Bad prints out of brand New Form 3

There are two very different issues here.

The first one is the ongoing but improving issues the Form 3 is having. There are problems with printing anything below 100 micron, because of software configuration issues that shouldn’t be there, and will hopefully be gone eventually. In that context, it is a big fat excuse.

The second issue is a more fundamental, but less serious than most people are making it sound. There is ALWAYS a chance for an artifact on any layer, so the more layers you have, the more chances of an artifact you have. But this whole idea that this makes 25 micron unsuitable for anything but tiny jewelry, just isn’t true. I have printed literally hundreds of parts at 25 micron between both the Form 2 and Form3, and outside of the first issue I mentioned, I have not noticed all these supposed problems with rampant layer shifting in 25 micron.

That said, 25 micron is slow. If I can get away with printing at a larger layer size, I do, but I have never once said to myself “oh, I wish I could use 25 micron, but then layers are going to shift.” I use it on every part that needs it, and on the Form 2, and now in black on the Form 3, have always been happy with the result.

1 Like

I agree with your points, @lmlloyd. Almost everything I print is at 25 microns with Grey v4 because the quality of the prints is better than 50 microns with the smaller scale models.

There is a third issue though. It relates to the ‘impression’ that potential buyers may get. Before purchasing a Form 3, I noted that some resins were listed as supporting prints down to 25 micron layers. This appealed because I knew in advance what type of models I was interested in printing. When I contacted the local distributor, they were very clear that:
a) anything below 100 microns was not supported at that time in the colour type that I was most interested in; and
b) going below that was rarely beneficial in any case

So I was able to make a decision based on a clear understanding of the known limits of the Form 3 at that time. The representative from the local distributor was incredibly helpful in this regard. She even offered to produce a sample print. Online video reviews were helpful as well.

Once I had purchased my Form 3, it was very clear from PreForm that 25 and 50 micron prints were in beta-phase only. It would have been good to see this stated more clearly in the marketing material but this would not have affected my decision.

Robert

I agree with you- I just mentioned jewelry because that was my discussion with formlabs and not exclusively for jewelry.

This is exactly where I think there is a huge divide between new buyers and previous owners of the Form 2. I am glad to hear that there are distributors out there properly setting expectations for the machine, but for previous Form 2 owners, the obvious expectation was that the new printer would be better than the old printer. I seriously doubt a single person would have “upgraded” to the Form 3, had the expectation been that in a year or so it might get to the point where it prints as well as what you already have, in some select resins.

I don’t want to just endlessly bag on Formlabs. They actually have done a good job fixing some of these problems, and are clearly still working hard to fix the remaining ones. Credit where credit is due. But it is difficult when you hear that your expectation were too high, if you expected a new $3,500 printer to work as well out of the box, as the several year old machine it is replacing.

4 Likes

I’ve had my Form 3 for 9 months now and printing in White V4 still sucks as much as it did 9 months ago.

Just to collate data for as many resins as possible-- I’ve been trying to address the same issues discussed in this thread (overcuring, layer-shifting, etc.) with Castable Wax resin at 25 microns to no avail.

I would be interested to hear from anyone successfully printing high resolution Castable Wax on the Form 3.

-JD
Lab Partners Jewelry

Have been having some success with the Form 3 in gray V4 and black V4 printing models that do not have openings. My Form 3 has achieved a 69% success rate since the 1.6.6 firmware update. This may sound very low but it had a 0% success rate previously. (My two Form 2s have a 96% success rate). So tried printing large, flat objects, to see how the Form 3 would perform, in this case printing nameplates for scale models. As you know, large flat objects are typically challenging for 3D printers to print. My Form 2s print them very well. The nameplates are 10 cm long (~4 inches).

Several nameplates were printed on the Form 3 in both black V4 and gray V4.

The good:

  • the models did print flat as intended
  • text printed distinctly with sharp edges
  • decorative rope border printed as intended

The bad: all test prints were noticeably warped, gray V4 very slightly, but the black V4 especially so. It was if the nameplate were bent in the shape of a frown. Also, the left and right edges were shorter dimensionally than the middle in some cases by as much as 3 mm.

The nameplates were warped such that the top and bottom edges printed curved although designed to be perfectly straight. Text followed the curve. There was some layer shifting present in black V4 but not gray V4.

The nameplates’ top edge, which suffered the most curvature, were oriented down for printing.

Changing orientation to attempt to eliminate the warp:

  • Changing the orientation on the build platform (left-right vs. front-back) did not improve or eliminate the warp.
  • Changing the nameplate’s position on the build platform did not improve or eliminate the warp.
  • Changing the orientation of a nameplate manually to increase the angle in two planes (printing it with a corner pointed down like a diamond rather than an edge pointed down) had no affect.

Below is a rendering of one as designed followed by photos of the test prints. Neither gray or black model is sellable in this condition.

  • PreForm 3.6.0
  • firmware 1.6.8
  • 50 microns
  • Gray V4 and Black V4 resins
  • orientation: “one click print” auto-orientation used





2 Likes

Ouch. Does it curve mostly in the Z plane or X/Y?

I know it would introduce other issues but I’d be curious to see what it looks like with the backside printed directly on the base (you’ll probably need to add some thickness).

Did you post cure these, or is this straight out of the wash?

I haven’t seen this problem before. I just printed several mostly flat pieces that were about 11cm, and they were close enough in the X and Y, that they all tile together just fine. There was some bowing in the Z, but shelling out the center helped with that.

1 Like

What was the support touch point size and density? Supports look pretty sparse.

Also curious how long it was washed for and if this is post cure. I haven’t seen issues like this on my F3.

I don’t think it’s a support issue. That looks like adequate supports for something that thickness. Moreover, at least as far as I can tell from that picture, the supports look straight, and there doesn’t seem to be any warping in them. So it’s not the part sagging against the supports.

1 Like

Relative to the nameplate, the warp occurs in the XY plane with dimensional inaccuracy at the ends of the nameplate (the ends are shorter than the middle, as if they ends have shrunk in the XY plane). The nameplates are uniformly thick in the Z plane, as intended.

The flat background portion of the nameplate is 1 mm thick. The rope border rises another 1 mm above the background for a total thickness of 2.0 mm. The text rises 0.75 mm above the flat background.

9 nameplates tested. All nine warped. Warp in black V4 much worse than in gray V4.

Yes, the supports are straight. I may be wrong, but the print raft does not appear to be warped, just the nameplates. No sag between supports. Once removed from the raft, the back of the nameplates are flat, as intended.

Touchpoint size 0.6 mm
Touchpoint density 1.0

Supports and orientation auto-generated by PreForm for 6 of the nine test prints. The last 3 nameplates were auto-generated in PreFrom then their orientation was manually changed to increase the angle of the nameplate relative to the build platform, and to increase the number of supports, and for test print number 9 to change the nameplate orientation on the build platform in the Z plane by 90 degrees. Manual changes did not improve the warp or eliminate it.

Warp noticeable before wash and cure, right out of the Form 3. FormWash and FormCure had no noticeable affect on the existing warp. In other words, the FormWash and FormCure did not create the warp nor worsen it.

Before photos taken:
7 minutes in Formwash with 99% IPA
30 minutes in FormCure at 60 degrees

I would not attempt to print nameplates directly on the build platform due to adhesion issues with the Form 3 (a different issue discussed in another thread and with FL support). Models are generally very difficult to remove from the Form 3 build platform without breaking them, especially with black V4. The Form 2 does not have this adhesion issue. Having broken several print rafts and models trying to remove them from the Form 3 build platform, I am quite certain I would break thin nameplates attempting to remove them from the build platform if printed directly on the platform. Regarding thickening the model so that it can be removed from the platform, in principle, a model that prints well on a Form 2 should not have to be redesigned changing its geometry, effectively making it a new model, in order to be printable on a Form 3. This would indicate that in order for a model to be printable on the Form 3, each model would require an iterative process to find by print experimentation the optimal orientation and support array until a printable, non-warped model was achieved. That is not practical or economical in terms of time and expense. The Form 2 does not require such an iterative process. A $240 Anycubic Photon does not require such a process.

Since the 1.6.6 firmware update, my Form 3 is now able to print some models well in gray and black at 50 microns and 25 microns, too, generally those that have a boxy geometry and do not have openings. Very thankful for those who worked hard to achieve that update. But the Form 3 continues to struggle with models that are flat or have small openings, especially if the openings are in the vertical plane. These nameplates are among those that my Form 3 cannot successfully print. Hopefully, future Form 3 firmware and PreForm updates will correct the warp exhibited by these nameplates.

3 Likes

I’ve never printed with the Form 2. I agree 100% with your observation that printing on the Form 3 requires far too much iteration. I have found that the quality of prints is very dependent on orientation and support settings. Even when Preform tells me that I have no issues, I have had very many prints that needed fine tuning of the orientation/supports. It is very frustrating and results in unnecessary waste of resin. Formlabs needs to address this issue.

I’m curious how this was positioned on the build platform. Let’s call the front side X, and the wiper side Y. Was this positioned running along X, along Y, or diagonally along both axis?

Just a quick edit for clarity, I know you said you rotated it 90° in Z, which is I think what I’m asking, but I’m not 100% sure from the way you phrased it.

Yes, you’re right. The nameplates were first oriented along the wiper side which is where PreForm put them. When 6 prints oriented that way failed, the next 3 nameplates were rotated 90 degrees so that their long side was oriented parallel to the front side of the build platform. Their upward angle was also increased so that a corner was down rather than a flat edge and the number of supports was increased. Those 3 nameplates also failed with the same kind of warp, much worse with black V4 than with gray V4. Change in orientation and increasing support density had no affect.

Since the nameplates print well on both Form 2s I own, with no warp or distortion, this tends to indicate that the Form 3’s inability to print them is not related to the design of the nameplate, but to problems with the Form 3. Since user-initiated changes in orientation, and the addition of more supports do not correct the warp, this tends to indicate that these problems cannot be resolved by the user/owner.

Frequent discussion with FL techs confirm that FL is well aware of these very real issues but despite their best efforts have not yet been able to resolve them. Very talented engineers are working hard to solve Form 3 print quality problems achieving some notable successes. But we’re more than a year from launch and the Form 3 continues to exhibit the most basic print capability issues for many owners. Their standard resins are still in beta at 50 and 25 microns! After this much time, I am deeply concerned that the Form 3’s problems cannot be solved. FL engineers aren’t stupid. If the Form 3’s issues could be solved, FL’s very talented engineers would have been able to solve them by now.

My first Form 3 only printed the rafts, but nothing above them, no models. Its “hot swapped” replacement cannot print a simple rectangle with dimensional accuracy as shown in the nameplate photos in the posts above. My second Form 3 had a 0% print success rate for the first 7 months after receipt with all models warped, deformed, or loaded with shifted layers. After firmware update 1.6.8, it has just now achieved a marginal 69% success rate, but only able to print boxy-shaped objects with no openings.

My two form 2s have a 96% success rate and print all of my designs perfectly, with no iterative processes required to find optimal orientations or support array. The only time the Form 2s suffer a print fail is when they need some routine cleaning or the resin tank is worn out and needs replacing.

I am truly happy to hear that the Form 3 is working for some customers. For me, the two Form 3s I have owned have experienced significant and costly print quality problems.

It is extremely interesting to me how it is failing in this case. It looks like an optical distortion, which would cause me to guess it would happen along the axis of the LPU mirror (wiper side axis). If it were a mechanical tracking distortion, along the axis of screw travel, I would expect a different kind of distortion, and that distortion would be minimized by placing the part along the wiper side axis. If you are getting the same distortion along both axis, that would almost seem to be a software issue, but the fact that you are using the same software to print just fine on the Form 2, rules that out. This one is a real head scratcher. Which of course does you no good at all.

This is odd.

I have a suspicion that if you increased your support density (or just manually added supports), this could prevent the warp.

You mentioned this is only in the XY plane. You also mentioned this is printed at 50um. It’s very possible that the thin 50um layers aren’t being supported appropriately and then are curling or warping. One possible reason why this might occur on the Form 3 and not the Form 2 is because the Form 3 has a smaller laser spot size which might mean a) longer raster time for warping, b) less forgiving in attaching to supports if there are inadequate supports placed.

I agree that you shouldn’t have to worry about this and the software should just do things appropriately automatically…however, I’ve worked in this industry enough to know that there are always things you need to tweak and adjust for because parts are always different.

Also, I’ve printed flat labels/decals like this in the past right on the build plate. It comes out pretty well. The only problem is that when you use side cutters to get under the one corner to pop it off, that corner gets a little ugly…but nothing you can’t fix with some light sanding! Again, side cutters + metal spatula has saved me countless hours of battling with really stuck on parts.

I appreciate the advice, especially the recommendation to use side cutters to get prints off the Form 3 build platform. As discussed in another thread, Form 3 print adhesion is much stronger than Form 2. FL support also recommended putting the platform with models still attached in the freezer as another technique that works, too. I have not tried the freezer technique because the time required would disrupt an already tight production schedule crippling my ability to meet customer orders. But for others, it may be a good method. Your recommended side cutter technique is working.

For the last 3 of 9 nameplate prints, I did increase support density by manually adding supports, over 100 more supports to each nameplate. Those 3 nameplates were essentially saturated with supports. I regret to report that the addition of supports did not eliminate the warp. This tends to indicate that support density is not the cause of the warp.

Several months ago, I first contacted FL support about terrible Form 3 print quality compared to Form 2, stating that unlike the Form 2, my Form 3 was not capable of printing a sellable model (see photo below). Models were warped, there was significant layer shifting present and openings were misshapen, etc. Models printed on my Form 2s never exhibited these artifacts.

At that time, FL support was in complete denial there was anything wrong with the Form 3. Reps repeatedly stated or implied that Form 3 print problems were caused by user error or user unfamiliarity with 3D printing. FL rep responses to my claims explicitly stated that Form 3 print failures could be easily corrected by the user by changing model orientation and/or increasing support density through an iterative process until a successful print was achieved.

With over 1500 successful prints on the Form 2, and having been specifically told that the concept of low-force SLA would mean fewer supports would be needed and even smaller attachment points would be needed compared to the Form 2, the advice seemed to be completely inconsistent with the concept and benefits of low-force SLA. If low-force SLA required fewer supports than the Form 2, why would the Form 3 need more supports than the Form 2 to achieve a successful print? But anxious to get something, anything, sellable out of the Form 3, I made numerous, costly attempts to follow FL advice precisely. Despite these efforts, there was no improvement.

Only after some months did an FL rep admit that FL had come to recognize that the problems my Form 3 was having were real, common, had become known to FL engineers, and that the problems could not be solved by the user. The problems could and would be resolved through future firmware and PreFrom updates.

Firmware and PreForm updates as the likely solution were again recommended by FL reps as late as March. At that time, a FL support rep contacted me concerned about my Form 3’s abysmal success rate and asked what was happening and if FL might be able to help (kudos to FL reps for monitoring customer success rates and reaching out). FL considered the possibility that the problems were specific to my machine and that a “hot swap” may be needed. But after evaluating photos of failed prints, the FL support rep stated that the serious print quality problems I was experiencing (warped prints and serious layer shifting) remained common and known, and a replacement printer was not likely to print any better. He stated these issues were being worked on by FL engineers and future firmware and PreFrom updates would come.

As promised, Preform update and firmware update to version 1.6.8 did in fact resolve many of the print quality problems my Form 3 was experiencing. But, the warped nameplates indicate not all Form 3 problems have yet been fixed.

The model on the right is how every model was coming out of my Form 3 prior to the update to firmware 1.6.8. Obviously, not sellable in that condition.

Whatever the cause of warped models, the user is not likely to be able to fix them. More firmware improvement is needed. That Form 3 printing at 50 microns and 25 microns is still in beta tends to indicate that FL engineers are not yet satisfied with Form 3 performance at those resolutions. Only FL can make the necessary improvements needed to achieve satisfactory quality and resolution certification.

I wish to emphasize yet again that the Form 2 does not have any of these issues and prints models I designed, including nameplates, perfectly at all available resolutions with no need for an iterative process. The print quality problems my Form 3 is experiencing are Form 3 specific, very real, and are serious. I anxiously await the firmware and/or PreForm update that will fix warping.

2 Likes

Hm interesting to hear that additional supports didn’t help. Do you want to DM me your .form file? I’m curious to print it on my F3.

It seems like the F2 is worse at printing miniatures than the F2. Funnily enough, the F3, for me at least, has proved to be much better than the F2 in handling larger parts designed for engineering applications.

The only thing I do wish though, is that the fine tuning and tweaking of resins across the board was faster, as I think there are still some engineering resins that don’t look as sharp as Black (although they are less frequently used).

It’s been pretty clear for quite some time that this is not an orientation or support issue. it has been stated multiple times in this thread, and has come directly from Formlabs’ mouth after initial denial of the fact.

And yes, Form 3 is considerably worse at miniatures than the Form 2. I started this entire thread because I bought a Form 3 after being beyond thrilled with my Form 2 and it’s results in miniatures, that I figured although an iterative upgrade, it had to be better than the Form 2 and thus I would see my good results from the 2 get just that much better. I was very, VERY wrong and, going on nearly a year later, there are only slight improvements. In my opinion it’s an unmitigated disaster.

2 Likes