread.


#274

Here is the bottom (blush):

Robert


#275

there are great results thank you for all this work and confirming my results with black resin!


#276

This is really interesting. I wonder if light scattering is playing a part here. Curious to see how this part would turn out on my F3 in clear or another light color other than gray.


#277

This is really interesting. I wonder if light scattering is playing a part here. Curious to see how this part would turn out on my F3 in clear or another light color other than gray.

I think you’re on to something with regards to light-control. Someone else has also mentioned this before, and it made me do a very simple test-print of small “centicubes” with Gray v4 @ 25micron on my Form 3. By far the biggest issue I saw with that test was potential “overcuring” of resin on the support side of the print. It seems as if the excess resin - that sits around on the “top” of the model (Facing the build plate) - gets cured by the laser as the laser works on layers underneath. I’ll post pictures of this sometime during the weekend, so you can see what I’m on about, but there definitely seems to be an issue with the laser “shining through” layers and curing resin that just sits around on “top” of the model. This might also explain why so many models look “soft”. Anyway, I’ll keep you posted on here.


#278

Makes perfect sense, thank you Oskar. And to leonhart88 for the initial thought.

Robert


#279

So i finally got around to taking some pictures of one of my “centicubes”. The design itself is pretty simple; It’s a 10mm x 10mm x 10mm cube with a notch and hole on two opposing sides. In theory you should be able to join up two or more centicubes by means of this noth and hole. The notch and the hole are 2mm x 2mm x 2mm in dimensions.
The idea behind making these models, where to make as simple a model as possible while still making the printer and software having to deal with overhang and holes. All this for the purpose of seeing which faults and failures would occour and maybe make them a bit easier to diagnose.
All the models where printed in a 3 x 3 grid all across the build plate, to be able to see whether or not the faults were the same or amplified across different areas of the tank.
The models where numbered and arranged in the order that you see in this first picture.

The second and the third picture shows you how the supports where attached to the models. The notch was supported, but the hole wasn’t supported internally.

Now for the interresting thing; the actual print!
In general the print turned out just fine. Almost every face was perfectly straight and was at most 0,04mm out from the original specification. Inspected under a microscope, there was a noticeable difference in surface “texture” on the different sides - but more about this in a later post.
But I wouldn’t be making this post if something wasn’t up, and two major issues did indeed arise;

1. Overcuring of resin on “top” side of model.
This is by far the biggest issue as far as I’m concerned! To give you an idea of how bad this defect is; the biggest deviation from original spec found elsewhere on the model is 0,04mm - the deviation from spec on the top side is 0,48mm!! Thats nearly 0,5mm deviation or 5% in relation to the models intended size. This is - as so many of you guys have already put it - unacceptable and unusable for anyone producing small scale models.
In this first picture I have used a set of callipers to show you the difference between the different sides and the “top” side.

As it can be clearly seen, something is wrong with the “top” side of the model. It has what looks like a slight bulge to it. The sides of the callipers are straight and are pretty much completely parralel to the sides of the model. The printer has done a fairly good job on these sides. It can also be seen from the picture that the notch has a bulge to it as well. The interresting thing about this is that this deviation is about the same as the one for the top side - it’s around 0,46-0,48 in deviation. Also do note that both the top of the notch and the cube itself “slopes” downwards towards the edges of the model. This indicates to me that the resin being cured on the “top” side of all faces is risidual resin - stuck there due to the surface tension of the resin itself - hence why it slopes towards the edge. This seems to back up the theory of this being a light control or a light-power control issue - as the laser cures resin beyond the actual current working layer.

On this second picture the “top” side of the cube is being compared to the straight edge of the calliper, and the deviation becomes very vissible. Another thing to notice is that the hole has also been affected by this error. The side of the hole that has been affected is also the “top” - in the sence that the laser would have been shinning underneath it to create the rest of the cube. Once againg, the deviation is almost spot on 0,48mm…
As can also be seen - although I can’t get a proper picture of this - the “Top” side of the hole also slopes outwards, giving the same impression as the two other examples of the viscous resin resting on the top surfaces.

In conclusion to this specific issue, I think it has something to do with the laser shinning through the current - and maybe even more layers - and hitting some small amount of resin that is naturally going to accumulate on top of the model while undergoing printing. Especially as resin is such a high viscosity liquid, it would make sense that some larger amount of it would stay on top of the mdoel during print. The slopes towards the edges also indicates something along those lines, as som of the resin would be able to run off of the edge while the resin further from the edge would stay due to a combination of high viscosity and surface tension.
So maybe a software tweak to the amount of power granted to the laser, as it works close to top surfaces might be some type of fix - but of course I don’t know what it would take to fix in reality.

2. Displacement / Artifacts along notch/hole
Another issue that I’ve noticed with my test-batch of centicubes, is that they all have similar artifacting/displacement about where the notch and hole is placed. This artifacting is horizontal with the build plate and seems to only occour in layers where the notch and hole is a part of the print proces. The rest of the surface seems to not experience any artifacts. The bottom of the model seems a tiny bit displaced from the top, and it seems to be worst on the first layer after the notch and hole is finished. The first picture - and excuse my ugly fingers - shows how a “dent” in the side of the model (The single arrow). This dent is persistant all across the model at that height. The other arrows shows where the “artifact” or deformation begins and ends. This is also persistant on all sides of the model.

I can’t quite work out how exactly the model is “warped” or deformed. By looking at the model through a microscope I can’t completely tell exactly what is happening, but it seems to be one of two cases - atleast to me;

Case 1; The bottom of the model, after the notch and hole, is warped or deformed towards the non-mixer side of the printer. This seems to be the most likely, however there seems to have been som deformation towards the front of the printer as well - so it might be the case that this deformation is at an angle to the model, being pushed from less than 45deg behind and from the mixer-side.

Case 2; The middle of the model - in the same layers as the notch and hole is present - warping occured in a direction similar to that of Case 1, although in this case the deformation is returned to normal after the layers containing the notch and hole have ceased to be printed. This seems less likely to me, but I wanted to include it here as I can’t quite eliminate one or the other.
I have made an illustration trying to visualize how the two cases looks to me under the microscope.

I’ve also made an illustration of the combined direction in which I think the model is being affected. Se the picture below;

As a final note;
These models and the support and likewise was made with the software version of both preform and firmware that was released around easter.
I’m completely new to everything going on with 3D printing and the Form 3 was something my company bougth as our first machine so that we could produce small scale models for H0 or 1:87 scale trains. We picked the Form 3 due to its promise of being extremly accurate and able to achieve great detail. When I did the first couple of prints on this machine I was actually impressed, and being the rookie I am, I thought that this was the peak of 3D printing capabilities and that it was just the “physical limitations” of the technology that I was seing in werid lines going across my prints and also the very soft undetalied “top”.
When I one day decided to go onto the forums, trying to se if Formlabs were working on a solution to some of the “artifacts” occuring, I stumbled upon this thread. Had it not been for you guys, I would have never realized that the printer was performing wildly under what whould be concidered to be acceptable by industry standard. And escpecially after seing the excellent comparrison photos between Form 2 & 3, I would have never thought that it was possible to get better prints - than what I was getting - with this technology.
So a massive thank you to all of you who have been sharing your perspective of this issue!

I am glad to see that Formlabs seems to be working on a solution for these issues, but I would have liked for them to act in a more professional “buisness-fashion” and giving a public statement on the issue a long time ago.


#280

Hello thanks for your great report. I also hope that Formlabs can do it. My Form3 has already been sent back twice. Because of the errors. I have not yet received a new one. You only tested your test with Black 04 see if it can be due to the material? One reads that black gives different results


#281

Wow! Thanks for the very detailed info. This really shows off the flaws well. After seeing such positive results for other folks on here with Black, do you have access to the black resin for a similar test? If black resin doesn’t have these issues, it seems like the grey could get “there” someday with firmware.


#282

O my gosh. This is so much worse than my (already sold) Epax X1. This is exactly the issues i wanted to avoid with form 3.


#283

Just to drop an update, I have a V2 tank, and some black resin coming, and will post comparison pictures after it gets here and I have a chance to print some of the same parts to compare. That said, the latest software/firmware combo, seems to show some improvement of overcuring with the grey resin, but I haven’t done any systematic testing yet.


#284

Next step was to take the models down another scale. This is a 1/600 scale set of the same model, with a 30mm x 30mm base.

The details are preserved. Form 3 seems to have few problems with models like this - flat as opposed to vertical. The underside of the base was relatively indistinct. I tried printing Steve’s model of the ship superstructure but even at 25 microns the same defects were visible with Gray V4.

As it stands at the moment, some models will print very well; others will not. There has been some improvement, as evidenced by the results with Black, for more complex ‘vertical’ models. This suggests that the problems with Gray will be correctable.

Robert


#285

And a painted version:

The unit ID marker (top right) did not come out well but that is down to my painting skills, or lack thereof. Next version will have the markers recessed into the marker label. That way it only needs a drop of panel line accent colour (like an ink) in the recessed areas to give clear black markers. That’s the theory anyway…

The base is 3cm x 3cm so that you get a sense of the scale.

Robert